v>EPA
EPA/635/R-23/056b
External Review Draft
www.epa.gov/iris
IRIS Toxicological Review of Perfluorodecanoic Acid [PFDA, CASRN 335-
76-2] and Related Salts
Supplemental Information
April 2023
Integrated Risk Information System
Center for Public Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
DISCLAIMER
This document is an external review draft for review purposes only. This information is
distributed solely for the purpose of predissemination peer review under applicable information
quality guidelines. It has not been formally disseminated by EPA. It does not represent and should
not be construed to represent any Agency determination or policy. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
This document is a draft for review purposes only and does not constitute Agency policy.
ii DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
CONTENTS
APPENDIX A. SYSTEMATIC REVIEW PROTOCOL FOR THE PFAS IRIS ASSESSMENTS A-l
APPENDIX B. LITERATURE SEARCH STRATEGY AND POPULATIONS, EXPOSURES, COMPARATORS,
AND OUTCOMES (PECO) CRITERIA B-l
B.l. LITERATURE SEARCH AND SCREENING STRATEGY B-l
APPENDIX C. BENCHMARK DOSE MODELING RESULTS C-l
C.l. BENCHMARK DOSE MODELING RESULTS FROM HUMAN STUDIES C-l
C.l.l. BENCHMARK DOSE MODELING APPROACHES FOR IMMUNE EFFECTS C-l
C.l.2. BENCHMARK DOSE MODELING APPROACHES FOR DEVELOPMENTAL EFFECTS C-16
C.2. BENCHMARK DOSE MODELING RESULTS FROM ANIMAL STUDIES C-24
C.2.1. BENCHMARK DOSE MODELING APPROACHES C-24
C.2.2. INCREASED AST—MALE RATS (NTP, 2018) C-26
C.2.3. INCREASED AST—FEMALE RATS (NTP, 2018) C-31
C.2.4. INCREASED ALP—FEMALE RAT (NTP, 2018) C-35
C.2.5. INCREASED RELATIVE LIVER WEIGHT—MALE RAT (NTP, 2018) C-39
C.2.6. INCREASED RELATIVE LIVER WEIGHT—FEMALE RAT (NTP, 2018) C-44
C.2.7. INCREASED RELATIVE LIVER WEIGHT (HISTO)—FEMALE RATS (Frawley et al.,
2018) C-51
C.2.8. INCREASED RELATIVE LIVER WEIGHT (MPS)—FEMALE RATS (Frawley et al.,
2018) C-55
C.2.9. INCREASED RELATIVE LIVER WEIGHT (TDAR)—FEMALE RATS (Frawley et al.,
2018) C-59
C.2.10.DECREASED FETAL WEIGHT—MALE AND FEMALE RATS (Harris and Birnbaum,
1989) C-63
C.2.11.DECREASED SPERM COUNT—MALE RATS (NTP, 2018) C-69
C.2.12.DECREASED ABSOLUTE TESTIS WEIGHT IN MALE RATS (NTP, 2018) C-71
C.2.13.DECREASED ABSOLUTE CAUDAL EPIDIDYMIS WEIGHT IN MALE RATS (NTP,
2018) C-75
C.2.14.DECREASED ABSOLUTE WHOLE EPIDIDYMIS WEIGHT IN MALE RATS (NTP, 2018) C-78
C.2.15.DECREASED DAYS IN ESTRUS—FEMALE RATS (Butenhoff et al., 2012; van
Otterdijk, 2007) C-82
C.2.16.INCREASED DAYS IN DIESTRUS—FEMALE RATS (Butenhoff et al., 2012; van
Otterdijk, 2007) C-86
This document is a draft for review purposes only and does not constitute Agency policy.
iii DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
C.2.17.DECREASED RELATIVE UTERINE WEIGHT—FEMALE RATS (Butenhoff et al.,
2012; van Otterdijk, 2007) C-90
C.2.18.DECREASED ABSOLUTE UTERINE WEIGHT—FEMALE RAT (Butenhoff et al.,
2012; van Otterdijk, 2007) C-94
APPENDIX D. ADVERSE OUTCOME PATHWAY/ MODE OF ACTION(AOP/MOA)-BASED APPROACH
FOR EVALUATING PFDA-INDUCED MECHANISM OF HEPATOXITY D-l
D.l. OBJECTIVE AND METHODOLOGY D-l
D.2. PROPOSED MOA/AOP APPROACH FOR EVALUATING PFAS-INDUCED LIVER TOXICITY D-2
D.3. SYNTHESIS OF MECHANISTIC STUDIES AND SUPPLEMENTAL INFORMATION FOR PFDA D-4
D.3.1. MOLECULAR INITIATING EVENTS D-4
D.3.2. CELLULAR EFFECTS D-8
D.3.3. ORGAN-LEVEL EFFECTS D-15
APPENDIX E. ANALYSIS OF RELEVANT HIGH-THROUGHPUT SCREENING ASSAYS FROM EPA'S
CHEMICALS DASHBOARD E-l
E.l. IN VITRO BIOACTIVITY DATA RELEVANT TO THE MECHANISMS OF PFDA-INDUCED
LIVER EFFECTS E-l
E.2. IN VITRO BIOACTIVITY DATA RELEVANT TO THE POTENTIAL MECHANISMS OF
REPRODUCTIVE TOXICITY E-16
APPENDIX F. ADDITIONAL CONFOUNDING CONSIDERATIONS F-23
F.l. SPECIFIC PFAS CONFOUNDING CONSIDERATIONS FOR FETAL GROWTH RESTRICTION F-23
F.2. PFAS COEXPOSURE STATISTICAL APPROACHES AND CONFOUNDING DIRECTIONALITY F-24
F.3. PFDA AND PFAS COEXPOSURE STUDY RESULTS F-25
APPENDIX G. DETAILED PHARMACOKINETIC ANALYSES G-l
G.l. PARTIAL POOLING OF PFDA PHARMACOKINETIC DATA FOR HIERARCHICAL BAYESIAN
ANALYSIS G-l
G.l.l. Pharmacokinetic model G-l
G.l.2. Bayesian inference G-3
G.1.3. Prior sensitivity analysis G-5
G.1.4. Study-specific Clearance Values and Model Fits G-6
G.2. DESCRIPTION AND EVALUATION OF A SINGLE-COMPARTMENT PK APPROACH G-10
APPENDIX H. SUMMARY OF PUBLIC AND EXTERNAL PEER REVIEW COMMENTS AND EPA'S
DISPOSITION H-l
APPENDIX I. QUALITY ASSURANCE FOR THE IRIS TOXICOLOGICAL REVIEW OF
PERFLUORODECANOIC ACID AND RELATED SALTS 1-1
This document is a draft for review purposes only and does not constitute Agency policy.
iv DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
TABLES
Table B-l. Summary of detailed search strategies for Perfluorodecanoic Acid and Related Salts
(PubMed, Web of Science, Toxline, TSCATS, Toxcenter) B-l
Table C-l. Results specific to the slope from the linear analyses of PFDA measured in serum at
age 5 years and log2(tetanus antibody concentrations) measured at age 7 years
in a single-PFAS model and in a multi-PFAS model from (Budtz-J0rgensen and
Grandjean, 2018b) C-l
Table C-2. BMDs and BMDLs for effect of PFDA at age five years on anti-tetanus antibody
concentrations at age seven years using a BMR of Vz SD change in log2(tetanus
antibodies concentration) and a BMR of 1 SD change in log2(tetanus antibodies
concentration) C-6
Table C-3. Results specific to the slope from the linear analyses of PFDA in serum measured at
age 5 years and log2(diphtheria antibodies) measured at age 7 years from Table
1 in a single-PFAS model and in a multi-PFAS model from (Budtz-J0rgensen and
Grandjean, 2018b) C-8
Table C-4. BMDs and BMDLs for effect of PFDA at age 5 years on anti-diphtheria antibody
concentrations at age 7 years using a BMR of Vz SD change in log2(diphtheria
antibodies concentration) and a BMR of 1 SD log2(diphtheria antibodies
concentration) C-10
Table C-5. Results of the linear analyses of PFDA measured perinatally in maternal serum and
tetanus antibodies measured at age 5 years in a single-PFAS model and in a
multi-PFAS model from (Budtz-J0rgensen and Grandjean, 2018b) C-ll
Table C-6. BMDs and BMDLs for effect of PFDA measured perinatally and anti-tetanus antibody
concentrations at age 5 years C-12
Table C-7. Results of the analyses of PFDA measured perinatally in maternal serum and
diphtheria antibodies measured at age 5 years in a single-PFAS model and in a
multi-PFAS model from (Budtz-J0rgensen and Grandjean, 2018b) C-14
Table C-8. BMDs and BMDLs for effect of PFDA measured perinatally and anti-diphtheria
antibody concentrations at age 5 years C-15
Table C-9. Selected BMDs and BMDLs and associated uncertainty for effect of PFDA on
decreased antibody responses in children from Budtz-J0rgensen and Grandjean
(2018a) C-16
Table C-10. BMDs and BMDLs for effect of PFDA on decreased birth weight, by using percentage
(8.27%) of live births falling below the public health definition of low birth
weight, or alternative study-specific tail probability C-22
Table C-ll. Sources of data used in benchmark dose modeling of PFDA endpoints from animal
studies C-25
Table C-12. Dose-response data for increased AST in male rats (NTP, 2018) C-26
Table C-13. Benchmark dose results for increased AST in male rats—constant variance, BMR = 1
standard deviation (NTP, 2018) C-27
Table C-14. Dose-response data for increased AST in female rats (NTP, 2018) C-31
Table C-15. Benchmark dose results for increased AST in female rats—constant variance,
BMR = 1 standard deviation (NTP, 2018) C-31
Table C-16. Benchmark dose results for increased AST in female rats—nonconstant variance,
BMR = 1 standard deviation (NTP, 2018) C-32
This document is a draft for review purposes only and does not constitute Agency policy.
v DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-17. Benchmark dose results for increased AST in female rats—log-normal, constant
variance, BMR = 1 standard deviation (NTP, 2018) C-33
Table C-18. Dose-response data for increased ALP in female rats (NTP, 2018) C-35
Table C-19. Benchmark dose results for increased ALP in female rats—BMR = constant variance,
1 standard deviation (NTP, 2018) C-35
Table C-20. Benchmark dose results for increased ALP in female rats—nonconstant variance,
BMR = 1 standard deviation (NTP, 2018) C-37
Table C-21. Benchmark dose results for increased ALP in female rats—log-normal, constant
variance, BMR = 1 standard deviation (NTP, 2018) C-38
Table C-22. Dose-response data for increased relative liver weight in male rats (NTP, 2018) C-39
Table C-23. Benchmark dose results for increased relative liver weight in male rats—constant
variance, BMR = 10% relative deviation (NTP, 2018) C-39
Table C-24. Benchmark dose results for increased relative liver weight in male rats—constant
variance, BMR = 1 standard deviation (NTP, 2018) C-43
Table C-25. Dose-response data for increased relative liver weight in female rats (NTP, 2018) C-44
Table C-26. Benchmark dose results for increased relative liver weight in female
rats—BMR = constant variance, 10% relative deviation (NTP, 2018) C-44
Table C-27. Benchmark dose results for increased relative liver weight in female
rats—nonconstant variance, BMR = 10% relative deviation (NTP, 2018) C-45
Table C-28. Benchmark dose results for increased relative liver weight in female rats—log-
normal, constant variance, BMR = 10% relative deviation (NTP, 2018) C-46
Table C-29. Benchmark dose results for increased relative liver weight in female rats, high dose
dropped—BMR = constant variance, 10% relative deviation (NTP, 2018) C-47
Table C-30. Benchmark dose results for increased relative liver weight in female rats, high dose
dropped—constant variance, BMR = 1 standard deviation (NTP, 2018) C-51
Table C-31. Dose-response data for increased relative liver weight (Histo) in female rats (Frawley
etal., 2018) C-51
Table C-32. Benchmark dose results for increased relative liver weight (Histo) in female
rats—constant variance, BMR = 10% relative deviation (Frawley et al., 2018) C-52
Table C-33. Benchmark dose results for increased relative liver weight (Histo) in female
rats—constant variance, BMR = 1 standard deviation (Frawley et al., 2018) C-54
Table C-34. Dose-response data for increased relative liver weight (MPS) in female rats (Frawley
etal., 2018) C-55
Table C-35. Benchmark dose results for increased relative liver weight (Histo) in female
rats—constant variance, BMR = 10% relative deviation (Frawley et al., 2018) C-55
Table C-36. Benchmark dose results for increased relative liver weight (MPS) in female rats —
constant variance, BMR = 1 standard deviation (Frawley et al., 2018) C-58
Table C-37. Dose-response data for increased relative liver weight (TDAR) in female rats (Frawley
etal., 2018) C-59
Table C-38. Benchmark dose results for increased relative liver weight (TDAR) in female
rats—constant variance, BMR = 10% relative deviation (Frawley et al., 2018) C-59
Table C-39. Benchmark dose results for increased relative liver weight (TDAR) in female
rats—non-constant variance, BMR = 10% relative deviation (Frawley et al., 2018) C-61
Table C-40. Benchmark dose results for increased relative liver weight (TDAR) in female
rats—log-normal, constant variance, BMR = 10% relative deviation (Frawley et
al., 2018) C-62
Table C-41. Dose-response data for decreased fetal weight in male and female rats (Harris and
Birnbaum, 1989) C-63
This document is a draft for review purposes only and does not constitute Agency policy.
vi DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-42. Benchmark dose results for decreased fetal weight in male and female
rats—constant variance, BMR = 5% relative deviation (Harris and Birnbaum,
1989) C-64
Table C-43. Benchmark dose results for decreased fetal weight in male and female
rats—nonconstant variance, BMR = 5% relative deviation (Harris and Birnbaum,
1989) C-65
Table C-44. Benchmark dose results for decreased fetal weight in male and female rats—log-
normal, constant variance, BMR = 5% relative deviation (Harris and Birnbaum,
1989) C-67
Table C-45. Dose-response data for decreased sperm counts in male rats (NTP, 2018) C-69
Table C-46. Benchmark dose results for decreased sperm counts in male rats, BMR = 1 standard
deviation (NTP, 2018) C-69
Table C-47. Dose-response data for decreased absolute testis weight in male rats (NTP, 2018) C-71
Table C-48. Benchmark dose results for decreased absolute testis weight in male rats—constant
variance, BMR = 1 standard deviation (NTP, 2018) C-72
Table C-49. Dose-response data for decreased absolute caudal epididymis weight in male rats
(NTP, 2018) C-75
Table C-50. Benchmark dose results for decreased absolute caudal epididymis weight in male
rats—constant variance, BMR = 1 standard deviation (NTP, 2018) C-75
Table C-51. Benchmark dose results for decreased absolute caudal epididymis weight in male
rats—nonconstant variance, BMR = 1 standard deviation (NTP, 2018) C-76
Table C-52. Dose-response data for decreased absolute whole epididymis weight in male rats
(NTP, 2018) C-78
Table C-53. Benchmark dose results for decreased whole caudal epididymis weight in male
rats—constant variance, BMR = 1 standard deviation (NTP, 2018) C-79
Table C-54. Benchmark dose results for decreased absolute whole epididymis weight in male
rats—nonconstant variance, BMR = 1 standard deviation (NTP, 2018) C-79
Table C-55. Dose-response data for decreased days in estrus in female rats (Butenhoff et al.,
2012; van Otterdijk, 2007) C-82
Table C-56. Benchmark dose results for decreased days in estrus in female rats—constant
variance, BMR = 5% relative deviation (Butenhoff et al., 2012; van Otterdijk,
2007) C-83
Table C-57. Benchmark dose results for decreased days in estrus in female rats—constant
variance, BMR = 1 standard deviation (Butenhoff et al., 2012; van Otterdijk,
2007) C-85
Table C-58. Dose-response data for increased days in diestrus in female rats (Butenhoff et al.,
2012; van Otterdijk, 2007) C-86
Table C-59. Benchmark dose results for increased days in diestrus in female rats—constant
variance, BMR = 5% relative deviation (Butenhoff et al., 2012; van Otterdijk,
2007) C-87
Table C-60. Benchmark dose results for increased days in diestrus in female rats—constant
variance, BMR = 1 standard deviation (Butenhoff et al., 2012; van Otterdijk,
2007) C-90
Table C-61. Dose-response data for decreased relative uterine weight in female rats (Butenhoff
et al., 2012; van Otterdijk, 2007)) C-90
Table C-62. Benchmark dose results for decreased relative uterine weight in female
rats—BMR = constant variance, 1 standard deviation (Butenhoff et al., 2012; van
Otterdijk, 2007) C-91
This document is a draft for review purposes only and does not constitute Agency policy.
vii DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-63. Benchmark dose results for decreased relative uterine weight in female rats —
nonconstant variance, BMR = 1 standard deviation (Butenhoff et al., 2012; van
Otterdijk, 2007) C-92
Table C-64. Benchmark dose results for decreased relative uterine weight in female rats—log-
normal, constant variance, BMR = 1 standard deviation (Butenhoff et al., 2012;
van Otterdijk, 2007) C-92
Table C-65. Dose-response data for decreased absolute uterine weight in female rats (NTP,
2018) C-94
Table C-66. Benchmark dose results for decreased absolute uterine weight in female
rats—BMR = constant variance, 1 standard deviation (Butenhoff et al., 2012; van
Otterdijk, 2007) C-94
Table C-67. Benchmark dose results for decreased absolute uterine weight in female
rats—nonconstant variance, BMR = 1 standard deviation (Butenhoff et al., 2012;
van Otterdijk, 2007) C-95
Table C-68. Benchmark dose results for decreased absolute uterine weight in female rats—log-
normal, constant variance, BMR = 1 standard deviation (Butenhoff et al., 2012;
van Otterdijk, 2007) C-96
Table E-l. Bioactivity summary for PFDA from in vitro HTS assays from ToxCast/Tox21 conducted
in human liver cell lines (HepG2 and HepaRG cells) and grouped by biological
response/target3,b E-6
Table E-2. Bioactivity summary for PFDA from in vitro HTS assays evaluating nuclear receptor-
related activities from ToxCast/Tox21 across multiple endpoints and cell
typesabc E-13
Table E-3. Bioactivity summary for PFDA from in vitro HTS assays evaluating activities for the AR,
ERab E-17
Table E-4. ToxCast model predictions for the ER and AR pathways for PFDAa E-20
Table E-5. Bioactivity summary for PFDA from in vitro HTS assays related to steroidogenesis3,15 E-21
Table F-l. PFAS correlation coefficients in mutually adjusted studies F-25
Table F-2. Impact of coexposure adjustment on estimated change in mean birth weight per unit
change (ng/mL) in PFDA levels3 F-28
Table G-l. Weakly informed prior distributions for pharmacokinetic parameters used in the
Bayesian analysis G-3
FIGURES
Figure C-l. Difference in population tail probabilities resulting from a one standard deviation
shift in the mean from a standard normal distribution, illustrating the
theoretical basis for a baseline BMR of 1 SD C-4
Figure C-2. Difference in population tail probabilities resulting from a Vz standard deviation shift
in the mean from an estimation of the distribution of log2(tetanus antibody
concentrations at age seven years) C-6
Figure C-3. Dose-response curve for the Hill model fit to increased AST in male rats (NTP, 2018) C-28
Figure C-4. User Input for dose-response modeling of increased AST in male rats (NTP, 2018) C-29
Figure C-5. Model Results for increased AST in male rats (NTP, 2018) C-30
This document is a draft for review purposes only and does not constitute Agency policy.
viii DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Figure C-6. Dose-response curve for the Hill model fit to increased relative liver weight in male
rats (NTP, 2018) C-40
Figure C-7. User Input for dose-response modeling of increased relative liver weight in male rats
(NTP, 2018) C-41
Figure C-8. Model Results for increased relative liver weight in male rats (NTP, 2018) C-42
Figure C-9. Dose-response curve for the Hill model fit to increased relative liver weight in female
rats with the highest dose dropped (NTP, 2018) C-48
Figure C-10. User input for dose-response modeling of increased relative liver weight in females
rats with highest dose dropped (NTP, 2018) C-49
Figure C-ll. Model results for increased relative liver weight in female rats with highest dose
dropped (NTP, 2018) C-50
Figure C-12. Dose-response curve for the Exponential 2 model fit to increased relative liver
weight (Histo) in female rats (Frawley et al., 2018) C-53
Figure C-13. User input for dose-response modeling of increased relative liver weight (Histo) in
female rats (Frawley et al., 2018) C-53
Figure C-14. Model results for increased relative liver weight (Histo) in female rats (Frawley et
al., 2018) C-54
Figure C-15. Dose-response curve for the Linear model fit to increased relative liver weight
(MPS) in female rats (Frawley et al., 2018) C-56
Figure C-16. User input for dose-response modeling of increased relative liver weight (MPS) in
female rats (Frawley et al., 2018) C-57
Figure C-17. Model results for increased relative liver weight (MPS) in female rats (Frawley et al.,
2018) C-58
Figure C-18. Dose-response curve for the Exponential 2 model fit to decreased sperm counts in
male rats (NTP, 2018) C-70
Figure C-19. User input for dose-response modeling of decreased sperm counts in male counts
(NTP, 2018) C-70
Figure C-20. Model results for decreased sperm counts in rat males (NTP, 2018) C-71
Figure C-21. Dose-response curve for the Linear model fit to decreased absolute testis weight in
male rats (NTP, 2018) C-73
Figure C-22. User input for dose-response modeling of decreased absolute testis weight in male
rats (NTP, 2018) C-73
Figure C-23. Model results for decreased absolute testis weight in male rats (NTP, 2018) C-74
Figure C-24. Dose-response curve for the Linear model fit to decreased absolute caudal
epididymis weight in male rats (NTP, 2018) C-77
Figure C-25. User Input for dose-response modeling of decreased caudal epididymis weight in
male rats (NTP, 2018) C-77
Figure C-26. Model results for decreased caudal epididymis weight in male rats (NTP, 2018) C-78
Figure C-27. Dose-response curve for the Linear model fit to decreased absolute whole
epididymis weight in male rats (NTP, 2018) C-80
Figure C-28. User input for dose-response modeling of decreased absolute whole epididymis
weight in male rats (NTP, 2018) C-81
Figure C-29. Model Results for decreased absolute whole epididymis weight in male rats (NTP,
2018) C-82
Figure C-30. Dose-response curve for the Polynomial 2 model fit to decreased days in estrus in
female rats (Butenhoff et al., 2012; van Otterdijk, 2007) C-84
Figure C-31. User input for dose-response modeling of decreased days in estrus in female rats
(NTP, 2018) C-84
This document is a draft for review purposes only and does not constitute Agency policy.
ix DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Figure C-32. Model results for decreased days in estrus in female rats (NTP, 2018) C-85
Figure C-33. Dose-response curve for the Exponential 2 model fit to increased days in diestrus in
female rats (Butenhoff et al., 2012; van Otterdijk, 2007) C-88
Figure C-34. User input for dose-response modeling of increased days in diestrus in female rats
(NTP, 2018) C-88
Figure C-35. Model results for increased days in diestrus in female rats (NTP, 2018) C-89
Figure D-l. This proposed MOA is based on previous analyses on PFAS-induced
(e.g., PFOA/PFOS) liver toxicity and the role of nuclear receptor pathways in
hepatotoxicity D-3
Figure E-l. Bioactivity data for PFDA from in vitro HTS ToxCast/Tox21 assays conducted in
human liver cell lines (HepG2 and HepaRG cells) E-3
Figure E-2. Analysis of PFDA-induced upregulation of transcriptional activity in ToxCast/Tox21
assays conducted in human liver cell lines (HepG2 and HepaRG cells) E-4
Figure E-3. Analysis of PFDA-induced nuclear receptor-related activities in ToxCast/Tox21 assays
across multiple endpoints and cell types E-5
Figure G-l. Prior predictive check to ensure equal-tailed interval from prior distributions
encompass the available time-course concentration data for fitting G-5
Figure G-2. Prior sensitivity on half-life, steady-state volume of distribution, and clearance to
ensure weakly informed priors do not bias posterior distributions of the
pharmacokinetic parameters G-6
Figure G-3. Predicted (black line with blue 90% credible interval) and observed (black circles)
serum time-courses for female (left) and male (right) rats after a 25 mg/kg IV
bolus of PFDA. Observed data from (Ohmori et al., 2003) G-7
Figure G-4. Predicted (black line with blue 90% credible interval) and observed (black circles)
serum time-courses for female (top 2 panels) and male (bottom 2 panels) rats
after a 1 mg/kg gavage or IV bolus of PFDA. Gavage exposures are on the left,
while IV exposures are on the left, while IV exposures are on the right. Observed
data from (Kim et al., 2019) G-8
Figure G-5. Predicted (black line with blue 90% credible interval) and observed (black circles)
serum time-courses for female rats after a 2 mg/kg IV or 2, 10, or 20 mg/kg
gavage bolus of PFDA. Observed data from (Dzierlenga et al., 2019) G-9
Figure G-6. Predicted (black line with blue 90% credible interval) and observed (black circles)
serum time-courses for male rats after a 2 mg/kg IV or 2, 10, or 20 mg/kg
gavage bolus of PFDA. Observed data from (Dzierlenga et al., 2019) G-10
Figure G-7. Male rat body weight changes during 28-day PFDA bioassay (NTP, 2018). Data sets
are identified by the dose (mg/kg-d) G-12
Figure G-8. Predicted accumulation and observed end-of-study of PFDA in male rats in the NTP
bioassay (NTP, 2018) as a function of dose. Predicted and measured
concentrations (mg/L) were normalized to respective doses (mg/kg-d) G-12
Figure G-9. Measured end-of-study of PFDA in male rats in the NTP bioassay (NTP, 2018) as a
function of dose G-13
This document is a draft for review purposes only and does not constitute Agency policy.
x DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
ABBREVIATIONS AND ACRONYMS
AC50 activity concentration at 50% HAP
ADME absorption, distribution, metabolism, HAWC
and excretion
AIC Akaike's information criterion Hb/g-A
ALT alanine aminotransferase Hb/g-H
AOP adverse outcome pathway HBCD
AST aspartate aminotransferase HEC
atm atmosphere HED
ATSDR Agency for Toxic Substances and HERO
Disease Registry
BMC benchmark concentration i.p.
BMCL benchmark concentration lower i.v.
confidence limit IAP
BMD benchmark dose IARC
B MD L benchmark dose lower confidence limit
BMDS Benchmark Dose Software IRIS
BMR benchmark response IUR
BUN blood urea nitrogen LCso
BW body weight LDso
BW3/4 body weight scaling to the 3/4 power LOAEL
CA chromosomal aberration LOEL
CAA Clean Air Act MeSH
CAS Chemical Abstracts Service MN
CASRN Chemical Abstracts Service registry MNPCE
number
CERCLA Comprehensive Environmental MOA
Response, Compensation, and Liability MTD
Act NCI
CHO Chinese hamster ovary (cell line cells) NMD
CI confidence interval NOAEL
CL confidence limit NOEL
CNS central nervous system NTP
COI conflict of interest NZW
CPAD Chemical and Pollutant Assessment OAR
Division OECD
CPHEA Center for Public Health and
Environmental Assessment OLEM
CYP450 cytochrome P450
DAF dosimetric adjustment factor ORD
DMSO dimethylsulfoxide OSF
DNA deoxyribonucleic acid PBPK
EPA Environmental Protection Agency PECO
ER extra risk
FDA Food and Drug Administration PK
FEVi forced expiratory volume of 1 second PND
GD gestation day POD
GDH glutamate dehydrogenase POD[adj]
GGT y-glutamyl transferase QSAR
GLP Good Laboratory Practice
GSH glutathione RD
GST glutathione-^"-transferase RfC
hazardous air pollutant
Health Assessment Workspace
Collaborative
animal blood: gas partition coefficient
human blood: gas partition coefficient
hexabromocyclododecane
human equivalent concentration
human equivalent dose
Health and Environmental Research
Online
intraperitoneal
intravenous
IRIS Assessment Plan
International Agency for Research on
Cancer
Integrated Risk Information System
inhalation unit risk
median lethal concentration
median lethal dose
lowest-observed-adverse-effect level
lowest-observed-effect level
Medical Subject Headings
micronuclei
micronucleated polychromatic
erythrocyte
mode of action
maximum tolerated dose
National Cancer Institute
normalized mean difference
no-observed-adverse-effect level
no-observed-effect level
National Toxicology Program
New Zealand White (rabbit breed)
Office of Air and Radiation
Organization for Economic
Co-operation and Development
Office of Land and Emergency
Management
Office of Research and Development
oral slope factor
physiologically based pharmacokinetic
populations, exposures, comparators,
and outcomes
pharmacokinetic
postnatal day
point of departure
duration-adjusted POD
quantitative structure-activity
relationship
relative deviation
inhalation reference concentration
This document is a draft for review purposes only and does not constitute Agency policy.
xi DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
RfD oral reference dose
RGDR regional gas dose ratio
RNA ribonucleic acid
ROBINS I Risk of Bias in Nonrandomized Studies
of Interventions
SAR structure-activity relationship
SCE sister chromatid exchange
SD standard deviation
SDH sorbitol dehydrogenase
SE standard error
SGOT serum glutamic oxaloacetic
transaminase, also known as AST
SGPT serum glutamic pyruvic transaminase,
also known as ALT
TK toxicokinetics
TSCATS Toxic Substances Control Act Test
Submissions
TWA time-weighted average
UF uncertainty factor
UFa animal-to-human uncertainty factor
UFd database deficiencies uncertainty factor
UFh human variation uncertainty factor
UFl LOAEL-to-NOAEL uncertainty factor
UFs subchronic-to-chronic uncertainty
factor
WOS Web of Science
This document is a draft for review purposes only and does not constitute Agency policy.
xii DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
APPENDIX A. SYSTEMATIC REVIEW PROTOCOL FOR
THE PFAS IRIS ASSESSMENTS
1 A single systematic review protocol was used to guide the development of five separate IRIS
2 PFAS [per- and polyfluoroalkyl substances] assessments (i.e., perfluorobutanoic acid [PFBA],
3 perfluorohexanoic acid [PFHxA], perfluorohexane sulfonate [PFHxS], perfluorononanoic acid
4 [PFNA], and perfluorodecanoic acid [PFDA]). This "Systematic Review Protocol for the PFAS IRIS
5 Assessments" was released for public comment and subsequently updated. The updated protocol
6 and prior revisions can be found at the following location:
7 http://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=345065
This document is a draft for review purposes only and does not constitute Agency policy.
A-l DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
APPENDIX B. LITERATURE SEARCH STRATEGY AND
POPULATIONS, EXPOSURES, COMPARATORS,
AND OUTCOMES (PECO) CRITERIA
B.l. LITERATURE SEARCH AND SCREENING STRATEGY
Table B-l. Summary of detailed search strategies for Perfluorodecanoic Acid
and Related Salts (PubMed, Web of Science, Toxline, TSCATS, Toxcenter)
Search
Search strategy
Dates of search
PubMed
Search
terms
335-76-2[rn] OR "Ndfda"[tw] OR "Nonadecafluoro-n-decanoic acid"[tw] OR
"Nonadecafluorodecanoic acid"[tw] OR "Perfluoro-n-decanoic acid"[tw] OR
"Perfluorodecanoic acid"[tw] OR
"2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-Decanoic acid"[tw]
OR "Decanoic acid,
2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-"[tw] OR "Decanoic
acid, nonadecafluoro-"[tw] OR "Perfluorodecanoate"[tw] OR "PFDeA"[tw] OR
"PFDcA"[tw] OR ("PFDA"[tw] AND (fluorocarbon*[tw]
OR fluorotelomer*[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] OR PFAS[tw] OR PFOS[tw] OR
PFOA[tw]))
No date
limit—7/26/2017
Literature
update
search
terms
((335-76-2[rn] OR "Ndfda"[tw] OR "Nonadecafluoro-n-decanoic acid"[tw] OR
"Nonadecafluorodecanoic acid"[tw] OR "Perfluoro-n-decanoic acid"[tw] OR
"Perfluorodecanoic acid"[tw] OR
"2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-Decanoic acid"[tw]
OR "Decanoic acid,
2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-"[tw] OR "Decanoic
acid, nonadecafluoro-"[tw] OR "Perfluorodecanoate"[tw] OR "PFDeA"[tw] OR
"PFDcA"[tw] OR ("PFDA"[tw] AND (fluorocarbon*[tw] OR fluorotelomer*[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] OR PFAS[tw] OR PFOS[tw] OR
PFOA[tw])) AND ("2017/08/01"[Date - Publication] :
"2018/03/01"[Date - Publication])
8/1/2017-2/14/2018
This document is a draft for review purposes only and does not constitute Agency policy.
B-l DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Search
Search strategy
Dates of search
Web of Science
Search
terms
TS="PFDeA" ORTS="PFDcA" ORTS="Ndfda" OR
TS="Nonadecafluoro-n-decanoic acid" OR TS="Nonadecafluorodecanoic acid"
OR TS="Perfluoro-n-decanoic acid" OR TS="Perfluorodecanoic acid" OR
TS="2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-Decanoic acid"
OR TS="Decanoic acid,
2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-" ORTS="Decanoic
acid, nonadecafluoro-" OR TS="Perfluorodecanoate" OR (TS=PFDA AND
TS=(fluorocarbon* OR fluorotelomer* OR polyfluoro* OR perfluoro-* OR
perfluoroa* OR perfluorob* OR perfluoroc* OR perfluorod* OR perfluoroe*
OR perfluoroh* OR perfluoron* OR perfluoroo* OR perfluorop* OR
perfluoros* OR perfluorou* OR perfluorinated OR fluorinated)) OR (TS=PFDA
AND TS=(fluorocarbon* OR fluorotelomer* OR polyfluoro* OR perfluoro-* OR
perfluoroa* OR perfluorob* OR perfluoroc* OR perfluorod* OR perfluoroe*
OR perfluoroh* OR perfluoron* OR perfluoroo* OR perfluorop* OR
perfluoros* OR perfluorou* OR perfluorinated OR fluorinated OR PFAS OR
PFOS OR PFOA
No date
limit—7/26/2017
Literature
update
search
terms
TS="PFDeA" ORTS="PFDcA" ORTS="Ndfda" OR
TS="Nonadecafluoro-n-decanoic acid" OR TS="Nonadecafluorodecanoic acid"
OR TS="Perfluoro-n-decanoic acid" OR TS="Perfluorodecanoic acid" OR
TS="2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-Decanoic acid"
OR TS="Decanoic acid,
2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-" ORTS="Decanoic
acid, nonadecafluoro-" OR TS="Perfluorodecanoate" OR (TS=PFDA AND
TS=(fluorocarbon* OR fluorotelomer* OR polyfluoro* OR perfluoro-* OR
perfluoroa* OR perfluorob* OR perfluoroc* OR perfluorod* OR perfluoroe*
OR perfluoroh* OR perfluoron* OR perfluoroo* OR perfluorop* OR
perfluoros* OR perfluorou* OR perfluorinated OR fluorinated)) OR (TS=PFDA
AND TS=(fluorocarbon* OR fluorotelomer* OR polyfluoro* OR perfluoro-* OR
perfluoroa* OR perfluorob* OR perfluoroc* OR perfluorod* OR perfluoroe*
OR perfluoroh* OR perfluoron* OR perfluoroo* OR perfluorop* OR
perfluoros* OR perfluorou* OR perfluorinated OR fluorinated OR PFAS OR
PFOS OR PFOA)) AND PY=2017-2018
2017-2018
Toxline
Search
terms
(335-76-2 [rn] OR
"2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluorodecanoic acid" OR
"2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-decanoic acid" OR
"decanoic acid 2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-" OR
"decanoic acid nonadecafluoro-" OR "nonadecafluoro-n-decanoic acid" OR
"nonadecafluorodecanoic acid" OR "perfluoro-l-nonanecarboxylic acid" OR
"perfluoro-n-decanoic acid" OR "perfluorocapric acid" OR
"perfluorodecanoate" OR "perfluorodecanoic acid" OR "ndfda" OR "PFDeA"
OR "PFDcA" OR ( pfda AND (fluorocarbon* OR fluorotelomer* OR polyfluoro*
OR perfluoro* OR perfluorinated OR fluorinated OR pfas OR pfos OR pfoa )))
AND ( ANEUPL [org] OR BIOSIS [org] OR CIS [org] OR DART [org] OR EMIC [org]
OR EPIDEM [org] OR HEEP [org] OR HMTC [org] OR IPA [org] OR RISKUNE [org]
OR MTGABS [org] OR NIOSH [org] OR NTIS [org] OR PESTAB [org] OR PPBIB
[org]) AND NOT PubMed [org] AND NOT pubdart [org]
No date
limit—7/21/2017
This document is a draft for review purposes only and does not constitute Agency policy.
B-2 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Search
Search strategy
Dates of search
Literature
update
search
terms
2017-2018
TSCATS
Search
terms
335-76-2[rn] AND TSCATS [org]
No date
limit—7/21/2017
This document is a draft for review purposes only and does not constitute Agency policy.
B-3 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
APPENDIX C. BENCHMARK DOSE MODELING
RESULTS
C.l. BENCHMARK DOSE MODELING RESOLTS FROM HUMAN STUDIES
The endpoints selected for benchmark dose (BMD) modeling include decreased serum
antibody concentrations fBudtz-largensen and Grandiean. 2018a: Grandiean etal.. 20121 and
decreased birth weight fLuo etal.. 2021: Yao etal.. 2021: Wikstrom etal.. 2020: Valvi etal.. 2017:
Lenters etal.. 20161. The internal doses reported in the human studies were used in the BMD
modeling and then converted to human equivalent doses (HEDs) using the estimated human
clearance as described in Section 3.7 of the main document, the modeling results are presented in
this appendix.
C.l.l. BENCHMARK DOSE MODELING APPROACHES FOR IMMUNE EFFECTS
Modeling Results for Decreased Tetanus Antibody Concentrations at 7 Years of Age and PFDA
Measured at 5 Years of Age
Budtz-l0rgensen and Grandiean (2018a) fit multivariate models of PFDA measured at age
five years, against log2-transformed anti-tetanus antibody concentrations measured at the 7-year-
old examination controlling for sex, exact age at the 7-year-old examination, and booster type at age
5 years. Models were evaluated with additional control for PFOS (as log2[PFOS]) and PFOA (as
log2[PFOA]), and without PFOS and PFOA. Three model shapes were evaluated by Budtz-largensen
and Grandiean f2018al using likelihood ratio tests: a linear model, a piecewise-linear model with a
knot at the median PFDA concentration, and a logarithmic function. The logarithmic functions did
not fit better than the piecewise-linear functions (Budtz-l0rgensen and Grandiean. 2018a). The
piecewise-linear model did not fit better than the linear model for the PFHxS exposure without
adjustment for PFOS and PFOA using a likelihood ratio test (p = 0.51; see Budtz-largensen and
Grandiean f2018al Table 3), or for the model that did adjust for PFOS and PFOA (log2[PFOS] and
log2[PFOA]) (p = 0.40).
Table C-l summarizes the results from Budtz-J0rgensen and Grandiean (2018a) for PFDA
at age 5 years and tetanus antibodies at age 7 years. These regression coefficients ([3], their
standard errors (SE), p-values, and the 90% lower confidence bounds were provided by Budtz-
J0rgensen and Grandiean (2018b).
Table C-l. Results specific to the slope from the linear analyses of PFDA
measured in serum at age 5 years and log2(tetanus antibody concentrations)
This document is a draft for review purposes only and does not constitute Agency policy.
C-l DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
measured at age 7 years in a single-PFAS model and in a multi-PFAS model
from fBudtz-lorgensen and Grandiean. 2018bl.
Exposure
Model
shape
PFOS &
PFOA
adjusted
Slope (P)
per ng/mL
in serum
SE(P)
ng/m Lin
serum
Slope (P) fit
Lower bound
slope (pLB) per
ng/mL in serum
PFDA at Age 5
Linear
No
-1.55
0.602
p = 0.01
-2.55
PFDA at Age 5
Linear
Yes
-0.98
0.681
p = 0.15
-2.10
Interpretation of results in Table C-l:
• PFDA is a significant predictor in the single-PFAS model ((3 = -1.55; p = 0.01)
• Effects of PFDA in the single-PFAS model are attenuated when log2[PFOS] and log2[PFOA]
are included in the model ((3 = -0.98; p = 0.15).
• The point estimate results for PFDA ((3) in the single-PFAS model are potentially confounded
by PFOS and/or PFOA since there was a 37% reduction in the effect size for PFDA from -
1.55 to -0.98 when controlling for PFOS and PFOA.
o One explanation is that PFOS and/or PFOA was a confounder of the PFDA effect and
controlling for those co-exposures removed confounding,
o Another possibility is that controlling for co-exposures like PFOS and PFOA actually
induced confounding (Weisskopf et al.. 2018; Weisskopf and Webster. 2017).
o The reasons for the change in main effect size for PFDA are not known. For this
reason, there is uncertainty in knowing which point estimate is the best
representation of any effect of PFDA.
• However, the lower bound on the point estimates ((3lb) for the single-PFAS is 21% lower
than the multi-PFAS model estimate for PFDA.
o The definition of the RfD, which is based upon the (3lb, includes allowing for an order
of magnitude (10-fold or 1,000%) uncertainty in the estimate and the uncertainty
for potential confounding in the BMD from including, or excluding, PFOS and PFOA
here is about 37%, while the uncertainty for potential confounding in the BMDL is
about 21%.
Selection of the Benchmark Response
The benchmark dose (BMD) approach involves dose-response modeling to obtain BMDs,
i.e., dose levels corresponding to specific response levels near the low end of the observable range
of the data and the lower limit of the BMD (BMDLs) to serve as potential PODs for deriving
quantitative estimates below the range of observation (U.S. EPA. 2012). Selecting a BMR to
estimate the BMDs and BMDLs involves making judgments about the statistical and biological
characteristics of the data set and about the applications for which the resulting BMDs and BMDLs
will be used. An extra risk of 10% is recommended as a standard reporting level for quantal data
for toxicological data. Biological considerations may warrant the use of a BMR of 5% or lower for
some types of effects as the basis of the POD for a reference value. However, a BMR of 1% has
typically been used for quantal human data from epidemiology studies (U.S. EPA. 2012). although
This document is a draft for review purposes only and does not constitute Agency policy.
C-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
this is more typically used for epidemiologic studies of cancer mortality within large cohorts of
workers which can support the statistical estimation of small BMRs.
A blood concentration for tetanus antibodies of 0.1 IU/mL is sometimes cited in the tetanus
literature as a 'protective level' and fGrandiean etal.. 20171 noted that the Danish vaccine producer
Statens Serum Institut recommended the 0.1 IU/mL "cutoff" level "to determine whether antibody
concentrations could be considered protective"; and Galazka and Kardymowicz fl9891mentions the
same concentration, but Galazka et al. f!9931argues:
"The amount of circulating antitoxin needed to ensure complete immunity against
tetanus is not known for certain. Establishment of a fixed level of tetanus antitoxin
does not take into consideration variable conditions of production and adsorption of
tetanus toxin in the anaerobic area of a wound or a necrotic umbilical stump. A given
serum level could be overwhelmed by a sufficiently large dose of toxin. Therefore, there
is no absolute protective level of antitoxin and protection results when there is
sufficient toxin-neutralizing antibody in relation to the toxin load fPassen and
Andersen. 19861."
In the absence of a clear definition of an adverse effect for a continuous endpoint like
antibody concentrations, a default BMR of one SD change from the control mean may be selected, as
suggested in EPA's draft Benchmark Dose Technical Guidance Document fU.S. EPA. 20121. As noted
above, a lower BMR can also be used if it can be justified on a biological and/or statistical basis.
Figure C-l replicates a figure in the Technical Guidance (page 23; fU.S. EPA. 20121 to show that in a
control population where 1.4% are considered to be at risk of having an adverse effect, a downward
shift in the control mean of one SD results in a ~10% extra risk of being at risk of having an adverse
effect.
This document is a draft for review purposes only and does not constitute Agency policy.
C-3 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Standard deviation units
Figure C-l. Difference in population tail probabilities resulting from a one
standard deviation shift in the mean from a standard normal distribution,
illustrating the theoretical basis for a baseline BMR of 1 SD.
Statistically, the Technical Guidance additionally suggests that studies of developmental
effects can support lower BMRs. Biologically, a BMR of Vi SD is a reasonable choice as anti-tetanus
antibody concentrations prevent against tetanus, which is a rare, but severe and sometimes fatal
infection, with a case-fatality rate in the U.S. of 13% during 2001-2008 (Liang etal.. 20181. The
case-fatality rate can be more than 80% for early lifestage cases fPatel and Mehta. 19991. Selgrade
f20071 suggests that specific immuno-toxic effects observed in children may be broadly indicative
of developmental immunosuppression impacting these children's ability to protect against a range
of immune hazards—which has the potential to be a more adverse effect than just a single immuno-
toxic effect. Thus, decrements in the ability to maintain effective levels of tetanus antitoxins
following immunization may be indicative of wider immunosuppression in these children exposed
to PFDA. By contrast, a BMR of one SD may be more appropriate for an effect that would be
considered 'minimally adverse.' A BMR smaller than Vi SD is generally selected for severe effects
(e.g., 1% extra risk of cancer mortality); decreased antibody concentrations offer diminished
protection from severe effects but are not themselves severe effects.
Following the technical guidance (U.S. EPA. 20121. EPA derived BMDs and BMDLs
associated with a one SD change in the distribution of log2(tetanus antibody concentrations), and Vi
SD change in the distribution of log2(tetanus antibody concentrations). The SD of the log2(tetanus
This document is a draft for review purposes only and does not constitute Agency policy.
C-4 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
antibody concentrations) at age 7 years was estimated from the distributional data presented in
Grandiean etal. (20121 as follows: the interquartile range (IQR) of the tetanus antibody
concentrations at age 7 years in IU/mL was (0.65, 4.6). Log2-tranforming these values provides the
IQRinlog2(IU/mL) as (-0.62, 2.20). Assuming that these log2-transformed values are reasonably
represented by a normal distribution, the width of the IQR is approximately 1.35 SDs. Thus, SD =
IQR/1.35, and the SD of tetanus antibodies inlog2(IU/mL) is (2.20 - (-0.62))/1.35 = 2.09
log2(IU/mL). To show the impact of the BMR on these results, Table E-2 presents the BMDs and
BMDLs at BMRs of % SD and 1 SD.
While there was not a clear definition of the size of an adverse effect for a continuous
endpointlike antibody concentrations, the value of 0.1 IU/mL is sometimes cited. As a check, EPA
evaluated how much extra risk would have been associated with a BMR set at a cutoff value of 0.1
IU/mL. Using the observed distribution of tetanus antibodies at age seven years in log2(IU/mL),
EPA calculated that 2.8% of those values would be below the cutoff value of 0.1 IU/mL which is -
3.32 log2(IU/mL). A BMR of xh SD resulted in 7.9% of the values being below that cutoff which is
5.1% extra risk and shows that the generic guidance that a BMR of Vi SD can provide a reasonably
good estimate of 5% extra risk. Figure C-2 shows an example of this.
This document is a draft for review purposes only and does not constitute Agency policy.
C-5 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Tetanus antibody concentrations in Log2(IU/ml)
Figure C-2. Difference in population tail probabilities resulting from a V2
standard deviation shift in the mean from an estimation of the distribution of
log2(tetanus antibody concentrations at age seven years).
Table C-2. BMDs and BMDLs for effect of PFDA at age five years on anti-tetanus
antibody concentrations at age seven years using a BMR of V2 SD change in
log2(tetanus antibodies concentration) and a BMR of 1 SD change in
log2(tetanus antibodies concentration).
Estimated without control of PFOS and PFOA
Estimated with control of PFOS and PFOA
BMR
BMD (ng/mL in serum)
P = -1.55 per ng/mL
BMDL (ng/mL in serum)
Plb = -2.55 per ng/mL
BMD (ng/mL in serum)
P = -0.98 per ng/mL
BMDL (ng/mL in serum)
Plb = -2.10 per ng/mL
% SD
0.673
0.411a
1.067
0.497
1 SD
1.346
0.821
2.135
0.994
a Denotes the selected POD.
1 The lowest serum PFDA concentration measured at age five years was 0.05 ng/mL, the 5th
2 percentile was 0.1 ng/mL, and the 10th percentile was 0.2 ng/mL fGrandiean and Bateson. 20211 so
3 the estimated BMDL for a BMR of xh SD (BMDL% sd) in the single-PFAS model is above the 10th
This document is a draft for review purposes only and does not constitute Agency policy.
C-6 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
percentile of the observed distribution. No information was available to judge the fit of the model
in the range of the BMDLs, but the BMD and BMDL were both within the range of observed values
and the model fit PFDA well.
The BMDy2 sd estimate from the multi-PFAS models is 59% higher than the BMDy2 sd estimate
from the models with just PFDA, and the BMDLy2 sd estimates is 21% higher. The change in BMD
estimates may, or may not, reflect control for any potential confounding of the regression effect
estimates. While it is not clear which PFAS model provided 'better' estimate of the point estimate of
the effect of PFDA, the two BMDLy2sD estimates are similar (0.411 ng/mL vs. 0.497 ng/mL) and EPA
advanced the derivation based on results that did not controls for PFOS and PFOA because this
model appeared to fit PFDA better (p = 0.01 vs. 0.15) and there was low uncertainty due to
potential confounding in the BMDL. However, confidence was somewhat diminished by the
potential confounding in the main effect—even though there was low confounding of the BMDL.
Overall confidence in the BMDLs for Tetanus was judged to be medium confidence.
For immunotoxicity related to tetanus associated with PFDA exposure measured at
age five years, the POD is based on a BMR of Vi SD and a BMDL%Sd of 0.411 ng/mL in
serum.
Modeling Results for Decreased Diphtheria Antibody Concentrations at 7 Years of Age and
PFDA Measured at 5 Years of Age
Budtz-J0rgensen and Grandjean (2018a) fit multivariate models of PFDA measured at age
5 years, against log2-transformed anti-diphtheria antibody concentrations measured at the seven-
year-old examination controlling for sex, exact age at the 7-year-old examination, and booster type
at age 5 years. Models were evaluated with additional control for PFOS (as log2[PFOS]) and PFOA
(as log2[PFOA]), and without PFOS and PFOA. Three model shapes were evaluated by Budtz-
J0rgensen and Grandjean (2018a) using likelihood ratio tests: a linear model of PFDA, a piecewise-
linear model with a knot at the median, and a logarithmic function. The logarithmic functions did
not fit better than the piecewise-linear functions (Budtz-J0rgensen and Grandjean. 2018a). The
piecewise-linear model did not fit better than the linear model for the PFHxS exposure without
adjustment for PFOS and PFOA using a likelihood ratio test (p = 0.55; see Budtz-J0rgensen and
Grandjean (2018a) Table 3), or for the model that did adjust for PFOS and PFOA (log2[PFOS] and
log2[PFOA]) (p = 0.73). Table C-3 summarizes the results from Budtz-J0rgensen and Grandjean
(2018a) for diphtheria in this exposure window. These regression coefficients ((3), their standard
errors (SE), p-values, and the 90% lower confidence bounds were provided by Budtz-J0rgensen
and Grandjean (2018b).
This document is a draft for review purposes only and does not constitute Agency policy.
C-7 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-3. Results specific to the slope from the linear analyses of PFDA in
serum measured at age 5 years and log2 (diphtheria antibodies) measured at
age 7 years from Table 1 in a single-PFAS model and in a multi-PFAS model
from (Budtz-lorgensen and Grandjean. 2018b).
Exposure
Model shape
PFOS & PFOA
adjusted
Slope (P) per
ng/mL in
serum
SE(P)
ng/mL in
serum
Slope (P) fit
Lower bound
slope (pLB)
per
ng/mL in
serum
PFDA at Age 5
Linear
No
-0.894
0.561
p = 0.11
-1.82
PFDA at Age 5
Linear
Yes
-0.297
0.635
p = 0.64
-1.35
Interpretation of results in Table C-3:
• PFDA is a non-significant predictor in the single-PFAS model ((3 = -0.894; p = 0.11)
• Effects are attenuated when log2[PF0S] and log2[PF0A] are included in the model ((3 = -
0.297; p = 0.64).
• The point estimate results for PFDA are potentially confounded by PFOS and/or PFOA since
there was a 67% reduction in the effect size for PFDA from -0.894 to -0.297 when
controlling for PFOS and PFOA.
• One explanation is that PFOS and/or PFOA was a confounder of the PFDA effect and
controlling for those co-exposures removed confounding.
• Another possibility is that controlling for co-exposures like PFOS and PFOA actually
induced confounding (Weisskopf et al.. 2018; Weisskopf and Webster. 2017).
• The reasons for the change in main effect size for PFDA are not known. For this
reason, there is uncertainty in knowing which point estimate is the best
representation of any effect of PFDA.
• However, the lower bound on the point estimates ((3lb) for the single-PFAS model is 35%
lower than the multi-PFAS model estimate for PFDA.
o The definition of the RfD, which is based upon the (3lb, includes allowing for an order
of magnitude (10-fold or 1,000%) uncertainty in the estimate and the uncertainty for
potential confounding in the BMD from including, or excluding, PFOS and PFOA here
is about 67%, while the uncertainty for potential confounding in the BMDL is about
35%.
Selection of the Benchmark Response
Following the technical guidance fU.S. EPA. 20121. EPA derived BMDs and BMDLs
associated with a one SD change in the distribution of log2(diphtheria antibody concentrations),
and Vi SD change in the distribution of log2(diphtheria antibody concentrations). A blood
concentration for diphtheria antibodies of 0.1 IU/mL is sometimes cited in the diphtheria literature
as a 'protective level' Grandiean et al. f20171 noted that the Danish vaccine producer Statens Serum
Institut recommended the 0.1IU/mL 'cutoff level; and Galazka et al. T19931 mentions the same
concentration), but Galazka et al. T19931 argues:
"However¦, it has also been shown that there is no sharply defined level of antitoxin that gives
complete protection from diphtheria (Ipsen, 1946). A certain range of variation must be
This document is a draft for review purposes only and does not constitute Agency policy.
C-8 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
accepted; the same degree of antitoxin may give an unequal degree of protection in different
persons. Other factors may influence the vulnerability to diphtheria including the dose and
virulence of the diphtheria bacilli and the general immune status of the person infected
fChristenson and Bottiger. 1986]. Thus, an antibody concentration between 0.01 and 0.09
IU/ml may be regarded as giving basic immunitywhereas a higher titer may be needed for full
protection. In some studies that used in vitro techniques; a level of 0.1 IU/ml was considered
protective fCellesi et al.. 1989; Galazka and Kardymowicz. 1989)."
Statistically, the Technical Guidance suggests that studies of developmental effects can
support lower BMRs. Biologically, a BMR of Vi SD is a reasonable choice as anti-diphtheria antibody
concentrations prevent against diphtheria, which is very rare in the U.S., but can cause life-
threatening airway obstruction, or cardiac failure fCollier. 19751. Among 13 cases reported in the
U.S. during 1996-2016, no deaths were mentioned fLiang etal.. 20181. However, diphtheria
remains a potentially fatal disease in other parts of the world (Galazka et al.. 19931 mentions a case
fatality rate of 5-10%) and PFDA-related changes in anti-diphtheria antibody concentrations
cannot be considered 'minimally adverse' given the historic lethality of diphtheria in the absence of
vaccination. Selgrade (20071 suggests that specific immuno-toxic effects observed in children may
be broadly indicative of developmental immunosuppression impacting these children's ability to
protect against a range of immune hazards—which has the potential to be a more adverse effect
that just a single immuno-toxic effect.
Following the technical guidance (U.S. EPA. 20121. EPA derived BMDs and BMDLs
associated with a one SD change in the distribution of log2(diphtheria antibody concentrations) as a
standard reporting level, and Vi SD change in the distribution of log2(diphtheria antibody
concentrations). The SD of the log2(diphtheria antibody concentrations) at age 7 years was
estimated from the distributional data presented in Grandiean etal. f20121 as follows: the
interquartile range (IQR) of the diphtheria antibody concentrations at age 7 years in IU/mL was
(0.4,1.6). Log2-tranforming these values provides the IQR inlog2(IU/mL) as (-1.32, 0.68).
Assuming that these log2-transformed values are similar to the normal distribution, the width of the
IQR is approximately 1.35 SDs, thus SD = IQR/1.35, and the SD of tetanus antibodies in log2(IU/mL)
is (0.68 - (-1.32))/1.35 = 1.48 log2(IU/mL). To showthe impactofthe BMR on these results, Table
E-4 presents the BMDs and BMDLs at BMRs of xh SD and 1 SD.
This document is a draft for review purposes only and does not constitute Agency policy.
C-9 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-4. BMDs and BMDLs for effect of PFDA at age 5 years on anti-
diphtheria antibody concentrations at age 7 years using a BMR of SD change
in log2(diphtheria antibodies concentration) and a BMR of 1 SD
log2(diphtheria antibodies concentration).
Estimated without control of PFOS and PFOA
Estimated with control of PFOS and PFOA
BMR
BMD (ng/mL in serum)
P = -0.894 per ng/mL
BMDL (ng/mL in serum)
Plb = -1.82 per ng/mL
BMD (ng/mL in serum)
P = -0.297 per ng/mL
BMDL (ng/mL in serum)
Plb = -1.35 per ng/mL
% SD
0.827
0.407a
2.488
0.550
1 SD
1.655
0.813
4.976
1.100
a Denotes the selected POD.
The lowest serum PFDA concentration measured at age five years was 0.05 ng/mL, the 5th
percentile was 0.1 ng/mL, and the 10th percentile was 0.2 ng/mL (Grandiean and Bateson. 20211 so
the estimated BMDL for a BMR of xh SD (BMDLy2sD) in the single-PFAS model is at the 10th
percentile of the observed distribution. No information was available to judge the fit of the model
in the range of the BMDLs, but the BMD and BMDL were both within the range of observed values
and the model fit PFDA well.
The BMDy2 sd estimate from the multi-PFAS models is 3-fold higher than the BMDy2 sd
estimate from the model with just PFDA, and the BMDLy2 sd is 35% higher. This may, or may not,
reflect control for any potential confounding of the regression effect estimates. While it is not clear
which PFAS model provided the 'better' estimate of the point estimate of the effect of PFDA, the two
BMDLy2sD estimates which serve as the PODs are comparable (0.407 ng/mL vs. 0.550 ng/mL) and
EPA advanced POD based on results that did not controls for PFOS and PFOA because this model
appeared to fit PFDA better (p = 0.11 vs. 0.64) and there was low uncertainty due to potential
confounding in the BMDL. However, confidence was diminished by the non-significant fit for PFDA
(p = 0.11) and stronger potential confounding in the main effect—even though there was low
confounding of the BMDL, and overall confidence in the BMDLs for diphtheria was judged to be low
confidence.
For immunotoxicity related to diphtheria, associated with PFDA measured at age 5
years, the POD is based on a BMR of Vi SD and a BMDLy2 Sd of 0.407 ng/mL in serum.
Modeling Results for Decreased Tetanus Antibody Concentrations at 5 Years of Age and
perinatal PFDA
Budtz-l0rgensen and Grandiean (2018a) fit multivariate models of PFDA measured
perinatally in maternal serum, against log2-transformed anti-tetanus antibody concentrations
measured at the 5-year-old examination controlling for sex, and exact age at the 5-year-old
examination, cohort, and interaction terms between cohort and sex, and between cohort and age.
Models were evaluated with additional control for PFOS (as log2[PFOS]) and PFOA (as log2[PFOA]),
and without PFOS and PFOA. Three model shapes of PFDA were evaluated by Budtz-l0rgensen and
This document is a draft for review purposes only and does not constitute Agency policy.
C-10 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Grandiean (2018a) using likelihood ratio tests: a linear model, a piecewise-linear model with a knot
at the median, and a logarithmic function. The logarithmic functions did not fit better than the
piecewise-linear functions Budtz-largensen and Grandiean f2018al Compared to the linear model,
the piecewise-linear model did not fit better than the linear model for either the PFDA exposure
without adjustment for PFOS and PFOA using a likelihood ratio test (p = 0.81; see Budtz-largensen
and Grandiean f2018al Table 3), or for the model that did adjust for PFOS and PFOA (log2[PFOS]
and log2[PFOA]) (p = 0.84).
Table C-5 summarizes the results from Budtz-J0rgensen and Grandiean (2018a) for
tetanus in this exposure window. These regression coefficients ((3), their standard errors (SE), p-
values, and the 90% lower confidence bounds were provided by Budtz-J0rgensen and Grandiean
(2018b).
Table C-5. Results of the linear analyses of PFDA measured perinatally in
maternal serum and tetanus antibodies measured at age 5 years in a single-
PFAS model and in a multi-PFAS model from fBudtz-lorgensen and Grandiean.
2018b)-
Exposure
Model shape
PFOS & PFOA
adjusted
Slope (P) per
ng/mL in
serum
SE(P)
ng/mL in
serum
Slope (P) fit
Lower bound
slope (Plb) per
ng/mL in
serum
Perinatal PFDA
Linear
No
-0.343
0.462
p = 0.46
-1.103
Perinatal PFDA
Linear
Yes
0.038
0.554
p = 0.95
-0.874
Interpretation of results in Table C-5:
• PFDA is a non-significant predictor in the single-PFAS model ((3 = -0.34; p = 0.46).
• Effects are attenuated when log2[PFOS] and log2[PFOA] are included in the model ((3 =
0.038; p = 0.55)
• Nevertheless, these data can be used to estimate a BMDL for completeness and to allow
comparisons across PFAS.
Selection of the Benchmark Response
Following the technical guidance fU.S. EPA. 20121. EPA derived BMDs and BMDLs
associated with a one SD change in the distribution of log2(tetanus antibody concentrations), and Vi
SD change in the distribution of log2(tetanus antibody concentrations). The SD of the log2(tetanus
antibody concentrations) at age 5 years was estimated from two sets of distributional data
presented from two different cohorts of 5-year-olds that were pooled in Budtz-l0rgensen and
Grandiean (2018a). Grandiean etal. (2012) reported on 587 5-year-olds from the cohort of
children born during 1997-2000 and in Grandiean etal. (2017) reported on 349 5-year-olds from
the cohort of children born during 2007-2009. The means and SDs were computed separately and
then pooled to describe the common SD. The IQR of the tetanus antibody concentrations in the
earlier birth cohort at age 5 years in IU/mL was (0.1, 0.51). Log2-tranforming these values provides
This document is a draft for review purposes only and does not constitute Agency policy.
C-ll DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
the IQR in log2(IU/mL) as (-3.32, -0.97). Assuming that these log2-transformed values are similar to
the normal distribution, the width of the IQR is approximately 1.35 SDs, thus SD = IQR/1.35, and the
SD of tetanus antibodies in log2(IU/mL) is (-0.97 - (-3.32))/1.35 = 1.74 log2(IU/mL). The IQR of the
tetanus antibody concentrations in the later birth cohort at age 5 years in IU/mL was (0.1, 0.3).
Log2-tranforming these values provides the IQR in log2(IU/mL) as (-3.32, -1.74), and the SD of
tetanus antibodies in log2(IU/mL) is (-1.74- (-3.32))/1.35 = 1.17 log2(IU/mL). The pooled variance
is a weighted sum of the independent SDs, and the pooled SD was estimated as 1.55 log2(IU/mL).1
To show the impact of the BMR on these results, Table E-6 presents the BMDs and BMDLs at BMRs
of% SD andl SD.
Table C-6. BMDs and BMDLs for effect of PFDA measured perinatally and anti-
tetanus antibody concentrations at age 5 years
Estimated without control of PFOS and PFOA
Estimated with control of PFOS and PFOA
BMR
BMD (ng/mL in serum)
P = -0.343 per ng/mL
BMDL (ng/mL in serum)
Plb = -1.103 per ng/mL
BMD (ng/mL in serum)
P = 0.038 per ng/mL
BMDL (ng/mL in serum)
Plb = -0.874 per ng/mL
% SD
2.260
0.702a
-
0.886
1 SD
4.520
1.405
-
1.773
a Denotes the POD that corresponds to the analyses of PFDA concentrations perinatally and tetanus antibodies at
age 5 years; - values can't be determined.
The lowest perinatal maternal serum PFDA concentration measured was 0.03 ng/mL, the
5th percentile was 0.1 ng/mL, and the 10th% was 0.2 ng/mL (Grandjean, 2021) so the estimated
BMDLs for a BMR of xh SD (BMDLy2 sd =0.702 ng/mL) in the single-PFAS model is well above the
10th% of the observed distribution. No information was available to judge the fit of the model in the
range of the BMDLs, but the BMD and BMDL were both within the range of observed values and the
model fit PFDA well. The BMDLy2 sd estimate from the single-PFAS models was 0.702 ng/mL in
serum. The BMDL estimates from the multi-PFAS models were about 26% higher than for the
single-PFAS model.
Low confidence in the BMDLs from the PFDA-only model (0.702 ng/mL in serum) and in the
multi-PFAS model (0.886 ng/mL in serum). Confidence is diminished by the low quality of the
model fit for PFDA in either model compared to the PFDA results from tetanus in the 5-year to 7-
year exposure-outcome window of time and there is some uncertainty regarding potential
confounding.
For immunotoxicity related to tetanus, associated with PFDA measured perinatally, the POD
is based on a BMR of Vi SD and a BMDLy2sDof 0.702 ng/mL in serum. Note that this result is based
on a poorly fit PFDA regression parameter ((3) estimated as -0.343 per ng/mL in serum (90% CI:
1 Pooled variance for tetanus in 5-year-olds = [(502-l)(1.74)A2+ (298-l)(1.17)A2]/[502+298-2] = 2.41. The pooled
SD is the square root of 2.41 which is 1.55 log2(IU/mL).
This document is a draft for review purposes only and does not constitute Agency policy.
C-12 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
-1.103, 0.417; p = 0.46) Budtz-l0rgensen and Grandiean (2018b). and thus this POD is identified
with low confidence.
For immunotoxicity related to tetanus associated with PFDA exposure measured at
age 5 years, the POD estimated for comparison purposes were based on a BMR of
SD and a BMDL%Sd of 0.702 ng/mL in serum.
Modeling Results for Decreased Diphtheria Antibody Concentrations at 5 Years of Age and
perinatal PFDA
Budtz-l0rgensen and Grandiean f2018al fit multivariate models of PFDA measured
perinatally, against log2-transformed anti-diphtheria antibody concentrations measured at the 5-
year-old examination controlling for sex and age. Models were evaluated with additional control
for PFOS (as log2[PF0S]) and PFOA (as log2[PF0A]), and without PFOS and PFOA. Three model
shapes were evaluated by Budtz-l0rgensen and Grandiean (2018a) using likelihood ratio tests: a
linear model of PFDA, a piecewise-linear model with a knot at the median, and a logarithmic
function. The logarithmic functions did not fit better than the piecewise-linear functions Budtz-
Targensen and Grandiean f2018al. There was evidence that the piecewise-linear model fit better
than the linear model for the PFDA exposure without adjustment for PFOS and PFOA (p = 0.05; see
in Budtz-largensen and Grandiean f2018al. Table 3), but not for the model that adjusted for PFOS
and PFOA (log2[PFOS] and log2[PFOA]) (p = 0.12). Table C-7 summarizes the results from Budtz-
l0rgensen and Grandiean (2018a) for diphtheria in this exposure window. These regression
coefficients ((3) and their standard errors (SE) were computed by EPA from the published BMDs
and BMDL based on a BMR of 5% change in diphtheria antibody concentrations in Table 2 of Budtz-
l0rgensen and Grandiean f2018al2.
2 (Budtz-J0rgensen and Grandiean, 2018a) computed BMDs and BMDLs using a BMR of 5% decrease in the
antibody concentrations. Their formula, BMD = log2(l-BMR)/p, can simply be reversed to solve for p = log2(l-
BMR)/BMD. For negative dose-response where more exposure results in lower antibody concentration, the BMDL
is based on the lower bound of p, (PLB). Thus, the PLB = log2(l-BMR)/BMDL The SE(P) = (P - PLB)/1.645. The p-
value is the two-sided probability that Z <= SE(P)/p.
This document is a draft for review purposes only and does not constitute Agency policy.
C-13 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-7. Results of the analyses of PFDA measured perinatally in maternal
serum and diphtheria antibodies measured at age 5 years in a single-PFAS
model and in a multi-PFAS model from fBudtz-lorgensen and Grandiean.
2018b)-
Exposure
Model shape
PFOS & PFOA
adjusted
Slope (P) per
ng/mL in
serum
SE(P)
Slope (P) fit
Lower bound
slope (Plb) per
ng/mL in
serum
Perinatal PFDA
Piecewise
No
-3.700
2.249
p = 0.100
-7.400
Perinatal PFDA
Piecewise
Yes
-2.467
0.750
p = 0.001
-3.700
Interpretation of results in Table C-7:
• PFDA is a non-significant predictor in the single-PFAS model ((3 = -3.700; p = 0.10)
• Effects of PFDA are attenuated when PFOA and PFOA are in the model ((3 = -2.467; p =
0.001).
• The point estimate results for PFDA are potentially confounded by PFOS and/or PFOA since
there was a 33% change in the effect size for PFDA from -3.700 to -2.467 when controlling
for PFOS and PFOA.
o One explanation is that PFOS and/or PFOA was a confounder of the PFDA effect and
controlling for those co-exposures removed confounding,
o Another possibility is that controlling for co-exposures like PFOS and PFOA actually
induced confounding (Weisskopf et al.. 2018; Weisskopf and Webster. 2017).
o The reasons for the change in main effect size for PFDA are not known. For this
reason, there is uncertainty in knowing which point estimate is the best
representation of any effect of PFDA.
• However, the lower bound on the point estimates ((3lb) for the single-PFAS model for PFDA
is 100% lower than the multi-PFAS model effect estimate for PFDA.
o The definition of the RfD, which is based upon the (3lb, includes allowing for an order
of magnitude (10-fold or 1,000%) uncertainty in the estimate and the uncertainty for
potential confounding in the BMD from including, or excluding, PFOS and PFOA here
is about 33%, while the uncertainty for potential confounding in the BMDL is about
100%.
Selection of the Benchmark Response
Following the technical guidance fU.S. EPA. 20121. EPA derived BMDs and BMDLs
associated with a one SD change in the distribution of log2(tetanus antibody concentrations) as a
standard reporting level, and Vi SD change in the distribution of log2(tetanus antibody
concentrations). The SD of the log2(diphtheria antibody concentrations) at age 5 years was
estimated from two sets of distributional data presented from two different birth cohorts of 5-year-
olds that were pooled in Budtz-l0rgensen and Grandiean (2018a). Grandiean etal. (2012) reported
on 587 5-year-olds from the cohort of children born during 1997-2000 and Grandiean etal. (2017)
reported on 349 5-year-olds from the cohort of children born during 2007-2009. The means and
SDs were computed separately and then pooled to describe the common SD. The IQR of the
diphtheria antibody concentrations in the earlier birth cohort at age 5 years in IU/mL was (0.05,
This document is a draft for review purposes only and does not constitute Agency policy.
C-14 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
0.4). Log2-tranforming these values provides the IQR in log2(IU/mL) as (-4.32, -1.32). Assuming
that these log2-transformed values are similar to the normal distribution, the width of the IQR is
approximately 1.35 SDs, thus SD = IQR/1.35, and the SD of diphtheria antibodies in log2(IU/mL) is
(-1.32 - (-4.32))/1.35 = 2.22 log2(IU/mL). The IQR of the diphtheria antibody concentrations in the
later birth cohort at age 5 years in IU/mL was (0.1, 0.3). Log2-tranforming these values provides
the IQR in log2(IU/mL) as (-3.32, -1.74), and the SD of diphtheria antibodies in log2(IU/mL) is (-1.74
- (-3.32))/1.35 = 1.17 log2(IU/mL). The pooled variance is a weighted sum of the independent SDs,
and the pooled SD was estimated as 1.90 log2(IU/mL)3. To showthe impactofthe BMR on these
results, Table C-8 presents the BMDs and BMDLs atBMRs of xh SD and 1 SD.
Table C-8. BMDs and BMDLs for effect of PFDA measured perinatally and anti-
diphtheria antibody concentrations at age 5 years.
Estimated without control of PFOS and PFOA
Estimated with control of PFOS and PFOA
BMR
BMD (ng/mL in serum)
P = -3.700 per ng/mL
BMDL (ng/mL in serum)
Plb = -7.400 per ng/mL
BMD (ng/mL in serum)
P = -2.467 per ng/mL
BMDL (ng/mL in serum)
Plb = -3.700 per ng/mL
% SD
0.257
0.128
0.385
0.257a
1 SD
0.514
0.257
0.770
0.514
a Denotes the POD that corresponds to the analyses of PFDA concentrations perinatally and diphtheria antibodies
at age 5 years.
The lowest serum PFDA concentration measured perinatally was 0.03 ng/mL, the 5th
percentile was 0.1 ng/mL, and the 10th% was 0.2 ng/mL (Grandiean and Bateson. 2021) so the
estimated BMD for a BMR of Vi SD (BMDLy2 sd) in the single-PFAS model is well within the observed
range. No information was available to judge the fit of the model in the range of the BMDLs, but the
BMD and BMDL were both within the range of observed values and the model fit PFDA well.
The BMD y2 sd estimate from the multi-PFAS models is 50% higher than the BMDy2 sd
estimated from the model with just PFDA, and the BMDLy2 sd is 100% higher. This may, or may not,
reflect control for any potential confounding of the regression effect estimates. The BMDLs which
serve as the PODs are two-fold different (0.128 ng/mL vs. 0.257 ng/mL) and EPA advanced the
derivation based on results that did control for PFOS and PFOA because this model appeared to fit
PFDA well (p = 0.001 vs. 0.10) and there was low uncertainty due to potential confounding in the
BMD and moderate uncertainty in the BMDL. Medium confidence in the BMDLs from PFDA linear
model (0.257 ng/mL in serum) with control of PFOS and PFOA since the model fit reasonably well
and these BMDLs show moderate uncertainty about confounding.
3 Pooled variance for diphtheria in 5-year-olds = [(502-l)(2.22)A2+ (298-l)(1.17)A2]/[502+298-2] = 3.60. The
pooled SD is the square root of 2.41 which is 1.90 log2(IU/mL).
This document is a draft for review purposes only and does not constitute Agency policy.
C-15 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
For immunotoxicity related to diphtheria, associated with PFDA measured at age 5
years, the POD is based on a BMR of SD and a BMDLy2 Sd of 0.257 ng/mL in serum.
Summary of Modeling Results for Decreased Antibody Responses in Children
Table C-9 presents the BMDs and BMDLs from Budtz-largensen and Grandiean f2018al
considered for POD derivation for reduced antibody responses across different combinations of
exposure timing and outcome measurement as detailed above. The BMDLs across the studies and
methods ranged from 0.257-0.702 ng/mL.
Table C-9. Selected BMDs and BMDLs and associated uncertainty for effect of
PFDA on decreased antibody responses in children from Budtz-lorgensen and
Grandiean f2018al
Endpoint
BMDi/2sd (ng/mL)
BMDLi/2sd (ng/mL)
Confidence
Decreased serum tetanus antibody
concentrations at 7 years of age and
PFDA measured at 5 years of age3
0.673
0.411
Medium
Decreased serum diphtheria antibody
concentrations at 7 years of age and
PFDA concentrations at 5 years of age3
0.827
0.407
Low
Decreased serum tetanus antibody
concentrations at 5 years of age and
perinatal PFDA (pregnancy week 32-2
weeks postpartum)3
2.260
0.702
Low
Decreased serum diphtheria antibody
concentrations at 5 years of age and
perinatal PFDA (pregnancy week 32-2
weeks postpartum)15
0.385
0.257
Medium
Estimated without control for PFOA and PFOS.
Estimated with control for PFOA and PFOS.
C.1.2. BENCHMARK DOSE MODELING APPROACHES FOR DEVELOPMENTAL EFFECTS
Modeling Results for Decreased Birth Weight
Five high confidence studies (Luo etal.. 2021: Yao etal.. 2021: Wikstrom etal.. 2020: Valvi
etal.. 2017: Division of Environmental Epidemiology etal.. 20161 reported decreased birth weight
in infants whose mothers were exposed to PFDA. All studies reported their exposure metric in
units of ng/mL and reported the (3 coefficients per ln(ng/mL) or per log2(ng/mL), along with 95%
confidence intervals, estimated from linear regression models. The logarithmic transformation of
exposure yields a negative value for low numbers, which can result in implausible results from
dose-response modeling (i.e., estimated risks are negative and unable to determine the responses at
zero exposure). EPA first re-expressed the reported (3 coefficients in terms of per ng/mL according
to Dzierlenga etal. (20201. Then EPA used the re-expressed (3 and the lower limit on the confidence
This document is a draft for review purposes only and does not constitute Agency policy.
C-16 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
interval to estimate BMD and BMDL values using the general equation y = mx + b, where y is birth
weight and x is exposure, substituting the re-expressed (3 values from these studies for m. The
intercept b represents the baseline value of birth weight in an unexposed population and it can be
estimated through y = mx + b using an average birth weight from an external population as y, an
average exposure as x and re-expressed (3 from the studies as m.
The CDC Wonder site fhttp s: //wo nder. cdc. gov /natality, html! provides vital statistics for
babies born in the United States. There were 3,791,712 live births in the U.S. in 2018 according to
final natality data. The mean and standard deviation for birth weight were 3,261.6 ± 590.7 g
(7.19 ± 1.30 lb), with 8.27% of live births falling below the public health definition of low birth
weight (i.e., <2,500 g, or 5.5 lb). The full natality data for the U.S. data on birth weight was used as it
is more relevant for deriving toxicity values for the U.S. public than the study-specific birth weight
data. Also, the CDC Wonder database may be queried to find the exact percentage of the population
falling below the cut-off value for clinical adversity. The CDC Fourth National Report on Human
Exposure to Environmental Chemicals (https://www.cdc.gov/exposurereport/index.html)
provides the median of serum PFDA concentrations (0.19 ng/mL) among NHANES females in
2011-2012. These values are subsequently used in the estimation of BMD and BMDL values from
the available five epidemiological studies.
fValvi etal.. 20171 reported a (3 coefficient of-41 gper log2(ng/mL) (95%CI: -102,18) for
the association between birth weight and maternal PFDA serum concentrations in a Denmark
cohort The reported (3 coefficient can be re-expressed in terms of per ng/mL according to
(Dzierlenga et al.. 2020). Given the reported study-specific median (0.28 ng/mL) and interquartile
range (IQR) (0.22-0.38 ng/mL) of the exposure from (Valvi etal.. 2017). EPA estimated the
distribution of exposure by assuming the exposure follows a log-normal distribution with mean and
standard deviation as:
Then, EPA estimated the 25th-75th percentiles at 10 percentile intervals of the exposure
distribution and corresponding responses of reported (3 coefficient. The re-expressed (3 coefficient
is determined by minimizing the sum of squared differences between the curves generated by the
re-expressed (3 and the reported p. This resulted in a re-expressed (3 coefficient of-2 07.7 gper
ng/mL (95% CI: -516.8, 91.2 g per ng/mL).
Typically, for continuous data, the preferred definition of the benchmark response (BMR) is
to have a basis for what constitutes a minimal level of change in the endpointthat is biologically
significant For birth weight, there is no accepted percent change that is considered adverse.
However, there is a clinical measure for what constitutes an adverse response: babies born
weighing less than 2,500 g are considered to have low birth weight, and further, low birth weight is
associated with a wide range of health conditions throughout life fTian etal.. 2019: Reyes and
jj. = ln(q50) = ln( 0.28) = —1.27
a = ln(q75/q25)/1.349 = Zn(0.38/0.22)/1.349 = 0.41
(C-l)
(C-2)
This document is a draft for review purposes only and does not constitute Agency policy.
C-17 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Manalich. 2005: Hack etal.. 19951. Given this clinical cut-off for adversity and that 8.27% of live
births in the U.S. in 2018 fell below this cut-off, the hybrid approach can be used to define the BMR.
The hybrid approach is advantageous in that it harmonizes the definition of the BMR for continuous
data with that for dichotomous data.4 Essentially, the hybrid approach involves the estimation of
the dose that increases the percentile of responses falling below (or above) some cut-off for
adversity in the tail of the response distribution. Application of the hybrid approach requires the
selection of an extra risk value for BMD estimation. In the case of birth weight, an extra risk of 5%
is selected given that this level of response is typically used when modeling developmental
responses from animal toxicology studies, and that low birthweight confers increased risk for
adverse health effects throughout life, thus supporting a BMR lower than the standard BMR of 10%
extra risk.
Therefore, given a background response and a BMR = 5% extra risk, the BMD would be the
dose that results in 12.86% of the responses falling below the 2,500 g cut-off value:
Extra Risk(ER) = (P(d) - P(0)) / (1 - P(0)) (C-3)
P(d) = ER(1 - P(0)) + P(0) = 0.05(1 - 0.0827) + 0.0827 = 0.1286 (C-4)
Based on the mean birth weight for all births in the United States of 3,261.6 g with a
standard deviation of 590.7 g, EPA calculated the mean response that would be associated with the
12.86th percentile of the distribution falling below 2,500 g. In this case, the mean birth weight
would be 3,169.2 g. Given the median exposure among NHANES females as x, the mean birth
weight in the United States as y and the re-expressed (3 as m term, the intercept b can be estimated
as:
b =y -nix = 3261.6 g- (-207.7 °-19ff = 3301.1 g (C-5)
The BMD was calculated by rearranging the equation y = mx + b and solving for x, using
3301.1 g for the b term and -207.7 for the m term. This resulted in a value of 0.63 ng/mL:
x = (y — b)/m = (3169.2 g - 3301.1 #)/(-207.7 = 0.63 ng/mL (C-6)
To calculate the BMDL, the method is essentially the same except that the lower limit (LL)
on the re-expressed (3 coefficient (-516.8 g per ng/mL) is used for the m term. However, (Valvi et
al.. 2017) reports a two-sided 95% confidence interval for the (3 coefficient, meaning that the lower
limit of that confidence interval corresponds to a 97.5% one-sided lower limit The BMDL is
defined as the 95% lower limit of the BMD (i.e., corresponds to a two-sided 90% confidence
4While the explicit application of the hybrid approach is not commonly used in IRIS
dose/concentration/exposure-response analyses, the more commonly used SD-definition of the BMR for
continuous data is simply one specific application of the hybrid approach. The SD-definition of the BMR
assumes that the cut-off for adversity is the 1.4th percentile of a normally distributed response and that
shifting the mean of that distribution by one standard deviation approximates an extra risk of 10%.
This document is a draft for review purposes only and does not constitute Agency policy.
C-18 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
interval), so the corresponding lower limit on the re-expressed (3 coefficient needs to be calculated
before calculating the BMDL. First, the standard error of the re-expressed (3 coefficient can be
calculated as:
ŁŁ _ Upper Limit-Lower Limit _ 91-2 1~(~516-8 1) _ ^ fQ y>
~~ 3.92 ~~ 3.92 ~~ ' ^ '
Then the corresponding 95% one-sided lower limit on the re-expressed (3 coefficient is
calculated as:
95% one - sided LL = /3 — 1.645(SŁ(/?)) = -207.7 g(^T1 - 1.645 (l55.1 = -462.9g^)'1 (C-8)
Using this value for the m term results in a BMDL value of 0.28 ng/mL maternal serum
concentration.
Valvi etal. f20171 also reported a (3 coefficient of-44 g per log2(ng/mL) (95%CI: -133, 44 g
per log2(ng/mL) for boys and -28 g per log2(ng/mL) (95%CI: -110, 54 g per log2(ng/mL)) for girls.
The re-expressed (3 coefficients are -222.9 g per ng/mL (95%CI: -673.9, 222.9 g per ng/mL) and
-141.9 g per ng/mL (95%CI: -557.3, 273.6 g per ng/mL), and the intercepts b are 3,304.0 g and
3,288.6 g for boys and girls, respectively. Using these sex-specific values, the estimated BMD values
are 0.60 ng/mL for boys and 0.84 ng/mL for girls.
To calculate the BMDL, the same procedure as above is used to calculate the corresponding
95% one-sided lower limit for the re-expressed (3 coefficient from the re-expressed lower limit on
the 95% two-sided confidence interval of-673.9 gper ng/mL for boys and -557.3 gper ng/mL for
girls. Using the corresponding lower limit (-599.2 g per ng/mL for boys and -490.5 g per ng/mL
for girls), the BMDLs of 0.22 ng/mL for boys and 0.24 ng/mL for girls are calculated.
Division of Environmental Epidemiology etal. (2016) reported a (3 coefficient of -43.9 g per
ln(ng/mL) (95%CI: -104.8,17.0 gper ln(ng/mL) for the association between birth weight and
maternal PFDA serum concentrations in a multi-country cohort. Given the reported study-specific
geometric mean (0.25) and standard deviation of In-transformed exposure (0.70), EPA estimated
the mean (-1.41) and standard deviation (0.70) of the log normally distributed exposure. The re-
expressed (3 coefficient is -122.2 g (95%CI: -291.5, 47.2) per ng/mL and the interceptb is
3,284.8 g. The 95% one-sided lower limits for the re-expressed (3 coefficient are -264.3 gper
ng/mL. The values of the BMD and BMDL are 0.95 ng/mL and 0.44 ng/mL, respectively.
Luo etal. (2021) reported a (3 coefficient of-96.8 gper ln(ng/mL) (95%CI: -178.0, -15.5 g
per ln(ng/mL)) for the association between birth weight and maternal PFDA serum concentrations
in a China cohort Given the reported study-specific median (0.48 ng/mL) and IQR (0.34-0.70
ng/mL) of the exposure, EPA estimated the mean (-0.73) and standard deviation (0.54) of the log
normally distributed exposure. The re-expressed (3 coefficient is -195.8 g per ng/mL (95%CI:
-360.2, -31.4 gper ng/mL) and the intercept b is 3,298.8 g. The 95% one-sided lower limits for the
This document is a draft for review purposes only and does not constitute Agency policy.
C-19 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
re-expressed (3 coefficient are -333.8 gper ng/mL. The values of the BMD and BMDL are 0.66
ng/mL and 0.39 ng/mL, respectively.
Wikstrom etal. f20201 reported a (3 coefficient of -58.0 g per ln(ng/mL) (95%CI: -103.0,
-13.0 g per ln(ng/mL)) for the association between birth weight and maternal PFDA serum
concentrations in a Swedish cohort Given the reported study-specific median (0.26 ng/mL) and
IQR (0.19-0.34 ng/mL) of the exposure, EPA estimated the mean (-1.35) and standard deviation
(0.43) of the log normally distributed exposure. The re-expressed (3 coefficient is -218.9 g per
ng/mL (95%CI: -388.7, -49.1 g per ng/mL) and the intercept b is 3303.2 g. The 95% one-sided
lower limits for the re-expressed (3 coefficient are -361.4 g per ng/mL. The values of the BMD and
BMDL are 0.61 ng/mL and 0.37 ng/mL, respectively.
Wikstrom etal. f20201 also reported (3 coefficients of-47 gper ln(ng/mL) (95%CI: -112,
17 g per ln(ng/mL)) for boys and -69 g per ln(ng/mL) (95%CI: -133, -6 g per ln(ng/mL)) for girls.
The re-expressed (3 coefficients are -177.4 g per (95%CI: -422.7, 64.2 g per ng/mL) and -260.4 g
per (95%CI: -501.9, -22.6 gper ng/mL), and the intercepts b are 3,295.3 g and 3,311.1 g for boys
and girls, respectively. Using these sex-specific values, the estimated BMD values are 0.71 ng/mL
for boys and 0.54 ng/mL for girls. The corresponding 95% one-sided lower limits for the re-
expressed (3 coefficient are -381.6 g per and -461.5 g per for boys and girls, respectively. The
BMDL values are 0.33 ng/mL for boys and 0.31 ng/mL for girls.
Yao etal. f20211 reported a (3 coefficient of-46.3 gper ln(ng/mL) (95%CI: -131.1, 38.5 g
per ln(ng/mL)) for the association between birth weight and maternal PFDA serum concentrations
in a China cohort Given the reported study-specific median (0.55 ng/mL) and IQR (0.37-0.74
ng/mL) of the exposure, EPA estimated the mean (-0.60) and standard deviation (0.51) of the log
normally distributed exposure. The re-expressed (3 coefficient is -82.0 gper (95%CI: -232.1, 68.1 g
per ng/mL) and the intercept b is 3277.2 g. The 95% one-sided lower limits for the re-expressed (3
coefficient are -208.0 gper ng/mL. The values of the BMD and BMDL are 1.32 ng/mL and 0.52
ng/mL, respectively.
For all the above calculations, EPA used the exact percentage (8.27%) of live births in the
U.S. in 2018 that fell below the cut-off of 2,500 g as the tail probability to represent the probability
of extreme ("adverse") response at zero dose (P(0)). However, this exact percentage of 8.27% was
calculated without accounting for the existence of background PFDA exposure in the U.S.
population (i.e., 8.27% is not the tail probability of extreme response at zero dose). Thus, EPA
considers an alternative control-group response distribution ac)), using the study-specific
intercept b obtained through equation (C-5) (representing the baseline value of birth weight in an
unexposed population) as and the standard deviation of the U.S. population as ac, to estimate the
tail probability that fell below the cut-off of 2,500 g. EPA estimated the study-specific tail
probability of live births falling below the public health definition of low birth weight (2,500 g) as:
1 r2500 j 1 f2500 (- (x~6)2?) j cn ^
P(0) =—= e 2ac dx = = ey 2*590.72J dx (C-9)
v J o-cV27TJ-o° 590.7V27TJ—CO 1
This document is a draft for review purposes only and does not constitute Agency policy.
C-20 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
b = y -mx = 3261.6 - (fire-expressed * °-19[C"10)
In this alternative approach, P(0) is 9.86% if there is no background exposure (x = 0). By
using the median serum PFDA concentrations (0.19 ng/mL) from NHANES females in 2011-2012
as background exposure (x), the tail probabilities using this alternative approach were study
specific and ranged from 8.48% to 9.41%. As such, the results from this alternative approach,
presented under the column of "Alternative Tail Probability" in Table C-8, are very similar to the
main results, presented under the column of "Exact Percentage" in Table C-8, when background
exposure was not accounted for while estimating the tail probability.
Table C-8 presents the BMDs and BMDLs for all studies considered for POD derivation, with
and without accounting for background exposure while estimating the percentage of the population
falling below the cut-off value. The BMDLs across the studies and methods ranged from 0.22 ng/mL
to 0.66 ng/mL.
This document is a draft for review purposes only and does not constitute Agency policy.
C-21 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-10. BMDs and BMDLs for effect of PFDA on decreased birth weight, by using percentage (8.27%) of live
births falling below the public health definition of low birth weight, or alternative study-specific tail probability
Study
Exposure
median
(IQR) or
GM (SD)
Exposure
distribution
Reported |3
(95%CI)
Re-expressed (3
(95%CI)
g/ng/mL
Intercept
b
SE of |3
95% one-
sided LL
of |3
Exact percentage
(P(0) =8.27%)
Alternative tail probability3
BMD
(ng/mL)
BMDL
(ng/mL)
P( 0)
BMD
(ng/mL)
BMDL
(ng/mL)
Valvi et al.
(2017)
0.28
(0.22-0.38)
(-1.27, 0.41)
-41.0
(-102.0, 18.0)
g/log2(ng/mL)
-207.7
(-516.8, 91.2)
3301.1
155.11
-462.9
0.63
0.28
8.75%
0.70
0.31
Valvi et al.
(2017) Bovs
0.28
(0.22-0.38)
(-1.27, 0.41)
-44.0
(-133.0, 44.0)
g/log2(ng/mL)
-222.9
(-673.9, 222.9)
3304.0
228.78
-599.2
0.60
0.22*
8.67%
0.65
0.24
Valvi et al.
(2017) Girls
0.28
(0.22-0.38)
(-1.27, 0.41)
-28.0
(-110.0, 54.0)
g/log2(ng/mL)
-141.9
(-557.3, 273.6)
3288.6
211.98
-490.5
0.84
0.24
9.09%
0.99
0.29
Division of
Environmental
Epidemiology et
al. (2016)
0.25
(0.70)b
(-1.41, 0.70)
-43.9
(-104.8, 17.0)
g/ln(ng/mL)
-122.2
(-291.5, 47.2)
3284.8
86.40
-264.3
0.95
0.44
9.20%
1.14
0.53
Luo et al. (2021)
0.48
(0.34-0.70)
(-0.73, 0.54)
-96.8
(-178.0, -15.5)
g/ln(ng/mL)
-195.8
(-360.2, -31.4)
3298.8
83.88
-333.8
0.66
0.39
8.81%
0.73
0.43
Wikstrom et al.
(2020)
0.26
(0.19-0.34)
(-1.35, 0.43)
-58.0
(-103.0, -13.0)
g/ln(ng/mL)
-218.9
(-388.7, -49.1)
3303.2
86.64
-361.4
0.61
0.37
8.69%
0.66
0.40
Wikstrom et al.
(2020) Bovs
0.26
(0.19-0.34)
(-1.35, 0.43)
-47.0
(-112.0, 17.0)
g/ln(ng/mL)
-177.4
(-422.7, 64.2)
3295.3
124.19
-381.6
0.71
0.33
8.91%
0.80
0.37
Wikstrom et al.
(2020) Girls
0.26
(0.19-0.34)
(-1.35, 0.43)
-69.0
(-133.0, -6.0)
g/ln(ng/mL)
-260.4
(-501.9, -22.6)
3311.1
122.26
-461.5
0.54
0.31
8.48%
0.57
0.32
This document is a draft for review purposes only and does not constitute Agency policy.
C-22 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Study
Exposure
median
(IQR) or
GM (SD)
Exposure
distribution
Reported |3
(95%CI)
Re-expressed (3
(95%CI)
g/ng/mL
Intercept
b
SE of |3
95% one-
sided LL
of |3
Exact percentage
(P(0) =8.27%)
Alternative tail probability3
BMD
(ng/mL)
BMDL
(ng/mL)
P( 0)
BMD
(ng/mL)
BMDL
(ng/mL)
Yao et al. (2021)
0.55
(0.37-0.74)
(-0.60, 0.51)
-46.3
(-131.1, 38.5)
g/ln(ng/mL)
-82.0
(-232.1, 68.1)
3277.2
76.58
-208.0
1.32
0.52
9.41%
1.68
0.66
*Smallest BMDL using the five individual studies.
aThe alternative study-specific tail probability of live births falling below the public health definition of low birth weight based on Normal distribution with
intercept b as mean and standard deviation of 590.7 based on U.S. population.
bDivision of Environmental Epidemiology et al. (2016) reports Geometric Mean (GM) and standard deviation (SD) of In-transformed concentrations.
This document is a draft for review purposes only and does not constitute Agency policy.
C-23 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
C.2. BENCHMARK DOSE MODELING RESULTS FROM ANIMAL STUDIES
C.2.1. BENCHMARK DOSE MODELING APPROACHES
The endpoints selected for benchmark dose (BMD) modeling are listed in Table C-ll. The
animal doses in the study were used in the BMD modeling and then converted to human equivalent
doses (HEDs) using data-derived extrapolation factors (DDEFs) described in Section 3.1.7 of the
main document; the modeling results are presented in this appendix.
Modeling Procedure for Dichotomous Noncancer Data
BMD modeling of dichotomous noncancer data was conducted using EPA's Benchmark Dose
Software (BMDS, version 3.2). For these data, the Gamma, Logistic, Log-Logistic, Log-Probit,
Multistage, Probit, Weibull, and Dichotomous Hill models available within the software were fit
using a benchmark response (BMR) of 10% extra risk (see Toxicological Review, Section 5.2.1 for
justification of selected BMRs). 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
onx2 goodness-of-fitp-value (p > 0.1), scaled residuals at the data point (except the control) closest
to the predefined benchmark response (absolute value <2.0), and visual inspection of the model fit.
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 presented with that of the full
data set In cases where a model with several parameters equal to the number of dose groups was
fit to the data set, all parameters were estimated, and no p-value was calculated, that model was not
considered for estimating a point of departure (POD) unless no other model provided adequate fit.
Among all models providing adequate fit, the benchmark dose lower confidence limit (BMDL) from
the model with the lowest Akaike's information criterion (AIC) was selected as a potential POD
when BMDL values were sufficiently close (within 3-fold). Otherwise, the lowest BMDL was
selected as a potential POD.
Modeling Procedure for Continuous Noncancer Data
BMD modeling of continuous noncancer data was conducted using EPA's Benchmark Dose
Software (BMDS, version 3.2). For these data, the Exponential, Hill, Polynomial, and Power models
available within the software are fit using a BMR of 1 standard deviation (SD) when no toxicological
information was available to determine an adverse level of response. When toxicological
information was available, the BMR was based on relative deviation, as outlined in the Benchmark
Dose Technical Guidance fU.S. EPA. 20121 (see Toxicological Review, Section 5.2.1 justification for
using BMRs); when a BMR based on relative deviation was used, modeling results using BMRs
based on SD are included for reference. An adequate fit is judged on the basis of x2 goodness-of-fit
p-value (p > 0.1), scaled residuals at the data point (except the control) closest to the predefined
benchmark response (absolute value <2.0), and visual inspection of the model fit In addition to
these three criteria for judging adequacy of model fit, a determination is made on whether the
This document is a draft for review purposes only and does not constitute Agency policy.
C-24 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
variance across dose groups is homogeneous. If a homogeneous variance model is deemed
appropriate on the basis of the statistical test provided by BMDS (i.e., Test 2), the final BMD results
are estimated from a homogeneous variance model. If the test for homogeneity of variance is
rejected (p < 0.05), the model is run again while modeling the variance as a power function of the
mean to account for this nonhomogeneous variance. If this nonhomogeneous variance model does
not adequately fit the data (i.e., Test 3; p < 0.05), alternative approaches are assessed on a case-by-
case basis. For example, in cases where neither variance model fit, or constant variance did not fit
(with adequate Test-4 p-value) and nonconstant variance did fit (with inadequate Test-4 p-value),
the log-normal distribution was attempted.
In cases where a model with several parameters equal to the number of dose groups was fit
to the data set, all parameters were estimated, and no p-value was calculated, that model was not
considered for estimating 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 estimates differed by less than 3-fold. When BMDL estimates differed by greater
than 3-fold, the model with the lowest BMDL was selected to account for model uncertainty.
Modeling Procedure for Continuous Noncancer Developmental Toxicity Data
For continuous developmental toxicity data, data for individual animals were requested
from the study authors when possible. The use of individual animal data allows for the correct
measure of variance to be calculated. When a biological rationale for selecting a benchmark
response level is lacking, a BMR equal to 0.5 SD was used. The use of 1 SD for the BMR for
continuous endpoints is based on the observation that shifting the distribution of the control group
by 1 SD results in ~10% of the animal data points falling beyond an adversity cutoff defined at the
~1.5 percentile fCrump. 19951. This approximates the 10% extra risk commonly used as the BMR
for dichotomous endpoints. Thus, the use of 0.5 SD for continuous developmental toxicity
endpoints approximates the extra risk commonly used for dichotomous developmental toxicity
endpoints.
Data Used for Modeling
The source of the data used for modeling endpoints from animal studies is provided in
Table C-ll. These data also are included in full in the tables below.
Table C-ll. Sources of data used in benchmark dose modeling of PFDA
endpoints from animal studies
Endpoint/reference
Reference
HAWC link
/T* AST - M
NTP (2018)
https://hawcprd.epa.gov/ani/endpoint/100506861/
/T* AST - F
NTP (2018)
https://hawcprd.epa.gov/ani/endpoint/100506957/
This document is a draft for review purposes only and does not constitute Agency policy.
C-25 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Endpoint/reference
Reference
HAWC link
¦f ALP- F
NTP (2018)
https://hawcprd.epa.gov/ani/endpoint/100506956/
'T* Relative Liver weight - M
NTP (2018)
https://hawcprd.epa.gov/ani/endpoint/100506814/
'T* Relative Liver weight - F
NTP (2018)
https://hawcprd.epa.gov/ani/endpoint/100506920/
'T* Relative Liver weight - F
(Histo)
Frawlev et al. (2018)
httDs://hawcDrd.eDa.sov/ani/endDoint/100506676/
'T* Relative Liver weight - F
(MPS)
Frawlev et al. (2018)
https://hawcprd.epa.gov/ani/endpoint/100506669/
'T* Relative Liver weight - F
(TDAR)
Frawlev et al. (2018)
https://hawcprd.epa.gov/ani/endpoint/100506677/
si Fetal Body Weight (GD6-15)
Harris and Birnbaum (1989)
httDs://hawcDrd.eDa.sov/ani/endDoint/100506643/
\1/ Caudal Epididymis Sperm
Count
NTP (2018)
https://hawcprd.epa.gov/ani/endpoint/100506879/
\1/ Absolute Testis Weight
NTP (2018)
httDs://hawcDrd.eDa.sov/ani/endDoint/100506820/
\1/ Absolute Cauda Epididymis
Weight
NTP (2018)
httDs://hawcDrd.eDa.sov/ani/endDoint/100506878/
\1/ Absolute Whole Epididymis
Weight
NTP (2018)
httDs://hawcDrd.eDa.sov/ani/endDoint/100506877/
\1/ Estrus Time
NTP (2018)
httDs://hawcDrd.eDa.Eov/ani/endooint/100524936/
'T* Diestrus Time
NTP (2018)
httDs://hawcDrd.eDa.Eov/ani/endooint/100524930/
\1/ Relative Uterus Weight
NTP (2018)
httDs://hawcDrd.eDa.sov/ani/endDoint/100506941/
\1/ Absolute Uterus Weight
NTP (2018)
httDs://hawcDrd.eDa.sov/ani/endDoint/100506940/
C.2.2. INCREASED AST-
Table C-12. Dose-i
MALE RATS fNTP. 20181
'esponse data for increase
d AST in male rats fNTP. 20181
Dose (mg/kg-d)
n
Mean
SD
0
10
65.3
10.18
0.156
10
74
9.55
0.312
10
77.3
16.98
0.625
10
81.3
9.84
1.25
10
87.5
14.61
2.5
9
92.67
8.04
This document is a draft for review purposes only and does not constitute Agency policy.
C-26 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-13. Benchmark dose results for increased AST in male rats—constant
variance, BMR = 1 standard deviation fNTP. 201811
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
1.3924
1.0640
0.1386
467.4755
Viable—Alternate
Exponential 3
(CV—normal)
Restricted
1.3924
1.0640
0.1386
467.4755
Viable—Alternate
Exponential 4
(CV—normal)
Restricted
0.3933
0.1723
0.8692
463.2441
Viable—Alternate
Exponential 5
(CV—normal)
Restricted
0.3949
0.1723
0.8692
463.2441
Viable—Alternate
Hill
(CV—normal)
Restricted
0.3266
0.1227
0.9560
462.8481
Viable-
Recommended
Lowest BMDL
Polynomial
(5 degree)
(CV—normal)
Restricted
1.2558
0.9260
0.1910
466.6376
Viable—Alternate
Polynomial
(4 degree)
(CV—normal)
Restricted
1.2558
0.9260
0.1910
466.6376
Viable—Alternate
Polynomial
(3 degree)
(CV—normal)
Restricted
1.2558
0.9260
0.1910
466.6376
Viable—Alternate
Polynomial
(2 degree)
(CV—normal)
Restricted
1.2558
0.9260
0.1910
466.6376
Viable—Alternate
Power
(CV—normal)
Restricted
1.2558
0.9260
0.1910
466.6376
Viable—Alternate
Linear
(CV—normal)
Unrestricted
1.2558
0.9260
0.1910
466.6376
Viable—Alternate
1 Throughout this section, in the Benchmark Dose results table, the "Restriction" column denotes the restriction
status of applied models, and the "Classification" column denotes whether a model can be considered for model
selection purposes. See BMDS User Guide: https://www.epa.gov/bmds. If a model was selected as appropriately
fitting the modeled data, that model's entries in the tables are in green shaded cells and the text is bolded.
This document is a draft for review purposes only and does not constitute Agency policy.
C-27 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Frequentist Hill Model with BMR of 1 Std. Dev. for the BMD and 0.95 Lower Confidence
Limit for the BMDL
no
Dose
Figure C-3. Dose-response curve for the Hill model fit to increased AST in male
rats fNTP. 20181.
This document is a draft for review purposes only and does not constitute Agency policy.
C-28 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
User Input
Info
Model
frequentist Hill vl.l
Dataset Name
AST M NTP
User notes
[Add user notes here]
Dose-Response Model
M[dose] = g + v*doseAn/(kAn + dose^n)
Variance Model
Var[i] = alpha
Model Options
BMR Type
Std. Dev.
BMRF
1
Tail Probability
-
Confidence Level
0.95
Distribution Type
Mormal
Variance Type
Constant
Model Data
Dependent Variable
[Dose]
Independent Variable
[Mean]
Total # of Observations
6
Adverse Direction
Automatic
Figure C-4. User Input for dose-response modeling of increased AST in male
rats fNTP. 20181.
This document is a draft for review purposes only and does not constitute Agency policy.
C-29 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Model Results
Benchmark Dose
BMD
0.32659537
BMDL
0.122653237
BMDU
0.926151614
AIC
462.8480778
Test 4 P-value
0.956041631
D.O.F.
3
Model Parameters
# of Parameters
5
Variable
Estimate
g
65.96003464
V
32.30491688
k
0.59693749
n
Bounded
alpha
130.5126471
Goodness of Fit
Dose
Size
Estimated
Median
Calc'd
Median
Observed
Mean
Estimated
SD
Calc'd SD
Observed
SD
Scaled
Residual
0
10
65.96003464
65.3
65.3
11.4242132
10.18
10.18
-0.182700792
0.156
10
72.65324238
74
74
11.4242132
9.55
9.55
0.372789045
0.312
10
77.0489535
77.3
77.3
11.4242132
16.98
16.98
0.06949089
0.625
10
82.48344375
81.3
81.3
11.4242132
9.84
9.84
-0.327582975
1.25
10
87.82387482
87.5
87.5
11.4242132
14.61
14.61
-0.089650123
2.5
9
92.03814971
92.67
92.67
11.4242132
8.04
8.04
0.165923977
Likelihoods of Interest
# of
Model
Log Likelihood*
Pa ra meters
AIC
A1
-227.2635646
7
468.527129
A2
-223.0848415
12
470.169683
A3
-227.2635646
7
468.527129
fitted
-227.4240389
4
462.848078
R
-241.1426777
2
486.285355
* Includes additive constant of -54.21737. This constant was not included in the LL derivation prior to BMDS 3.0.
Tests of Interest
-2*Log(Likelihood
Test
Ratio)
Test df
p-value
1
36.11567239
10
<0.0001
2
8.357446131
5
0.13760531
3
8.357446131
5
0.13760531
4
0.320948692
3
0.95604163
Figure C-5. Model Results for increased AST in male rats (NTP. 2018).
This document is a draft for review purposes only and does not constitute Agency policy.
C-30 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
C.2.3. INCREASED AST-FEMALE RATS fNTP. 20181
Table C-14. Dose-response data for increased AST in female rats fNTP. 20181
Dose (mg/kg-d)
n
Mean
SD
0
10
62.6
10.75
0.156
9
60.44
6.51
0.312
10
57.9
4.11
0.625
10
63.3
5
1.25
10
81.9
8.29
2.5
7
112.57
22.54
Table C-15. Benchmark dose results for increased AST in female
rats—constant variance, BMR = 1 standard deviation (NTP. 2018)
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.6219
0.5312
0.1426
427.8867
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Exponential 3
(CV—normal)
Restricted
0.8024
0.5551
0.1375
428.5314
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Exponential 4
(CV—normal)
Restricted
0.5006
0.0000
0.0153
433.4316
Unusable
BMD computation failed;
lower limit includes zero
BMDL not estimated
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual at control | > 2
Exponential 5
(CV—normal)
Restricted
0.1055
0.1048
<0.0001
553.6193
Questionable
Constant variance test failed
(Test 2 5-value < 0.05)
Goodness of fit p-value < 0.1
| Residual at control | > 2
Modeled control response
std. dev. >11.51 actual
response std. dev.
Hill
(CV—normal)
Restricted
0.9445
0.6992
0.5341
426.2660
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Polynomial
(5 degree)
(CV—normal)
Restricted
0.8055
0.5285
0.1331
428.6052
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Polynomial
(4 degree)
(CV—normal)
Restricted
0.8055
0.5285
0.1331
428.6052
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Polynomial
(3 degree)
(CV—normal)
Restricted
0.8055
0.5285
0.1331
428.6052
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
This document is a draft for review purposes only and does not constitute Agency policy.
C-31 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Polynomial
(2 degree)
(CV—normal)
Restricted
0.8055
0.5285
0.1331
428.6052
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Power
(CV—normal)
Restricted
0.8126
0.5686
0.2122
427.5127
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Linear
(CV—normal)
Unrestricted
0.5006
0.4134
0.0339
431.4316
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual at control | > 2
Table C-16. Benchmark dose results for increased AST in female
rats—nonconstant variance, BMR = 1 standard deviation (NTP. 2018)
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Non-constant variance
Exponential 2
(NCV—
normal)
Restricted
0.4683
0.3822
0.0006
417.7886
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual at control | > 2
Exponential 3
(NCV—
normal)
Restricted
0.7433
0.5327
0.0048
413.2499
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Exponential 4
(NCV—
normal)
Restricted
0.4044
0.3201
<0.0001
425.5227
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual at control | > 2
Exponential 5
(NCV—
normal)
Restricted
0.9173
0.6965
0.0484
408.4035
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Hill (NCV—
normal)
Restricted
1.1570
0.6738
0.0375
408.9143
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Polynomial
(5 degree)
(NCV—
normal)
Restricted
0.8488
0.5738
0.0172
410.3710
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Polynomial
(4 degree)
(NCV—
normal)
Restricted
0.8488
0.5738
0.0172
410.3710
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
This document is a draft for review purposes only and does not constitute Agency policy.
C-32 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Non-constant variance
Polynomial
(3 degree)
(NCV—
normal)
Restricted
0.8488
0.5738
0.0172
410.3710
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Polynomial
(2 degree)
(NCV—
normal)
Restricted
0.8488
0.5738
0.0172
410.3710
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Power (NCV—
normal)
Restricted
0.7553
0.5621
0.0104
411.6066
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Linear (NCV—
normal)
Unrestricted
0.4052
0.3203
<0.0001
423.4964
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual at control | > 2
Table C-17. Benchmark dose results for increased AST in female rats—log-
normal. constant variance. BMR = 1 standard deviation (NTP. 2018)
Models
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
Restriction
BMD
BMDL
Constant variance
Exponential 2
(CV—log-
normal)
Restricted
0.4981
0.4114
0.0353
410.1569
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Exponential 3
(CV- log-
normal)
Restricted
0.7017
0.4707
0.0518
409.5663
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Exponential 4
(CV- log-
normal)
Restricted
0.4173
0.0000
0.0061
414.2361
Unusable
BMD computation failed;
lower limit includes zero
BMDL not estimated
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
This document is a draft for review purposes only and does not constitute Agency policy.
C-33 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 5
(CV- log-
normal)
Restricted
-9999.00
00
0.0000
<0.0001
482.3726
Unusable
BMD computation failed;
lower limit includes zero
BMD not estimated
BMDL not estimated
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual at control | > 2
Hill (CV— log-
normal)
Restricted
0.8526
0.6413
0.4051
405.6388
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
| Residual at control | > 2
Polynomial
(5 degree)
(CV- log-
normal)
Restricted
0.7220
0.4645
0.0501
409.6412
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Polynomial
(4 degree)
(CV- log-
normal)
Restricted
0.7220
0.4645
0.0501
409.6412
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Polynomial
(3 degree)
(CV- log-
normal)
Restricted
0.7220
0.4645
0.0501
409.6412
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Polynomial
(2 degree)
(CV- log-
normal)
Restricted
0.7220
0.4645
0.0501
409.6412
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Power (CV—
log-normal)
Restricted
0.7158
0.5034
0.0953
408.1933
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Linear (CV—
log-normal)
Unrestricted
0.4170
0.3303
0.0061
414.2360
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
This document is a draft for review purposes only and does not constitute Agency policy.
C-34 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
C.2.4. INCREASED ALP-FEMALE RAT fNTP. 20181
Table C-18. Dose-response data for increased ALP in female rats fNTP. 20181
Dose (mg/kg-d)
n
Mean
SD
0
9
136.4
18.6
0.156
9
156.1
24
0.312
10
182.8
36.68
0.625
10
184.2
33.2
1.25
10
281.1
72.42
2.5
7
262.4
60.06
Table C-19. Benchmark dose results for increased ALP in female rats—
BMR = constant variance, 1 standard deviation (NTP. 2018)
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
1.2058
0.9747
<0.0001
598.0449
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
| Residual for Dose Group
Near BMD| >2
Modeled control response
std. dev. >| 1.51 actual
response std. dev.
Exponential 3
(CV—normal)
Restricted
1.2058
0.9747
<0.0001
598.0449
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
| Residual for Dose Group
Near BMD| >2
Modeled control response
std. dev. >| 1.51 actual
response std. dev.
Exponential 4
(CV—normal)
Restricted
0.3043
0.1894
0.0206
585.6900
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response
std. dev. >| 1.51 actual
response std. dev.
Exponential 5
(CV—normal)
Restricted
0.6977
0.3389
0.0530
583.7962
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response
std. dev. >| 1.51 actual
response std. dev.
This document is a draft for review purposes only and does not constitute Agency policy.
C-35 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Hill
(CV—normal)
Restricted
0.6547
0.6162
0.1011
582.1450
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response
std. dev. >| 1.51 actual
response std. dev.
Polynomial
(5 degree)
(CV—normal)
Restricted
0.9018
0.6940
0.0005
594.1122
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response
std. dev. >| 1.51 actual
response std. dev.
Polynomial
(4 degree)
(CV—normal)
Restricted
0.9018
0.6940
0.0005
594.1122
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response
std. dev. >| 1.51 actual
response std. dev.
Polynomial
(3 degree)
(CV—normal)
Restricted
0.9018
0.6940
0.0005
594.1122
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response
std. dev. >| 1.51 actual
response std. dev.
Polynomial
(2 degree)
(CV—normal)
Restricted
0.9018
0.6940
0.0005
594.1122
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response
std. dev. >| 1.51 actual
response std. dev.
Power
(CV—normal)
Restricted
0.9018
0.6941
0.0005
594.1122
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response
std. dev. >| 1.51 actual
response std. dev.
Linear
(CV—normal)
Unrestricted
0.9018
0.6940
0.0005
594.1122
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response
std. dev. >| 1.51 actual
response std. dev.
This document is a draft for review purposes only and does not constitute Agency policy.
C-36 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-20. Benchmark dose results for increased ALP in female rats—
nonconstant variance, BMR = 1 standard deviation fNTP. 20181
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Non-constant variance
Exponential 2
(NCV—
normal)
Restricted
0.3761
0.2620
<0.0001
578.1584
Questionable
Goodness of fit p-value < 0.1
Exponential 3
(NCV—
normal)
Restricted
0.3761
0.2620
<0.0001
578.1584
Questionable
Goodness of fit p-value < 0.1
Exponential 4
(NCV—
normal)
Restricted
0.1191
0.0720
0.0174
565.0835
Questionable
Goodness of fit p-value < 0.1
Exponential 5
(NCV—
normal)
Restricted
0.1556
0.0758
0.0083
566.5363
Questionable
Goodness of fit p-value < 0.1
Hill (NCV—
normal)
Restricted
0.1501
0.0700
0.0056
567.3018
Questionable
Goodness of fit p-value < 0.1
Polynomial
(5 degree)
(NCV—
normal)
Restricted
0.2457
0.1655
0.0012
570.9484
Questionable
Goodness of fit p-value < 0.1
Polynomial
(4 degree)
(NCV—
normal)
Restricted
0.2457
0.1655
0.0012
570.9484
Questionable
Goodness of fit p-value < 0.1
Polynomial
(3 degree)
(NCV—
normal)
Restricted
0.2457
0.1655
0.0012
570.9484
Questionable
Goodness of fit p-value < 0.1
Polynomial
(2 degree)
(NCV—
normal)
Restricted
0.2457
0.1655
0.0012
570.9484
Questionable
Goodness of fit p-value < 0.1
Power (NCV—
normal)
Restricted
0.2457
0.1655
0.0012
570.9484
Questionable
Goodness of fit p-value < 0.1
Linear (NCV—
normal)
Unrestricted
0.2457
0.1655
0.0012
570.9484
Questionable
Goodness of fit p-value < 0.1
This document is a draft for review purposes only and does not constitute Agency policy.
C-37 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-21. Benchmark dose results for increased ALP in female rats—log-
normal, constant variance, BMR = 1 standard deviation fNTP. 20181
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—log-
normal)
Restricted
0.8447
0.6570
0.0001
575.0495
Questionable
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Exponential 3
(CV- log-
normal)
Restricted
0.8447
0.6570
0.0001
575.0495
Questionable
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Exponential 4
(CV- log-
normal)
Restricted
0.2215
0.1355
0.0337
563.1028
Questionable
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Exponential 5
(CV- log-
normal)
Restricted
0.3331
0.1470
0.0200
564.2382
Questionable
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Hill (CV— log-
normal)
Restricted
0.2860
0.1283
0.0121
565.2461
Questionable
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Polynomial
(5 degree)
(CV- log-
normal)
Restricted
0.5606
0.4106
0.0017
569.7238
Questionable
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Polynomial
(4 degree)
(CV- log-
normal)
Restricted
0.5606
0.4106
0.0017
569.7238
Questionable
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Polynomial
(3 degree)
(CV- log-
normal)
Restricted
0.5606
0.4106
0.0017
569.7238
Questionable
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Polynomial
(2 degree)
(CV- log-
normal)
Restricted
0.5606
0.4106
0.0017
569.7238
Questionable
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Power (CV—
log-normal)
Restricted
0.5606
0.4107
0.0017
569.7238
Questionable
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
| Residual at control | > 2
Linear (CV—
log-normal)
Unrestricted
0.5606
0.4106
0.0017
569.7238
Questionable
Goodness of fit p-value < 0.1
| Residual for Dose Group
This document is a draft for review purposes only and does not constitute Agency policy.
C-38 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Near BMD| >2
| Residual at control | > 2
C.2.5. INCREASED RELATIVE LIVER WEIGHT-MALE RAT fNTP. 20181
Table C-22. Dose-response data for increased relative liver weight in male rats
fNTP. 20181
Dose (mg/kg-d)
n
Mean
SD
0
10
35.5
3.07
0.156
10
39.32
1.68
0.312
10
42.61
1.77
0.625
10
45.56
2.66
1.25
10
54.77
2.15
2.5
10
67.9
3.76
Table C-23. Benchmark dose results for increased relative liver weight in male
rats—constant variance, BMR = 10% relative deviation fNTP. 20181
Models
Restriction
10% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.4081
0.3852
<0.0001
314.8501
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Exponential 3
(CV—normal)
Restricted
0.4081
0.3852
<0.0001
314.8501
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Exponential 4
(CV—normal)
Restricted
0.2116
0.1764
0.2654
291.5391
Viable—Alternate
Exponential 5
(CV—normal)
Restricted
0.2112
0.1764
0.2653
291.5398
Viable—Alternate
Hill
(CV—normal)
Restricted
0.2078
0.1710
0.2774
291.4313
Viable-
Recommended
Lowest AIC
Polynomial
(5 degree)
(CV—normal)
Restricted
0.2978
0.2836
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Polynomial
(4 degree)
(CV—normal)
Restricted
0.2978
0.2778
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
This document is a draft for review purposes only and does not constitute Agency policy.
C-39 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
10% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Polynomial
(3 degree)
(CV—normal)
Restricted
0.2978
0.2775
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Polynomial
(2 degree)
(CV—normal)
Restricted
0.2978
0.2775
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Power
(CV—normal)
Restricted
0.2978
0.2775
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Linear
(CV—normal)
Unrestricted
0.2978
0.2775
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Frequentist Hill Model with BMR of 0.1 Rel. Dev. for the BMD and 0.95 Lower Confidence
Limit for the BMDL
80
Figure C-6. Dose-response curve for the Hill model fit to increased relative
liver weight in male rats (NTP. 2018).
This document is a draft for review purposes only and does not constitute Agency policy.
C-40 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
User Input
Info
Model
frequentist Hill vl.l
Dataset Name
LiverWt Rel M NTP
User notes
[Add user notes here]
Dose-Response Model
M[dose] = g + v*doseAn/(kAn + doseAn)
Variance Model
Var[i] = alpha
Model Options
BMR Type
Rel. Dev.
BMRF
0.1
Tail Probability
-
Confidence Level
0.95
Distribution Type
Normal
Variance Type
Constant
Model Data
Dependent Variable
[Dose]
Independent Variable
[Mean]
Total # of Observations
6
Adverse Direction
Automatic
Figure C-7. User Input for dose-response modeling of increased relative liver
weight in male rats fNTP. 20181.
This document is a draft for review purposes only and does not constitute Agency policy.
C-41 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Model Results
Benchmark Dose
BMD
0.207847359
BMDL
0.170963922
BMDU
0.269772648
AIC
291.4312778
Test 4 P-value
0.277392913
D.O.F.
3
Model Parameters
# of Parameters
5
Variable
Estimate
g
36.19093843
V
106.3618737
k
5.900597337
n
Bounded
alpha
6.592795984
Goodness of Fit
Dose
Size
Estimated
Median
Calc'd
Median
Observed
Mean
Estimated
SD
Calc'd SD
Observed
SD
Scaled
Residual
0
10
36.19093843
35.5
35.5
2.56764405
3.07
3.07
-0.850950961
0.156
10
38.93050512
39.32
39.32
2.56764405
1.68
1.68
0.479696924
0.312
10
41.53248929
42.61
42.61
2.56764405
1.77
1.77
1.32704845
0.625
10
46.37792479
45.56
45.56
2.56764405
2.66
2.66
-1.007345743
1.25
10
54.78411826
54.77
54.77
2.56764405
2.15
2.15
-0.017387866
2.5
10
67.84400709
67.9
67.9
2.56764405
3.76
3.76
0.068960153
Likelihoods of Interest
# of
Model
Log Likelihood*
Parameters
AIC
A1
-139.7874356
7
293.574871
A2
-134.7721348
12
293.54427
A3
-139.7874356
7
293.574871
fitted
-141.7156389
4
291.431278
R
-229.7698577
2
463.539715
* Includes additive constant of -55.13631. This constant was not included in the LL derivation prior to BMDS 3.0.
Tests of Interest
-2*Log(Likelihood
Test
Ratio)
Test df
p-value
1
189.9954459
10
<0.0001
2
10.03060162
5
0.07437279
3
10.03060162
5
0.07437279
4
3.856406652
3
0.27739291
Figure C-8. Model Results for increased relative liver weight in male rats (NTP.
2018).
This document is a draft for review purposes only and does not constitute Agency policy.
C-42 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-24. Benchmark dose results for increased relative liver weight in male
rats—constant variance, BMR = 1 standard deviation fNTP. 20181
Models
Restriction
10% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.3381
0.2930
<0.0001
314.8501
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Exponential 3
(CV—normal)
Restricted
0.3381
0.2930
<0.0001
314.8501
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Exponential 4
(CV—normal)
Restricted
0.1486
0.1209
0.2654
291.5391
Viable—Alternate
Exponential 5
(CV—normal)
Restricted
0.1485
0.1209
0.2653
291.5398
Viable—Alternate
Hill
(CV—normal)
Restricted
0.1460
0.1169
0.2774
291.4313
Viable-
Recommended
Lowest AIC
Polynomial
(5 degree)
(CV—normal)
Restricted
0.2202
0.1909
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Polynomial
(4 degree)
(CV—normal)
Restricted
0.2202
0.1976
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Polynomial
(3 degree)
(CV—normal)
Restricted
0.2202
0.1894
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Polynomial
(2 degree)
(CV—normal)
Restricted
0.2202
0.1894
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Power
(CV—normal)
Restricted
0.2202
0.1894
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Linear
(CV—normal)
Unrestricted
0.2202
0.1894
0.0115
298.5321
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
This document is a draft for review purposes only and does not constitute Agency policy.
C-43 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
C.2.6. INCREASED RELATIVE LIVER WEIGHT-FEMALE RAT fNTP. 20181
Table C-25. Dose-response data for increased relative liver weight in female
rats fNTP. 20181
Dose (mg/kg-d)
n
Mean
SD
0
10
33.52
2.37
0.156
10
37.66
2.81
0.312
10
40.08
1.77
0.625
10
44.25
2.59
1.25
10
50.84
2.12
2.5
10
67.75
2.85
Table C-26. Benchmark dose results for increased relative liver weight in
female rats—BMR = constant variance, 10% relative deviation (NTP. 2018)
Models
Restriction
10% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.3761
0.3585
0.0005
297.3583
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Exponential 3
(CV—normal)
Restricted
0.3761
0.3585
0.0005
297.3583
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Exponential 4
(CV—normal)
Restricted
0.2457
0.2042
0.0512
287.1715
Questionable
Goodness of fit p-value < 0.1
Exponential 5
(CV—normal)
Restricted
0.2456
0.2042
0.0512
287.1717
Questionable
Goodness of fit p-value < 0.1
Hill
(CV—normal)
Restricted
0.2446
0.2018
0.0518
287.1453
Questionable
Goodness of fit p-value < 0.1
Polynomial
(5 degree)
(CV—normal)
Restricted
0.2688
0.2545
0.0764
285.8573
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Polynomial
(4 degree)
(CV—normal)
Restricted
0.2688
0.2528
0.0764
285.8573
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Polynomial
(3 degree)
(CV—normal)
Restricted
0.2688
0.2524
0.0764
285.8573
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Polynomial
(2 degree)
(CV—normal)
Restricted
0.2688
0.2524
0.0764
285.8573
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Power
(CV—normal)
Restricted
0.2688
0.2524
0.0764
285.8573
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Linear
(CV—normal)
Unrestricted
0.2688
0.2524
0.0764
285.8573
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
This document is a draft for review purposes only and does not constitute Agency policy.
C-44 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-27. Benchmark dose results for increased relative liver weight in
female rats—nonconstant variance, BMR = 10% relative deviation fNTP. 20181
Models
Restriction
10% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Non-constant variance
Exponential 2
(NCV—
normal)
Restricted
0.3779
0.3586
0.0005
299.1741
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Exponential 3
(NCV—
normal)
Restricted
0.3779
0.3586
0.0005
299.1741
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Exponential 4
(NCV—
normal)
Restricted
0.2443
0.2017
0.0468
289.1376
Questionable
Goodness of fit p-value < 0.1
Exponential 5
(NCV—
normal)
Restricted
0.2464
0.2016
0.0466
289.1432
Questionable
Goodness of fit p-value < 0.1
Hill (NCV—
normal)
Restricted
0.2431
0.1997
0.0474
289.1075
Questionable
Goodness of fit p-value < 0.1
Polynomial
(5 degree)
(NCV—
normal)
Restricted
0.2688
0.2519
0.0695
287.8570
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Polynomial
(4 degree)
(NCV—
normal)
Restricted
0.2688
0.2519
0.0695
287.8570
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Polynomial
(3 degree)
(NCV—
normal)
Restricted
0.2688
0.2521
0.0695
287.8570
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Polynomial
(2 degree)
(NCV—
normal)
Restricted
0.2688
0.2521
0.0695
287.8570
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Power (NCV—
normal)
Restricted
0.2688
0.2521
0.0695
287.8570
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
Linear (NCV—
normal)
Unrestricted
0.2688
0.2521
0.0695
287.8570
Questionable
Goodness of fit p-value < 0.1
| Residual at control | > 2
This document is a draft for review purposes only and does not constitute Agency policy.
C-45 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-28. Benchmark dose results for increased relative liver weight in
female rats—log-normal, constant variance, BMR = 10% relative deviation
fNTP. 20181
Models
Restriction
10% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—log-
normal)
Restricted
0.3617
0.3404
<0.0001
304.9243
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near
BMD| >2
| Residual at control | > 2
Exponential 3
(CV- log-
normal)
Restricted
0.3617
0.3404
<0.0001
304.9243
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near
BMD| >2
| Residual at control | > 2
Exponential 4
(CV- log-
normal)
Restricted
0.2228
0.1850
<0.0001
291.5746
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near
BMD| >2
| Residual at control | > 2
Exponential 5
(CV- log-
normal)
Restricted
0.2228
0.1850
<0.0001
291.5746
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near
BMD| >2
| Residual at control | > 2
Hill (CV— log-
normal)
Restricted
0.2200
0.1800
<0.0001
291.4503
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near
BMD| >2
| Residual at control | > 2
Polynomial
(5 degree)
(CV- log-
normal)
Restricted
0.2622
0.2441
<0.0001
291.8437
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near
BMD| >2
| Residual at control | > 2
Polynomial
(4 degree)
(CV- log-
normal)
Restricted
0.2622
0.2454
<0.0001
291.8437
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near
BMD| >2
| Residual at control | > 2
This document is a draft for review purposes only and does not constitute Agency policy.
C-46 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
10% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Polynomial
(3 degree)
(CV- log-
normal)
Restricted
0.2622
0.2433
<0.0001
291.8437
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near
BMD| >2
| Residual at control | > 2
Polynomial
(2 degree)
(CV- log-
normal)
Restricted
0.2622
0.2433
<0.0001
291.8437
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near
BMD| >2
| Residual at control | > 2
Power (CV—
log-normal)
Restricted
0.2622
0.2433
<0.0001
291.8437
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near
BMD| >2
| Residual at control | > 2
Linear (CV—
log-normal)
Unrestricted
0.2622
0.2433
<0.0001
291.8437
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near
BMD| >2
| Residual at control | > 2
Table C-29. Benchmark dose results for increased relative liver weight in
female rats, high dose dropped—BMR = constant variance, 10% relative
deviation fNTP. 20181
Models
Restriction
10% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.3195
0.2902
0.0031
242.3745
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Exponential 3
(CV—normal)
Restricted
0.3195
0.2902
0.0031
242.3745
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Exponential 4
(CV—normal)
Restricted
0.1611
0.1214
0.5849
231.5654
Viable-
Alternate
Exponential 5
(CV—normal)
Restricted
0.1610
0.1214
0.5849
231.5654
Viable-
Alternate
Hill
(CV—normal)
Restricted
0.1544
0.1117
0.6566
231.3342
Viable-
Recommended
Lowest AIC
This document is a draft for review purposes only and does not constitute Agency policy.
C-47 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
10% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Polynomial
(5 degree)
(CV—normal)
Restricted
0.2659
0.2374
0.0308
237.3809
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Polynomial
(4 degree)
(CV—normal)
Restricted
0.2659
0.2374
0.0308
237.3809
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Polynomial
(3 degree)
(CV—normal)
Restricted
0.2659
0.2374
0.0308
237.3809
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Polynomial
(2 degree)
(CV—normal)
Restricted
0.2659
0.2374
0.0308
237.3809
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Power
(CV—normal)
Restricted
0.2659
0.2374
0.0308
237.3809
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Linear
(CV—normal)
Unrestricted
0.3195
0.2902
0.0031
242.3745
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Frequentist Hill Model with BMR of 0.1 Rel. Dev. for the BMD and 0.95 Lower Confidence
Limit for the BMDL
55
Dose
Figure C-9. Dose-response curve for the Hill model fit to increased relative
liver weight in female rats with the highest dose dropped fNTP. 20181.
This document is a draft for review purposes only and does not constitute Agency policy.
C-48 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
User Input
Info
Model
frequentist Hill vl.l
Dataset Name
LiverWt Rel F NTP hdd
User notes
[Add user notes here]
Dose-Response Model
M[dose] = g + v*doseAn/(kAn + doseAn)
Variance Model
Var[i] = alpha
Model Options
BMR Type
Rel. Dev.
BMRF
0.1
Tail Probability
-
Confidence Level
0.95
Distribution Type
Normal
Variance Type
Constant
Model Data
Dependent Variable
[Custom]
Independent Variable
[Custom]
Total # of Observations
5
Adverse Direction
Automatic
Figure C-10. User input for dose-response modeling of increased relative liver
weight in females rats with highest dose dropped fNTP. 20181.
This document is a draft for review purposes only and does not constitute Agency policy.
C-49 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Model Results
Benchmark Dose
BMD
0.154369377
BMDL
0.111740633
BMDU
0.218901711
AIC
231.3341743
Test 4 P-value
0.656565161
D.O.F.
2
Model Parameters
# of Parameters
5
Variable
Estimate
g
33.78210999
V
38.98056451
k
1.626870887
n
Bounded
alpha
5.097775081
Goodness of Fit
Dose
Size
Estimated
Median
Calc'd
Median
Observed
Mean
Estimated
SD
Calc'd SD
Observed
SD
Scaled
Residual
0
10
33.78210999
33.52
33.52
2.2578253
2.37
2.37
-0.367107486
0.156
10
37.19288309
37.66
37.66
2.2578253
2.81
2.81
0.654237225
0.312
10
40.05480005
40.08
40.08
2.2578253
1.77
1.77
0.035294693
0.625
10
44.60104858
44.25
44.25
2.2578253
2.59
2.59
-0.491673589
1.25
10
50.71915985
50.84
50.84
2.2578253
2.12
2.12
0.169246978
Likelihoods of Interest
# of
Model
Log Likelihood*
Pa ra meters
AIC
A1
-111.2463538
6
234.492708
A2
-110.0141933
10
240.028387
A3
-111.2463538
6
234.492708
fitted
-111.6670871
4
231.334174
R
-163.1738575
2
330.347715
* Includes additive constant of -45.94693. This constant was not included in the LL derivation prior to BMDS 3.0.
Tests of Interest
-2*Log(Likelihood
Test
Ratio)
Test df
p-value
1
106.3193285
8
<0.0001
2
2.464321029
4
0.65103586
3
2.464321029
4
0.65103586
4
0.84146667
2
0.65656516
Figure C-ll. Model results for increased relative liver weight in female rats
with highest dose dropped fNTP. 20181.
This document is a draft for review purposes only and does not constitute Agency policy.
C-50 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-30. Benchmark dose results for increased relative liver weight in
female rats, high dose dropped—constant variance, BMR = 1 standard
deviation fNTP. 20181
Models
Restriction
10% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.2341
0.1980
0.0031
242.3745
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Exponential 3
(CV—normal)
Restricted
0.2341
0.1980
0.0031
242.3745
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Exponential 4
(CV—normal)
Restricted
0.1050
0.0785
0.5849
231.5654
Viable-
Alternate
Exponential 5
(CV—normal)
Restricted
0.1049
0.0785
0.5849
231.5654
Viable-
Alternate
Hill
(CV—normal)
Restricted
0.1000
0.0722
0.6566
231.3342
Viable-
Recommended
Lowest AIC
Polynomial
(5 degree)
(CV—normal)
Restricted
0.1854
0.1675
0.0308
237.3809
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Polynomial
(4 degree)
(CV—normal)
Restricted
0.1854
0.1553
0.0308
237.3809
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Polynomial
(3 degree)
(CV—normal)
Restricted
0.1854
0.1553
0.0308
237.3809
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Polynomial
(2 degree)
(CV—normal)
Restricted
0.1854
0.1553
0.0308
237.3809
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Power
(CV—normal)
Restricted
0.1854
0.1553
0.0308
237.3809
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Linear
(CV—normal)
Unrestricted
0.2341
0.1980
0.0031
242.3745
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
C.2.7. INCREASED RELATIVE LIVER WEIGHT (HISTO)—FEMALE RATS fFrawlev etal.. 20181
Table C-31. Dose-response data for increased relative liver weight (Histo) in
female rats fFrawlev etal.. 20181
Dose (mg/kg-d)
n
Mean
SD
0
8
4.02
0.28
0.125
8
4.06
0.28
0.25
8
4.35
0.28
0.5
8
4.68
0.34
This document is a draft for review purposes only and does not constitute Agency policy.
C-51 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-32. Benchmark dose results for increased relative liver weight (Histo)
in female rats—constant variance, BMR = 10% relative deviation fFrawlev et
al.. 20181
Models
Restriction
10% Relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.2929
0.2224
0.6024
15.6701
Viable-
Recommended
Lowest AIC
Exponential 3
(CV—normal)
Restricted
0.3215
0.2240
0.3551
17.5116
Viable—Alternate
Exponential 4
(CV—normal)
Restricted
0.2823
0.1647
0.2944
17.7557
Viable—Alternate
Exponential 5
(CV—normal)
Restricted
0.2729
0.1840
NA
18.6564
Questionable
d.f. = 0, saturated model
(Goodness of fit test
cannot be calculated)
Hill
(CV—normal)
Restricted
0.2777
0.1901
NA
18.6564
Questionable
d.f. = 0, saturated model
(Goodness of fit test
cannot be calculated)
Polynomial
(3 degree)
(CV—normal)
Restricted
0.3170
0.2099
0.3338
17.5904
Viable—Alternate
Polynomial
(2 degree)
(CV—normal)
Restricted
0.3170
0.2099
0.3338
17.5904
Viable—Alternate
Power
(CV—normal)
Restricted
0.3195
0.2113
0.3675
17.4686
Viable—Alternate
Linear
(CV—normal)
Unrestricted
0.2824
0.2081
0.5775
15.7543
Viable—Alternate
This document is a draft for review purposes only and does not constitute Agency policy.
C-52 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Frequentist Exponential Degree 2 Model with BMR of 0.1 Rel. Dev. for the BMD and 0.95
Lower Confidence Limit for the BMDL
5.5
Dose
Figure C-12. Dose-response curve for the Exponential 2 model fit to increased
relative liver weight (Histo) in female rats (Frawlev et al.. 2018).
User Input
Info
Model
frequentist Exponential degree 2 vl.l
Dataset Name
LiverWt_Rel_Frawley_Histo
User notes
[Add user notes here]
Dose-Response Model
M[dose] = a * exp(±l * b * dose)
Variance Model
Var[i] = alpha
Model Options
BMRType
Rel. Dev.
BMRF
0.1
Tail Probability
-
Confidence Level
0.95
Distribution Type
Normal
VarianceType
Constant
Model Data
Dependent Variable
[Dose]
Independent Variable
[Mean]
Total # of Observations
4
Adverse Direction
Automatic
Figure C-13. User input for dose-response modeling of increased relative liver
weight (Histo) in female rats (Frawlev et al.. 2018).
This document is a draft for review purposes only and does not constitute Agency policy.
C-53 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Model Results
Benchmark Dose
BMD
0.292874336
BMDL
0.222375421
BMDU
0.429901615
AIC
15.67013988
Test 4 P-value
0.602376128
D.O.F.
2
Model Parameters
# of Parameters
3
Variable
Estimate
a
3.97629556
b
0.325430373
log-alpha
-2.536765652
Goodness of Fit
Dose
Size
Estimated
Median
Calc'd
Median
Observed
Mean
Estimated
SD
Calc'd SD
Observed
SD
Scaled
Residual
0
8
3.97629556
4.02
4.02
0.28128614
0.28
0.28
0.439462902
0.125
8
4.141381462
4.06
4.06
0.28128614
0.28
0.28
-0.818318074
0.25
8
4.31332132
4.35
4.35
0.28128614
0.28
0.28
0.368816506
0.5
8
4.678912956
4.68
4.68
0.28128614
0.34
0.34
0.01093059
Likelihoods of Interest
# of
Model
Log Likelihood*
Parameters
AIC
Al
-4.328196707
5
18.6563934
A2
-4.087877276
8
24.1757546
A3
-4.328196707
5
18.6563934
fitted
-4.835069939
3
15.6701399
R
-14.72410737
2
33.4482147
* Includes additive constant of-29.40603. This constant was not included in the LL derivation prior to BMDS 3.0.
Tests of Interest
Test
-2*Log( Likelihood
Ratio)
Test df
p-value
1
21.2724602
6
0.00163883
2
0.480638862
3
0.92312391
3
0.480638862
3
0.92312391
4
1.013746464
2
0.60237613
Figure C-14. Model results for increased relative liver weight (Histo) in female
rats (Frawlev et al.. 2018).
Table C-33. Benchmark dose results for increased relative liver weight (Histo)
in female rats—constant variance, BMR = 1 standard deviation fFrawlev et al..
20181
Models
Restriction
1 standard deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.2100
0.1561
0.6024
15.6701
Viable-
Recommended
Lowest AIC
Exponential 3
(CV—normal)
Restricted
0.2405
0.1572
0.3551
17.5116
Viable—Alternate
Exponential 4
(CV—normal)
Restricted
0.2003
0.1453
0.2944
17.7557
Viable—Alternate
This document is a draft for review purposes only and does not constitute Agency policy.
C-54 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 5
(CV—normal)
Restricted
0.2332
0.1314
NA
18.6564
Questionable
d.f. = 0, saturated model
(Goodness of fit test
cannot be calculated)
Hill
(CV—normal)
Restricted
0.2310
0.1312
NA
18.6564
Questionable
d.f. = 0, saturated model
(Goodness of fit test
cannot be calculated)
Polynomial
(3 degree)
(CV—normal)
Restricted
0.2343
0.1467
0.3338
17.5904
Viable—Alternate
Polynomial
(2 degree)
(CV—normal)
Restricted
0.2343
0.1467
0.3338
17.5904
Viable—Alternate
Power
(CV—normal)
Restricted
0.2394
0.1476
0.3675
17.4686
Viable—Alternate
Linear
(CV—normal)
Unrestricted
0.2005
0.1455
0.5775
15.7543
Viable—Alternate
C.2.8. INCREASED RELATIVE LIVER WEIGHT (MPS)—FEMALE RATS fFrawlev etal.. 20181
Table C-34. Dose-response data for increased relative liver weight (MPS) in
female rats (Frawlev etal.. 2018)
Dose (mg/kg-d)
n
Mean
SD
0
8
3.42
0.26
0.125
8
3.77
0.28
0.25
8
3.86
0.26
0.5
8
4.19
0.17
Table C-35. Benchmark dose results for increased relative liver weight (Histo)
in female rats—constant variance, BMR = 10% relative deviation fFrawlev et
al.. 20181
Models
Restriction
10% Relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.2575
0.2036
0.2714
5.4499
Viable—Alternate
Exponential 3
(CV—normal)
Restricted
0.2575
0.2044
0.2714
5.4499
Viable—Alternate
Exponential 4
(CV—normal)
Restricted
0.1644
0.0852
0.3121
5.8634
Viable—Alternate
This document is a draft for review purposes only and does not constitute Agency policy.
C-55 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
10% Relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Exponential 5
(CV—normal)
Restricted
0.1646
0.0851
0.3121
5.8634
Viable—Alternate
Hill
(CV—normal)
Restricted
0.1587
0.0730
0.3336
5.7766
Viable—Alternate
Polynomial
(3 degree)
(CV—normal)
Restricted
0.2419
0.1864
0.3283
5.0691
Viable—Alternate
Polynomial
(2 degree)
(CV—normal)
Restricted
0.2419
0.1864
0.3283
5.0691
Viable—Alternate
Power
(CV—normal)
Restricted
0.2419
0.1864
0.3283
5.0691
Viable—Alternate
Linear
(CV—normal)
Unrestricted
0.2419
0.1864
0.3283
5.0691
Viable-
Recommended
Lowest AIC
Frequentist Linear Model with BMR of 0.1 Rel. Dev. for the BMD and 0.95 Lower
Confidence Limit for the BMDL
4.5
0.25
Dose
Estimated Probability
Response at BMD
O Data
BMD
BMDL
Figure C-15. Dose-response curve for the Linear model fit to increased relative
liver weight (MPS) in female rats fFrawlev et al.. 20181.
This document is a draft for review purposes only and does not constitute Agency policy.
C-56 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
User Input
Info
Model
frequentist Linear vl.l
Dataset Name
LiverWt_Rel_Frawley_MPS
User notes
[Add user notes here]
Dose-Response Model
M[dose] = g + bl*dose
Variance Model
Var[i] = alpha
Model Options
BMRType
Rel. Dev.
BMRF
0.1
Tail Probability
-
Confidence Level
0.95
Distribution Type
Normal
VarianceType
Constant
Model Data
Dependent Variable
[Dose]
Independent Variable
[Mean]
Total # of Observations
4
Adverse Direction
Automatic
Figure C-16. User input for dose-response modeling of increased relative liver
weight (MPS) in female rats fFrawlev et al.. 20181.
This document is a draft for review purposes only and does not constitute Agency policy.
C-57 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Model Results
Benchmark Dose
BMD
0.24187088
BMDL
0.186409723
BMDU
0.337253407
AIC
5.069125072
Test 4 P-value
0.328309463
D.O.F.
2
Model Parameters
# of Parameters
3
Variable
Estimate
g
3.49399996
betal
1.444571613
alpha
0.05687116
Goodness of Fit
Dose
Size
Estimated
Median
Calc'd
Median
Observed
Mean
Estimated
SD
Calc'd SD
Observed
SD
Scaled
Residual
0
8
3.49399996
3.42
3.42
0.23847675
0.26
0.26
-0.877668349
0.125
8
3.674571412
3.77
3.77
0.23847675
0.28
0.28
1.13182022
0.25
8
3.855142863
3.86
3.86
0.23847675
0.26
0.26
0.057607531
0.5
8
4.216285767
4.19
4.19
0.23847675
0.17
0.17
-0.31175943
Likelihoods of Interest
# of
Model
Log Likelihood*
Pa ra meters
AIC
Al
1.579236095
5
6.84152781
A2
2.643027712
8
10.7139446
A3
1.579236095
5
6.84152781
fitted
0.465437464
3
5.06912507
R
-12.53902329
2
29.0780466
* Includes additive constant of -29.40603. This constant was not included in the LL derivation prior to BMDS 3.0.
Tests of Interest
Test
-2*Log(Likelihood
Ratio)
Test df
p-value
1
30.364102
6
<0.0001
2
2.127583234
3
0.54635267
3
2.127583234
3
0.54635267
4
2.227597262
2
0.32830946
Figure C-17. Model results for increased relative liver weight (MPS) in female
rats fFrawlev et al.. 20181.
Table C-36. Benchmark dose results for increased relative liver weight (MPS)
in female rats — constant variance, BMR = 1 standard deviation (Frawlev et al..
2018)
Models
Restriction
1 standard deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.1788
0.1367
0.2714
5.4499
Viable—Alternate
Exponential 3
(CV—normal)
Restricted
0.1788
0.1367
0.2714
5.4499
Viable—Alternate
This document is a draft for review purposes only and does not constitute Agency policy.
C-58 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 4
(CV—normal)
Restricted
0.1046
0.0549
0.3121
5.8634
Viable—Alternate
Exponential 5
(CV—normal)
Restricted
0.1048
0.0549
0.3121
5.8634
Viable—Alternate
Hill
(CV—normal)
Restricted
0.0994
0.0450
0.3336
5.7766
Viable-
Recommended
Lowest BMDL
Polynomial
(3 degree)
(CV—normal)
Restricted
0.1651
0.1238
0.3283
5.0691
Viable—Alternate
Polynomial
(2 degree)
(CV—normal)
Restricted
0.1651
0.1238
0.3283
5.0691
Viable—Alternate
Power
(CV—normal)
Restricted
0.1651
0.1238
0.3283
5.0691
Viable—Alternate
Linear
(CV—normal)
Unrestricted
0.1651
0.1238
0.3283
5.0691
Viable—Alternate
C.2.9. INCREASED RELATIVE LIVER WEIGHT (TDAR)—FEMALE RATS fFrawlev etal.. 20181
Table C-37. Dose-response data for increased relative liver weight (TDAR) in
female rats (Frawlev etal.. 2018)
Dose (mg/kg-d)
n
Mean
SD
0
8
3.85
0.14
0.125
8
3.94
0.11
0.25
8
4.6
0.37
0.5
8
5.21
0.28
Table C-38. Benchmark dose results for increased relative liver weight (TDAR)
in female rats—constant variance, BMR = 10% relative deviation fFrawlev et
al.. 20181
Models
Restriction
10% Relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.1478
0.1295
0.0284
10.5539
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Modeled control response
std. dev. >11.51 actual
response std. dev.
This document is a draft for review purposes only and does not constitute Agency policy.
C-59 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
10% Relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 3
(CV—normal)
Restricted
0.1541
0.1297
0.0077
12.5248
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Modeled control response
std. dev. >11.51 actual
response std. dev.
Exponential 4
(CV—normal)
Restricted
0.1294
0.0935
0.0073
12.6257
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
Modeled control response
std. dev. >11.51 actual
response std. dev.
Exponential 5
(CV—normal)
Restricted
0.1951
0.1458
NA
7.4299
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response
std. dev. >11.51 actual
response std. dev.
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Hill
(CV—normal)
Restricted
0.1904
0.1497
NA
7.4299
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response
std. dev. >11.51 actual
response std. dev.
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Polynomial
(3 degree)
(CV—normal)
Restricted
0.1419
0.1108
0.0079
12.4766
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Modeled control response
std. dev. >11.51 actual
response std. dev.
Polynomial
(2 degree)
(CV—normal)
Restricted
0.1419
0.1108
0.0079
12.4766
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Modeled control response
std. dev. >11.51 actual
response std. dev.
Power
(CV—normal)
Restricted
0.1556
0.1124
0.0103
12.0114
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Modeled control response
std. dev. >11.51 actual
response std. dev.
This document is a draft for review purposes only and does not constitute Agency policy.
C-60 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
10% Relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Linear
(CV—normal)
Unrestricted
0.1295
0.1103
0.0274
10.6256
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
Modeled control response
std. dev. >11.51 actual
response std. dev.
Table C-39. Benchmark dose results for increased relative liver weight (TDAR)
in female rats—non-constant variance, BMR = 10% relative deviation (Frawlev
etal.. 20181
Models
Restriction
10% Relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Non-constant variance
Exponential 2
(NCV—
normal)
Restricted
0.1478
0.1284
0.0012
10.0543
Questionable
Goodness of fit p-value <0.1
| Residual for Dose Group
Near BMD| >2
Exponential 3
(NCV—
normal)
Restricted
0.1607
0.1292
0.0003
11.8202
Questionable
Goodness of fit p-value <0.1
| Residual for Dose Group
Near BMD| >2
Exponential 4
(NCV—
normal)
Restricted
0.1333
0.1030
0.0002
12.4411
Questionable
Goodness of fit p-value <0.1
| Residual for Dose Group
Near BMD| >2
Exponential 5
(NCV—
normal)
Restricted
0.1937
0.1654
NA
0.5572
Questionable
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Hill (NCV—
normal)
Restricted
0.1880
0.1653
NA
0.5577
Questionable
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Polynomial
(3 degree)
(NCV—
normal)
Restricted
0.1507
0.1144
0.0002
11.9784
Questionable
Goodness of fit p-value <0.1
| Residual for Dose Group
Near BMD| >2
Polynomial
(2 degree)
(NCV—
normal)
Restricted
0.1507
0.1144
0.0002
11.9784
Questionable
Goodness of fit p-value <0.1
| Residual for Dose Group
Near BMD| >2
Power (NCV—
normal)
Restricted
0.1628
0.1183
0.0004
11.0771
Questionable
Goodness of fit p-value <0.1
Linear (NCV—
normal)
Unrestricted
0.1334
0.1127
0.0010
10.4397
Questionable
Goodness of fit p-value <0.1
| Residual for Dose Group
Near BMD| >2
This document is a draft for review purposes only and does not constitute Agency policy.
C-61 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-40. Benchmark dose results for increased relative liver weight (TDAR)
in female rats—log-normal, constant variance, BMR = 10% relative deviation
fFrawlev et al.. 20181
Models
Restriction
10% Relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Log-normal Constant variance
Exponential 2
(CV—log-
normal)
Restricted
0.1478
0.1295
0.0172
7.4633
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response
std. dev.
Exponential 3
(CV- log-
normal)
Restricted
0.1639
0.1304
0.0050
9.2051
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response
std. dev.
Exponential 4
(CV- log-
normal)
Restricted
0.1315
0.1026
0.0033
9.9692
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response
std. dev.
Exponential 5
(CV- log-
normal)
Restricted
0.1644
0.1111
NA
10.6210
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response
std. dev.
d.f. = 0, saturated model
(Goodness of fit test cannot be
calculated)
Hill (CV— log-
normal)
Restricted
0.1918
0.1692
NA
3.3425
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response
std. dev.
d.f. = 0, saturated model
(Goodness of fit test cannot be
calculated)
Polynomial
(3 degree)
(CV- log-
normal)
Restricted
0.1541
0.1143
0.0046
9.3729
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response
std. dev.
This document is a draft for review purposes only and does not constitute Agency policy.
C-62 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
10% Relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Log-normal Constant variance
Polynomial
(2 degree)
(CV- log-
normal)
Restricted
0.1541
0.1143
0.0046
9.3729
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response
std. dev.
Power (CV—
log-normal)
Restricted
0.1649
0.1176
0.0070
8.6207
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response
std. dev.
Linear (CV—
log-normal)
Unrestricted
0.1315
0.1122
0.0134
7.9687
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response
std. dev.
C.2.10. DECREASED FETAL WEIGHT-MALE AND FEMALE RATS fHarris and Birnhaum. 19891
Table C-41. Dose-response data for decreased fetal weight in male and female
rats (Harris and Birnbaum. 1989)
Dose (mg/kg-d)
n
Mean
SD
0
86.4
1.17
0.09
0.03
85.8
1.16
0.02
0.1
94.8
1.13
0.2
0.3
102
1.16
0.3
1
103.6
1.12
0.2
3
87.6
1.1
0.09
6.4
75.4
0.9
0.26
12.8
32.2
0.59
0.11
This document is a draft for review purposes only and does not constitute Agency policy.
C-63 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-42. Benchmark dose results for decreased fetal weight in male and
female rats—constant variance, BMR = 5% relative deviation fHarris and
Birnbaum. 19891
Models
Restriction
5% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
1.1862
1.0702
0.0010
-303.6182
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Exponential 3
(CV—normal)
Restricted
2.4486
1.8922
0.3529
-318.5263
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Exponential 4
(CV—normal)
Restricted
1.1862
1.0702
0.0010
-303.6182
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Exponential 5
(CV—normal)
Restricted
3.0401
2.0145
0.3470
-317.6098
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Hill
(CV—normal)
Restricted
3.0451
2.0215
0.3383
-317.5367
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Polynomial
(7 degree)
(CV—normal)
Restricted
1.9190
1.4664
0.1942
-316.6978
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Polynomial
(6 degree)
(CV—normal)
Restricted
1.9190
1.4668
0.1942
-316.6978
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Polynomial
(5 degree)
(CV—normal)
Restricted
1.9190
1.4667
0.1942
-316.6978
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
This document is a draft for review purposes only and does not constitute Agency policy.
C-64 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
5% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Polynomial
(4 degree)
(CV—normal)
Restricted
1.9190
1.4667
0.1942
-316.6978
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Polynomial
(3 degree)
(CV—normal)
Restricted
1.9190
1.4681
0.1942
-316.6978
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Polynomial
(2 degree)
(CV—normal)
Restricted
1.9190
1.4884
0.1942
-316.6978
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Power
(CV—normal)
Restricted
2.1795
1.6300
0.2568
-317.5277
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Linear
(CV—normal)
Unrestricted
1.3815
1.2741
0.0441
-313.1368
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Table C-43. Benchmark dose results for decreased fetal weight in male and
female rats—nonconstant variance, BMR = 5% relative deviation fHarris and
Birnbaum. 19891
Models
Restriction
5% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Non-constant variance
Exponential 2
(NCV— normal)
Restricted
1.2032
1.0775
0.0012
-302.0911
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Exponential 3
(NCV— normal)
Restricted
2.4989
1.9388
0.4468
-317.3295
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
This document is a draft for review purposes only and does not constitute Agency policy.
C-65 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
5% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Non-constant variance
Exponential 4
(NCV— normal)
Restricted
1.2031
1.0775
0.0012
-302.0911
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Exponential 5
(NCV— normal)
Restricted
2.4942
1.9392
0.3140
-315.3322
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Hill (NCV—
normal)
Restricted
2.9282
1.9155
0.3696
-315.8031
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Polynomial
(7 degree)
(NCV— normal)
Restricted
1.9751
1.6128
0.2753
-315.7500
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Polynomial
(6 degree)
(NCV— normal)
Restricted
1.9716
1.4955
0.2749
-315.7461
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Polynomial
(5 degree)
(NCV— normal)
Restricted
1.9712
1.4921
0.2749
-315.7460
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Polynomial
(4 degree)
(NCV— normal)
Restricted
1.9751
1.4965
0.2753
-315.7500
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Polynomial
(3 degree)
(NCV— normal)
Restricted
1.9751
1.4973
0.2753
-315.7500
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Polynomial
(2 degree)
(NCV— normal)
Restricted
1.9751
1.5263
0.2753
-315.7500
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
This document is a draft for review purposes only and does not constitute Agency policy.
C-66 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
5% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Non-constant variance
Power (NCV—
normal)
Restricted
2.2422
1.6842
0.3562
-316.5655
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Linear (NCV—
normal)
Unrestricted
1.3772
1.2719
0.0450
-311.2042
Questionable
Nonconstant variance test failed
(Test 3 p-value < 0.05)
Goodness of fit p-value <0.1
Modeled control response std.
dev. >| 1.51 actual response std.
dev.
Table C-44. Benchmark dose results for decreased fetal weight in male and
female rats—log-normal, constant variance, BMR = 5% relative deviation
fHarris and Birnbaum. 19891
Models
Restriction
5% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Log-normal, constant variance
Exponential 2
(CV—log-
normal)
Restricted
1.0479
0.9755
<0.0001
-307.8546
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Modeled control response
std. dev. >11.51 actual
response std. dev.
Exponential 3
(CV- log-
normal)
Restricted
2.1631
1.7042
0.0286
-326.0092
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Modeled control response
std. dev. >11.51 actual
response std. dev.
Exponential 4
(CV- log-
normal)
Restricted
1.0479
0.9755
<0.0001
-307.8546
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Modeled control response
std. dev. >11.51 actual
response std. dev.
Exponential 5
(CV- log-
normal)
Restricted
3.4280
2.4438
0.1216
-329.2234
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response
std. dev. >11.51 actual
response std. dev.
This document is a draft for review purposes only and does not constitute Agency policy.
C-67 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
5% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Log-normal, constant variance
Hill (CV— log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(7 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(6 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(5 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(4 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(3 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(2 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Power (CV—
log-normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Linear (CV—
log-normal)
Unrestricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
This document is a draft for review purposes only and does not constitute Agency policy.
C-68 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
C.2.11. DECREASED SPERM COUNT-MALE RATS fNTP. 20181
Table C-45. Dose-response data for decreased sperm counts in male
rats fNTP. 20181
Dose (mg/kg-d)
n
Mean
SD
0
10
136.3
32.26
0.625
10
120.8
17.39
1.25
10
112.9
23.09
2.5
10
95.7
36.37
Table C-46. Benchmark dose results for decreased sperm counts in male rats,
BMR = 1 standard deviation (NTP. 2018)
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
1.5928
0.9634
0.9331
382.8116
Viable-
Recommended
Lowest AIC
Exponential 3
(CV—normal)
Restricted
1.5928
0.9634
0.9331
382.8116
Viable-
Recommended
Lowest AIC
Exponential 4
(CV—normal)
Restricted
1.4241
0.5083
0.8023
384.7359
Viable—Alternate
Exponential 5
(CV—normal)
Restricted
1.4241
0.5083
0.8023
384.7359
Viable—Alternate
Hill
(CV—normal)
Restricted
1.4208
0.4347
0.8120
384.7298
Viable—Alternate
Polynomial
(3 degree)
(CV—normal)
Restricted
1.7202
1.1328
0.8756
382.9388
Viable—Alternate
Polynomial
(2 degree)
(CV—normal)
Restricted
1.7202
1.1328
0.8756
382.9388
Viable—Alternate
Power
(CV—normal)
Restricted
1.7202
1.1329
0.8756
382.9388
Viable—Alternate
Linear
(CV—normal)
Unrestricted
1.7202
1.1328
0.8756
382.9388
Viable—Alternate
This document is a draft for review purposes only and does not constitute Agency policy.
C-69 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Frequentist Exponential Degree 2 Model with BMR of 1 Std. Dev. for the BMD and 0.95
Lower Confidence Limit for the BMDL
180
Dose
Figure C-18. Dose-response curve for the Exponential 2 model fit to decreased
sperm counts in male rats (NTP. 2018).
User Input
Info
Model
frequentist Exponential degree 2 vl.l
Dataset Name
Sperm_Count_NTP
User notes
[Add user notes here]
Dose-Response Model
M[dose] = a * exp(±l * b * dose)
Variance Model
Var[i] = alpha
Model Options
BMRType
Std. Dev.
BMRF
1
Tail Probability
-
Confidence Level
0.95
Distribution Type
Normal
Variance Type
Constant
Model Data
Dependent Variable
[Dose]
Independent Variable
[Mean]
Total # of Observations
4
Adverse Direction
Automatic
Figure C-19. User input for dose-response modeling of decreased sperm
counts in male counts (NTP. 2018).
This document is a draft for review purposes only and does not constitute Agency policy.
C-70 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Model Results
Benchmark Dose
BMD
1.592768431
BMDL
0.963412903
BMDU
3.624046063
AIC
382.8116246
Test 4 P-value
0.933123027
D.O.F.
2
Model Parameters
# of Parameters
3
Variable
Estimate
a
134.5572517
b
0.139886976
log-alpha
6.582413542
Goodness of Fit
Dose
Size
Estimated
Median
Calc'd
Median
Observed
Mean
Estimated
SD
Calc'd SD
Observed
SD
Scaled
Residual
0
10
134.5572517
136.3
136.3
26.8752764
32.26
32.26
0.205060364
0.625
10
123.2926024
120.8
120.8
26.8752764
17.39
17.39
-0.293291902
1.25
10
112.9709891
112.9
112.9
26.8752764
23.09
23.09
-0.008352922
2.5
10
94.84768903
95.7
95.7
26.8752764
36.37
36.37
0.100287116
Likelihoods of Interest
# of
Model
Log Likelihood*
Pa ra meters
AIC
A1
-188.3365941
5
386.673188
A2
-185.2790038
8
386.558008
A3
-188.3365941
5
386.673188
fitted
-188.4058123
3
382.811625
R
-193.5430425
2
391.086085
* Includes additive constant of -36.75754. This constant was not included in the LL derivation prior to BMDS 3.0.
Tests of Interest
Test
-2*Log(Likelihood
Ratio)
Test df
p-value
1
16.52807739
6
0.01118344
2
6.115180405
3
0.10613895
3
6.115180405
3
0.10613895
4
0.138436451
2
0.93312303
Figure C-20. Model results for decreased sperm counts in rat males (NTP.
2018).
C.2.12. DECREASED ABSOLUTE TESTIS WEIGHT IN MALE RATS fNTP. 20181
Table C-47. Dose-response data for decreased absolute testis weight in male
rats fNTP. 20181
Dose (mg/kg-d)
n
Mean
SD
0
9
1.777
0.17
0.156
10
1.797
0.15
0.312
10
1.742
0.12
0.625
10
1.74
0.1
1.25
10
1.695
0.11
2.5
10
1.553
0.2
This document is a draft for review purposes only and does not constitute Agency policy.
C-71 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-48. Benchmark dose results for decreased absolute testis weight in
male rats—constant variance, BMR = 1 standard deviation fNTP. 20181
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
1.4763
1.0220
0.9324
-59.4936
Viable—Alternate
Exponential 3
(CV—normal)
Restricted
1.7052
1.0373
0.8973
-57.7417
Viable—Alternate
Exponential 4
(CV—normal)
Restricted
1.4763
1.0220
0.9324
-59.4936
Viable—Alternate
Exponential 5
(CV—normal)
Restricted
1.7049
0.8202
0.7420
-55.7409
Viable—Alternate
Hill
(CV—normal)
Restricted
1.7088
0.8010
0.7448
-55.7486
Viable—Alternate
Polynomial
(5 degree)
(CV—normal)
Restricted
1.7976
1.0880
0.9114
-57.8041
Viable—Alternate
Polynomial
(4 degree)
(CV—normal)
Restricted
1.7750
1.0878
0.9107
-57.8008
Viable—Alternate
Polynomial
(3 degree)
(CV—normal)
Restricted
1.7482
1.0873
0.9089
-57.7926
Viable—Alternate
Polynomial
(2 degree)
(CV—normal)
Restricted
1.7214
1.0861
0.9046
-57.7738
Viable—Alternate
Power
(CV—normal)
Restricted
1.7089
1.0848
0.8995
-57.7514
Viable—Alternate
Linear
(CV—normal)
Unrestricted
1.5110
1.0742
0.9430
-59.5723
Viable-
Recommended
Lowest AIC
This document is a draft for review purposes only and does not constitute Agency policy.
C-72 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Frequentist Linear Model with BMR of 1 Std. Dev. for the BMD and 0.95 Lower
Confidence Limit for the BMDL
0 0.5 1 1.5 2 2.5
Dose
Figure C-21. Dose-response curve for the Linear model fit to decreased
absolute testis weight in male rats (NTP. 2018).
User Input
Info
Model
frequentist Linear vl.l
Dataset Name
TestisWt Abs NTP
User notes
[Add user notes here]
Dose-Response Model
M[dose] = g + bl*dose
Variance Model
Var[i] = alpha
Model Options
BMRType
Std. Dev.
BMRF
1
Tail Probability
-
Confidence Level
0.95
Distribution Type
Normal
VarianceType
Constant
Model Data
Dependent Variable
[Dose]
Independent Variable
[Mean]
Total # of Observations
6
Adverse Direction
Automatic
Figure C-22. User input for dose-response modeling of decreased absolute
testis weight in male rats (NTP. 2018).
This document is a draft for review purposes only and does not constitute Agency policy.
C-73 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Model Results
Benchmark Dose
BMD
1.511042118
BMDL
1.074196873
BMDU
2.542202182
AIC
-59.57226688
Test 4 P-value
0.943009409
D.O.F.
4
Model Parameters
# of Parameters
3
Variable
Estimate
g
1.791729181
betal
-0.091864992
alpha
0.019268735
Goodness of Fit
Dose
Size
Estimated
Median
Calc'd
Median
Observed
Mean
Estimated
SD
Calc'd SD
Observed
SD
Scaled
Residual
0
9
1.791729181
1.777
1.777
0.13881187
0.17
0.17
-0.318326837
0.156
10
1.777398242
1.797
1.797
0.13881187
0.15
0.15
0.446548271
0.312
10
1.763067304
1.742
1.742
0.13881187
0.12
0.12
-0.479934917
0.625
10
1.734313561
1.74
1.74
0.13881187
0.1
0.1
0.129542952
1.25
10
1.676897941
1.695
1.695
0.13881187
0.11
0.11
0.412383595
2.5
10
1.5620667
1.553
1.553
0.13881187
0.2
0.2
-0.206548786
Likelihoods of Interest
# of
Model
Log Likelihood*
Pa ra meters
AIC
A1
33.16889532
7
-52.3377906
A2
36.76108906
12
-49.5221781
A3
33.16889532
7
-52.3377906
fitted
32.78613344
3
-59.5722669
R
24.53190731
2
-45.0638146
* Includes additive constant of -54.21737. This constant was not included in the LL derivation prior to BMDS 3.0.
Tests of Interest
Test
-2*Log(Likelihood
Ratio)
Test df
p-value
1
24.4583635
10
0.0064724
2
7.184387472
5
0.20728439
3
7.184387472
5
0.20728439
4
0.765523758
4
0.94300941
Figure C-23. Model results for decreased absolute testis weight in male rats
fNTP. 20181.
This document is a draft for review purposes only and does not constitute Agency policy.
C-74 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
C.2.13. DECREASED ABSOLUTE CAUDAL EPIDIDYMIS WEIGHT IN MALE RATS fNTP. 20181
Table C-49. Dose-response data for decreased absolute caudal epididymis
weight in male rats fNTP. 20181
Dose (mg/kg-d)
n
Mean
SD
0
10
0.184
0.02
0.625
10
0.178
0.01
1.25
10
0.164
0.02
2.5
10
0.138
0.03
Table C-50. Benchmark dose results for decreased absolute caudal epididymis
weight in male rats—constant variance, BMR = 1 standard deviation (NTP.
20181
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.9906
0.7014
0.6614
-192.1231
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Exponential 3
(CV—normal)
Restricted
1.2840
0.7347
0.7934
-190.8813
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Exponential 4
(CV—normal)
Restricted
0.9906
0.7014
0.6614
-192.1231
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Exponential 5
(CV—normal)
Restricted
1.2550
0.6841
NA
-188.9499
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Hill
(CV—normal)
Restricted
1.2551
0.6802
NA
-188.9499
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Polynomial
(3 degree)
(CV—normal)
Restricted
1.2961
0.8004
0.6972
-190.7984
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Polynomial
(2 degree)
(CV—normal)
Restricted
1.2961
0.8004
0.6972
-190.7984
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Power
(CV—normal)
Restricted
1.2924
0.8027
0.7563
-190.8535
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Linear
(CV—normal)
Unrestricted
1.0647
0.7868
0.7835
-192.4618
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
This document is a draft for review purposes only and does not constitute Agency policy.
C-75 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-51. Benchmark dose results for decreased absolute caudal epididymis
weight in male rats—nonconstant variance, BMR = 1 standard deviation fNTP.
20181
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(NCV—
normal)
Restricted
0.7898
0.5327
0.3071
-193.9474
Viable-
Alternate
Exponential 3
(NCV—
normal)
Restricted
1.1440
0.6331
0.5123
-193.8789
Viable-
Alternate
Exponential 4
(NCV—
normal)
Restricted
0.7902
0.5326
0.3070
-193.9463
Viable-
Alternate
Exponential 5
(NCV—
normal)
Restricted
1.1558
0.6708
NA
-192.3083
Questionable
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Hill (NCV—
normal)
Restricted
1.1495
0.6702
NA
-192.3080
Questionable
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Polynomial
(3 degree)
(NCV—
normal)
Restricted
1.1618
0.6304
0.4150
-193.6438
Viable-
Alternate
Polynomial
(2 degree)
(NCV—
normal)
Restricted
1.1618
0.6304
0.4150
-193.6438
Viable-
Alternate
Power (NCV—
normal)
Restricted
1.1497
0.6390
0.4771
-193.8028
Viable-
Alternate
Linear (NCV—
normal)
Unrestricted
0.8363
0.5824
0.4086
-194.5183
Viable-
Recommended
Lowest AIC
This document is a draft for review purposes only and does not constitute Agency policy.
C-76 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Frequentist Linear Model with BMR of 1 Std. Dev. for the BMD and 0.95 Lower
Confidence Limit for the BMDL
0.22
Dose
Figure C-24. Dose-response curve for the Linear model fit to decreased
absolute caudal epididymis weight in male rats (NTP. 2018).
User Input
Info
Model
frequentist Linear vl.l
Dataset Name
Cauda EpiWt_Abs_NTP
User notes
[Add user notes here]
Dose-Response Model
M[dose] = g + bl*dose
Variance Model
Var[i] = alpha * mean[i] A rho
Model Options
BMRType
Std. Dev.
BMRF
1
Tail Probability
-
Confidence Level
0.95
Distribution Type
Normal
VarianceType
Non-Constant
Model Data
Dependent Variable
[Dose]
Independent Variable
[Mean]
Total # of Observations
4
Adverse Direction
Automatic
Figure C-25. User Input for dose-response modeling of decreased caudal
epididymis weight in male rats (NTP. 2018).
This document is a draft for review purposes only and does not constitute Agency policy.
C-77 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Model Results
Benchmark Dose
BMD
0.836267471
BMDL
0.582449886
BMDU
1.345231202
AIC
-194.5182635
Test 4 P-value
0.408601663
D.O.F.
2
Model Parameters
# of Parameters
4
Variable
Estimate
g
0.186188825
betal
-0.018332295
rho
-3.81191884
alpha
-14.76361024
Goodness of Fit
Dose
Size
Estimated
Median
Calc'd
Median
Observed
Mean
Estimated
SD
Calc'd SD
Observed
SD
Scaled
Residual
0
10
0.186188825
0.184
0.184
0.01533068
0.02
0.02
-0.451491469
0.625
10
0.174731141
0.178
0.178
0.01730351
0.01
0.01
0.59739552
1.25
10
0.163273457
0.164
0.164
0.01969127
0.02
0.02
0.116677641
2.5
10
0.140358089
0.138
0.138
0.02626961
0.03
0.03
-0.283861514
Likelihoods of Interest
# of
Model
Log Likelihood*
Pa ra meters
AIC
A1
99.47492849
5
-188.949857
A2
104.7074099
8
-193.41482
A3
102.1541463
6
-192.308293
fitted
101.2591317
4
-194.518263
R
87.99544268
2
-171.990885
* Includes additive constant of -36.75754. This constant was not included in the LL derivation prior to BMDS 3.0.
Tests of Interest
-2*Log(Likelihood
Test
Ratio)
Test df
p-value
1
33.42393448
6
<0.0001
2
10.46496286
3
0.01500047
3
5.106527298
2
0.07782725
4
1.790029051
2
0.40860166
Figure C-26. Model results for decreased caudal epididymis weight in male
rats fNTP. 20181.
C.2.14. DECREASED ABSOLUTE WHOLE EPIDIDYMIS WEIGHT IN MALE RATS fNTP. 20181
Table C-52. Dose-response data for decreased absolute whole epididymis
weight in male rats (NTP. 2018)
Dose (mg/kg-d)
n
Mean
SD
0
10
0.528
0.05
0.625
10
0.508
0.03
1.25
10
0.474
0.04
2.5
10
0.407
0.08
This document is a draft for review purposes only and does not constitute Agency policy.
C-78 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-53. Benchmark dose results for decreased whole caudal epididymis
weight in male rats—constant variance, BMR = 1 standard deviation fNTP.
20181
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.9572
0.6866
0.7614
-118.5715
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Exponential 3
(CV—normal)
Restricted
1.2024
0.7076
0.8891
-117.0973
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Exponential 4
(CV—normal)
Restricted
0.9572
0.6866
0.7614
-118.5715
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Exponential 5
(CV—normal)
Restricted
1.2024
0.7076
0.8891
-117.0973
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Hill
(CV—normal)
Restricted
1.1911
0.6254
NA
-115.1168
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Polynomial
(3 degree)
(CV—normal)
Restricted
1.2061
0.7720
0.7980
-117.0513
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Polynomial
(2 degree)
(CV—normal)
Restricted
1.2061
0.7720
0.7980
-117.0513
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Power
(CV—normal)
Restricted
1.2076
0.7732
0.8530
-117.0825
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Linear
(CV—normal)
Unrestricted
1.0266
0.7639
0.8678
-118.8333
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Table C-54. Benchmark dose results for decreased absolute whole epididymis
weight in male rats—nonconstant variance, BMR = 1 standard deviation (NTP.
20181
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(NCV— normal)
Restricted
0.7358
0.5033
0.3609
-121.3235
Viable-
Alternate
Exponential 3
(NCV— normal)
Restricted
1.0959
0.5980
0.7979
-121.2963
Viable-
Alternate
Exponential 4
(NCV— normal)
Restricted
0.7360
0.5033
0.3609
-121.3235
Viable-
Alternate
This document is a draft for review purposes only and does not constitute Agency policy.
C-79 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 5
(NCV— normal)
Restricted
1.0960
0.5986
NA
-119.2989
Questionable
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Hill (NCV—
normal)
Restricted
1.1035
0.6011
NA
-119.3619
Questionable
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Polynomial
(3 degree)
(NCV— normal)
Restricted
1.1012
0.5975
0.6702
-121.1805
Viable-
Alternate
Polynomial
(2 degree)
(NCV— normal)
Restricted
1.1012
0.5974
0.6702
-121.1805
Viable-
Alternate
Power (NCV—
normal)
Restricted
1.0965
0.6018
0.7557
-121.2651
Viable-
Alternate
Linear (NCV—
normal)
Unrestricted
0.7766
0.5458
0.4809
-121.8975
Viable-
Recommended
Lowest AIC
0.6
0.55
0.5 L
0.45
a;
Ł
g. 0.4
(/>
cc
0.35
0.3
0.25
0.2
0
Figure C-27. Dose-response curve for the Linear model fit to decreased
absolute whole epididymis weight in male rats fNTP. 20181.
Frequentist Linear Model with BMR of 1 Std. Dev. for the BMD and 0.95 Lower
Confidence Limit for the BMDL
Estimated Probability
Response at BMD
O Data
BMD
BMDL
This document is a draft for review purposes only and does not constitute Agency policy.
C-80 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
User Input
Info
Model
frequentist Linear vl.l
Dataset Name
EpididymisWt_Abs_NTP
User notes
[Add user notes here]
Dose-Response Model
M[dose] = g + bl*dose
Variance Model
Var[i] = alpha * mean[i] A rho
Model Options
BMRType
Std. Dev.
BMRF
1
Tail Probability
-
Confidence Level
0.95
Distribution Type
Normal
VarianceType
Non-Constant
Model Data
Dependent Variable
[Dose]
Independent Variable
[Mean]
Total # of Observations
4
Adverse Direction
Automatic
Figure C-28. User input for dose-response modeling of decreased absolute
whole epididymis weight in male rats fNTP. 20181.
This document is a draft for review purposes only and does not constitute Agency policy.
C-81 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Model Results
Benchmark Dose
BMD
0.776560307
BMDL
0.545815255
BMDU
1.227214732
AIC
-121.8975001
Test 4 P-value
0.48085367
D.O.F.
2
Model Parameters
# of Parameters
4
Variable
Estimate
g
0.532146909
betal
-0.048115367
rho
-4.500456294
alpha
-9.413118476
Goodness of Fit
Dose
Size
Estimated
Median
Calc'd
Median
Observed
Mean
Estimated
SD
Calc'd SD
Observed
SD
Scaled
Residual
0
10
0.532146909
0.528
0.528
0.03736447
0.05
0.05
-0.350966522
0.625
10
0.502074805
0.508
0.508
0.0425899
0.03
0.03
0.439942633
1.25
10
0.472002701
0.474
0.474
0.0489403
0.04
0.04
0.129055502
2.5
10
0.411858493
0.407
0.407
0.06650773
0.08
0.08
-0.231009273
Likelihoods of Interest
# of
Model
Log Likelihood*
Pa ra meters
AIC
Al
62.55839468
5
-115.116789
A2
67.81861539
8
-119.637231
A3
65.68094232
6
-119.361885
fitted
64.94875004
4
-121.8975
R
50.54148697
2
-97.0829739
* Includes additive constant of -36.75754. This constant was not included in the LL derivation prior to BMDS 3.0.
Tests of Interest
-2*Log(Likelihood
Test
Ratio)
Test df
p-value
1
34.55425682
6
<0.0001
2
10.52044141
3
0.01462287
3
4.275346136
2
0.11792894
4
1.46438455
2
0.48085367
Figure C-29. Model Results for decreased absolute whole epididymis weight in
male rats fNTP. 20181.
C.2.15. DECREASED DAYS IN ESTRUS-FEMALE RATS fButenhoff etal.. 2012: van Otterdiik.
20071
Table C-55. Dose-response data for decreased days in estrus in female
rats fButenhoff et al.. 2012: van Otterdiik. 20071
Dose (mg/kg-d)
n
Mean
SD
0
10
5.5
1.5092
0.625
10
4.3
2.0575
1.25
10
3.2
1.8136
2.5
10
0.9
0.9944
This document is a draft for review purposes only and does not constitute Agency policy.
C-82 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-56. Benchmark dose results for decreased days in estrus in female
rats—constant variance, BMR = 5% relative deviation fButenhoff et al.. 2012:
van Otterdiik. 20071
Models
Restriction
5% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.0923
0.0687
0.3592
157.0377
Viable—Alternate
BMD 3x lower than lowest
non-zero dose
BMDL 3x lower than lowest
non-zero dose
Exponential 3
(CV—normal)
Restricted
0.2611
0.0778
0.6119
157.2473
Viable—Alternate
BMDL 3x lower than lowest
non-zero dose
Exponential 4
(CV—normal)
Restricted
0.0923
0.0687
0.3592
157.0377
Viable—Alternate
BMD 3x lower than lowest
non-zero dose
BMDL 3x lower than lowest
non-zero dose
Exponential 5
(CV—normal)
Restricted
0.2608
0.0776
0.6119
157.2473
Viable—Alternate
BMDL 3x lower than lowest
non-zero dose
Hill
(CV—normal)
Restricted
0.1487
0.0739
NA
158.9967
Questionable
BMD 3x lower than lowest
non-zero dose
BMDL 3x lower than lowest
non-zero dose
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Polynomial
(3 degree)
(CV—normal)
Restricted
0.1495
0.1283
0.9965
154.9969
Viable—Alternate
BMD 3x lower than lowest
non-zero dose
BMDL 3x lower than lowest
non-zero dose
Polynomial
(2 degree)
(CV—normal)
Restricted
0.1495
0.1283
0.9965
154.9969
Viable-
Recommended
Lowest AIC
BMD 3x lower than lowest
non-zero dose
BMDL 3x lower than lowest
non-zero dose
Power
(CV—normal)
Restricted
0.1495
0.1283
0.9965
154.9969
Viable—Alternate
BMD 3x lower than lowest
non-zero dose
BMDL 3x lower than lowest
non-zero dose
Linear
(CV—normal)
Unrestricted
0.1495
0.1283
0.9965
154.9969
Viable—Alternate
BMD 3x lower than lowest
non-zero dose
BMDL 3x lower than lowest
non-zero dose
This document is a draft for review purposes only and does not constitute Agency policy.
C-83 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Frequentist Polynomial Degree 2 Model with BMR of 0.05 Rel. Dev. for the BMD and 0.95
Lower Confidence Limit for the BMDL
Dose
Figure C-30. Dose-response curve for the Polynomial 2 model fit to decreased
days in estrus in female rats (Butenhoff et al.. 2012: van Otterdijk. 2007).
User Input
Info
Model
frequentist Polynomial degree 2 vl.l
Dataset Name
Estrus_Days_NTP
User notes
[Add user notes here]
Dose-Response Model
M[dose] = g + bl*dose + b2*doseA2 + ...
Variance Model
Var[i] = alpha
Model Options
BMRType
Rel. Dev.
BMRF
0.05
Tail Probability
-
Confidence Level
0.95
Distribution Type
Normal
VarianceType
Constant
Model Data
Dependent Variable
[Dose]
Independent Variable
[Mean]
Total # of Observations
4
Adverse Direction
Automatic
Figure C-31. User input for dose-response modeling of decreased days in
estrus in female rats (NTP. 2018).
This document is a draft for review purposes only and does not constitute Agency policy.
C-84 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Model Results
Benchmark Dose
BMD
0.149469972
BMDL
0.128321644
BMDU
0.470941256
AIC
154.9969111
Test 4 P-value
0.996475595
D.O.F.
2
Model Parameters
# of Parameters
4
Variable
Estimate
g
5.479999986
betal
-1.833142847
beta2
Bounded
alpha
2.427946195
Goodness of Fit
Dose
Size
Estimated
Median
Calc'd
Median
Observed
Mean
Estimated
SD
Calc'd SD
Observed
SD
Scaled
Residual
0
10
5.479999986
5.5
5.5
1.55818683
1.5092
1.5092
0.040589227
0.625
10
4.334285706
4.3
4.3
1.55818683
2.0575
2.0575
-0.069581465
1.25
10
3.188571426
3.2
3.2
1.55818683
1.8136
1.8136
0.023193832
2.5
10
0.897142867
0.9
0.9
1.55818683
0.9944
0.9944
0.005798436
Likelihoods of Interest
# of
Model
Log Likelihood*
Pa ra meters
AIC
Al
-74.49492494
5
158.98985
A2
-71.87802546
8
159.756051
A3
-74.49492494
5
158.98985
fitted
-74.49845557
3
154.996911
R
-90.10938562
2
184.218771
* Includes additive constant of -36.75754. This constant was not included in the LL derivation prior to BMDS 3.0.
Tests of Interest
-2*Log(Likelihood
Test
Ratio)
Test df
p-value
1
36.46272031
6
<0.0001
2
5.233798961
3
0.15545622
3
5.233798961
3
0.15545622
4
0.007061261
2
0.9964756
Figure C-32. Model results for decreased days in estrus in female rats (NTP.
2018).
Table C-57. Benchmark dose results for decreased days in estrus in female
rats—constant variance, BMR = 1 standard deviation (Butenhoff et al.. 2012:
van Otterdijk. 2007)
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.5895
0.3889
0.3592
157.0377
Viable—Alternate
This document is a draft for review purposes only and does not constitute Agency policy.
C-85 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 3
(CV—normal)
Restricted
0.8806
0.4576
0.6119
157.2473
Viable—Alternate
Exponential 4
(CV—normal)
Restricted
0.5895
0.3889
0.3592
157.0377
Viable—Alternate
Exponential 5
(CV—normal)
Restricted
0.8804
0.4576
0.6119
157.2473
Viable—Alternate
Hill
(CV—normal)
Restricted
0.8393
0.4491
NA
158.9967
Questionable
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Polynomial
(3 degree)
(CV—normal)
Restricted
0.8500
0.6520
0.9965
154.9969
Viable—Alternate
Polynomial
(2 degree)
(CV—normal)
Restricted
0.8500
0.6520
0.9965
154.9969
Viable-
Recommended
Lowest AIC
Power
(CV—normal)
Restricted
0.8500
0.6520
0.9965
154.9969
Viable—Alternate
Linear
(CV—normal)
Unrestricted
0.8500
0.6520
0.9965
154.9969
Viable—Alternate
C.2.16. INCREASED DAYS IN DIESTRUS-FEMALE RATS fButenhoff etal.. 2012: van Otterdiik.
20071
Table C-58. Dose-response data for increased days in diestrus in female
rats (Butenhoff et al.. 2012: van Otterdijk. 2007)
Dose (mg/kg-d)
n
Mean
SD
0
10
9.2
1.874
0.625
10
10.1
2.1833
1.25
10
11.7
2.2632
2.5
10
15
1.0541
This document is a draft for review purposes only and does not constitute Agency policy.
C-86 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-59. Benchmark dose results for increased days in diestrus in female
rats—constant variance, BMR = 5% relative deviation fButenhoff et al.. 2012:
van Otterdiik. 20071
Models
Restriction
5% relative
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.2430
0.2000
0.9231
167.0076
Viable-
Recommended
Lowest AIC
BMDL 3x lower than lowest
non-zero dose
Exponential 3
(CV—normal)
Restricted
0.2891
0.2006
0.7433
168.9548
Viable—Alternate
BMDL 3x lower than lowest
non-zero dose
Exponential 4
(CV—normal)
Restricted
0.1870
0.1136
0.4064
169.5368
Viable—Alternate
BMD 3x lower than lowest
non-zero dose
BMDL 3x lower than lowest
non-zero dose
Exponential 5
(CV—normal)
Restricted
0.4063
0.1241
NA
170.8476
Questionable
BMDL 3x lower than lowest
non-zero dose
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Hill
(CV—normal)
Restricted
0.4079
0.1226
NA
170.8476
Questionable
BMDL 3x lower than lowest
non-zero dose
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Polynomial
(3 degree)
(CV—normal)
Restricted
0.2770
0.1470
0.7388
168.9588
Viable—Alternate
BMDL 3x lower than lowest
non-zero dose
Polynomial
(2 degree)
(CV—normal)
Restricted
0.2770
0.1470
0.7388
168.9588
Viable—Alternate
BMDL 3x lower than lowest
non-zero dose
Power
(CV—normal)
Restricted
0.3283
0.1475
0.8200
168.8993
Viable—Alternate
BMDL 3x lower than lowest
non-zero dose
Linear
(CV—normal)
Unrestricted
0.1872
0.1427
0.7099
167.5330
Viable—Alternate
BMD 3x lower than lowest
non-zero dose
BMDL 3x lower than lowest
non-zero dose
This document is a draft for review purposes only and does not constitute Agency policy.
C-87 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Frequentist Exponential Degree 2 Model with BMR of 0.05 Rel. Dev. for the BMD and
0.95 Lower Confidence Limit for the BMDL
18
16
Dose
Figure C-33. Dose-response curve for the Exponential 2 model fit to increased
days in diestrus in female rats (Butenhoff et al.. 2012: van Otterdijk. 2007).
User Input
Info
Model
frequentist Exponential degree 2 vl.l
Dataset Name
Diestrus_Days_NTP
User notes
[Add user notes here]
Dose-Response Model
M[dose] = a * exp(±l * b * dose)
Variance Model
Var[i] = alpha
Model Options
BMRType
Rel. Dev.
BMRF
0.05
Tail Probability
-
Confidence Level
0.95
Distribution Type
Normal
VarianceType
Constant
Model Data
Dependent Variable
[Dose]
Independent Variable
[Mean]
Total # of Observations
4
Adverse Direction
Automatic
Figure C-34. User input for dose-response modeling of increased days in
diestrus in female rats (NTP. 2018).
This document is a draft for review purposes only and does not constitute Agency policy.
C-88 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Model Results
Benchmark Dose
BMD
0.242986679
BMDL
0.200009167
BMDU
0.309079273
AIC
167.0076126
Test 4 P-value
0.923101914
D.O.F.
2
Model Parameters
# of Parameters
3
Variable
Estimate
a
9.070650097
b
0.200793313
log-alpha
1.187317248
Goodness of Fit
Dose
Size
Estimated
Median
Calc'd
Median
Observed
Mean
Estimated
SD
Calc'd SD
Observed
SD
Scaled
Residual
0
10
9.070650097
9.2
9.2
1.81060062
1.874
1.874
0.225914154
0.625
10
10.28349063
10.1
10.1
1.81060062
2.1833
2.1833
-0.320472836
1.25
10
11.65850059
11.7
11.7
1.81060062
2.2632
2.2632
0.072480181
2.5
10
14.98466312
15
15
1.81060062
1.0541
1.0541
0.026786401
Likelihoods of Interest
# of
Model
Log Likelihood*
Pa ra meters
AIC
A1
-80.42379068
5
170.847581
A2
-77.43412842
8
170.868257
A3
-80.42379068
5
170.847581
fitted
-80.50380632
3
167.007613
R
-98.71832217
2
201.436644
* Includes additive constant of -36.75754. This constant was not included in the LL derivation prior to BMDS 3.0.
Tests of Interest
Test
-2*Log(Likelihood
Ratio)
Test df
p-value
1
42.56838749
6
<0.0001
2
5.97932452
3
0.11262048
3
5.97932452
3
0.11262048
4
0.160031269
2
0.92310191
Figure C-35. Model results for increased days in diestrus in female rats (NTP.
2018).
This document is a draft for review purposes only and does not constitute Agency policy.
C-89 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-60. Benchmark dose results for increased days in diestrus in female
rats—constant variance, BMR = 1 standard deviation fButenhoff et al.. 2012:
van Otterdiik. 20071
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.9064
0.7377
0.9231
167.0076
Viable-
Recommended
Lowest AIC
Exponential 3
(CV—normal)
Restricted
0.9766
0.7391
0.7433
168.9548
Viable—Alternate
Exponential 4
(CV—normal)
Restricted
0.7661
0.5970
0.4064
169.5368
Viable—Alternate
Exponential 5
(CV—normal)
Restricted
0.9947
0.5599
NA
170.8476
Questionable
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Hill
(CV—normal)
Restricted
0.9936
0.5580
NA
170.8476
Questionable
d.f. = 0, saturated model
(Goodness of fit test cannot
be calculated)
Polynomial
(3 degree)
(CV—normal)
Restricted
0.9687
0.6117
0.7388
168.9588
Viable—Alternate
Polynomial
(2 degree)
(CV—normal)
Restricted
0.9687
0.6117
0.7388
168.9588
Viable—Alternate
Power
(CV—normal)
Restricted
0.9805
0.6134
0.8200
168.8993
Viable—Alternate
Linear
(CV—normal)
Unrestricted
0.7667
0.5963
0.7099
167.5330
Viable—Alternate
C.2.17. DECREASED RELATIVE UTERINE WEIGHT-FEMALE RATS fButenhoff etal.. 2012: van
Otterdiik. 20071
Table C-61. Dose-response data for decreased relative uterine weight in
female rats (Butenhoff et al.. 2012: van Otterdijk. 2007))
Dose (mg/kg-d)
n
Mean
SD
0
10
3.26
1.3
0.156
10
2.73
0.41
0.312
10
2.94
0.79
0.625
10
3.65
1.68
1.25
10
2.05
0.61
2.5
10
1.81
0.32
This document is a draft for review purposes only and does not constitute Agency policy.
C-90 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-62. Benchmark dose results for decreased relative uterine weight in
female rats—BMR = constant variance, 1 standard deviation fButenhoffetal..
2012: van Otterdiik. 20071
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
Classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
1.6357
0.9728
0.0296
178.4420
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Exponential 3
(CV—normal)
Restricted
1.8431
1.0220
0.0170
179.8915
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Exponential 4
(CV—normal)
Restricted
1.6357
0.9728
0.0296
178.4420
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Exponential 5
(CV—normal)
Restricted
1.2312
0.7036
0.1496
175.0232
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Hill
(CV—normal)
Restricted
1.2139
0.7285
0.1496
175.0233
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Polynomial
(5 degree)
(CV—normal)
Restricted
1.8244
1.2032
0.0147
180.2109
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Polynomial
(4 degree)
(CV—normal)
Restricted
1.8244
1.2032
0.0147
180.2109
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Polynomial
(3 degree)
(CV—normal)
Restricted
1.8244
1.2032
0.0147
180.2109
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Polynomial
(2 degree)
(CV—normal)
Restricted
1.8244
1.2032
0.0147
180.2109
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Power
(CV—normal)
Restricted
1.8813
1.2094
0.0153
180.1247
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Linear
(CV—normal)
Unrestricted
1.7547
1.2018
0.0324
178.2308
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
This document is a draft for review purposes only and does not constitute Agency policy.
C-91 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table C-63. Benchmark dose results for decreased relative uterine weight in
female rats — nonconstant variance, BMR = 1 standard deviation fButenhoff et
al.. 2012: van Otterdiik. 20071
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Non-constant variance
Exponential 2
(NCV— normal)
Restricted
2.3599
1.4658
<0.0001
168.8763
Questionable
Goodness of fit p-value <0.1
Exponential 3
(NCV— normal)
Restricted
2.4946
1.8929
<0.0001
167.1138
Questionable
Goodness of fit p-value <0.1
Exponential 4
(NCV— normal)
Restricted
2.3592
1.4658
<0.0001
168.8763
Questionable
Goodness of fit p-value <0.1
Exponential 5
(NCV— normal)
Restricted
1.2787
1.1724
0.0011
157.4375
Questionable
Goodness of fit p-value <0.1
Hill (NCV—
normal)
Restricted
1.3094
1.1258
0.0011
157.4376
Questionable
Goodness of fit p-value <0.1
Polynomial
(5 degree)
(NCV— normal)
Restricted
2.5118
1.9996
<0.0001
165.4887
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum
dose
Polynomial
(4 degree)
(NCV— normal)
Restricted
2.5118
1.9997
<0.0001
165.4887
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum
dose
Polynomial
(3 degree)
(NCV— normal)
Restricted
2.5118
1.9997
<0.0001
165.4887
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum
dose
Polynomial
(2 degree)
(NCV— normal)
Restricted
2.5118
1.9997
<0.0001
165.4887
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum
dose
Power (NCV—
normal)
Restricted
2.5092
1.9643
<0.0001
167.4725
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum
dose
Linear (NCV—
normal)
Unrestricted
2.4008
1.7105
<0.0001
167.5269
Questionable
Goodness of fit p-value <0.1
Table C-64. Benchmark dose results for decreased relative uterine weight in
female rats—log-normal, constant variance, BMR = 1 standard deviation
fButenhoff et al.. 2012: van Otterdiik. 20071
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—log-
normal)
Restricted
1.9961
0.9991
0.0518
147.6232
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
This document is a draft for review purposes only and does not constitute Agency policy.
C-92 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 3
(CV- log-
normal)
Restricted
2.0457
1.0012
0.0249
149.5811
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Exponential 4
(CV- log-
normal)
Restricted
1.9491
0.6763
0.0246
149.6001
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Exponential 5
(CV- log-
normal)
Restricted
1.2532
0.6896
0.2457
145.0275
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Hill (CV— log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(5 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(4 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(3 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(2 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Power (CV—
log-normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Linear (CV—
log-normal)
Unrestricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
This document is a draft for review purposes only and does not constitute Agency policy.
C-93 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
C.2.18. DECREASED ABSOLUTE UTERINE WEIGHT-FEMALE RAT fButenhoff etal.. 2012: van
Otterdiik. 20071
Table C-65. Dose-response data for decreased absolute uterine weight in
female rats (NTP. 2018)
Dose (mg/kg-d)
n
Mean
SD
0
10
0.731
0.27
0.156
10
0.646
0.09
0.312
10
0.691
0.18
0.625
10
0.818
0.35
1.25
10
0.409
0.13
2.5
10
0.26
0.03
Table C-66. Benchmark dose results for decreased absolute uterine weight in
female rats—BMR = constant variance, 1 standard deviation fButenhoff etal..
2012: van Otterdiik. 20071
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—normal)
Restricted
0.8877
0.5920
0.0083
-6.1338
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
Exponential 3
(CV—normal)
Restricted
1.2592
0.7971
0.0140
-7.2318
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Exponential 4
(CV—normal)
Restricted
0.8877
0.5920
0.0083
-6.1338
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
Exponential 5
(CV—normal)
Restricted
1.2039
0.9713
0.2538
-13.7789
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Hill
(CV—normal)
Restricted
1.1828
0.8675
0.1306
-11.7788
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Polynomial
(5 degree)
(CV—normal)
Restricted
1.2569
0.8354
0.0076
-5.9234
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Polynomial
(4 degree)
(CV—normal)
Restricted
1.2569
0.8354
0.0076
-5.9234
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
This document is a draft for review purposes only and does not constitute Agency policy.
C-94 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Polynomial
(3 degree)
(CV—normal)
Restricted
1.2569
0.8354
0.0076
-5.9234
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Polynomial
(2 degree)
(CV—normal)
Restricted
1.2569
0.8354
0.0076
-5.9234
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Power
(CV—normal)
Restricted
1.3086
0.8477
0.0088
-6.2298
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Linear
(CV—normal)
Unrestricted
1.0823
0.8275
0.0163
-7.7099
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Table C-67. Benchmark dose results for decreased absolute uterine weight in
female rats—nonconstant variance, BMR = 1 standard deviation fButenhoff et
al.. 2012: van Otterdiik. 20071
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Non-constant variance
Exponential 2
(NCV—
normal)
Restricted
1.3500
0.9186
<0.0001
-25.2943
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Exponential 3
(NCV—
normal)
Restricted
1.8175
1.3964
<0.0001
-33.2616
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
Exponential 4
(NCV—
normal)
Restricted
1.3502
0.9186
<0.0001
-25.2943
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Exponential 5
(NCV—
normal)
Restricted
1.2424
1.1367
0.0036
-42.1526
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Hill (NCV—
normal)
Restricted
1.2387
1.1069
0.0103
-44.1525
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Polynomial
(5 degree)
(NCV—
normal)
Restricted
2.0088
1.5693
0.0001
-33.9754
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
This document is a draft for review purposes only and does not constitute Agency policy.
C-95 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Non-constant variance
Polynomial
(4 degree)
(NCV—
normal)
Restricted
2.0088
1.5692
0.0001
-33.9754
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Polynomial
(3 degree)
(NCV—
normal)
Restricted
2.0088
1.5692
0.0001
-33.9754
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Polynomial
(2 degree)
(NCV—
normal)
Restricted
2.0088
1.5692
0.0001
-33.9754
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Power (NCV—
normal)
Restricted
1.9555
1.5188
<0.0001
-32.0845
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
Linear (NCV—
normal)
Unrestricted
1.6526
1.2761
<0.0001
-30.8879
Questionable
Nonconstant variance test
failed (Test 3 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group
Near BMD| >2
Table C-68. Benchmark dose results for decreased absolute uterine weight in
female rats—log-normal, constant variance, BMR = 1 standard deviation
(Butenhoff et al.. 2012: van Otterdijk. 2007)
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
Exponential 2
(CV—log-
normal)
Restricted
1.0282
0.5795
0.0129
-43.7584
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Modeled control response
std. dev. >11.51 actual
response std. dev.
Exponential 3
(CV- log-
normal)
Restricted
1.2617
0.6141
0.0101
-43.1248
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Modeled control response
std. dev. >11.51 actual
response std. dev.
Exponential 4
(CV- log-
normal)
Restricted
1.0282
1.0189
0.0129
-43.7584
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
Modeled control response
This document is a draft for review purposes only and does not constitute Agency policy.
C-96 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Models
Restriction
1 standard
deviation
p-Value
AIC
BMDS
classification
BMDS notes
BMD
BMDL
Constant variance
std. dev. >11.51 actual
response std. dev.
Exponential 5
(CV- log-
normal)
Restricted
1.2149
0.9197
0.3929
-50.5863
Questionable
Constant variance test failed
(Test 2 p-value < 0.05)
Modeled control response
std. dev. >11.51 actual
response std. dev.
Hill (CV— log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(5 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(4 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(3 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Polynomial
(2 degree)
(CV- log-
normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Power (CV—
log-normal)
Restricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
Linear (CV—
log-normal)
Unrestricted
Unusable
BMD computation failed
Model was not run.
Adverse direction "down"
not compatible with
lognormal distribution
1
This document is a draft for review purposes only and does not constitute Agency policy.
C-97 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
APPENDIX D. ADVERSE OUTCOME PATHWAY/
MODE OF ACTION(AOP/MOA)-BASED
APPROACH FOR EVALUATING PFDA-INDUCED
MECHANISM OF HEPATOXITY
D.l. OBJECTIVE AND METHODOLOGY
The goal of the qualitative analysis described here is to evaluate the available mechanistic
evidence for PFDA-induced liver effects to assess the biological plausibility of effects observed in
animal models and identify mechanistic pathways that are conserved across species and strains of
animals and liver cell culture models and are therefore more relevant to human health. The
available mechanistic and toxicological evidence was organized and evaluated in concordance with
the frameworks used for mode of action (MOA) analysis for non-cancer effects and development of
adverse outcome pathways (AOP)1 fEdwards etal.. 2016: Boobis etal.. 2008: IPCS. 20071. PFDA-
induced hepatic effects reported in in vivo and cell culture studies were organized according to the
following levels of biological organization: molecular interactions, cellular effects, organ effects, and
organism effects. The analysis described here was focused on the concordance of key events and
adverse responses across species to obtain clarification on the relevance of animal studies to
human health.
In addition to analyzing the available evidence published in the peer-reviewed literature,
EPA also considered mechanistic evidence from in vitro high throughput screening (HTS) assays on
PFDA available from the EPA's CompTox Chemicals Dashboard
fhttps://co mptox.epa.gov/dashboardl fU.S. EPA. 20191. Bioactivity data from the ToxCast and
Tox21 collaborative projects were also considered at the same levels of biological organization
described below. A more detailed description of the HTS analysis and results is provided in
Appendix E.
Although the World Health Organization [WHO]-Inter national Programme on Chemical Safety (IPCS]-M0A
and the Organization for Economic Co-operation and Development (0ECD]-A0P frameworks are similar in
the identification and analysis of key events following modified Bradford-Hill criteria fMeek et al., 20141.
AOPs are chemically agnostic, whereas MOA analyses are intended to inform health assessments of individual
(or groups of) chemical(s) fEdwards et al., 20161.
This document is a draft for review purposes only and does not constitute Agency policy.
D-l DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
D.2. PROPOSED MOA/AOP APPROACH FOR EVALUATING PFAS-INDUCED
LIVER TOXICITY
The proposed MOA displayed in Figure D-l is based on molecular initiating events, key
events, and adverse outcomes identified in previous mechanistic evaluations and reviews on PFOS
and PFOA fATSDR. 2018: Li etal.. 2017: U.S. EPA. 2016a. b), which are structurally related to PFDA
and among the most well-studied PFAS. Additional reviews on biological pathways associated with
chemical-induced cancer and noncancer liver effects were also consulted (see citations below). A
summary of the MOA is presented below.
At the molecular level, experimental studies using in vivo and cell culture models have
shown that perfluorinated compounds such as PFOS and PFOA can activate several nuclear
receptor pathways including the constitutive androstane receptor (CAR), the pregnane X receptor
(PXR), the farnesoid X receptor (FXR), the peroxisome proliferator activated receptor alpha
(PPARa) and gamma (PPARy), estrogen receptor alpha (ERa) and other receptor-independent cell
signaling pathways (e.g., phosphatidylinositol 3-kinase-serine/threonine protein kinase (PI3K-Akt)
signal transduction pathway, and the nuclear factor kappa B pathway [NFkB]) fATSDR. 2018: Li et
al.. 2017: U.S. EPA. 2016a. b). PFOS- and PFOA-induced activation of PPARa is associated with
hepatocellular hypertrophy caused by peroxisome proliferation, and increased peroxisomal fatty
acid p oxidation and cytochrome P450 4A (CYP4A) expression and activity fATSDR. 2018: U.S. EPA.
2016a. b), and altered cholesterol metabolism (Li etal.. 2017). Increased PPARa activity can lead to
oxidative stress via induction of acyl CoA oxidase expression and activity and to H2O2 production in
peroxisomes fHall etal.. 20121. Several studies have used genetically modified animal and cell
culture models and immortalized human cell lines to evaluate potential PFOS or PFOA activation of
the human PPARa. COS-1 cells transfected with the murine or human PPARa were responsive to
PFAS exposure (U.S. EPA. 2016a. b), and F1 generation PPARa-humanized mice were responsive to
PFOA-induced expression responsive genes on GD 18, but unlike wild type animals this response
was not apparent on PND 20 (U.S. EPA. 2016b: Takacs and Abbott. 2007). Studies using human
liver cell lines or humanized animal models suggest that humans are less sensitive to PPARa
activation by the perfluorinated compounds PFOS and PFOA (reviewed in Li etal. f 20171 and U.S.
EPA f2016all. PPARa has also been shown to be activated by exposure to several PFAS, including
PFOS, PFOA, PFNA, and PFHxS fATSDR. 2018: Li etal.. 20171. Although PPARa is not expressed in
high levels in the liver, its activation by pharmaceuticals and xenobiotic compounds has been
proposed to be associated with hepatic steatosis caused by lipid accumulation (Angrish etal.. 2016:
Mellor etal.. 2016).
As described above, exposure to perfluorinated compounds such as PFOS and PFOA has also
been shown to activate other nuclear receptor and cell signaling pathways including the CAR, PXR,
FXR, ERa, NFkB, and the oxidative stress responsive nuclear factor erythroid 2 related factor 2
(Nrf2) fATSDR. 2018: Li etal.. 2017: U.S. EPA. 2016al. Furthermore, experiments using null animal
models exposed to several PFAS suggest that activation of CAR/PXR occurs independently of PPARa
This document is a draft for review purposes only and does not constitute Agency policy.
D-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
fATSDR. 2018: Li et al.. 20171. Previous analyses of chemical-induced hepatotoxicity suggest that
activation of these cell signaling pathways in experimental models is associated with increased
expression and activity of xenobiotic metabolizing enzymes (XMEs) floshi-Barve etal.. 2015: Hall et
al.. 20121. formation of reactive metabolites, alterations in cellular lipid metabolism fAngrish etal..
20161. and endoplasmic reticulum damage fToshi-Barve etal.. 20151.
At the cellular level, exposure to PFAS such as PFOS and PFOA has been shown to increase
reactive oxygen species production and oxidative damage to cellular macromolecules fATSDR.
2018: Li etal.. 2017: U.S. EPA. 2016al: promote mitochondrial damage, inhibit mitochondrial
function, activate mitochondrial-mediatedcell death fLi etal.. 2017: U.S. EPA. 2016bl: increase
endoplasmic reticulum stress fU.S. EPA. 2016bl: induce DNA damage fATSDR. 2018: U.S. EPA.
2016b); disrupt intercellular gap junction communication fATSDR. 20181: elevate
production/levels of proinflammatory cytokines fU.S. EPA. 2016bl: alter lipid and glucose
metabolism and bile acid biosynthesis (U.S. EPA. 2016a. b); and increase hepatocellular death fLi et
al.. 2017: U.S. EPA. 2016al. These path ways/mechanisms are associated with toxicant-induced
liver disease and can promote steatohepatitis and fibrosis fAngrish etal.. 2016: Cao etal.. 2016:
Toshi-Barve etal.. 2015: Wahlangetal.. 20131.
Figure D-l. This proposed MOA is based on previous analyses on PFAS-
induced (e.g., PFOA/PFOS) liver toxicity and the role of nuclear receptor
pathways in hepatotoxicity.
This document is a draft for review purposes only and does not constitute Agency policy.
D-3 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
D.3. SYNTHESIS OF MECHANISTIC STUDIES AND SUPPLEMENTAL
INFORMATION FOR PFDA
As mentioned previously, mechanistic evidence from peer-reviewed studies and HTS assays
from EPA's ToxCast/Tox21 database were organized and evaluated according to the proposed MOA
for the noncancer-liver effects associated with exposure to PFAS (see Figure D-l). The evidence
consists primarily of in vitro and in vivo studies conducted in liver tissues derived from human and
animal models. When available, cell-free receptor binding studies and gene reporter assays
profiling different key events in receptor signaling pathways in other cell tissue models
(e.g., receptor dimerization, cofactor recruitment, DNA binding and gene transactivation) were
included in the analysis to provide additional information on the activation of nuclear receptor
pathways and on potential species-specific differences in receptor sensitivity relevant to the
mechanisms of liver toxicity for PFDA and other PFAS.
D.3.1. MOLECULAR INITIATING EVENTS
As discussed below, the available studies have examined several nuclear receptor and cell
signaling pathways associated with chemical-induced liver toxicity.
PPARa
PPARa is involved in a variety of processes, including nutrient metabolism, tissue
development, cell differentiation, xenobiotic biotransformation and inflammation fLi etal.. 20171.
Induction of PPARa activity is primarily associated with increased CYP450 activity, peroxisomal
proliferation and hepatomegaly (liver enlargement) fHall etal.. 20121 and has been implicated in
the mechanisms of hepatotoxicity of PFAS such as PFOA and PFOS (ATSDR. 2018: U.S. EPA. 2016a).
Several experimental studies have evaluated PFDA-induced activation of the PPARa in vivo in the
rat and mouse liver, and in human and rodent hepatocyte cell cultures. PFDA exposure was
associated with increased hepatic expression of PPARa-responsive genes in Sprague Dawley rats
fNTP. 2018: Sterchele etal.. 19961. C57BL/6J mice fAbe etal.. 2017: Cheng and Klaassen. 2008a. b;
Maher etal.. 20081 and SV129 mice fLuo etal.. 20171. PFDA treatment has also been shown to
increase hepatic PPARa mRNA levels fSterchele etal.. 19961 and activity of the PPARa-responsive
enzyme acyl-CoA oxidase in Sprague Dawley rats (NTP. 20181. Chinie etal. (19941 exposed male
Wistar rats and Harley Guinea pigs to PFDA and reported increased CYP4A1 mRNA levels
(indicative of PPARa activation) in rats, but no effects in Guinea pigs. These findings are consistent
with analyses, which conclude that Guinea pigs, along with Syrian hamsters and non-human
primates, are less responsive to PPARa activation than other rodent models fCorton etal.. 2018:
Hall etal.. 20121.
Several cell culture and in vitro studies also report evidence considered supportive of the
in vivo findings. PFDA exposure increased mRNA levels of PPARa and PPARa-responsive genes in
rat hepatoma FaO cells (Sterchele etal.. 19961. Two studies evaluated PFDA-induced effects on
This document is a draft for review purposes only and does not constitute Agency policy.
D-4 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
PPARa-responsive genes in human hepatic progenitor cells (HepaRG). One study was unable to
measure activation of PPARa or other nuclear receptors due to PFDA exposure associated with
cytotoxicity (100 |iM] but detected gene reported activity in non-human primate kidney cells
transfected with the mouse PPARa (COS-1) fAbe etal.. 20171. The other study that tested a lower
PFDA concentration (45 |iM] confirmed PPARa activation fLim etal.. 20211. Rosen etal. f20131
analyzed gene expression changes in response to PFDA treatment and reported higher
transcriptional activity in cultured primary human versus mouse hepatocytes, including the
induction of PPARa-dependent and PPARa-independent genes. The lower than expected pattern of
transcriptional activity for PFDA and other PFAS in cultured primary mouse hepatocytes compared
to previous in vivo studies was attributed to cell culture conditions and the absence of hepatic
non-parenchymal cells fRosen et al.. 20131. The authors also noted inconsistencies in the dose-
response patterns of transcriptional activity in human hepatocytes across PFAS that could be due to
inter individual variation in donor cells or inherent differences in the pattern of gene expression of
tested chemicals (Rosen etal.. 20131. PPARa-dependent reporter gene expression was also
induced after PFDA treatment in human hepatoma HepG2 cells (Rosenmai etal.. 20181 and human
embryonic kidney HEK293 cells (Buhrke etal.. 20131. HTS assays showed induction of PPARa
transactivation in HepG2 cells but no activity in a binding reporter assay for the human PPARa (see
Table E-2). However, a recent in vitro study in the peer-reviewed literature reported that PFDA can
bind to the human PPARa ligand binding domain, albeit with lower affinity than the Baikal seal
PPARa (Ishibashi et al.. 20191. Potential interspecies differences in PPARa activation were also
described by Routti etal. (20191: Wolf etal. (20121: Wolf etal. (20081. showing induction of
transcriptional activity of the mouse and polar PPARa isoforms but minimal or no activity towards
the human PPARa in non-human primate kidney cells (COS-1 and COS-7) exposed to PFDA.
Overall, the available evidence suggests that PFDA can activate hepatic PPARa in rats and
mice in vivo and in cell culture models. There are inconsistencies with respect to the activation of
PPARa in in vitro human models possibly due to differences in experimental design and/or
potential confounding with PFDA-induced cytotoxicity. However, some evidence indicates that
PFDA interacts with the human PPARa in immortalized and primary cells derived from liver tissue.
The data also suggest potential species differences in the binding affinity and activity of PPARa with
the human isoform being potentially less sensitive compared to other mammalian species. In vivo
studies with genetically modified animals in which the gene encoding PPARa is inactivated are
needed to further characterize these differences.
Other PPARs (PPARy and PPARfi/S)
Two other PPAR subtypes have been characterized, PPARy and PPAR(3/S, that play an
essential role in energy homeostasis and metabolism. PPARy is known to regulate adipogenesis,
lipid and glucose metabolism and inflammatory pathways and its hepatic upregulation has been
proposed as a key mechanism in the pathogenesis of non-alcoholic fatty liver disease (NAFLD) (Al
Sharif etal.. 20141. PFDA-induced transactivation of human PPARy was observed in HEK263
This document is a draft for review purposes only and does not constitute Agency policy.
D-5 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
(Buhrke etal.. 20131 and HepG2 cells (Zhang etal.. 20141 and HTS results from the EPA's
ToxCast/Tox21 database displayed in Table F-2). PFDA also showed affinity for the human PPARy
in receptor-ligand binding assays fZhang etal. f 20141 and Table E-2) but displayed no activity in
agonist/antagonist or cofactor recruitment assays related to this receptor conducted in HEK293T
cells (see Table E-2). Further, PFDA upregulated the expression of the PPARy gene in primary
humanhepatocytes fRosenetal.. 20131.
PPARp/6 is involved in fatty acid metabolism and suppression of macrophage-derived
inflammation (Barish etal.. 20061. Studies examining potential interaction between PFAS and
PPARp/6 are limited. In vitro evidence showed that PFDA is capable of binding to the human
PPARp/6 and activating its transcriptional activity in HEK293 cells at non-cytotoxic concentrations
(< 100 [J.M) fLi etal.. 20191. In contrast, PFDA was inactive in ToxCast/Tox21 assays (see Table F-
2), evaluating human PPAR(3/5 transactivation in HEK293 and HepG2 cells at concentrations up to
200 [iM. Differences in experimental design (e.g., reporter system) could account for discrepancies
in the results.
There is in vitro evidence that suggests potential activation of other human PPAR subtypes
after PFDA treatment, primarily PPARy and possibly PPAR(3/S. Experimental studies in animals
and humanized models would be critical to confirming and better characterizing the potential role
of these receptors in the mechanism(s) of hepatotoxicity from PFDA exposure.
CAR/PXR
Chemical-induced activation of CAR and PXR leads to increased expression and activity of
xenobiotic metabolizing enzymes (XMEs) (Li etal.. 2017: Hall etal.. 20121 and drug transport
proteins (Mackowiak etal.. 20181. In addition to metabolism and excretion of xenobiotic
compounds (and endogenous substrates such as steroids and fatty acids), CAR/PXR-induced
xenobiotic enzyme activities have been proposed to promote formation of reactive metabolites
fWang etal.. 2014: Li etal.. 20121. alter drug interactions fMackowiak etal.. 20181. and increase
oxidative stress, immune responses, and mitochondrial disfunction (Wang etal.. 20141. CAR/PXR
activation can also alter lipid homeostasis and promote hepatic steatosis (Mackowiak etal.. 2018:
Mellor etal.. 20161.
Experimental studies have evaluated PFDA-mediated activation of CAR and PXR in rodents.
PFDA exposure led to increase in CAR mRNA levels, nuclear translocation of CAR, and increased
mRNA and/or protein levels of CAR- and PXR-responsive genes such as Cyp2B10 and Cyp3All in
C57BL6/6J mice fAbe etal.. 2017: Cheng and Klaassen. 2008bl. NTP f20181 also reported
increased in the mRNA levels of CAR-responsive genes, CyplBl and cyplB2, in Sprague-Dawley
rats. Further evaluation of the effects ofPFDA on CYP450s in genetically modified mice devoid of
function of specific nuclear receptors revealed that PFDA-mediated Cyp2B10 mRNA expression is
regulated by CAR and independent of PPARa, PXR or FXR (Cheng and Klaassen. 2008b). PXR was
also not required for the induction of Cyp3All mRNA after PFDA exposure fCheng and Klaassen.
2008b).
This document is a draft for review purposes only and does not constitute Agency policy.
D-6 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Cell culture studies and HTS assays from the ToxCast database have also evaluated PFDA-
induced activation of CAR and PXR. PFDA exposure resulted in increased mRNA and protein levels
of PXR but did not affect the expression of the PXR target gene, Cyp3A23, in primary rat
hepatocytes fMa etal.. 20051. PXR-dependent CYP3A4 activation by PFDA was reported in HepG2
cells transfected with the human PXR fZhang etal.. 20171. and increased mRNA levels of CAR/PXR-
responsive genes, CYP2B6 and CYP3A4, were detected in primary human hepatocytes after PFDA
treatment (Rosen etal.. 20131. In primary mouse hepatocytes, PFDA treatment had no effect on
CAR-responsive genes, but according to the study authors this may have been caused by cell culture
conditions and time in culture before and during exposure (Rosen etal.. 20131. An additional study
reported no effects on the induction of the mouse or human CAR in gene reporter assays using
nonhuman primate kidney COS-1 cells but failed to assess PFDA-induced expression of CAR-
responsive genes in HepaRG cells due to increased cytotoxicity after chemical exposure (100 [J.M)
(Abe etal.. 20171. Using a lower PFDA concentration (45 [J.M), Lim etal. (20211 showed
upregulation of the CAR-target gene, CYP2B6. Gene reporter activity measured in HTS assays
conducted in HepG2 cells revealed PFDA-induced activation of the human PXR in 1 of 3 assays but
no activation of the human CAR across 4 assays (see Table E-2). PFDA also demonstrated binding
activity for the human PXR (see Table E-2).
Overall, the available evidence suggests that PFDA exposure can activate the murine CAR
resulting in altered levels of CYP450s in vivo and, although not all of the available experiments
were clearly positive, PFDA appears to interact with PXR in in vitro rodent and human model
systems. Future studies focusing on the potential involvement of these receptors in the
mechanisms of PFDA-induced liver effects would be informative.
FXR
FXR is a key regulator of bile acid synthesis and lipid metabolism fRussell. 20031. Deletion
of the mouse FXR gene (Nrlh4) leads to fatty liver and insulin resistance fMa etal.. 20061 and
exacerbation of chemical-induced acute liver injury (Takahashi etal.. 20171. while activation of FXR
in response to liver injury and disease may have a protective role (Han. 20181. PFDA was evaluated
in HTS from EPA's ToxCast/Tox21 database (see Analysis of relevant high throughput screening
assays from the EPA's CompTox Chemicals Dashboard in Appendix E for more details). No FXR
activity was detected in assays related to receptor/cofactor interaction or agonist/antagonist
transactivation in human embryonic kidney HEK293 cells (see Table E-2). Conversely, PFDA
displayed agonist activity in a cell-free receptor-ligand binding assay and was active in one of two
assays profiling transcriptional activity of this receptor in a human liver cell line (HepG2) (see
Table E-2). Importantly, PFDA exhibited high potency for the human FXR compared to other
nuclear receptors (e.g., PPARa/y and CAR/PXR) based on estimated effective concentrations
(i.e., AC50 values) (see Figure F-2B). In summary, FXR appears to be a sensitive target ofPFDA in
HTS assays and thus, similar to CAR above, experiments specifically targeting the potential role of
this receptor in the liver effects ofPFDA would be informative.
This document is a draft for review purposes only and does not constitute Agency policy.
D-7 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Other Pathways
Additional cell signaling pathways have been evaluated in vivo and in liver cells in vitro. In
Wistar rats and SV129 mice, PFDA exposure had no effects on mRNA levels of c-Jun/c-Fos fLuo et
al.. 20171 fOguro etal.. 19981. Similarly, PFDA exposure had no significant effects on aryl
hydrocarbon receptor (AHR)-inducible P450 activity in C57BL/6J mice fBrewster and Birnbaum.
1989) or mRNA expression of AHR-responsive genes (CyplAl/2) in C57BL/6J mice (Cheng and
Klaassen. 2008b] and HepaRG cells (Lim etal.. 2021], However, PFDA increased 2,3,7,8-
Tetrachlorodibenzo-p-dioxin (TCDD)-induced AHR transactivation in an antagonist assay
conducted in mouse hepatoma Hepa 1.12cR cells fLong etal.. 2013], Effects on inflammatory and
oxidative/cellular stress signaling involving the nuclear factor erythroid 2 related factor 2 (Nrf2),
nuclear factor kappa B pathway (NFkB), tumor necrosis factor alpha (TNFa), c-Jun-N-terminal
kinase (JNK) and activating transcription factor 2 (ATF-2) were reported following PFDA exposure
in rodents (see synthesis on Inflammation and Cellular Stress for more details).
In vitro HTS assays from ToxCast/Tox21 showed induction of target gene pathways in
HepG2 and HepaRG cells (measured as gene reporter activity) (see Table F-l), including several
nuclear receptors discussed previously. According to estimated AC 50 values (concentration at half
maximal response), gene-specific activities occurred upstream but were closely associated with
responses indicative of cellular stress/cytotoxicity (see Figure E-l). Specifically, PFDA was active in
all three assays measuring Nrf2 transcriptional or agonist activity but was inactive in
transactivation assays for NFkB and AHR in HepG2 and HepRG cells (see Table E-l). Induction of
transcriptional activity for JUN/FOS was demonstrated in HepaRG cells but not HepG2 cells with
PFDA exposure (see Table F-l).
Overall, the available experimental studies suggest that in addition to activation of PPARa
and CAR/PXR nuclear receptor pathways (and possibly PPARy and FXR based on limited in vitro
studies in human cells), exposure to PFDA may also promote activation of other cell signaling
pathways associated with inflammatory and oxidative/cellular stress responses (see synthesis on
Inflammation and Cellular stress in this Appendix for more details).
D.3.2. CELLULAR EFFECTS
As discussed below, the available studies provide evidence on potential PFDA-induced
alterations in hepatic expression and/or activity of XMEs, oxidative stress, cell and mitochondrial
damage, inflammation, and alterations in liver metabolic functions.
Expression and Activity of XMEs
Several in vivo studies have evaluated PFDA-induced effects on the expression and activity
of XMEs. In Wistar rats, PFDA exposure was associated with increased cytochrome P450 content
and activity of NADPH-cytochrome c (P-450) reductase (Yamamoto and Kawashima. 1997) and
decreased GST protein levels and activity (Oguro etal.. 1998: Kawashima etal.. 1995: Schramm et
This document is a draft for review purposes only and does not constitute Agency policy.
D-8 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
al.. 19891. Furthermore, PFDA exposure altered bilirubin glucuronosyltransferase activities and
bilirubin, morphine, testosterone, and naphthol glucuronidation (Arand etal.. 19911. In Fischer
rats, PFDA treatment resulted in decreased sulfotransferase protein levels fWitzmann etal.. 19961
and microsomal carboxylesterase activity fDerbel et al.. 19961. A study using SV129 mice found
that PFDA exposure decreased hepatic mRNA levels of CYP450s, and organic-anion-transporting
polypeptides (OATPs) involved in the bile acid synthesis and uptake, while increasing mRNA levels
of UDP-glucuronosyltransferases (UGT) enzymes fLuo etal.. 20171. PPARa-null mice were mostly
resistant to these effects fLuo etal.. 20171. Similarly, Cheng and Klaassen (2008b) reported that
PFDA-mediated downregulation of hepatic bile acid uptake transporters (OATPs and the Na+-
taurocholate cotransporting peptide) is notably disrupted in PPARa-null mice but not in CAR-,
PXR-, Nrf2- or FXR- null counterparts. As such, PPARa appears to be involved in the modulation of
metabolizing enzymes and transport mechanisms important for bile acid homeostasis.
Several in vivo studies evaluated the effects of PFDA exposure on multidrug resistance
proteins, which play important roles in hepatic metabolic and detoxifying functions, including bile
acid excretion (Roth etal.. 2019: Yang etal.. 20141. In Sprague Dawley rats, PFDA exposure was
associated with decreased mRNA and protein levels of the hepatic multidrug resistance protein 2
(Mrp2), albeit effects were not statically significant flohnson and Klaassen. 20021. A separate study
reported that PFDA exposure significantly increased Mrp2 mRNA levels in SV129 mice and that
PPARa-null animals were resistant to this effect fLuo etal.. 20171. Two studies using wild type and
PPARa-null mice evaluated PFDA-induced changes in hepatic levels of Mrp3 and Mrp4 fLuo etal..
2017: Maher etal.. 20081. Both studies report that PFDA treatment increased Mrp4 mRNA levels in
wild type SV129 or C57BL/6J mice, but the responses in PPARa-null animals differed: Maher et al.
(20081 observed that elimination of PPARa ameliorated this effect, while Luo etal. (20171 reported
that PPARa-nulls were as responsive as wild type animals. Maher etal. f20081 observed that unlike
wild type mice, PPARa-null animals were resistant to PFDA induction of Mrp3, and Luo etal. T20171
reported no exposure-related effects on Mrp3 levels in either wild type or null animals. Luo et al.
(20171 and Maher etal. (20081 used a similar dose regimen (single i.p. injection of 80 mg/kg) but
Luo etal. (20171 sampled animals on day 5 post exposure whereas Maher etal. (20081 sampled
animals 48 hours postexposure) and test mouse strain (SV129 and C57BL/6, respectively) differed
between studies. These differences in experimental model and/or design features could account
for the perceived discrepancies in the results. Maher etal. f20081 also reported that Nrf2-null mice
were resistant to PFDA-induced expression of Mrp3 and Mrp4, and that pretreatment with
gadolinium chloride ameliorated PFDA-induction of Mrp4 mRNA levels but had no effect on Mrp3.
Overall, the results suggest that PPARa and other signaling pathways (i.e., Nrf2 and Kupffer cell
activation) participate in PFDA-mediated disruption of hepatic efflux Mrp transporters.
A study evaluating transcriptomic changes in HepaRG cells with exposure to PFDA and
other long-chain PFAS observed enrichment of gene pathways involved in phase I and phase II
metabolism, transporters, bile acid metabolism, amino acid metabolism and carbohydrate
This document is a draft for review purposes only and does not constitute Agency policy.
D-9 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
metabolism fLim etal.. 20211. An increase in transcriptomic response was reported with increasing
carbon chain length with PFDA being the most potent PFAS tested. Specifically with respect to
transporters, PFDA exposure was associated with the upregulation of xenobiotic efflux transporters
(e.g., ABCA3, ABCC3/MRP3, ABCC10/MRP7, and ABCG2/BCRP) and amino acid transporters
involved in protein synthesis (e.g., SLC1A4, SLC1A5, SLC6A9, SLC7A1, SLC7A2, SLC7A5, SLC7A11,
and SLC43A1), as well as the downregulation of bile acid or xenobiotic uptake transporters (e.g.,
SLC10A1/NTCP, SLC02B1 and SLC04C1). These observations are consistent with a potential
compensatory mechanism against chemical-induced injury. The authors also noted that PFDA-
mediated regulation of transporters appeared to be associated with the induction of Nrf2 rather
than PPARa or CAR fLim etal.. 20211 Similarly, HTS ToxCast/Tox21 assays showed PFDA-
mediated induction of gene pathways associated with xenobiotic metabolism and transport (i.e.,
CYP1A1, CYP2C19, CYP4A11, CYP4A22, ABCC3 andABCG2,) inHepaRG cells (see Figure E-2 and
Table E-l).
The findings described above suggest that exposure to PFDA results in increased XME levels
and activity in animal models, which is supported by evidence on PFDA-induced activation of the
CAR/PXR signaling pathways, two key regulators of XMEs. Furthermore, evidence from
experiments using null animals suggest that PPARa is important for PFDA-induced regulation of a
number of XMEs and transporters involved in bile acid homeostasis (e.g., CYP450, UGT OATP, and
Mrp proteins). Additional mechanisms involving Nrf2 and Kupffer cell-mediated inflammatory
responses appear to also play a role in regulating the expression of hepatic transporters in
response to chemical-induced toxicity. The disruption of bile acid synthesis and transport
mechanisms is consistent with the observed increases in markers of hepatobiliary function/injury
in mice following PFDA exposure (see synthesis on Cellular stress and Metabolic effects below).
Further studies are necessary to clarify inconsistencies in the results described above and to
characterize the specific role of PPARa, Nrf2 and other cell signaling pathways (e.g., CAR/PXR) in
modulating XME expression and activity and associated downstream effects that could contribute
to the observed hepatic effects ofPFDA exposure.
Oxidative Stress
Increased production of reactive oxygen species (ROS) can lead to hepatocellular toxicity as
it can result in cellular damage (e.g., increase lipid peroxidation, protein oxidation, and oxidative
DNA damage) floshi-Barve etal.. 2015: Wahlang etal.. 20131 and activation of proinflammatory cell
signaling cascades floshi-Barve etal.. 20151.
Several in vivo and cell culture studies have evaluated PFDA-induced oxidative stress. In
CD-I mice, PFDA decreased the activity of antioxidant enzymes such as total superoxide dismutase
(T-SOD), catalase (CAT), and glutathione peroxidase (GPx) activities, while increasing the level of
hepatic oxidative markers including ROS, thiobarbituric acid reactive substances (TBARS) and
malondialdehyde (MDA) in hepatic tissue fWang etal.. 20201. Likewise, PFDA exposure increased
hepatic expression of ROS-responsive genes fMaher etal.. 2008: Permadi etal.. 19931 and
This document is a draft for review purposes only and does not constitute Agency policy.
D-10 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
microsomal lipid peroxidation fCai etal.. 19951 in C57BL/6J mice. In Sprague Dawley and Wistar
rats, PFDA exposure consistently altered expression of ROS-sensitive proteins known to respond to
increased ROS including, glutathione-S-transferase, catalase, and glutathione reductase f Chen etal..
2001: Kim etal.. 1998: Glauertetal.. 1992: Ikeda etal.. 19851. These findings are supported by the
observation that PFDA exposure results in the activation of the ROS-sensitive transcription factor,
Nrf2, in C57BL/6J mice (as indicated by the increase in the hepatic expression of the Nrf2 gene
marker, Nqol) fMaher et al.. 20081. Studies in PPARa -null mice determined that PFDA-mediated
activation of the mouse Nrf2 was independent of PPARa fMaher etal.. 20081. Moreover, PFDA was
associated with an increase in oxidative DNA damage in rat liver (Huang etal.. 1994: Takagi etal..
19911 in studies with repeated-dose exposure up to 54 weeks, while no alterations in oxidative
DNA damage fKim etal.. 19981. lipid peroxidation f Glauertetal.. 19921. or changes in cellular
antioxidant levels f Glauertetal.. 19921 were reported in single exposure studies in rats. Notably,
induction of microsomal lipid peroxidation in mice was also achieved after repeated-dose exposure
to PFDA for 2 weeks (Cai etal.. 19951.
PFDA exposure induced ROS levels (Oio etal.. 2021: Wiels0e etal.. 20151 and reduced
intracellular glutathione (GSH) (Oio etal.. 20211 in HepG2 cells but did not affect the total cellular
antioxidant capacity fWielsae etal.. 20151.
The available evidence suggests that PFDA exposure increases ROS production in animal
models and in HepG2 cells and may also promote ROS-related cellular damage (e.g., DNA oxidation
and lipid peroxidation) in rodent species after prolonged or repeated exposure. The specific
involvement of Nrf2 and other cell signaling pathways in PFDA-induced ROS and potential effects
on cellular antioxidant capacity and oxidative cellular and tissue damage with prolonged chemical
exposure remains to be elucidated.
Mitochondrial Damage
Mitochondrial damage is a mechanism associated with toxicant-induced alterations in
hepatocellular lipid balance (Angrish etal.. 20161 and increased liver toxicity (Wahlang etal..
20131. Damage to mitochondria caused by oxidative stress, attenuation in mitochondrial
transmembrane potential, and alterations in membrane permeability, electron transport and
calcium fluxes are considered stimuli that induce hepatic steatosis (Kaiser etal.. 20121 and
mitochondrial-mediatedliver cell death fLi etal.. 2017: Cao etal.. 20161.
Several in vivo studies using different animal species and strains have evaluated PFDA-
induced responses in hepatic mitochondria. In Sprague Dawley rats, exposure to PFDA led to
reduced cytochrome c oxidase activity (Harrison etal.. 19881 and increased mitochondrial swelling
(Harrison etal.. 19881. a response that can lead to disruption of the mitochondrial membrane
(Taeschke etal.. 20121. Consistent with this, PFDA exposure led to increased swelling and structural
alterations in liver mitochondria in CF-1 mice, Fischer rats, Syrian hamsters, and Guinea pigs;
responses varied across species with rats being most sensitive fVan Rafelghem et al.. 19871. In
C57BL/6J mice and Fischer rats, PFDA treatment caused alterations in mitochondrial protein
This document is a draft for review purposes only and does not constitute Agency policy.
D-ll DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
content and increased mitochondrial enzyme activity (Permadi etal.. 19931: (Witzmann and
Parker. 1991: Kelling etal.. 19871. In vitro studies reported that isolated rat liver mitochondria
exposed to PFDA display uncoupling of electron transport and oxidative phosphorylation fLanglev.
19901 and induction of mitochondrial permeability transition fWallace etal.. 20131. In primary
Sprague Dawley rat hepatocytes, PFDA treatment resulted in decreased mitochondrial metabolic
functions fVanden Heuvel et al.. 19911. In vitro HTS data showed changes in mitochondrial mass
but no effects on mitochondrial membrane potential in HepG2 cells after PFDA exposure (see
Table E-l).
Overall, in vivo and in vitro studies suggest that PFDA exposure disrupts hepatic
mitochondrial proteins, integrity and function, and some of the observed effects appeared to be
conserved across different species of animals, including Syrian hamsters and Guinea pigs, known to
be low PPARa responders compared to other rodent models fCorton etal.. 2018: Hall etal.. 20121.
Additional studies assessing the potential mitochondrial effects of PFDA in human primary and
immortalized liver cells would help clarify the potential human relevance and essentiality of the
apparent PFDA-induced disruptions of mitochondrial pathways in PFDA-induced hepatotoxicity.
Inflammation
Hepatic inflammation is a mechanism associated with toxicant-induced liver injury (Angrish
etal.. 2016: Wahlang etal.. 20131. Activated macrophages and Kupffer cells produce cytokines
(e.g., TNFa, interleukin-6 [IL-6] and interleukin-10 [IL-10]) that activate hepatic stellate cells and
contribute toxicant-induced liver damage (Toshi-Barve etal.. 2015: Malhi and Gores. 20081.
PFDA-induced markers of hepatic inflammation and related mechanisms were evaluated in
studies using rodent models. PFDA increased hepatic and/or serum protein levels of the
proinflammatory cytokine TNFa in C57BL/6J mice fMaher et al.. 20081. CD-I mice fWang etal..
20201 and Fisher-344 rats fAdinehzadeh and Reo. 19981. Induction of hepatic TNF-a levels were
accompanied by increases in other proinflammatory cytokines such as IL-ip, IL-18 and IL-6 and
increases in Nod-like receptor family, pyrin domain containing 3 (NLRP3) inflammasome activation
markers such as NLRP3, adaptor apoptosis-associated speck-like protein (ASC) and caspase-1 in
CD-I mice (Wang etal.. 20201. Maher etal. (20081 also reported thatpretreatmentwith
gadolinium chloride, an anti-inflammatory agent that suppresses Kupffer cell responses,
ameliorated induction of TNFa levels in PFDA-exposed C57BL/6J mice. These results suggest that
Kupffer cells may play a role in pro-inflammatory responses following PFDA exposure. Another
study evaluated the involvement of PPARa on PFDA-induced responses related to hepatic
inflammation. Luo etal. (20171 reported that exposure to PFDA induced anti-inflammatory
responses such as increased IL-10 mRNA levels and decreased phosphorylation of NFkB in SV129
mice and that these effects did not occur in exposed PPARa-null animals. Hepatic TNFa and IL-6
mRNA levels were unaffected by exposure regardless of the genetic background of the animals.
Similarly, Li etal. (20221 showed enrichment of gene pathways associated with anti-inflammatory
This document is a draft for review purposes only and does not constitute Agency policy.
D-12 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
responses in the liver of female C57BL/6J mice exposed to PFDA. Specifically, mRNA expression of
cytokines IL-ip and IL-18, caspase-1, inflammasome-related genes (NLRP1, NLRP3, and NLRC4)
and key regulators of inflammasome assembly (e.g., cellular inhibitor of apoptosis 2 [cIAP2]) were
suppressed. The data also showed inhibition of T helper cell type 1 (Thl) differentiation in mouse
livers treated with PFDA.
The inconsistent responses on TNFa levels between Luo et al. f20171 versus Maher et al.
(20081. Adinehzadeh and Reo (19981 and Wangetal. (20201 may have been due to differences in
experimental design. Adinehzadeh and Reo (19981 and Maher etal. (20081 measured protein
levels 24 and 48 hours, respectively, after a single dose of 50-80 mg/kg via i.p. injection, whereas
Luo etal. f 20171 measured transcription (i.e., mRNA levels) on day 5 after a single i.p. injection of
80 mg/kg. The negative response on TNFa in the Luo etal. f 20171 study is consistent with the
observed anti-inflammatory response (i.e., inhibition NFkB and IL-10) and may reflect a
compensatory mechanism following initial acute hepatic injury fLuo etal.. 20171. Furthermore,
Wang etal. f20201 evaluated protein levels of TNFa after oral administration ofPFDA (13 mg/kg)
for 12 days, demonstrating induction of TNF-a and other pro-inflammatory markers with sustained
PFDA exposure.
In summary, although uncertainties remain, PFDA exposure appears capable of promoting
both pro- and anti-inflammatory responses in rodents, and PPARa may be involved in some of
these effects.
Cellular Stress
Several in vivo studies have evaluated markers of cellular stress after exposure to PFDA. As
described in the Animal Studies section for liver effects in the main assessment document (see
Section 3.2.1), short-term oral exposure to PFDA has been shown to promote degenerative changes
such as necrosis (Frawlev etal.. 2018: NTP. 20181 and increase in serum biomarkers ofhepatocyte
damage in Sprague Dawley rats fNTP. 20181 and CD-I mice fWang etal.. 20201. Liver cell necrosis
can promote steatohepatitis and fibrosis by exacerbating tissue damage via increased release of
cellular contents which in turn trigger proinflammatory responses and death of neighboring
hepatocytes (Cattlev and Cullen. 2018: Toshi-Barve etal.. 20151. One study using Wistar rats
evaluated PFDA-induced effects on cytoskeletal proteins and reported no exposure related
alterations (Witzmann and Parker. 19911. Additional effects indicative of cell damage/stress
include PFDA-induced disruptions to the endoplasmic reticulum in the livers of Fischer or Sprague-
Dawley rats, CD-I mice, Syrian hamsters, and Guinea pigs fHarrison et al.. 1988: Van Rafelghem et
al.. 19871. and dysregulation in intercellular gap junctions in Fischer rat and WB-F344 liver
epithelial cells fSovadinova etal.. 20151. Wang etal. f20201 also reported increased expression of
proapoptotic protein markers, Bax and cleaved caspase-3, in the liver of CD-I mice exposed to
PFDA. Furthermore, PFDA exposure was associated with increases in serum markers of hepatocyte
and biliary damage (ALT, AST, and ALP) in wildtype SV129 mice that corresponded with the
This document is a draft for review purposes only and does not constitute Agency policy.
D-13 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
activation of responses indicative of cellular stress signaling, including phosphorylation of JNK and
its downstream target, ATF-2 fLuo etal.. 20171. Notably, PPARa-null animals did not show these
effects fLuo etal.. 20171.
Cell viability and DNA damage were not affected in HepG2 cells exposed to PFDA
concentrations of up to 100 [J.M across two studies fRosenmai et al.. 2 018: Wielsae etal.. 20151 but
three other studies reported that PFDA induced cytotoxicity in HepG2 cells in a concentration-
dependent manner (effective concentrations causing 50% cytotoxicity [IC50] were 14.10-15 [J.M)
(Oio etal.. 2021: Oio etal.. 2020: Buhrke etal.. 20131. Similarly, PFDA elevated markers of cellular
stress and cytotoxicity in HTS assays conducted in HepG2 cells at higher concentrations (AC50
values ranging from 106.54 to 122.76 |iM). PFDA-induced cytotoxicity was also reported in
HepaRG cells ffAbe etal.. 20171 and Table E-l of the ToxCast/Tox21 data summary), primary rat
and human hepatocytes fRosen etal.. 20131. immortalized human fetal liver cells (HL-7702) fHu et
al.. 20141.
Overall, the available evidence suggests that PFDA exposure increases hepatocyte
cytotoxicity in in vitro and in vivo animal models, including species considered less sensitive to
PPARa activation (i.e., Syrian hamsters and Guinea pigs). Studies using null animals suggest that
stress responses related to disruption of bile acid homeostasis in mice may be mediated, at least in
part, by PPARa. However, the potential involvement of other cellular signaling pathways in
PFDA-induced liver cell stress has not been investigated.
Metabolic Effects
Toxicant-induced alterations in hepatocyte function can result in abnormal metabolism and
accumulation of cholesterol, fatty acids and triglycerides, and exacerbate effects caused by steatosis
fAngrish etal.. 20161. which in turn may increase susceptibility to other insults or progress to
steatohepatitis fYangetal.. 2014: Wahlangetal.. 20131.
PFDA-induced effects on liver metabolic function have been evaluated in multiple rodent
models. In Wistar, Fischer, and Sprague-Dawley rats PFDA exposure was associated with
alterations in lipid composition (Adinehzadeh etal.. 1999: Yamamoto and Kawashima. 1997: Olson
and Andersen. 19831. fatty acid transport fVanden Heuvel et al.. 19931 and metabolism (Reo etal..
1994: Davis etal.. 19911: and increased fatty acid and triglyceride accumulation (Kudo and
Kawashima. 2003: Adinehzadeh and Reo. 1998: Kawashima etal.. 1995: Sterchele etal.. 1994:
Harrison et al.. 1988: Van Rafelghem etal.. 19881. Rat studies have also reported increased hepatic
levels of cholesterol fKawashima et al.. 19951. bilirubin, and bile acids fNTP. 20181: decreased
microsomal electron transport (Kawashima etal.. 1995: Van Rafelghem and Andersen. 19881:
alterations in hepatic cholesterol metabolism (Davis etal.. 19911: glucose transport (Goecke-Flora
etal.. 19951 and metabolism (Goecke etal.. 19941: and decreased albumin levels (NTP. 2018:
Witzmann and Parker. 19911. PFDA also increases peroxisomal proliferation (Van Rafelghem et al..
19871. activity of responsive enzymes such as acyl-CoA oxidases fNTP. 2018: Kim etal.. 1998:
Huang etal.. 1994: Borges etal.. 1993: Vanden Heuvel etal.. 1993: Borges etal.. 1992: Glauertetal..
This document is a draft for review purposes only and does not constitute Agency policy.
D-14 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
1992: Intrasuksri and Feller. 1991: Kozuka etal.. 1991a: Borges etal.. 19901. and p-oxidation (Kudo
and Kawashima. 2003: Kudo etal.. 2000: Adinehzadeh etal.. 1999: Kawashima etal.. 1995: Kozuka
et al.. 1991b], which are consistent with the evidence of PPARa activation in experimental animal
models (see synthesis on Molecular Initiating Events above). As mentioned previously, PPARs,
including PPARa, regulate genes involved in lipid and cholesterol metabolism and promote (3-
oxidation of fatty acids fXu etal.. 20051. The findings from in vivo studies are supported by cell
culture studies using primary rat hepatocytes that report alterations in fatty acid metabolism
(Vanden Heuvel etal.. 19911 and increased peroxisomal p-oxidation (Kudo etal.. 20001.
Mice exposed to PFDA also demonstrate alterations in hepatic metabolic functions. PFDA
exposure increased activity of fatty acid metabolizing enzymes (Permadi et al.. 19931 and increased
hepatic lipid accumulation in C57BL/6J mice (Brewster and Birnbaum. 19891. an initial
manifestation of fatty liver disease that may progress to fibrosis (Wahlang etal.. 20131. PFDA
exposure caused alterations in the levels of bile acid metabolizing enzymes and transporters and
increased serum levels of several indicators of cholestasis (including bile acids and their
components and bilirubin) in mice fLuo etal.. 2017: Maher etal.. 20081 but PPARa-null animals
were resistant to these effects fLuo etal.. 20171. Finally, Van Rafelghem et al. f 19871 reported
extensive hepatic lipid vacuolization in hamsters and guinea pigs (and to a lesser extent in rats or
mice) after PFDA treatment
Studies examining PFDA-mediated liver metabolic effects in human models are mostly
lacking. A study by Zhang etal. f20131 showed binding affinity towards the human liver fatty acid
protein by multiple PFAS, including PFDA, which may disrupt fatty acid uptake and transport
The available evidence suggests that PFDA exposure alters liver metabolic functions across
multiple rodent species, and studies using genetically modified animals suggest that PFDA-induced
disruption of bile acid homeostasis is at least partially mediated by PPARa. More studies are
needed to understand the specific role that PPARa and other cell signaling pathways play in PFDA-
induced alterations in liver metabolic functions involving bile acid, glucose, lipid and cholesterol
metabolism and under what conditions these alterations might lead to steatohepatitis and other
liver pathologies in humans following prolonged chemical exposure.
D.3.3. ORGAN-LEVEL EFFECTS
Animal toxicity studies via the oral route have reported effects on histological and clinical
markers and organ weight measures, which are indicative of adverse responses in the liver. These
include changes in the incidence of hepatocellular necrosis, serum biomarkers of hepatobiliary and
liver damage and increased liver weights (see synthesis of Animal studies). A study by fNTP. 20181
compared liver effects in rats after short-term exposure between PFDA (and other PFAS) and
Wyeth-14,643, which was used as a positive control for PPARa activation. Much like PFDA,
Wyeth-14,643 caused increases in liver weights, changes in liver biomarkers in the blood and
hepatocyte hypertrophy; however, no evidence of necrosis or other degenerative lesions were
This document is a draft for review purposes only and does not constitute Agency policy.
D-15 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
associated with Wyeth-14,643 exposure. The findings provide support for the hypothesis that
some PFDA-induced liver responses are mediated by mechanisms independent of PPARa.
Additional evidence of PFDA-induced liver weight changes from i.p. injection studies is
described herein. Several studies using rats and mice support increases in liver weight following
PFDA exposure fAbe etal.. 2017: Luo etal.. 2017: Maher etal.. 2008: Kim etal.. 1998: Chen etal..
1994: Chinie etal.. 1994: Borges etal.. 1993: Borges etal.. 1992: Kozuka etal.. 1991b: Borges etal..
1990: Brewster and Birnbaum. 1989: Schramm etal.. 1989: Van Rafelghem and Andersen. 1988:
Kelling etal.. 1987: Van Rafelghem etal.. 1987: Kelling etal.. 1986: Powers and Aust. 1986: Ikeda et
al.. 1985: Olson and Andersen. 19831. One study in particular used wild type and PPARa-null mice
and reported that PFDA exposure led to increases in liver weight regardless of the genetic
background of the exposed animals fLuo etal.. 20171. Two other studies evaluated PFDA-induced
effects in Guinea pigs and Syrian hamsters. In Guinea pigs, exposure to PFDA did not have a
significant impact on relative liver weight fChinie etal.. 1994: Van Rafelghem et al.. 19871. while in
Syrian hamsters treatment was associated with increased liver weight (Van Rafelghem etal.. 19871.
As described above, Guinea pigs and Syrian hamsters are less responsive to PPARa activation when
compared to other rodent models. However, the observation that PFDA exposure caused increases
in liver weights in Syrian hamsters and PPARa-null mice suggests that other cell signaling pathways
may be contributing to PFDA-induced hepatomegaly in hamsters.
Overall, the available evidence from in vivo studies reports that PFDA exposure results in
organ-level effects, such as increases in liver weights that are consistently observed across multiple
species and may be mediated, at least in part, by PPARa-independent mechanisms.
This document is a draft for review purposes only and does not constitute Agency policy.
D-16 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
APPENDIX E. ANALYSIS OF RELEVANT HIGH-
THROUGHPUT SCREENING ASSAYS FROM EPA'S
CHEMICALS DASHBOARD
E.l. IN VITRO BIOACTIVITY DATA RELEVANT TO THE MECHANISMS OF
PFDA-INDUCED LIVER EFFECTS
In vitro high throughput screening (HTS) assays for PFDA were downloaded from EPA's
CompTox Chemicals Dashboard fhttps://comptox.epa.gov/dashboard) ffU.S. EPA. 20191. accessed
November 3, 2022) which provides bioactivity data from the ToxCast and Tox21 collaborative
projects. Available information most pertinent to the analysis of the potential mechanisms of
PFDA-induced liver effects was extracted to supplement and augment mechanistic findings from
studies in the peer-reviewed literature previously described. Results (active/inactive, AC50 values,
and scaled activity) from in vitro assays in human hepatoma HepG2 cells and metabolically
competent human hepatic progenitor cells (HepaRG) cells were obtained, filtering out background
control assays and nonspecific responses from inducible reporter gene assays analyzed in the
negative fitting direction relative to the control ("_dn"). Bioactivity data were analyzed based on the
type of biological response or gene target using the annotation structure within the ToxCast assay
summary information ffU.S. EPA. 20191. accessed November 3, 2022).
PFDA was active in 74 of 238 unique assay endpoints (~31%) in HepG2 and HepaRG cells,
inducing a range of cell- and gene-specific changes (see Figure E-l and Table E-l). PFDA was
associated with cell cycle arrest and proliferation responses and induction of markers of oxidative
stress and cell death (see Table E-l). Alterations in nuclear size and mitochondrial mass were also
observed in HTS assays for PFDA with no apparent changes in microtubule conformation and
mitochondrial membrane potential and respiration (see Table E-l). Further, PFDA caused
upregulation of transcriptional activity that occurred generally at lower effective concentrations
(i.e., AC50) compared to the cell-based responses (see Figure E-l). Specifically, PFDA induced the
expression of CYP450 enzymes, growth factors, transporters and transcriptional factors, including
several xenobiotic-sensing nuclear receptors previously implicated in the mechanisms of liver
toxicity ofPFDA or other PFAS (i.e., PPARa/y, PXR, and FXR) (see Figure E-2 and Table E-l).
In summary, PFDA elicited in vitro responses in HTS assays conducted in HepG2 and
HepaRG cells most consistently for cellular stress and cytotoxicity. Additionally, induction of gene
target pathways corresponding to several transcriptional factor/nuclear receptor activities
occurred upstream of the cell-mediated responses, albeit at similar effective concentrations.
This document is a draft for review purposes only and does not constitute Agency policy.
E-l DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Nuclear receptor activities were investigated more closely to provide further insights into
the putative interaction ofPFDA with these receptor-mediated signaling pathways in
ToxCast/Tox21 assays profiling multiple endpoints (e.g., receptor binding, coregulator recruitment,
and gene transactivation) and cell types (see Table E-2). As mentioned above, PFDA induced
activity of specific steroid/xenobiotic sensing receptors, most notably FXR, PPAR and PXR (see
Figure E-2A). PFDA interacted with the human FXR in a receptor-ligand binding assay evaluating
agonist activity and in one of two independent assays measuring transcriptional activity in HepG2
cells but was inactive in four FXR-related assays in human embryonic kidney cells (HEK293T),
targeting receptor/cofactor recruitment and agonist/antagonist activities (see Table E-2).
Upregulation of transcriptional activity for PPARa and PPARy but not PPAR(3/S (PPARD) was
demonstrated in HepG2 cells, and PFDA was found to interact with the human PPARy (but not
human PPARa) in a receptor-ligand binding assay (see Table E-2). No activity was detected in
assays conducted in HEK293T cells profiling agonist/antagonist activities for PPARy or PPAR(3/S or
receptor/cofactor recruitment for PPARy (see Table E-2). PFDA was active in two of four assays for
PXR, showing transcriptional induction in HepG2 cells (one of two independent assays) and direct
binding to the human PXR but no activity in an agonist assay using HepG2 cells (see Table E-2).
HNF4A, NURR1, RAR, ROR, RXR, and VDR were also targets of PFDA in reporter gene assays using
HepG2 cells and antagonist activity toward ERR was reported in HEK293T cells (see Table E-2).
PFDA targeted the ER and AR in in vitro HTS assays; however, overall activity for these receptors
was low (refer to Appendix E.2 for additional details on the HTS results for the ER and AR). PFDA
showed no appreciable activity in assays for GR, CAR, LXR, TR, and PR (Figure E-2A). Comparison
of AC50 values across the nuclear receptor assays indicate that PFDA exerts the highest potency
toward the human FXR with the lowest AC50 of 0.52 [J.M in a cell-free receptor binding assay
(Figure E-2B), which is below the lower bound of the ToxCast cytotoxicity limit estimated for this
chemical (7.108 [iM) ffU.S. EPA. 20191. accessed November 3, 2022).
Altogether, the results of the ToxCast/Tox21 HTS analysis provide some mechanistic
support for the PFDA-induced liver effects. PFDA caused upregulation of transcriptional activity in
human hepatoma HepG2 cells involving multiple nuclear receptor pathways previously implicated
in the MOA for PFDA-induced liver toxicity, namely PXR, FXR, and PPARa/y. These target gene
responses were associated with the induction of cellular stress/cytotoxicity. PFDA also interacted
directly with the human PXR, FXR, and PPARy in receptor binding assays, demonstrating particular
sensitivity for the human FXR at concentrations below those associated with cytotoxicity and
suggesting that FXR may be an important target for this chemical.
This document is a draft for review purposes only and does not constitute Agency policy.
E-2 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
12
10
>
4-»
'>
u
6
¦D
0)
0.1
0
10
AC50 (|iM)
.* H
100
• Cell cycle
• Cellular/organelle
conformation
• Cellular
stress/cytotoxicity
• Mitochondrial toxicity
• Upregulation of
transcriptional activity
1000
Figure E-l. Bioactivity data for PFDA from in vitro HTS ToxCast/Tox21 assays
conducted in human liver cell lines (HepG2 and HepaRG cells).
Scatterplots show AC50 and scaled activity values from assays visualized according to the type of biological
response. AC50 values refer to the concentration that elicits half maximal response and the scaled activity refers
to the response value divided by the activity cutoff. Assays for which chemicals were inactive are not displayed.
Additional information on all tested assays in HepG2 and HepaRG cells can be found in Table E-l.
This document is a draft for review purposes only and does not constitute Agency policy.
E-3 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
c
5 t
i
i Ł
n I—
5 O
<<<<<<
cc < < o
~ 2 Z n
Ol.
3
a
a
3
a
3
*
u
z
to
U,
l/l
o
UJ
cc
K
UJ
tc
cc
a
cc
O
-Q
cc
2
>\
o
<
o1
<
u u u u u
5 <.
~ z
cc o —
a X
UJ o
I . J
< < < <
I
< 3
e> < <|
S5 ^ o o o
CD CD
LL. |
e> o
—i cc
<3 Q.
I x X x <
H I I | UJ
N h
tc lu i.. r*
i I
< < s «
CYPs
Growth
factors
Nuclear receptors
Transcriptional
factors
Transporters
Other
Figure E-2. Analysis of PFDA-induced upregulation of transcriptional activity in ToxCast/Tox21 assays conducted
in human liver cell lines (HepG2 and HepaRG cells).
Bar graph compares AC50 values (concentration at half maximal response) for active assays. The scale for the AC50 values is shown in reverse order to
visualize the most sensitive assays (the higher bar indicates a lower AC50 value). Additional information on the transcriptional activity assays can be found in
Table E-l.
This document is a draft for review purposes only and does not constitute Agency policy,
E-4 DRAFT-DO NOT CITE OR QUOTE '
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Figure E-3. Analysis of PFDA-induced nuclear receptor-related activities in ToxCast/Tox21 assays across multiple
endpoints and cell types.
Panel A summarizes active/inactive calls from nuclear receptor assays mapped to specific target genes. Panel B compares AC50 values (concentration at half
maximal response) for active assays. The scale for the AC50 values is shown in reverse order to visualize the most sensitive nuclear receptor activities (the
higher bar indicates a lower AC50 value). Additional information on all tested nuclear receptor-related assays can be found in Table E-2.
Abbreviations: AR, androgen receptor; CAR, constitutive androgen receptor; ER, estrogen receptor; ERR, estrogen-related receptor; FXR, farnesoid X receptor;
GR, glucocorticoid receptor; HNF4A, hepatocyte nuclear factors 4 alpha; LXR, liver X receptor; NURR1, nuclear receptor related-1 protein; PPAR, peroxisome
proliferator-activated receptor; PXR, pregnane X receptor; RAR, retinoid acid receptor; ROR, RAR-related orphan receptor; RXR, retinoid X receptor; TR,
thyroid hormone receptor; VDR, vitamin D receptor.
This document is a draft for review purposes only and does not constitute Agency policy,
E-5 DRAFT-DO NOT CITE OR QUOTE '
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table E-l. Bioactivity summary for PFDA from in vitro HTS assays from
ToxCast/Tox21 conducted in human liver cell lines (HepG2 and HepaRG cells)
and grouped by biological response/targetab
Assay name
Activity call
Scaled
activity
AC50 (nM)
Assay design
Cell line
Cell cycle
APR_HepG2_CellCycleArrest_72h_dn
Active
1.23
69.51
morphology reporter
HepG2
APR_HepG2_MitoticArrest_24h_up
Active
2.25
107.91
morphology reporter
HepG2
APR_HepG2_MitoticArrest_72h_up
Active
2.44
98.57
morphology reporter
HepG2
APR_HepG2_CellCycleArrest_24h_dn
Inactive
NA
NA
morphology reporter
HepG2
APR_HepG2_CellCycleArrest_24h_up
Inactive
NA
NA
morphology reporter
HepG2
APR_HepG2_CellCycleArrest_72h_up
Inactive
NA
NA
morphology reporter
HepG2
APR_HepG2_MitoticArrest_24h_dn
Inactive
NA
NA
morphology reporter
HepG2
APR_HepG2_MitoticArrest_72h_dn
Inactive
NA
NA
morphology reporter
HepG2
Cellular/organelle conformation
APR_HepG2_NuclearSize_24h_dn
Active
1.33
128.23
morphology reporter
HepG2
APR_HepG2_NuclearSize_72h_dn
Active
1.51
121.20
morphology reporter
HepG2
APR_HepG2_MicrotubuleCSK_24h_dn
Inactive
NA
NA
conformation reporter
HepG2
APR_HepG2_MicrotubuleCSK_24h_up
Inactive
NA
NA
conformation reporter
HepG2
APR_HepG2_MicrotubuleCSK_72h_dn
Inactive
NA
NA
conformation reporter
HepG2
APR_HepG2_MicrotubuleCSK_72h_up
Inactive
NA
NA
conformation reporter
HepG2
APR_HepG2_NuclearSize_24h_up
Inactive
NA
NA
morphology reporter
HepG2
APR_HepG2_NuclearSize_72h_up
Inactive
NA
NA
morphology reporter
HepG2
Cellular stress/cytotoxicity
APR_HepG2_CellLoss_24h_dn
Active
3.75
108.88
viability reporter
HepG2
APR_HepG2_CellLoss_72h_dn
Active
3.63
106.54
viability reporter
HepG2
APR_HepG2_p53Act_24h_up
Active
1.61
107.89
viability reporter
HepG2
APR_HepG2_p53Act_72h_up
Active
2.28
113.49
viability reporter
HepG2
APR_HepG2_P-H2AX_24h_up
Active
2.35
112.97
viability reporter
HepG2
APR_HepG2_P-H2AX_72h_up
Active
2.88
108.81
viability reporter
HepG2
APR_HepG2_StressKinase_72h_up
Active
1.50
122.76
enzyme reporter
HepG2
LTEA_HepaRG_LDH_cytotoxicity
Active
7.31
66.39
viability reporter
HepaRG
APR_HepG2_CellLoss_24h_up
Inactive
NA
NA
viability reporter
HepG2
APR_HepG2_CellLoss_72h_up
Inactive
NA
NA
viability reporter
HepG2
APR_HepG2_p53Act_24h_dn
Inactive
NA
NA
viability reporter
HepG2
APR_HepG2_p53Act_72h_dn
Inactive
NA
NA
viability reporter
HepG2
APR_HepG2_P-H2AX_24h_dn
Inactive
NA
NA
viability reporter
HepG2
APR_HepG2_P-H2AX_72h_dn
Inactive
NA
NA
viability reporter
HepG2
APR_HepG2_StressKinase_24h_dn
Inactive
NA
NA
enzyme reporter
HepG2
APR_HepG2_StressKinase_24h_up
Inactive
NA
NA
enzyme reporter
HepG2
APR_HepG2_StressKinase_72h_dn
Inactive
NA
NA
enzyme reporter
HepG2
ATG_XTT_Cytotoxicity_up
Inactive
NA
NA
viability reporter
HepG2
This document is a draft for review purposes only and does not constitute Agency policy.
E-6 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Assay name
Activity call
Scaled
activity
AC50 (nM)
Assay design
Cell line
CCTE_Simmons_MITO_viability
Inactive
NA
NA
viability reporter
HepG2
TOX21_AhR_LUC_Agonist_viability
Inactive
NA
NA
viability reporter
HepG2
TOX21_ARE_BLA_agonist_viability
Inactive
NA
NA
viability reporter
HepG2
TOX21_CAR_Agonist_viabillity
Inactive
NA
NA
viability reporter
HepG2
TOX21_CAR_Antagonist_viability
Inactive
NA
NA
viability reporter
HepG2
TOX21_CASP3_H E PG 2
Inactive
NA
NA
inducible reporter
HepG2
TOX2 l_CASP3_HEPG2_viability
Inactive
NA
NA
viability reporter
HepG2
TOX21_MMP_viability
Inactive
NA
NA
viability reporter
HepG2
TOX21_PXR_viability
Inactive
NA
NA
viability reporter
HepG2
TOX21 RT HEPG2 FLO OOhr Ctrl viability
Inactive
NA
NA
viability reporter
HepG2
TOX2 l_RT_HEPG2_FLO_08hr_viability
Inactive
NA
NA
viability reporter
HepG2
TOX21 RT HEPG2 FLO 16hr Ctrl viability
Inactive
NA
NA
viability reporter
HepG2
TOX2 l_RT_HEPG2_FLO_24hr_viability
Inactive
NA
NA
viability reporter
HepG2
TOX21 RT HEPG2 FLO 32hr Ctrl viability
Inactive
NA
NA
viability reporter
HepG2
TOX21 RT HEPG2 FLO 40hr Ctrl viability
Inactive
NA
NA
viability reporter
HepG2
TOX21 RT HEPG2 GLO OOhr Ctrl viability
Inactive
NA
NA
viability reporter
HepG2
TOX21 RT HEPG2 GLO 08hr Ctrl viability
Inactive
NA
NA
viability reporter
HepG2
TOX21 RT HEPG2 GLO 16hr Ctrl viability
Inactive
NA
NA
viability reporter
HepG2
TOX21 RT HEPG2 GLO 24hr Ctrl viability
Inactive
NA
NA
viability reporter
HepG2
TOX21 RT HEPG2 GLO 32hr Ctrl viability
Inactive
NA
NA
viability reporter
HepG2
TOX21_RT_H EPG2_G L0_40h r_viabil ity
Inactive
NA
NA
viability reporter
HepG2
Mitochondrial toxicity
APR_HepG2_MitoMass_24h_dn
Active
4.72
117.36
morphology reporter
HepG2
APR_HepG2_MitoMass_72h_dn
Active
4.83
113.92
morphology reporter
HepG2
APR_HepG2_MitoMass_24h_up
Inactive
NA
NA
morphology reporter
HepG2
APR_HepG2_MitoMass_72h_up
Inactive
NA
NA
morphology reporter
HepG2
APR_HepG2_MitoMembPot_24h_dn
Inactive
NA
NA
membrane potential
reporter
HepG2
APR_HepG2_MitoMembPot_24h_up
Inactive
NA
NA
membrane potential
reporter
HepG2
APR_HepG2_MitoMembPot_72h_dn
Inactive
NA
NA
membrane potential
reporter
HepG2
APR_HepG2_MitoMembPot_72h_up
Inactive
NA
NA
membrane potential
reporter
HepG2
CCTE_Simmons_MITO_basal_resp_rate_OCR_dn
Inactive
NA
NA
respirometric reporter
HepG2
CCTE_Simmons_MITO_basal_resp_rate_OCR_up
Inactive
NA
NA
respirometric reporter
HepG2
CCTE_Simmons_MITO_inhib_resp_rate_OCR_dn
Inactive
NA
NA
respirometric reporter
HepG2
CCTE_Simmons_MITO_inhib_resp_rate_OCR_up
Inactive
NA
NA
respirometric reporter
HepG2
CCTE_Simmons_MITO_max_resp_rate_OCR_dn
Inactive
NA
NA
respirometric reporter
HepG2
CCTE_Simmons_MITO_max_resp_rate_OCR_up
Inactive
NA
NA
respirometric reporter
HepG2
TOX21_MMP_ratio_down
Inactive
NA
NA
membrane potential
reporter
HepG2
This document is a draft for review purposes only and does not constitute Agency policy.
E-7 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Assay name
Activity call
Scaled
activity
AC50 (nM)
Assay design
Cell line
TOX21_M M P_rati o_u p
Inactive
NA
NA
membrane potential
reporter
HepG2
Upregulation of transcriptional activity
ATG_EGR_CIS_up
Active
1.19
19.92377
inducible reporter
HepG2
ATG_ERa_TRANS_up
Active
1.50
16.43561
inducible reporter
HepG2
ATG_FXR_TRANS_up
Active
2.28
18.99931
inducible reporter
HepG2
ATG_HNF4a_TRANS_up
Active
1.59
80.32058
inducible reporter
HepG2
ATG_HSE_CIS_up
Active
2.31
28.98294
inducible reporter
HepG2
ATG_MRE_CIS_up
Active
1.78
12.43083
inducible reporter
HepG2
ATG_NRF2_ARE_CIS_up
Active
3.54
20.6361
inducible reporter
HepG2
ATG_NURRl_TRANS_up
Active
1.87
25.56622
inducible reporter
HepG2
ATG_Pax6_CIS_up
Active
1.56
29.70391
inducible reporter
HepG2
ATG_PPARa_TRANS_up
Active
1.30
18.12921
inducible reporter
HepG2
ATG_PPARg_TRANS_up
Active
1.31
11.97573
inducible reporter
HepG2
ATG_PPRE_CIS_up
Active
2.29
25.89358
inducible reporter
HepG2
ATG_PXR_TRANS_up
Active
1.42
30.14653
inducible reporter
HepG2
ATG_RARg_TRANS_up
Active
1.50
21.20087
inducible reporter
HepG2
ATG_RORE_CIS_up
Active
1.41
21.068
inducible reporter
HepG2
ATG_RXRb_TRANS_up
Active
4.26
16.95397
inducible reporter
HepG2
ATG_TGFb_CIS_up
Active
2.94
14.44227
inducible reporter
HepG2
ATG_VDRE_CIS_up
Active
1.25
19.38327
inducible reporter
HepG2
ATG_Xbpl_CIS_up
Active
2.05
31.73703
inducible reporter
HepG2
LTEA_HepaRG_ABCC3_up
Active
1.71
17.53302
inducible reporter
HepaRG
LTEA_HepaRG_ABCG2_up
Active
1.08
11.2217
inducible reporter
HepaRG
LTEA_HepaRG_BAX_up
Active
3.20
22.88926
inducible reporter
HepaRG
LTEA_HepaRG_BCL2_up
Active
6.13
14.76859
inducible reporter
HepaRG
LTEA_HepaRG_BCL2Lll_up
Active
3.41
22.55949
inducible reporter
HepaRG
LTEA_HepaRG_CASP8_up
Active
2.45
33.09058
inducible reporter
HepaRG
LTEA_HepaRG_CCNDl_up
Active
3.50
21.35921
inducible reporter
HepaRG
LTEA_HepaRG_CDKNlA_up
Active
2.49
13.57402
inducible reporter
HepaRG
LTEA_HepaRG_CFLAR_up
Active
3.93
23.40259
inducible reporter
HepaRG
LTEA_He pa RG_CYP1Al_u p
Active
1.40
37.12706
inducible reporter
HepaRG
LTEA_HepaRG_CYP2C19_up
Active
1.08
0.911362
inducible reporter
HepaRG
LTEA_He pa RG_CYP4A1l_u p
Active
3.00
4.084149
inducible reporter
HepaRG
LTEA_HepaRG_CYP4A22_up
Active
2.39
5.093503
inducible reporter
HepaRG
LTEA_HepaRG_DDIT3_up
Active
9.91
24.56621
inducible reporter
HepaRG
LTEA_HepaRG_EGRl_up
Active
2.35
27.13929
inducible reporter
HepaRG
LTEA_HepaRG_EZR_up
Active
2.29
20.2641
inducible reporter
HepaRG
LTEA_HepaRG_FAS_up
Active
2.46
23.51647
inducible reporter
HepaRG
LTEA_HepaRG_F0X03_up
Active
1.08
17.79771
inducible reporter
HepaRG
This document is a draft for review purposes only and does not constitute Agency policy.
E-8 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Assay name
Activity call
Scaled
activity
AC50 (nM)
Assay design
Cell line
LTEA_HepaRG_GADD45B_up
Active
1.37
316.2278
inducible reporter
HepaRG
LTEA_HepaRG_GADD45G_up
Active
3.77
16.26879
inducible reporter
HepaRG
LTEA_HepaRG_GCLC_up
Active
2.58
13.26529
inducible reporter
HepaRG
LTEA_HepaRG_HSPAlA_up
Active
2.48
86.07431
inducible reporter
HepaRG
LTEA_HepaRG_ICAMl_up
Active
1.37
16.93707
inducible reporter
HepaRG
LTEA_HepaRG_IGFBPl_up
Active
5.77
24.20317
inducible reporter
HepaRG
LTEA_HepaRG_IL6_up
Active
4.33
39.10404
inducible reporter
HepaRG
LTEA_HepaRG_JUN_up
Active
1.15
13.67962
inducible reporter
HepaRG
LTEA_HepaRG_KCNKl_up
Active
1.37
31.6189
inducible reporter
HepaRG
LTEA_HepaRG_KRT19_up
Active
1.75
13.95732
inducible reporter
HepaRG
LTEA_HepaRG_LPL_up
Active
3.94
20.11038
inducible reporter
HepaRG
LTEA_HepaRG_MMPl_up
Active
3.30
38.55908
inducible reporter
HepaRG
LTEA_HepaRG_MMP10_up
Active
3.14
35.00735
inducible reporter
HepaRG
LTEA_HepaRG_MYC_up
Active
3.67
17.50487
inducible reporter
HepaRG
LTEA_HepaRG_NFE2L2_up
Active
1.16
18.16403
inducible reporter
HepaRG
LTEA_HepaRG_PDK4_up
Active
4.93
24.64551
inducible reporter
HepaRG
LTEA_HepaRG_PEG10_up
Active
2.01
12.83903
inducible reporter
HepaRG
LTEA_HepaRG_PPP2R4_up
Active
3.31
23.18532
inducible reporter
HepaRG
LTEA_HepaRG_TGFA_up
Active
3.96
21.42175
inducible reporter
HepaRG
LTEA_HepaRG_TGFBl_up
Active
1.48
18.53422
inducible reporter
HepaRG
LTEA_HepaRG_TP53_up
Active
5.61
13.70365
inducible reporter
HepaRG
TOX21_AR E_B LA_ago n ist_ratio
Active
4.79
39.41989
inducible reporter
HepG2
ATG_Ahr_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_AP_l_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_AP_2_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_AR_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_BRE_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_C_EBP_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_CAR_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_CRE_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
AT G_D R4_LXR_CI S_u p
Inactive
NA
NA
inducible reporter
HepG2
ATG_DR5_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
AT G_E_Box_CI S_u p
Inactive
NA
NA
inducible reporter
HepG2
ATG_E2F_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_ERE_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_ERRa_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_ERRg_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_Ets_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_FoxA2_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_FoxO_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
This document is a draft for review purposes only and does not constitute Agency policy.
E-9 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Assay name
Activity call
Scaled
activity
AC50 (nM)
Assay design
Cell line
ATG_GATA_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_GLI_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_GR_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_GRE_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_HIFla_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_HNF6_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_IRl_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_ISRE_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_LXRa_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
AT G_LXR b_TR ANS_u p
Inactive
NA
NA
inducible reporter
HepG2
ATG_Myb_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_Myc_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_N F_kB_CIS_u p
Inactive
NA
NA
inducible reporter
HepG2
ATG_NFI_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_NRFl_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_Oct_MLP_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_p53_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_PBREM_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_PPARd_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_PXRE_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_RARa_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_RARb_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_RORb_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_RORg_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_RXRa_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_Sox_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_Spl_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_SREBP_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_STAT3_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_TCF_b_cat_CIS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_THRal_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
ATG_VDR_TRANS_up
Inactive
NA
NA
inducible reporter
HepG2
LTEA_HepaRG_ABCBl_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_ABCBll_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_ABCC2_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_ACLY_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_ACOXl_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_ADK_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_ALPP_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_APOA5_up
Inactive
NA
NA
inducible reporter
HepaRG
This document is a draft for review purposes only and does not constitute Agency policy.
E-10 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Assay name
Activity call
Scaled
activity
AC50 (nM)
Assay design
Cell line
LTEA_HepaRG_BAD_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_BID_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_CASP3_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_CAT_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_He pa RG_CYP1A2_u p
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_CYP24Al_l_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_CYP2B6_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_CYP2C8_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_CYP2C9_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_CYP2El_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_CYP3A4_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_CYP3A5_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_CYP3A7_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_CYP7Al_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_EGF_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_FABPl_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_FASN_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_FM03_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_F0X01_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_GADD45A_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_GSTA2_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_GSTM3_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_HGF_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_HIFlA_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_HMGCS2_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_IGFl_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_IL6R_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_LIPC_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_MIR122_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_MMP3_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_NFKBl_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_NQ01_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_PTEN_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_SDHB_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_SLC10Al_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_SLC22Al_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_SLC22A6_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_SLC01Bl_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_STAT3_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_SULT2Al_up
Inactive
NA
NA
inducible reporter
HepaRG
This document is a draft for review purposes only and does not constitute Agency policy.
E-ll DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Assay name
Activity call
Scaled
activity
AC50 (nM)
Assay design
Cell line
LTEA_HepaRG_THRSP_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_TIMPl_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_TNFRSFlA_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_UGTlAl_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_UGTlA6_up
Inactive
NA
NA
inducible reporter
HepaRG
LTEA_HepaRG_XBPl_up
Inactive
NA
NA
inducible reporter
HepaRG
TOX2 l_Ah R_LU C_Ago n i st
Inactive
NA
NA
inducible reporter
HepG2
TOX2 l_CAR_Ago n i st
Inactive
NA
NA
inducible reporter
HepG2
TOX2 l_CAR_Antago n i st
Inactive
NA
NA
inducible reporter
HepG2
TOX21_PXR_Agonist
Inactive
NA
NA
inducible reporter
HepG2
aData were sourced from EPA's CompTox Chemicals Dashboard ((U.S. EPA, 2019), accessed November 3, 2022).
background control assays and nonspecific responses from inducible reporter gene assays analyzed in the
negative fitting direction relative to the control ("_dn") are not presented herein.
NA = not applicable.
This document is a draft for review purposes only and does not constitute Agency policy.
E-12 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table E-2. Bioactivity summary for PFDA from in vitro HTS assays evaluating nuclear receptor-related activities
from ToxCast/Tox21 across multiple endpoints and cell typesab c
Assay name
Activity call
Scaled
Activity
AC50 (nM)
Biological target
Assay design
Organism
Tissue
Cell line
ATG_CAR_TRANS_up
Inactive
NA
NA
CAR (NR1I3)
inducible reporter
human
liver
HepG2
ATG_PBREM_CIS_up
Inactive
NA
NA
CAR (NR1I3)
inducible reporter
human
liver
HepG2
TOX2 l_CAR_Ago n i st
Inactive
NA
NA
CAR (NR1I3)
inducible reporter
human
liver
HepG2
TOX21_CAR_Antagonist
Inactive
NA
NA
CAR (NR1I3)
inducible reporter
human
liver
HepG2
TOX21_ERR_Antagonist
Active
1.31
6.62
ERR (ESRRA)
inducible reporter
human
kidney
HEK293T
AT G_ERRa_TRANS_u p
Inactive
NA
NA
ERR (ESRRA)
inducible reporter
human
liver
HepG2
ATG_ERRg_TRANS_up
Inactive
NA
NA
ERR (ESRRA)
inducible reporter
human
liver
HepG2
TOX21_ERR_Agonist
Inactive
NA
NA
ERR (ESRRA)
inducible reporter
human
kidney
HEK293T
TOX21_PGC_ERR_Agonist
Inactive
NA
NA
ERR (ESRRA)
inducible reporter
human
kidney
HEK293T
TOX21_PGC_ERR_Antagonist
Inactive
NA
NA
ERR (ESRRG)
inducible reporter
human
kidney
HEK293T
AT G_FXR_TRANS_u p
Active
2.28
19.00
FXR (NR1H4)
inducible reporter
human
liver
HepG2
N VS_N R_h FXR_Ago n i st
Active
5.52
0.52
FXR (NR1H4)
binding reporter
human
NA
NA
ATG_IRl_CIS_up
Inactive
NA
NA
FXR (NR1H4)
inducible reporter
human
liver
HepG2
OT_FXR_FXRSRC1_0480
Inactive
NA
NA
FXR (NR1H4)
binding reporter
human
kidney
HEK293T
OT_FXR_FXRSRC1_1440
Inactive
NA
NA
FXR (NR1H4)
binding reporter
human
kidney
HEK293T
TOX2 l_FXR_BLA_ago n i st_ra ti o
Inactive
NA
NA
FXR (NR1H4)
inducible reporter
human
kidney
HEK293T
TOX2 l_FXR_BLA_a n tago n ist_ratio
Inactive
NA
NA
FXR (NR1H4)
inducible reporter
human
kidney
HEK293T
ATG_GR_TRANS_up
Inactive
NA
NA
GR (NR3C1)
inducible reporter
human
liver
HepG2
ATG_GRE_CIS_up
Inactive
NA
NA
GR (NR3C1)
inducible reporter
human
liver
HepG2
NVS_NR_hGR
Inactive
NA
NA
GR (NR3C1)
binding reporter
human
NA
NA
TOX21_GR_BLA_Agonist_ratio
Inactive
NA
NA
GR (NR3C1)
inducible reporter
human
cervix
HeLa
TOX21_G R_BLA_Antago n i st_rat io
Inactive
NA
NA
GR (NR3C1)
inducible reporter
human
cervix
HeLa
This document is a draft for review purposes only and does not constitute Agency policy.
E-13 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Assay name
Activity call
Scaled
Activity
AC50 (nM)
Biological target
Assay design
Organism
Tissue
Cell line
AT G_H N F4a_TRANS_up
Active
1.59
80.32
HNF4A
inducible reporter
human
liver
HepG2
AT G_LXR b_TRANS_u p
Inactive
NA
NA
LXR (NR1H2)
inducible reporter
human
liver
HepG2
AT G_D R4_LXR_CI S_u p
Inactive
NA
NA
LXR (NR1H2|NR1H3)
inducible reporter
human
liver
HepG2
ATG_LXRa_TRANS_up
Inactive
NA
NA
LXR (NR1H3)
inducible reporter
human
liver
HepG2
ATG_NURRl_TRANS_up
Active
1.87
25.57
NURR1 (NR4A2)
inducible reporter
human
liver
HepG2
ATG_PPARa_TRANS_up
Active
1.30
18.13
PPAR (PPARA)
inducible reporter
human
liver
HepG2
NVS_NR_hPPARa
Inactive
NA
NA
PPAR (PPARA)
binding reporter
human
NA
NA
AT G_PPARd_TRANS_u p
Inactive
NA
NA
PPAR (PPARD)
inducible reporter
human
liver
HepG2
TOX21_P PARd_B LA_ago n ist_rat io
Inactive
NA
NA
PPAR (PPARD)
inducible reporter
human
kidney
HEK293T
TOX21_PPARd_BLA_antagonist_ratio
Inactive
NA
NA
PPAR (PPARD)
inducible reporter
human
kidney
HEK293T
AT G_PPARg_TRANS_u p
Active
1.31
11.98
PPAR (PPARG)
inducible reporter
human
liver
HepG2
NVS_NR_hPPARg
Active
5.15
13.73
PPAR (PPARG)
binding reporter
human
NA
NA
OT_PPARg_PPARgSRCl_0480
Inactive
NA
NA
PPAR (PPARG)
binding reporter
human
kidney
HEK293T
OT_PPARg_PPARgSRCl_1440
Inactive
NA
NA
PPAR (PPARG)
binding reporter
human
kidney
HEK293T
TOX21_P PARg_B LA_Ago n ist_ratio
Inactive
NA
NA
PPAR (PPARG)
inducible reporter
human
kidney
HEK293T
TOX21_P PARg_B LA_a ntago n ist_rati o
Inactive
NA
NA
PPAR (PPARG)
inducible reporter
human
kidney
HEK293
ATG_PPRE_CIS_up
Active
2.29
25.89
PPAR (PPARA| PPARD | PPARG)
inducible reporter
human
liver
HepG2
TOX21_PR_BLA_Agonist_ratio
Inactive
NA
NA
PR (PGR)
inducible reporter
human
kidney
HEK293T
TOX21_PR_BLA_Antagonist_ratio
Inactive
NA
NA
PR (PGR)
inducible reporter
human
kidney
HEK293T
ATG_PXR_TRANS_up
Active
1.42
30.15
PXR (NR1I2)
inducible reporter
human
liver
HepG2
NVS_NR_hPXR
Active
2.34
32.07
PXR (NR1I2)
binding reporter
human
NA
NA
ATG_PXRE_CIS_up
Inactive
NA
NA
PXR (NR1I2)
inducible reporter
human
liver
HepG2
TOX2 l_PXR_Ago n ist
Inactive
NA
NA
PXR (NR1I2)
inducible reporter
human
liver
HepG2
AT G_RARa_TRANS_u p
Inactive
NA
NA
RAR (RARA)
inducible reporter
human
liver
HepG2
TOX2 l_RAR_LUC_Ago n ist
Inactive
NA
NA
RAR (RARA)
inducible reporter
mouse
embryo
C3H10T1/2
This document is a draft for review purposes only and does not constitute Agency policy.
E-14 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Assay name
Activity call
Scaled
Activity
AC50 (nM)
Biological target
Assay design
Organism
Tissue
Cell line
TOX21_RAR_LUC_Antagonist
Inactive
NA
NA
RAR (RARA)
inducible reporter
mouse
embryo
C3H10T1/2
ATG_RARb_TRANS_up
Inactive
NA
NA
RAR (RARB)
inducible reporter
human
liver
HepG2
ATG_RARg_TRANS_up
Active
1.50
21.20
RAR (RARG)
inducible reporter
human
liver
HepG2
ATG_DR5_CIS_up
Inactive
NA
NA
RAR(RARA|RARB|RARG)
inducible reporter
human
liver
HepG2
ATG_RO R b_TRANS_u p
Inactive
NA
NA
ROR(RORB)
inducible reporter
human
liver
HepG2
ATG_RO Rg_TRANS_u p
Inactive
NA
NA
ROR(RORC)
inducible reporter
human
liver
HepG2
TOX21_RORg_LUC_CHO_Antagonist
Inactive
NA
NA
ROR(RORC)
inducible reporter
Chinese
hamster
ovary
CHO-K1
ATG_RORE_CIS_up
Active
1.41
21.07
ROR (RORA| RORB| RORC)
inducible reporter
human
liver
HepG2
AT G_RXRa_TRANS_u p
Inactive
NA
NA
RXR (RXRA)
inducible reporter
human
liver
HepG2
OT_NURRl_NURRlRXRa_0480
Inactive
NA
NA
RXR (RXRA)
binding reporter
human
kidney
HEK293T
OT_NURRl_NURRlRXRa_1440
Inactive
NA
NA
RXR (RXRA)
binding reporter
human
kidney
HEK293T
ATG_RXRb_TRANS_up
Active
4.26
16.95
RXR (RXRB)
inducible reporter
human
liver
HepG2
AT G_TH Ra1_TRANS_u p
Inactive
NA
NA
TR (THRA)
inducible reporter
human
liver
HepG2
TOX21_TR_LU C_G H 3_Ago n i st
Inactive
NA
NA
TR (THRA|THRB)
inducible reporter
rat
pituitary
gland
GH3
TOX21_TR_LUC_GH3_Antagonist
Inactive
NA
NA
TR (THRA|THRB)
inducible reporter
rat
pituitary
gland
GH3
ATG_VDRE_CIS_up
Active
1.25
19.38
VDR
inducible reporter
human
liver
HepG2
ATG_VDR_TRANS_up
Inactive
NA
NA
VDR
inducible reporter
human
liver
HepG2
TOX21_VDR_BLA_agonist_ratio
Inactive
NA
NA
VDR
inducible reporter
human
kidney
HEK293T
TOX21_VDR_BLA_antagonist_ratio
Inactive
NA
NA
VDR
inducible reporter
human
kidney
HEK293T
aData were sourced from EPA's CompTox Chemicals Dashboard ((U.S. EPA, 2019), accessed November 3, 2022).
bNonspecific responses from inducible reporter gene assays analyzed in the negative fitting direction relative to the control ("_dn") are not presented herein.
cln vitro bioactivity data for the AR and ER are summarized in detail in Appendix E.2 and, therefore, are not presented herein.
NA = not applicable.
This document is a draft for review purposes only and does not constitute Agency policy.
E-15 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
E.2. IN VITRO BIOACTIVITY DATA RELEVANT TO THE POTENTIAL
MECHANISMS OF REPRODUCTIVE TOXICITY
HTS screening ToxCast assays profiling in vitro activities for the AR, ER and steroid
hormone biosynthesis were sourced from EPA's CompTox Chemicals Dashboard ((U.S. EPA. 20191.
accessed November 3, 2022) to investigate potential mechanisms of disruption of steroid hormone
receptor activation and steroidogenesis that may be important for the reproductive toxicity of
PFDA.
The suite of ToxCast assays and model predictions for the ER and AR encompass several
endpoints in the signaling pathway of these receptors (e.g., receptor binding, receptor dimerization,
cofactor recruitment, DNA binding, gene expression, and cell proliferation) across multiple in vitro
models. PFDA was active in 2 of 17 AR assays (13%), demonstrating binding to the AR in rat
prostrate tissue and AR-induced cell proliferation in a human prostate carcinoma cell line (22Rvl),
but no activity in assays for cofactor recruitment and AR agonist/antagonist transactivation
conducted primarily in human cell lines (see Table E-3). In ER assays, PFDA was active in 2 of 21
assays (11%), demonstrating activity for the ERa (ESR1) in 1 of 2 assays measuring RNA
transcription in human hepatoma HepG2 cells and in an antagonist transactivation assays
measuring protein expression in human embryonic kidney HEK293T cells (see Table E-3). PFDA
was inactive in receptor binding assays for the ERa in human, bovine, and mouse tissues and in ER
a/p assays for receptor dimerization, transcription factor-DNA binding, agonist transactivation, and
ER-induced cell proliferation in different human cell lines. The AC50 values for the active ER and
AR assays ranged from 8.40 to 62.3 [J.M, which are above the lower bound of the estimated ToxCast
cytotoxicity limit (7.108 [iM) ((U.S. EPA. 2019). accessed November 3, 2022). ToxCast model
predictions incorporating in vitro assay results and nonspecific responses such as cytotoxicity
suggest that PFDA is inactive for both ER/AR agonist and antagonist pathways (AUC = 0) (see
Table E-4).
The ToxCast database also included in vitro assays related to the regulation of
steroidogenesis. PFDA showed a lack of activity in a single assay measuring inhibition of
transcriptional activity for the aromatase gene (CYP19A1) in human breast cancer MCF-7 cells and
several assays measuring biosynthesis of steroid hormones including glucocorticoids, androgens,
estrogens and progestogens in adrenal gland H295R cells (see Table E-5).
This document is a draft for review purposes only and does not constitute Agency policy.
E-16 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table E-3. Bioactivity summary for PFDA from in vitro HTS assays evaluating activities for the AR, ERab
Assay name
Activity call
Scaled
activity
AC50
(HM)
Biological
target
Assay design
Organism
Tissue
Cell line
ACEA_AR_a ntagon ist_80h r
Active
9.34
62.3
AR
growth
reporter
human
prostate
22Rvl
NVS_NR_rAR
Active
2.47
8.40
AR
binding
reporter
rat
prostate
NA
ACEA_AR_agonist_80hr
Inactive
NA
NA
AR
growth
reporter
human
prostate
22Rvl
ATG_AR_TRANS_up
Inactive
NA
NA
AR
inducible
reporter
human
liver
HepG2
OT_AR_ARELUC_AG_1440
Inactive
NA
NA
AR
inducible
reporter
Chinese
hamster
ovary
CHO-K1
OT_AR_ARSRC1_0480
Inactive
NA
NA
AR
binding
reporter
human
kidney
HEK293T
OT_AR_ARSRC1_0960
Inactive
NA
NA
AR
binding
reporter
human
kidney
HEK293T
TOX21_AR_BLA_Agonist_ratio
Inactive
NA
NA
AR
inducible
reporter
human
kidney
HEK293T
TOX21_AR_BLA_Antagonist_ratio
Inactive
NA
NA
AR
inducible
reporter
human
kidney
HEK293T
TOX21_AR_LUC_MDAKB2_Agonist
Inactive
NA
NA
AR
inducible
reporter
human
breast
MDA-kb2
TOX21_AR_LUC_MDAKB2_Agonist_3uM_Nilutamide
Inactive
NA
NA
AR
inducible
reporter
human
breast
MDA-kb2
TOX21_AR_LUC_MDAKB2_Antagonist_0.5nM_R1881
Inactive
NA
NA
AR
inducible
reporter
human
breast
MDA-kb2
TOX21_AR_LUC_MDAKB2_Antagonist_10nM_R1881
Inactive
NA
NA
AR
inducible
reporter
human
breast
MDA-kb2
UPITT_HCI_U20S_AR_TIF2_Nucleoli_Agonist
Inactive
NA
NA
AR
binding
reporter
human
bone
U20S
This document is a draft for review purposes only and does not constitute Agency policy.
E-17 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Assay name
Activity call
Scaled
activity
AC50
(HM)
Biological
target
Assay design
Organism
Tissue
Cell line
UPITT_HCI_U20S_AR_TIF2_Nucleoli_Antagonist
Inactive
NA
NA
AR
binding
reporter
human
bone
U20S
UPITT_HCI_U20S_AR_TIF2_Nucleoli_Cytoplasm_Ratio_Agonist
Inactive
NA
NA
AR
binding
reporter
human
bone
U20S
UPITT_HCI_U20S_AR_TIF2_Nucleoli_Cytoplasm_Ratio_Antagonist
Inactive
NA
NA
AR
binding
reporter
human
bone
U20S
ATG_ERa_TRANS_up
Active
1.50
16.44
ER(ESRl)
inducible
reporter
human
liver
HepG2
TOX21_ERa_BLA_Antagonist_ratio
Active
3.32
22.7
ER(ESRl)
inducible
reporter
human
kidney
HEK293T
ACEA_ER_80hr
Inactive
NA
NA
ER(ESRl)
growth
reporter
human
breast
T47D
ATG_ERE_CIS_up
Inactive
NA
NA
ER(ESRl)
inducible
reporter
human
liver
HepG2
NVS_NR_bER
Inactive
NA
NA
ER(ESRl)
binding
reporter
bovine
uterus
NA
NVS_NR_hER
Inactive
NA
NA
ER(ESRl)
binding
reporter
human
NA
NA
NVS_NR_mERa
Inactive
NA
NA
ER (Esrl)
binding
reporter
mouse
NA
NA
OT_ER_ERaERa_0480
Inactive
NA
NA
ER(ESRl)
binding
reporter
human
kidney
HEK293T
OT_ER_ERaERa_1440
Inactive
NA
NA
ER(ESRl)
binding
reporter
human
kidney
HEK293T
OT_ERa_EREGFP_0120
Inactive
NA
NA
ER(ESRl)
inducible
reporter
human
cervix
HeLa
OT_ERa_EREGFP_0480
Inactive
NA
NA
ER(ESRl)
inducible
reporter
human
cervix
HeLa
TOX21_E Ra_BLA_Ago n ist_ratio
Inactive
NA
NA
ER(ESRl)
inducible
reporter
human
kidney
HEK293T
This document is a draft for review purposes only and does not constitute Agency policy.
E-18 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Assay name
Activity call
Scaled
activity
AC50
(HM)
Biological
target
Assay design
Organism
Tissue
Cell line
TOX21_E Ra_LUC_VM7_Ago n ist
Inactive
NA
NA
ER(ESRl)
inducible
reporter
human
ovary
VM7
TOX21_ERa_LUC_VM7_Antagonist_0.1nM_E2
Inactive
NA
NA
ER(ESRl)
inducible
reporter
human
ovary
VM7
TOX21_ERa_LUC_VM7_Antagonist_0.5nM_E2
Inactive
NA
NA
ER(ESRl)
inducible
reporter
human
ovary
VM7
OT_ER_ERbERb_0480
Inactive
NA
NA
ER (ESR2)
binding
reporter
human
kidney
HEK293T
OT_ER_ERbERb_1440
Inactive
NA
NA
ER (ESR2)
binding
reporter
human
kidney
HEK293T
TOX21_E Rb_B LA_Ago n ist_ratio
Inactive
NA
NA
ER (ESR2)
inducible
reporter
human
kidney
HEK293T
TOX21_E Rb_B LA_An tago n ist_ratio
Inactive
NA
NA
ER (ESR2)
inducible
reporter
human
kidney
HEK293T
OT_ER_ERaERb_0480
Inactive
NA
NA
ER
(ESR11 ESR2)
binding
reporter
human
kidney
HEK293T
OT_ER_ERaERb_1440
Inactive
NA
NA
ER
(ESR11 ESR2)
binding
reporter
human
kidney
HEK293T
aData were sourced from EPA's CompTox Chemicals Dashboard ((U.S. EPA, 2019). accessed November 3, 2022).
bNonspecific responses from inducible reporter gene assays analyzed in the negative fitting direction relative to the control ("_dn") are not presented herein.
NA = not applicable.
This document is a draft for review purposes only and does not constitute Agency policy.
E-19 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table E-4. ToxCast model predictions for the ER and AR pathways for PFDAa
Agonist AUC values (95% CI)
Antagonist AUC values (95% CI)
ER pathway
0 (0-0.0051)
0 (0-0.019)
AR pathway
0 (0-0.063)
0 (0-0.00016)
aData for ER and AR pathways were sourced from Judson et al. (2015) and Kleinstreuer et al. (2017), respectively,
b95% CI for the ER activity model were sourced from a subsequent publication to the Judson et al. (2015) study
(Watt and Judson, 2018).
AUC = area under the curve score ranging from 0 to 1. An AUC value of 0 indicates that the chemical is inactive.
CI = confidence interval.
This document is a draft for review purposes only and does not constitute Agency policy.
E-20 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table E-5. Bioactivity summary for PFDA from in vitro HTS assays related to steroidogenesis3^
Assay name
Activity
call
Scaled
activity
AC50
(HM)
Biological target
Assay design
Organism
Tissue
Cell
line
CEETOX_H295R_llDCORT_noMTC_dn
Inactive
NA
NA
11-Deoxycortisol
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_llDCORT_noMTC_up
Inactive
NA
NA
11-Deoxycortisol
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_ANDR_noMTC_dn
Inactive
NA
NA
Androstenedione
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_ANDR_noMTC_up
Inactive
NA
NA
Androstenedione
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_CORTIC_noMTC_dn
Inactive
NA
NA
Corticosterone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_CORTIC_noMTC_up
Inactive
NA
NA
Corticosterone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_CORTISOL_noMTC_dn
Inactive
NA
NA
Cortisol
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_CORTISOL_noMTC_up
Inactive
NA
NA
Cortisol
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_DOC_noMTC_dn
Inactive
NA
NA
11-Deoxycorticosterone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_DOC_noMTC_up
Inactive
NA
NA
11-Deoxycorticosterone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_ESTRADIOL_noMTC_dn
Inactive
NA
NA
Estradiol
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_ESTRADIOL_noMTC_up
Inactive
NA
NA
Estradiol
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_ESTRONE_noMTC_dn
Inactive
NA
NA
Estrone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_ESTRONE_noMTC_up
Inactive
NA
NA
Estrone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_OHPREG_noMTC_dn
Inactive
NA
NA
17alpha-
hydroxypregnenolone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_OHPREG_noMTC_up
Inactive
NA
NA
17alpha-
hydroxypregnenolone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_OHPROG_noMTC_dn
Inactive
NA
NA
17alpha-
hydroxyprogesterone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_OHPROG_noMTC_up
Inactive
NA
NA
17alpha-
hydroxyprogesterone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_PROG_noMTC_dn
Inactive
NA
NA
Progesterone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_PROG_noMTC_up
Inactive
NA
NA
Progesterone
inducible reporter
human
adrenal gland
H295R
CEETOX_H295R_TESTO_noMTC_dn
Inactive
NA
NA
Testosterone
inducible reporter
human
adrenal gland
H295R
This document is a draft for review purposes only and does not constitute Agency policy.
E-21 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Assay name
Activity
call
Scaled
activity
AC50
(HM)
Biological target
Assay design
Organism
Tissue
Cell
line
CEETOX_H295R_TESTO_noMTC_up
Inactive
NA
NA
Testosterone
inducible reporter
human
adrenal gland
H295R
TOX21_Aromatase_lnhibition
Inactive
NA
NA
CYP19A1
inducible reporter
human
breast
MCF7
aData were sourced from EPA's CompTox Chemicals Dashboard (U.S. EPA, 2019). accessed November 3, 2022).
NA = not applicable.
This document is a draft for review purposes only and does not constitute Agency policy.
E-22 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
APPENDIX F. ADDITIONAL CONFOUNDING
CONSIDERATIONS
F.l. SPECIFIC PFAS CONFOUNDING CONSIDERATIONS FOR FETAL
GROWTH RESTRICTION
As noted, in the PFAS protocol, the potential for bias in effect estimates due to confounding
is a concern in epidemiological studies and was a focus during study evaluation. Hemodynamic
changes occur during pregnancy, such as increased blood plasma volume as a result of decreased
mean arterial pressure, increased cardiac output, and systemic vasodilation (Sagiv etal.. 2018:
Sanghavi and Rutherford. 2014: Chapman et al.. 19981. These changes could lead to lower PFAS
levels in plasma, due to dilution and increased renal filtration. A decrease in PFAS levels has been
noted in serial measurements of some PFAS during pregnancy, namely PFOA, PFOS, and PFNA
(Glynn etal.. 20121. These hemodynamic changes have been proposed as a potential confounder
for associations between PFDA and neonatal and early childhood growth measures. This is
suggested by the association between glomerular filtration rate (GFR), a marker of renal function
and, indirectly, of plasma volume expansion, and fetal growth independent of gestational age and
other maternal covariates fMorken et al.. 2014: Gibson. 19731. Because PFDA concentration in
serum is expected to decrease during pregnancy due to plasma volume expansion, increased renal
excretion, and transplacental transfer, time windows earlier in pregnancy prior to this decrease
may reflect the largest insult to a developing fetus. Potential confounding is one possible
explanation for the effects of pregnancy hemodynamics, but in their meta-analysis of PFOA
Steenland etal. (20181 also proposed that GFR may lead to reverse causality if increased fetal
growth leads to increased maternal blood expansion and glomerular filtration rate. This potential
source of bias related to pregnancy hemodynamics are anticipated to be of greater concern when
maternal serum PFAS samples are collected later in pregnancy. Therefore, as part of the study
quality evaluations, more confidence was placed in studies that adjusted for different pregnancy
hemodynamic markers or if they considered this potential source of confounding by sampling PFAS
levels earlier in pregnancy. As noted in the syntheses, pattern analyses of study results were also
considered according to biomarker sampling timing to determine pregnancy hemodynamics may
be a source of between-study heterogeneity.
Only 1 of the 22 PFDA birth weight-related studies included in the Developmental Effects
section collected and analyzed maternal hemodynamic data such as GFR and/or albumin (i.e., a
marker of plasma volume expansion). Gvllenhammar et al. (20181 did not find any evidence of
confounding following statistical adjustment of different GFR measures for any of the PFAS
examined. Outside of one study that showed some differences in PFOA results following
This document is a draft for review purposes only and does not constitute Agency policy.
F-23 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
adjustment for albumin, the Gvllenhammar et al. (20181 results are consistent with a lack of
confounding demonstrated by either adjustment for albumin (Sagiv etal.. 20181 or different GFR
measures fManzano-Salgado etal.. 2017: Whitworth etal.. 20121 for different PFAS examined in
other studies. Nonetheless, existing meta-analyses for both PFOA fSteenland etal.. 20181 and PFOS
fDzierlenga et al.. 20201 only detected birth weight deficits for later trimester sampling
(e.g., beyond trimester one). One limitation of these meta-analyses is that they did not have the
ability to differentiate late pregnancy from post-partum measures. Only 5 of the 22 PFDA studies of
mean BWT in the overall population examined any first trimester measures, which precluded a
more detailed examination here. Overall, there was limited evidence of any patterns of larger birth
weight associations with sample timing for PFDA. However, the ability to more fully evaluate this
further was limited given the available data as well as disparate exposure measures, distributions,
and contrasts being examined.
F.2. PFAS COEXPOSURE STATISTICAL APPROACHES AND CONFOUNDING
DIRECTIONALITY
In general, an additional source of uncertainty in epidemiological is the potential for
confounding by other PFAS (and other co-occurring contaminants). Although scientific consensus
on how best to address PFAS co-exposures remains elusive, this was considered in the study quality
evaluations and as part of the overall weight of evidence determination. To be a confounder, the co-
occurring PFAS would need to be associated with both the PFAS of interest and the outcome, but
not an intermediate in the causal pathway; such PFAS would be considered positive confounders if
their effect estimate with the endpoint of interest is in the same direction as the primary PFAS of
interest If positive confounders are not accounted for, the anticipation is that any resultant bias
would be away from the null.
Certain statistical approaches can help address the challenges of evaluating the associations
between health endpoints and numerous (often correlated) PFAS that may be present in the
environment. For example, multipollutant models (i.e., those that adjust for at least one co-
occurring exposure) can provide an estimate of the independent association for specific pollutants
with the endpoint of interest. However, these models may not perform well when co-occurring
exposures are highly correlated. Such correlation can lead to collinearity concerns and instability of
modeling results. When exposures are highly correlated and additionally subject to different
potential confounding factors (which may occur, e.g., when PFAS arise from different sources), co-
exposure amplification bias may be a concern (Weisskopf et al.. 20181. Under this scenario,
estimated associations from multi-PFAS adjusted models would be subject to greater bias
compared with results from single-PFAS models. A different approach is to instead 'screen' large
groups of exposures to determine which are associated with the outcome of interest and important
to retain in further analyses. These dimension-reducing statistical approaches (e.g., principal
component analysis, penalized modeling based on elastic net regression, Bayesian kernel machine
This document is a draft for review purposes only and does not constitute Agency policy.
F-24 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
regression, etc.) are increasingly being used for screening large groups of chemical exposures and
help prioritize specific mixtures. However, as noted by Meng etal. (20181. these approaches might
be better suited as "prediction models to screen for a wide range of chemicals from different
sources, and the interpretation of results might become less straightforward due to the necessary
standardization of exposure values." Given these interpretation difficulties and potential for co-
exposure amplification bias, it is not clear which statistical approach best represents independent
effects of specific pollutants within complex PFAS mixtures.
The objective of this part of the appendix is to assess whether there is any direct evidence
for confounding in the studies comparing results from multipollutant (mutually adjusted for other
PFAS) models and results from single pollutant (i.e., PFDA alone with other confounders adjusted
for) models. A second objective is to compare relationships between co-occurring PFAS and
evaluate the extent to which these PFAS may be associated with the primary endpoints of interest
(e.g., birth weight-related measures).
F.3. PFDA AND PFAS COEXPOSURE STUDY RESULTS
In general, the stronger an association between coexposures, and the larger the effect sizes
seen for the coexposure of interest, the more concern there would be for potential confounding.
Table F-l shows correlations between PFAS coexposures and PFDA reported from five studies with
mutually adjusted PFAS data, including four medium confidence fMeng etal.. 2018: Woods etal..
2017: Lenters etal.. 2016: Robledo etal.. 2015) and one high confidence study (Starling etal..
2017). As shown in the PFAS Systematic Review Protocol (see Appendix A) and in Table F-l, PFNA
and PFDA often co-occur (as expected given some similar anticipated sources) across studies with a
consistent correlation of 0.6 or higher. These results also show that other PFAS may not
consistently co-occur with PFDA, as the magnitude of these relationships can vary significantly
across studies.
Table F-l. PFAS correlation coefficients in mutually adjusted studies
Reference
Study Setting
Confidence
Correlations with PFDA
PFOS
PFOA
PFNA
PFHxS
Woods et al.
(2017)
Cincinnati, Ohio,
USA
Medium
0.3
0.1
0.6
0.1
Lenters et al.
(2016)
Greenland;
Kharkiv, Ukraine;
Warsaw, Poland
Medium
0.78
0.50
0.60
0.35
Luoetal. (2021)
Guangzhou, China
High
0.68
0.13
0.85
-0.03
Meng et al. (2018)
Denmark
Medium
0.48
0.28
0.73
0.17
Robledo et al.
(2015)
Michigan and Texas,
USA
Medium
N/A
N/A
N/A
N/A
This document is a draft for review purposes only and does not constitute Agency policy.
F-25 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Starling et al.
Colorado, USA
Low
0.49
0.56
0.65
0.27
(2017)
The results for the six studies based on continuous PFDA data (expressed as change in mean
birth weight per unit change in exposure) are compared and summarized below in Table F-2.
Three of the studies included multiple PFAS as predictors in ordinary least squares regression
models fMeng etal.. 2018: Woods etal.. 2017: Robledo etal.. 20151. Two studies fStarling etal..
2017: Lenters etal.. 20161 examined multiple PFAS using elastic net regression models. Elastic net
regression is a modeling approach to select independent predictors (from an initial group of
potentially correlated predictors) for inclusion in the model using penalized shrinkage methods
(Lenters etal.. 20161. As shown in Table F-2, two of the six studies (Luo etal.. 2021: Lenters etal..
20161 reported nonsignificant birth weight deficits for PFDA from single-pollutant models.
However, PFDA was not associated with birth weight changes in multipollutant models for either
study. For example, Lenters etal. f20161 reported null results for PFDA in both their single-
pollutant model and elastic net regression model, with only PFOA retained in the latter model.
Starling etal. f20171 did not report birth weight deficits associated with PFDA based on either
single-pollutant or multipollutant models nor was PFDA selected for inclusion using elastic net
regression. Mengetal. (20181 reported largely null results for PFDA in single-pollutant models but
detected increases in mean birth weight with adjustment for PFOS, PFOA, PFNA, perfluoroheptane
sulfonic acid (PFHpS), and PFHxS. Luo etal. (20211 reported large birth weight deficits (-97 g; -178,
-16 per each ln-unit PFDA increase) in single-pollutant PFDA model, but results were null in the
multipollutant model. Lastly, Robledo etal. f20151 did not report results from single pollutant
models (or correlations) but did find birth weight deficits associated with PFDA in female neonates
only.
Given the moderate and strong correlations between PDFA and other PFAS, the magnitude
of any associations may exist between these co-occurring PFAS and birth-weight related measures
(and other developmental effects) may inform the potential for confounding ofPFDA associations.
For example, Lenters etal. f 20161 reported birth weight deficits associated with increased levels of
PFNA ((3 =-44.7 g; 95%CI: -92.0, 2.7 per each 2SD ln-unit PFDA increase), PFOS ((3 =-68.8 g;
95%CI: -152.9, 15.2) and PFOA ((3 =-78.5 g; 95%CI: -137.01, -20.0) in single-pollutant models
although only PFOA ((3 =-63.8 g; 95%CI: -122.8, -4.7) was retained in the elastic net regression
model. Although birth weight deficits were not seen for PFDA in any of the regression models used
by Starling etal. (20171. there were large mean birth weight deficits associated with increased
exposure evaluated in single pollutant models for both PFNA ((3 =-58 g; 95%CI: -104, -11 per each
ln-unit PFDA increase) and PFOA ((3 = -51 g; 95%CI: -97, -6). These deficits were larger in
multipollutant models for both PFNA ((3 =-92 g; 95%CI: -167, -18) and PFOA ((3 =-70 g;
95%CI: -148, -9) but were attenuated when included in a penalized elastic net regression model ((3
= -33 g and -14 g, respectively). Mengetal. (20181 reported similar deficits in birth weight
associated with increased exposure to PFNA ((3 =-54.2 g; 95%CI: -105.8, -2.7 per each log2-unit
This document is a draft for review purposes only and does not constitute Agency policy.
F-26 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
PFDA increase) and PFOS ((3 =-55.5 g; 95%CI: -145.6, 34.5) in their model containing mutually
adjusted PFAS; however, effects were seen in the opposite direction (increase in mean birth weight)
for PFDA (p = 48.0 g; 95%CI: -0.6, 96.5) and PFOA (p = 49.5 g; 95%CI: -8.7,107.9) in the same
model. In the Woods etal. f20171 study, none of the five PFAS examined contributed greatly to the
overall changes in mean birth weight when other environmental contaminants were considered in
their elastic net model. Based on their multi-pollutant model, Luo etal. f20211 reported only large
birth weight deficits for PFOA (in excess of-100 g for each PFDA tertile. Finally, Robledo et al.
(2015) reported that only PFOA was associated with large deficits in mean birth weight (P=-61.6 g;
95%CI: -159.2, 35.9 per each SD ln-unitPFDA increase) in girls, while among boys deficits were
only seen for perfluorooctane sulfonamide (PFOSA) ((3 =-104.2 g; 95%CI: -194.2, -14.3) and PFDA
(P =-53.4 g; 95%CI: -161.0, 54.2). In contrast, increased birth weight in boys was reported for
PFNA (p = 62.7 g; 95%CI: -32.1, 157.4) and PFOS (p = 38 g; 95%CI: -73.5, 148.5).
In the six studies using mutually adjusted PFAS approaches to address coexposures, there
was not consistent evidence for birth weight deficits associated with increased exposure to PFDA.
Among the five studies that examined both single and multipollutant models, none of studies that
showed birth weight deficits in single-pollutant models reported greater or more precise
associations following statistical adjustment for other PFAS. Of the three studies showing some
adverse effects fLuo etal.. 2021: Lenters etal.. 2016: Robledo etal.. 20151. only one fRobledo etal..
20151 showed deficits in multipollutant models and this was limited to females only. Among the
three studies that provided correlations among co-occurring PFAS and showed some evidence of
adverse effects for any PFAS, the largest birth weight deficits were seen for PFNA (Meng et al..
2018: Starling etal.. 20171. PFOA (Robledo etal.. 20151. and PFOS (Luo etal.. 20211. The correlation
coefficients for PFDA and these three co-exposures across these studies were all at least 0.50.
As noted in the Developmental Effects section, 11 of 22 studies showed evidence of some
associations with PFDA and mean birth weight in the overall population. Among these 11 studies,
which included the 3 highlighted above fLuo etal.. 2021: Lenters etal.. 2016: Robledo etal.. 20151.
7 showed deficits comparable in magnitude for PFNA and PFDA. Two studies showed larger
deficits for PFDA compared to PFNA, and three studies showed larger deficits for PFNA compared
to PFDA. Given these comparable results seen in most of these studies for both PFNA and PFDA and
the moderately high correlations consistently reported between PFDA and PFNA, there is
considerable uncertainty due to potential confounding by co-occurring PFAS in the existing
literature. It remains unclear, however, if the consistency of birth weight deficits demonstrated
from (categorical and continuous) results in the full set of 22 mean birth weight PFDA studies could
be fully attributed to confounding by PFAS coexposures.
This document is a draft for review purposes only and does not constitute Agency policy.
F-27 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Table F-2. Impact of coexposure adjustment on estimated change in mean
birth weight per unit change (ng/mL) in PFDA levels3
Reference
Study
Confidence
Single
PFAS
Model
Results (in
grams)
with
95%Clsa
Multi-PFAS
Results (in
grams) with
95%Clsa
Elastic Net
Regression
Results
Exposure
Comparison11
Effect of
adjustment on
PFDA birth
weight results
PFAS
adjustments
Starling et
al. (2017)
High
11.5
(-37.3,
60.4)
97.5 (31.5,
163.6)
15.7
In-unit (ng/mL)
increase
Slightly
Strengthened
PFOS, PFOA,
PFNA, PFHxS
Lenters et
al. (2016)
Medium
-43.9
(-104.8,
17.0)
N/A
N/S
2 SD In-unit
(ng/mL)
increase
Attenuated
PFOS, PFOA,
PFNA, PFUnDA,
PFDoDA, PFHxS
Luo et al.
(2021)
High
-96.8
(-178.0,
-15.5)
6.6 (95%CI: -
84.2, 97.3)b
N/A
In-unit ()
increase
Attenuated
PFOA, PFOS,
PFBA, PFBS,
PFHxS, PFNA,
PFUnDA,
PFDoDA, PFTrDA,
6:2 CI-PFESA, 8:2
CI-PFESA
Meng et al.
(2018)
Medium
-9.0
(-43.2,
35.2)
48.0 (-0.6,
96.5)
N/A
log2-unit
(ng/mL)
increase
Changed from
Null to Positive
PFOS, PFOA,
PFNA, PFHxS,
PFHpS
Robledo et
al. (2015)
Medium
N/A
-53.4
(-161.0, 54.2)
Girls-1.8
(-90.6, 87.1)
Boysc
N/A
1 SD In-unit
(ng/mL)
increase
N/A
PFOA, PFOS,
PFNA, PFOSA, Et-
PFOSA-AcOH,
Me-PFOSA-AcOH
Woods et
al. (2017)
Medium
-12.6
(-56.8,
40.4)d
N/A
N/S
logiounit
(ng/mL)
increase
Attenuated
PFOS, PFOA,
PFNA, PFUnDA,
PFDoDA, PFHxS
Abbreviations: N/A: Not available; N/S: PFAS not selected in elastic net regression model.
aModels were based on ordinary least squares regression.
bBeta and 95%Cls estimated from Figure 3 of (Luo et al., 2021).
cThe birth weight results tabulated here are all for the overall population (i.e., male, and female
neonates combined), except for Robledo, which only reported sex-specific findings.
dThe Posterior 95% credible intervals reported for Woods et al. (2017) based on a Bayesian
hierarchical linear model.
This document is a draft for review purposes only and does not constitute Agency policy.
F-28 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
APPENDIX G. DETAILED PHARMACOKINETIC
ANALYSES
This appendix provides two detailed pharmacokinetic analyses. The first is a Bayesian
analysis ofPFDA pharmacokinetics in laboratory animals to estimate key pharmacokinetic
parameters. The second is the description and evaluation of a one-compartment PK modeling
approach for estimating internal doses, evaluated against rat PFDA PK data using the mean
parameter values estimated for male rats in the Bayesian estimation.
G.l. PARTIAL POOLING OF PFDA PHARMACOKINETIC DATA FOR
HIERARCHICAL BAYESIAN ANALYSIS
We estimated the sex-specific pharmacokinetic parameters (half-life, volume of
distribution, and clearance) of PFDA in rats by fitting one- and two-compartment models to the
available concentration vs. time data. A Bayesian hierarchical methodology was developed to fit
these models because of the need to pool time-course concentration data across numerous studies
with varying exposure scenarios within each study. This allowed for each concentration vs. time
dataset to be fit to each pharmacokinetic model where fitted parameters for each dataset are
sampled from a population-level distribution which models the similarities between each dataset.
In addition, the Bayesian analysis allowed for the generation of central estimates and credible
intervals for the pharmacokinetic parameter of interest e.g., half-life, volume of distribution and
clearance, using posterior distributions from the estimated variables. Finally, the Bayesian
methodology allowed for hypothesis testing of the 1- and 2-compartment formulations to decide
which model more appropriately fit the data.
G.l.l. Pharmacokinetic model
To determine pharmacokinetic parameters for PFDA, we estimated constants for both one-
and two-compartment model assumptions. For a one-compartment model assumption, the
following exponential decay functions were fit to the available data
rIV
°1 -crrvpt\L/J y c
cr Snp,m = (e"M -e_M)
This document is a draft for review purposes only and does not constitute Agency policy.
G-l DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
where D represents the administered dose and V, ke, and ka represent the central compartment
volume, elimination constant, and absorption constant (for oral only) to be fit From these fitted
constants, pharmacokinetic parameters are derived:
V
Vd =
BW
In 2
ti =-j—
2 ke
CLC = Vd*ke
where Vd, ti/2, and CLC represent the volume of distribution, terminal half-life, and clearance
respectively and BW represents the animal body weight.
For the two-compartment model assumption, the following exponential decay functions
were fit to available data
Ct — krlr , I krlr — Ct
AIV _ a kdc . A oral _ I. ( kdc a \
a-fi ' a\(ka-a)(p -a))
dIV _ @ kdc _ noral _ l, ( ^dc P \
B — a' a\(kn — B)(a — B)/
Cllcmpt(.Q=f(A're-'a + B're-/t)
Ci-clnpttt) =^(Aorale~at + Borale~^ - (Aoral + Boral)e~k«c)
where D represents the administered dose and V, a, (3, kdC, and ka represent central compartment
volume, alpha-phase elimination constant, beta-phase elimination constant, deep-to-central
compartment rate constant, and absorption constant (for oral only) to be fit From these fitted
constants, the remaining two-compartment constants (kCd: central-to-deep compartment rate
constant and ke: elimination constant) and the deep compartment volume (Vdeep) are derived by
solving:
a + p = kcd + kdc + ke
a * jl = kdc * ke
kcd
kdc
Vd=V
which allows for the desired pharmacokinetic parameters to be derived using the following
equations:
„ _ ^ Vdeep _ V fkcd + kdc\
d~ss ~ BW ~BW\ k^c /
In 2
% = ~
This document is a draft for review purposes only and does not constitute Agency policy.
G-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
V
where Vd-ss, ti/2, and CLC represent the steady-state volume of distribution, terminal half-life, and
clearance respectively and BW represents the animal body weight.
G.1.2. Bayesian inference
The fitted constants for each model structure (described above) were estimated using
available time-course concentration data reported in rats with parameters for each model
estimated using a hierarchical Bayesian calibration approach. This hierarchical Bayesian approach
pooled the time-course concentration data for male and female rats from multiple studies Ohmori
etal. (20031. Kim etal. (20191. Dzierlenga etal. (20191. For the two-compartment model, to ensure
parameter identifiability, a and /? were constrained to be ordered such that a > /?. This constraint
ensures the exponential terms are identifiable and don't "flip" while exploring the parameter space
during Markov-chain Monte-Carlo (MCMC) sampling. Finally, priors for each pharmacokinetic
parameter were chosen to be "weakly informative" based on prior knowledge of PFAS
pharmacokinetics fATSDR. 20211 with 95% equal-tailed intervals spanning multiple order of
magnitude.
Priors for pharmacokinetic parameters are presented in Table G-l with corresponding
model-specific parameter prior distributions presented below. Finally, a sensitivity analysis on the
model priors is shown in the Prior sensitivity analysis section.
Table G-l. Weakly informed prior distributions for pharmacokinetic
parameters used in the Bayesian analysis
median
mad
eti 3%
eti 97%
Half-life (d)
15
12
0.88
250
Clearance
(mL/kg-d)
50
49
0.32
6,000
Vd-ss (ml/kg)
900
811
9.3
32,822
For the hierarchical approach, the concentration vs. time data comprised a population- and
dataset-level for which model parameters were estimated. Here, each dataset represented each
study/sex/dose concentration vs. time dataset extracted from the literature and were fit using the
model
(C\°%eipt for 1-compartment model,
lJ [exempt for 2-compartment model
Cik ~LN (X;j, (T/j)
This document is a draft for review purposes only and does not constitute Agency policy.
G-3 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
where X;j is the sample mean of the observed concentrations at time ty for dataset j and ak is
study-level log-transformed standard deviation for the relative errors based on study k. Study-level
priors for ^e> a> P> k-dc)j
This document is a draft for review purposes only and does not constitute Agency policy.
G-4 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
per chain. Posterior parameter distributions were determined using the final 5,000 iterations of
each chain ensuring an effective sample size (ESS) greater than 10,000 fKruschke, 20211.
Convergence was assessed using a potential scale reduction factor with a maximum threshold of
R = 1.05 fKruschke. 20211.
G.1.3. Prior sensitivity analysis
To investigate the impact of prior selection on posterior pharmacokinetic parameter
estimation, we conducted a sensitivity analysis on the priors used in the Bayesian analysis. Priors
were classified into three categories: weakly informed, broad, and uninformed. Weakly informed
priors are defined using the half-life, clearance, and volume of distribution described above based
on reported ranges ofPFDA pharmacokinetics with a prior predictive check demonstrating
available data for fitting fall within the prior 90% credible interval.
Hero; 59X6078, PFDA gavage 20 00 mg/kg Hero: 5916078, PFDA gavage 10 00 mg/kg
time [days] time [days]
Hero: 5916078, PFDA gavage 2.00 mg/kg Hero: 5916078, PFDA iv 2.00 mg/kg
time [days] time [days]
Figure G-l. Prior predictive check to ensure equal-tailed interval from prior
distributions encompass the available time-course concentration data for
fitting.
In addition to these weakly informed priors, we also characterized a set of broad priors,
defined as uniform distributions spanning the 3% and 97% ETI from the weakly informed priors,
This document is a draft for review purposes only and does not constitute Agency policy.
G-5 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
and completely uninformed priors, representing uniform priors spanning multiple orders of
magnitude i.e., flat priors. Figure G-2 (prior sensitivity) compares these three classes of priors and
their impact on the posterior pharmacokinetic parameter distributions,
94.0% HDI
broad
Vdss (ml/kg)
Clearance (ml/kg/d)
Half-life (days)
10° 101 102 103
Figure G-2. Prior sensitivity on half-life, steady-state volume of distribution,
and clearance to ensure weakly informed priors do not bias posterior
distributions of the pharmacokinetic parameters.
Based on these findings, we used the weakly informed pharmacokinetic priors for fitting
available time-course concentration data.
G.1.4. Study-specific Clearance Values and Model Fits
Three data sets were used for the sex-specific parameter estimation, which had a mixture of
gavage and iv exposure routes and follow-up times extending up to 150 days (Dzierlenga etal..
2019: Kim etal.. 2019: Ohmori et al.. 20031. The sex-specific clearance value distribution obtained
from fitting the three data sets together had a mean and 90% credible interval of 4.06 (2.05-6.05)
mL/kg-day in female rats and 4.14 (0.68-7.02) mL/kg-day in male rats. For these data, a 2-
compartment PK model was deemed superior. Visual inspection shows some of the data have a
distinguishable distribution and excretion phase, which is appropriate for a 2-compartment model
(see Figure G-3). A 2-compartment model is also able to fit data that appear linear as is evidenced in
This document is a draft for review purposes only and does not constitute Agency policy.
G-6 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
fits to other data sets (see Figure G-4). Credible intervals for the fits to individual data sets are
qualitatively small showing good model fits to the data from individual studies. The relatively large
credible interval for the pooled data is due to the large variation between studies. For example, in
male rats the mean clearance values for individual studies ranged from 1.51 to 7.45 mL/kg-day, and
a similar range was seen in female rats.
Trends comparing the terminal clearance following IV and gavages doses appeared within
studies but did not hold for the whole data set. For example, in Kim etal. (20191IV doses resulted
in smaller, but similar clearance to gavage doses (see Figure G-4). However, these clearance values
were consistently smaller than clearance values calculated from the two other data sets. In the
analysis of the Dzierlenga etal. f20191 dataset, IV doses resulted in clearly greater clearance than
the three dose levels administered by gavage, which all had similar clearance within each sex (see
Figure G-5,6). There was a difference in clearance between sexes in this study, but only for gavage
doses. In this study, the gavage doses resulted in mean clearance values between 3.57 and 3.77
mL/kg-day in female rats and 5.12 and 5.74 mL/kg-day in male rats. However, the clearance
calculated from the single IV dose was similar between female and male rats. Likewise, the two
other studies showed similar mean clearance values for male and female rats (see Figure G-3 and
Figure G-4). It is possible that most of the difference in PFDA PK between male and female rats is
related to a difference in absorption, which can be moderated by active transport. Additional
experiments designed to carefully evaluate these factors would be needed to resolve this question.
Hero: 3858670 (0), PFDA iv 25.00 mg/kg
CLC (ml/kg/day): 5.31 (4.44 - 6.30)
Hero: 3858670 (0), PFDA iv 25.00 mg/kg
CLC (ml/kg/day): 5.32 (3.66 - 6.77)
Ol
E
U 6 x 101
c
« 4 x 101
2 x 101
0
20 40
time [days]
60
0
20
40
60
time [days]
Figure G-3. Predicted (black line with blue 90% credible interval) and
observed (black circles) serum time-courses for female (left) and male (right)
rats after a 25 mg/kg IV bolus of PFDA. Observed data from fOhmori et al..
20031.
This document is a draft for review purposes only and does not constitute Agency policy.
G-7 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Hero: 5063958 (1), PFDA gavage 1.00 img/kg
CLC (ml/kg/day): 2.61 (2.13 - 3.09)
50 100
time [days]
Hero: 5063958 (2), PFDA iv 1.00 img/kg
CLC (ml/kg/day): 2.07 (1.83 - 2.31)
o
u
10° -
50 100
time [days]
Hero: 5063958 (1), PFDA gavage 1.00 mg/kg Hero: 5063958 (2), PFDA iv 1.00 mg/kg
CLC (ml/kg/day): 2.94 (2.34 - 3.52) CLC (ml/kg/day): 1.61 (1.08 - 2.10)
Ol
Ł
^ ^ C~
2 x 10'
6 x 10
time [days]
time [days]
Figure G-4. Predicted (black line with blue 90% credible interval) and
observed (black circles) serum time-courses for female (top 2 panels) and
male (bottom 2 panels) rats after a 1 mg/kg gavage or IV bolus of PFDA.
Gavage exposures are on the left, while IV exposures are on the left, while IV
exposures are on the right. Observed data from (Kim etal.. 2019).
This document is a draft for review purposes only and does not constitute Agency policy.
G-8 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Hero: 5916078 (6), PFDA gavage 20.00 mg/kg Hero: 5916078 (5), PFDAgavage 10.00 mg/kg
CLC (ml/kg/day): 3.77 (3.40 - 4.14) CLC (ml/kg/day): 3.57 (3.22 - 3.93)
time [days] time [days]
CJ|
E
Hero: 5916078 (4), PFDA gavage 2.00 mg/kg
CLC (ml/kg/day): 3.78 (3.41 - 4.16)
Hero: 5916078 (3), PFDA iv 2.00 mg/kg
CLC (ml/kg/day): 7.40 (6.67 - 8.19)
10'
25 50 75 100 125 150
time [days]
25 50 75 100 125 150
time [days]
Figure G-5. Predicted (black line with blue 90% credible interval) and
observed (black circles) serum time-courses for female rats after a 2 mg/kg IV
or 2,10, or 20 mg/kg gavage bolus of PFDA. Observed data from (Dzierlenga et
al.. 201,91.
This document is a draft for review purposes only and does not constitute Agency policy.
G-9 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
time [days] time [days]
Figure G-6. Predicted (black line with blue 90% credible interval) and
observed (black circles) serum time-courses for male rats after a 2 mg/kg IV
or 2,10, or 20 mg/kg gavage bolus of PFDA. Observed data from (Dzierlenga et
al.. 20191.
G.2. DESCRIPTION AND EVALUATION OF A SINGLE-COMPARTMENT PK
APPROACH
1 For PFDA, the clearance values obtained in the preceding Bayesian analysis are low enough
2 that internal doses will not reach steady-state for shorter-term studies, in particular for
3 developmental studies where dosing may only be for a few weeks. In this case a PK model can
4 potentially be used to account for the growth of the animal, the intrinsic elimination, and the
5 accumulation of PFDA over the period of dosing. The single-compartment PK model is given by:
6 dA/dt = Fabs x dose x BW - CUt x A / Vd, (G-l)
7 where A is the total amount of PFDA in the animal (mg), Fabs is the fraction absorbed for an oral
8 dose (bioavailability), BW is the body-weight (kg), and CLtot is the total clearance, and Vd is the
9 volume of distribution. Implicit in this model is an assumption of rapid distribution of PFDA in the
10 body (relative to the clearance), in which case the concentration in plasma is:
11 Cplasma — A/ (Vd x BW). (G-2)
This document is a draft for review purposes only and does not constitute Agency policy.
G-10 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
The differential equation for the amount of chemical in the body can then be re-written:
dA/dt = Fabs x dose x BW - CLtot x BW x CpiaSma, (G-3)
which leads to the interpretation that the clearance or volume of blood cleared of the chemical per
unit time per kg BW is CLtot-
While Fabs is shown in equations (G-l) and (G-3) for completeness, the available data could
not be used to identify a value for Fabs independent of other parameters in the Bayesian PK analysis
and given the observations of generally high uptake (see the section on Absorption in the
Toxicological Review) it was set to a value of 1 (i.e., 100%) for this analysis, and hence is not
included in the subsequent description.
PK parameters for rats (CLtot, and Vd) are taken from the preceding Bayesian analysis
(values listed in Table 3-3). Given the slow clearance ofPFDA, the growth of rats during toxicity
studies lasting multiple weeks can be a significant factor as increases in BW dilute the body burden
from earlier exposures. The highest doses tested in the NTP bioassay significantly reduced animal
BW, which compounds this effect. Therefore, time-dependence in BW based on the empirical data
for BW at the doses evaluated was incorporated into the model evaluation, to account for this time-
and dose-dependence. For illustration, the change in male rat BW observed in the NTP bioassay
(28-day exposure fNTP. 201811 is shown in Figure G-7. Doses of 0.625 mg/kg-day and below did
not significantly affect BW gain during the bioassay, but higher dose levels caused a significant
decline after 7 days of exposure.
The internal dose of PFDA predicted by the PK model as a function of exposure day,
normalized to the dose for comparison, is shown in Figure G-8. For example, the model simulated
concentrations obtained using a dose of 0.625 mg/kg-day were divided by 0.625 before plotting. If
the BW curve was the same for all doses, all the resulting normalized curves would lie on top of
each other. The predicted concentration increases steadily throughout the study for all dose levels,
showing no sign of saturation. However, the increase in animals receiving the highest doses
becomes relatively faster after day 7, deflecting above the lower-dose curves. This occurs because
the decreasing BW at these doses concentrates the PFDA already administered into a smaller total
animal mass. For model simulations the dose is assumed to be adjusted continuously based on the
interpolated weights as shown in Figure 3-3. (The study report states that animals were weighed
daily, but only weekly values are provided there.) For example, if an animal loses weight between
day 7 and 21, the daily dose is assumed to be adjusted accordingly. Since the animals were
necropsied on day 29,1 day after the final dose, the model simulations include a final day with zero
exposure. Mean serum PFDA concentrations from the NTP study, collected at time of necropsy, are
shown for comparison.
This document is a draft for review purposes only and does not constitute Agency policy.
G-ll DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
360
340
320
3 300
_C
ttf
280
>
"O
O 260
CQ
240
220
200
Male Rat Body Weight
~
~ ^
— 1 **
- o
i-''
A -
—•— Control
' —"A
~ 0.156 mg/kg/d
- ¦- -0.312 mg/kg/d
o 0.625 mg/kg/d
—* ¦ 1.25 mg/kg/d
•-G-
¦ 2.5 mg/kg/d
10
15
Study day
20
25
30
Figure G-7. Male rat body weight changes during 28-day PFDA bioassay (NTP.
20181. Data sets are identified by the dose (mg/kg-d).
O
CJ
CM
03 'l_
LO
O CNI
C *-
O
O
"O o
o
o m
~
Simulations
- 0 156 mg/kg/d
0.312 mg/kg/d
- 0.625 mg/kg/d
* 1 25 mg/kg/d
- 2 5 mg/kg/d
Male rats
Data
0.156 mg/kg/d
0 312 mg/kg/d
0 625 mg/kg/d
1,25 mg/kg/d
2.5 mg/kg/d
10
—r~
15
Days
l
20
[
25
30
Figure G-S. Predicted accumulation and observed end-of-study of PFDA in
male rats in the NTP bioassay (NTP. 2018) as a function of dose. Predicted and
measured concentrations (mg/L) were normalized to respective doses (mg/kg-d).
1 In Figure G-8 the model consistently over-predicts the data by a factor of about 1.5. While
2 the EPA general considers this much discrepancy acceptable for a comparison of PK model
3 predictions to data, the fact that there is systematic bias, rather than some predictions being above
4 and some below the data raises concern. The direction of the error indicates that the model will
This document is a draft for review purposes only and does not constitute Agency policy.
G-12 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
over-predict internal doses in rats, and hence the corresponding HEDs. One might also note that the
data point for 0.625 mg/kg-day is less than that for 0.312 mg/kg-day, whereas the model
simulations show only increasing normalized concentration with dose. The pattern in the data
(which points are more closely clustered vs. farther apart) is a bit different from that predicted by
the model. To further evaluate the extent of nonlinearity, the end-of-study plasma concentrations
from NTP f20181 are plotted against the dose in Figure G-9. The exposure-dose relationship is seen
to be essentially linear for the three lowest doses (to 0.625 mg/kg-day), with some variation, and
then to increase a bit faster than linear with dose above that As indicated by the BW data in Figure
G-7 and resulting simulations in Figure G-8, this upward inflection could be due to dose-related BW
loses, which are predicted to concentrate the previously administered PFDA into a smaller total
volume. However, there is no evidence of saturation of renal resorption, which would result in
downward curvature in the exposure-dose relationship. Instead, the discrepancy between the NTP
data and the model simulations can be mostly explained if rat clearance is about three times higher
than estimated from the PK studies.
in
CM
—I
a;-
E
O
c
o
a
CM
-1—'
ro
±3
c=
O
o
LO
o
t—
c=
o
o
TO
o
E
o
V)
¦*—
TO
a
CD
CM
o
ITS
>.
TO
0
0 0 0 5 1.0 1.5 20 2.5
Dose (mg/kg/d)
Figure G-9. Measured end-of-study of PFDA in male rats in the NTP bioassay
fNTP. 20181 as a function of dose.
This document is a draft for review purposes only and does not constitute Agency policy.
G-13 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
APPENDIX H. SUMMARY OF PUBLIC AND
EXTERNAL PEER REVIEW COMMENTS AND EPA'S
DISPOSITION
1
This document is a draft for review purposes only and does not constitute Agency policy.
H-l DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
APPENDIX I. QUALITY ASSURANCE FOR THE IRIS
TOXICOLOGICAL REVIEW OF
PERFLUORODECANOIC ACID AND RELATED SALTS
This assessment is prepared under the auspices of the U.S. Environmental Protection
Agency's (EPA's) Integrated Risk Information System (IRIS) Program. The IRIS Program is housed
within the Office of Research and Development (ORD) in the Center for Public Health and
Environmental Assessment (CPHEA). EPA has an agency-wide quality assurance (QA) policy that is
outlined in the EPA Quality Manual for Environmental Programs (see CIO 2105-P-01.11 and follows
the specifications outlined in EPA Order CIO 2105.1.
As required by CIO 2105.1, ORD maintains a Quality Management Program, which is
documented in an internal Quality Management Plan (QMP). The latest version was developed in
2013 using Guidance for Developing Quality Systems for Environmental Programs fOA/G-11. An
NCEA/CPHEA-specific QMP was also developed in 2013 as an appendix to the ORD QMP. Quality
assurance for products developed within CPHEA is managed under the ORD QMP and applicable
appendices.
The IRIS Toxicological Review of Perfluorodecanoic acid (PFDA) is designated as Influential
Scientific Information (ISI) and is classified as QA Category A. Category A designations require
reporting of all critical QA activities, including audits. The development of IRIS assessments is done
through a seven-step process. Documentation of this process is available on the IRIS website:
https://www.epa.gOv/iris/basic-information-about-integrated-risk-information-system#process.
Specific management of quality assurance within the IRIS Program is documented in a
Programmatic Quality Assurance Project Plan (PQAPP). A PQAPP is developed using the EPA
Guidance for Quality Assurance Project Plans fOA/G-51 All IRIS assessments follow the IRIS
PQAPP, and all assessment leads and team members are required to receive QA training on the IRIS
PQAPP. During assessment development, additional QAPPs may be applied for quality assurance
management. They include:
Title
Document number
Date
Program Quality Assurance Project Plan
(PQAPP) for the Integrated Risk
Information System (IRIS) Program
L-CPAD-0030729-QP-1-5
June 2022
An Umbrella Quality Assurance Project
Plan (QAPP) for Dosimetry and
Mechanism-Based Models (PBPK)
L-CPAD-0032188-QP-1-2
December 2020
This document is a draft for review purposes only and does not constitute Agency policy.
1-1 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Quality Assurance Project Plan (QAPP)
for Enhancements to Benchmark Dose
Software (BMDS)
L-H EEAD-0032189-QP-1-2
October 2020
Umbrella Quality Assurance Project Plan
for CPHEA PFAS Toxicity Assessments
L-CPAD-0031652-QP-1-5
February 2023
1 During assessment development, this project undergoes four quality audits during
2 assessment development including:
Date
Type of audit
Major findings
Actions taken
August 2019
Technical system audit
None
None
August 2020
Technical system audit
None
None
July 2021
Technical system audit
None
None
August 2022
Technical system audit
None
None
3 During Step 3 and Step 6 of the IRIS process, the IRIS toxicological review is subjected to
4 external reviews by other federal agency partners, including the Executive Offices of the White
5 House. Comments during these IRIS process steps are available in the Docket EPA-HQ-ORD-2019-
6 0287 on http://www.regulations.gov.
This document is a draft for review purposes only and does not constitute Agency policy.
1-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
REFERENCES
Abe. T; Takahashi. M; Kano. M; Amaike. Y; Ishii. C; Maeda. K; Kudoh. Y; Morishita. T; Hosaka. T;
Sasaki. T; Kodama. S; Matsuzawa. A; Kojima. H; Yoshinari. K. (2017). Activation of
nuclear receptor CAR by an environmental pollutant perfluorooctanoic acid. Arch
Toxicol 91: 2365-2374. http://dx.doi.org/10.1007/sQ0204-016-1888-3.
Adinehzadeh. M; Reo. NV. (1998). Effects of peroxisome proliferators on rat liver phospholipids:
sphingomyelin degradation may be involved in hepatotoxic mechanism of
perfluorodecanoic acid. Chem Res Toxicol 11: 428-440.
http://dx.doi.org/10.1021/tx97Q155t.
Adinehzadeh. M; Reo. NV: Jarnot. BM; Taylor. CA; Mattie. DR. (1999). Dose-response
hapatotoxicity of the peroxisome proliferator, perfluorodecanoic acid and the
relationship to phospholipid metabolism in rats. Toxicology 134: 179-195.
http://dx.doi.org/10.1016/S0300-483X(99)00038-4.
Al Sharif. M; Alov. P; Vitcheva. V; Pajeva. I: Tsakovska. I. (2014). Modes-of-action related to
repeated dose toxicity: tissue-specific biological roles of PPAR y ligand-dependent
dysregulation in nonalcoholic fatty liver disease [Review]. PPAR Research 2014: 432647.
http://dx.doi.org/10.1155/2014/432647.
Angrish. MM: Kaiser. JP; Mcqueen. CA: Chorley. BIN. (2016). Tipping the balance: Hepatotoxicity
and the 4 apical key events of hepatic steatosis [Review]. Toxicol Sci 150: 261-268.
http://dx.doi.org/10.1093/toxsci/kfw018.
Arand. M; Coughtrie. MW; Burchell. B; Oesch. F; Robertson. LW. (1991). Selective induction of
bilirubin UDP-glucuronosyl-transferase by perfluorodecanoic acid. Chem Biol Interact
77: 97-105. http://dx.doi.org/10.1016/0009-2797(91)90008-U.
ATSDR (Agency for Toxic Substances and Disease Registry). (2018). Toxicological profile for
perfluoroalkyls. Draft for public comment [ATSDR Tox Profile]. Atlanta, GA: U.S.
Department of Health and Human Services, Centers for Disease Control and Prevention.
https://www.atsdr.cdc.gov/toxprofiles/tp200.pdf.
ATSDR (Agency for Toxic Substances and Disease Registry). (2021). Toxicological profile for
perfluoroalkyls [ATSDR Tox Profile]. Atlanta, GA: U.S. Department of Health and Human
Services, Public Health Service. http://dx.doi.Org/10.15620/cdc:59198.
Barish. GD; Narkar. VA; Evans. RM. (2006). PPAR delta: a dagger in the heart of the metabolic
syndrome. J Clin Invest 116: 590-597. http://dx.doi.org/10.1172/JCI27955.
Betancourt. MJ; Girolami. M. (2013). Hamiltonian monte carlo for hierarchical models.
Betancourt, MJ; Girolami, M. http://dx.doi.org/10.48550/arXiv.1312.0906.
Boobis. AR; Doe. JE; Heinrich-Hirsch. B; Meek. ME: Munn. S; Ruchirawat. M; Schlatter. J: Seed. J:
Vickers. C. (2008). IPCS framework for analyzing the relevance of a noncancer mode of
action for humans [Review]. Crit Rev Toxicol 38: 87-96.
http://dx.doi.org/10.1080/104084407Q1749421.
This document is a draft for review purposes only and does not constitute Agency policy.
R-l DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Borges, T; Glauert, HP; Chen, LC; Chow, CK; Robertson, LW. (1990). Effect of the peroxisome
proliferator perfluorodecanoic acid on growth and lipid metabolism in Sprague Dawley
rats fed three dietary levels of selenium. Arch Toxicol 64: 26-30.
http://dx.doi.org/10.1007/BF01973372.
Borges, T; Glauert, HP; Robertson, LW. (1993). Perfluorodecanoic acid noncompetitively inhibits
the peroxisomal enzymes enoyl-CoA hydratase and 3-hydroxyacyl-CoA dehydrogenase.
Toxicol Appl Pharmacol 118: 8-15. http://dx.doi.org/10.1006/taap.1993.10Q3.
Borges, T; Robertson, LW; Peterson, RE; Glauert, HP. (1992). Dose-related effects of
perfluorodecanoic acid on growth, feed intake and hepatic peroxisomal beta-oxidation.
Arch Toxicol 66: 18-22. http://dx.doi.org/10.1007/BF023Q7265.
Brewster. DW; Birnbaum. LS. (1989). The biochemical toxicity of perfluorodecanoic acid in the
mouse is different from that of 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicol Appl
Pharmacol 99: 544-554. http://dx.doi.org/10.1016/0041-008X(89)90161-0.
Budtz-J0rgensen, E; Grandjean. P. (2018a). Application of benchmark analysis for mixed
contaminant exposures: Mutual adjustment of perfluoroalkylate substances associated
with immunotoxicity. PLoS ONE 13: e0205388.
http://dx.doi.org/10.1371/iournal.pone.0205388.
Budtz-J0rgensen, E; Grandjean. P. (2018b). Computational details for the paper "Application of
benchmark analysis for mixed contaminant exposures: Mutual adjustment of
perfluoroalkylate substances associated with immunotoxicity".
Buhrke. T; Kibellus. A: Lampen. A. (2013). In vitro toxicological characterization of
perfluorinated carboxylic acids with different carbon chain lengths. Toxicol Lett 218: 97-
104. http://dx.doi.Org/10.1016/i.toxlet.2013.01.025.
Butenhoff. JL; Bjork. JA; Chang. SC; Ehresman. DJ; Parker. GA; Das. K; Lau. C; Lieder. PH; van
Otterdijk. FM; Wallace. KB. (2012). Toxicological evaluation of ammonium
perfluorobutyrate in rats: twenty-eight-day and ninety-day oral gavage studies. Reprod
Toxicol 33: 513-530. http://dx.doi.Org/10.1016/i.reprotox.2011.08.004.
Cai. Y; Appelkvist. EL: Depierre. JW. (1995). Hepatic oxidative stress and related defenses during
treatment of mice with acetylsaIicylic acid and other peroxisome proliferators. J
Biochem Toxicol 10: 87-94. http://dx.doi.org/10.1002/ibt.25701002Q5.
Cao. L; Quan. XB; Zeng. WJ; Yang. XO; Wang. MJ. (2016). Mechanism of hepatocyte apoptosis
[Review]. Journal of Cell Death 9: 19-29. http://dx.doi.org/10.4137/JCD.S39824.
Cattley. RC; Cullen. JM. (2018). Chapter 8. Liver and gall bladder. In MA Wallig; WM Haschek;
CG Rousseaux; B Bolon (Eds.), Fundamentals of toxicologic pathology (3rd ed., pp. 125-
151). Cambridge, MA: Academic Press. http://dx.doi.org/10.1016/B978-0-12-8Q9841-
7.00008-3.
Cellesi. C; Michelangeli. C; Rossolini. GM; Giovannoni. F; Rossolini. A. (1989). Immunity to
diphtheria, six to 15 years after a basic three-dose immunization schedule. Journal of
Biological Standardization 17: 29-34. http://dx.doi.org/10.1016/0092-1157(89)90025-5.
Chapman. AB; Abraham. WT; Zamudio. S; Coffin. C; Merouani. A: Young. D; Johnson. A: Osorio.
F; Goldberg. C; Moore. LG; Dahms. T; Schrier. RW. (1998). Temporal relationships
between hormonal and hemodynamic changes in early human pregnancy. Kidney Int 54:
2056-2063. http://dx.doi.Org/10.1046/i.1523-1755.1998.00217.x.
This document is a draft for review purposes only and does not constitute Agency policy.
R-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Chen, H; Huang, CY; Wilson, MW; Lay, LT; Robertson, LW; Chow, CK; Glauert, HP. (1994). Effect
of the peroxisome proliferators ciprofibrate and perfluorodecanoic acid on hepatic cell
proliferation and toxicity in Sprague-Dawley rats. Carcinogenesis 15: 2847-2850.
http://dx.doi.org/10.1093/carcin/15.12.2847.
Chen. LC; Tatum. V; Glauert. HP: Chow. CK. (2001). Peroxisome proliferator perfluorodecanoic
acid alters glutathione and related enzymes. J Biochem Mol Toxicol 15: 107-113.
http://dx.doi.Org/10.1002/ibt.6.
Cheng. X: Klaassen. CD. (2008a). Critical role of PPAR-alpha in perfluorooctanoic acid- and
perfluorodecanoic acid-induced downregulation of Oatp uptake transporters in mouse
livers. Toxicol Sci 106: 37-45. http://dx.doi.org/10.1093/toxsci/kfnl61.
Cheng. X: Klaassen. CD. (2008b). Perfluorocarboxylic acids induce cytochrome P450 enzymes in
mouse liver through activation of PPAR-alpha and CAR transcription factors. Toxicol Sci
106: 29-36. http://dx.doi.org/10.1093/toxsci/kfnl47.
Chinje. E; Kentish. P; Jarnot. B; George. M; Gibson. G. (1994). Induction of the CYP4A subfamily
by perfluorodecanoic acid: The rat and the guinea pig as susceptible and non-
susceptible species. Toxicol Lett 71: 69-75. http://dx.doi.org/10.1016/Q378-
4274(94)90200-3.
Christenson. B; Bottiger. M. (1986). Serological immunity to diphtheria in Sweden in 1978 and
1984. Scand J Infect Dis 18: 227-233. http://dx.doi.org/10.3109/00365548609Q32331.
Collier. RJ. (1975). Diphtheria toxin: Mode of action and structure [Review]. Bacteriol Rev 39:
54-85. http://dx.doi.Org/10.1128/br.39.l.54-85.1975.
Corton. JC; Peters. JM; Klaunig. JE. (2018). The PPARa-dependent rodent liver tumor response is
not relevant to humans: addressing misconceptions [Review]. Arch Toxicol 92: 83-119.
http://dx.doi.org/10.1007/sQ0204-017-2094-7.
Crump. KS. (1995). Calculation of benchmark doses from continuous data. Risk Anal 15: 79-89.
http://dx.doi.Org/10.llll/i.1539-6924.1995.tb00095.x.
Davis. JW; Vanden Heuvel. JP; Peterson. RE. (1991). Effects of perfluorodecanoic acid on de
novo fatty acid and cholesterol synthesis in the rat. Lipids 26: 857-859.
http://dx.doi.org/10.1007/BF0253617Q.
Derbel. M; Hosokawa. M; Satoh. T. (1996). Differences in the induction of carboxylesterase RL4
in rat liver microsomes by various perfluorinated fatty acids, metabolically inert
derivatives of fatty acids. Biol Pharm Bull 19: 765-767.
http://dx.doi.org/10.1248/bpb.19.765.
Division of Environmental Epidemiology. I. forRAS; Portengen. L; Rignell-Hydbom. A: Jonsson. B.
oAG; Lindh. CH; Piersma. AH: Toft. G; Bonde. JP: Heederik. D; Rylander. L; Vermeulen. R.
(2016). Prenatal Phthalate, Perfluoroalkyl Acid, and Organochlorine Exposures and Term
Birth Weight in Three Birth Cohorts: Multi-Pollutant Models Based on Elastic Net
Regression. Environ Health Perspect 124: 365-372.
Dzierlenga. AL; Robinson. VG; Waidyanatha. S; Devito. MJ; Eifrid. MA: Gibbs. ST: Granville. CA;
Blystone. CR. (2019). Toxicokinetics of perfluorohexanoic acid (PFHxA),
perfluorooctanoic acid (PFOA) and perfluorodecanoic acid (PFDA) in male and female
Hsd:Sprague dawley SD rats following intravenous or gavage administration.
Xenobiotica 50: 1-11. http://dx.doi.org/10.1080/00498254.2Q19.1683776.
This document is a draft for review purposes only and does not constitute Agency policy.
R-3 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Dzierlenga. M. ,W.; Crawford. L. Longnecker. M. .P. (2020). Birth weight and perfluorooctane
sulfonic acid: a random-effects meta-regression analysis. Environmental Epidemiology 4:
e095. http://dx.doi.org/10.1097/EE9.0000000000000Q95.
Edwards. SW; Tan. YM; Villeneuve. PL: Meek. ME: McQueen. CA. (2016). Adverse outcome
pathways—Organizing toxicological information to improve decision making [Review]. J
Pharmacol Exp Ther 356:170-181. http://dx.doi.org/10.1124/jpet.115.228239.
Frawley. RP; Smith. M; Cesta. MF; Hayes-Bouknight. S; Blystone. C; Kissling. GE; Harris. S;
Germolec. D. (2018). Immunotoxic and hepatotoxic effects of perfluoro-n-decanoic acid
(PFDA) on female Harlan Sprague-Dawley rats and B6C3F1/N mice when administered
by oral gavage for 28 days. J Immunotoxicol 15: 41-52.
http://dx.doi.org/10.1080/1547691X.2018.1445145.
Galazka. A: Kardymowicz. B. (1989). Immunity against diphtheria in adults in Poland. Epidemiol
Infect 103: 587-593. http://dx.doi.org/10.1017/s095026880003Q983.
Galazka. AM: Milstien. JB; Robertson. SE; Cutts. FT. (1993). The immunological basis for
immunization module 2 : Diphtheria. (WHO/EPI/Gen/93.11-18). Galazka, AM; Milstien,
JB; Robertson, SE; Cutts, FT.
http://apps.who.int/iris/bitstream/handle/10665/58891/WHO-EPI-GEN-93.12-mod2-
eng.pdf?sequence=38&isAllowed=y.
Gelman. A: Lee. D; Guo. L. (2015). Stan: a probabilistic programming language for bayesian
inference and optimization. American Educational Research Journal 40.
http://dx.doi.org/10.3102/1076998615606113.
Gibson. H. ,M. (1973). Plasma volume and glomerular filtration rate in pregnancy and their
relation to differences in fetal growth. Br J Obstet Gynaecol 80: 1067-1074.
http://dx.doi.Org/10.llll/i.1471-0528.1973.tb02981.x.
Glauert. HP: Srinivasan. S; Tatum. VL; Chen. LC; Saxon. DM: Lav. LT; Borges. T; Baker. M; Chen.
LH; Robertson. LW; Chow. CK. (1992). Effects of the peroxisome proliferators
ciprofibrate and perfluorodecanoic acid on hepatic cellular antioxidants and lipid
peroxidation in rats. Biochem Pharmacol 43: 1353-1359.
http://dx.doi.org/10.1016/0006-2952(92)90513-1.
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/es301168c.
Goecke-Flora, CM; Wyman, JF; Jarnot, BM; Reo, NV. (1995). Effect of the peroxisome
proliferator perfluoro-n-decanoic acid on glucose transport in the isolated perfused rat
liver. Chem Res Toxicol 8: 77-81. http://dx.doi.org/10.1021/txQ0043a010.
Goecke, CM; Jarnot, BM; Reo, NV. (1994). Effects of the peroxisome proliferator perfluoro-n-
decanoic acid on hepatic gluconeogenesis and glycogenesis: a 13C NMR investigation.
Chem Res Toxicol 7:15-22. http://dx.doi.org/10.1021/tx00037a003.
Grandjean. P; Andersen. EW; Budtz-J0rgensen, E; Nielsen. F; M0lbak. K; Weihe. P; Heilmann. C.
(2012). Serum vaccine antibody concentrations in children exposed to perfluorinated
compounds. JAMA 307: 391-397. http://dx.doi.org/10.1001/iama.2011.2Q34.
This document is a draft for review purposes only and does not constitute Agency policy.
R-4 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Grandjean. P; Bateson. T. (2021). RE: Benchmark analysis for PFAS immunotoxicity. Available
online at (accessed
Grandjean. P; Heilmann. C; Weihe. P; Nielsen. F; Mogensen. UB; Timmermann. A; Budtz-
J0rgensen. E. (2017). Estimated exposures to perfluorinated compounds in infancy
predict attenuated vaccine antibody concentrations at age 5-years. J Immunotoxicol 14:
188-195. http://dx.doi.org/10.1080/1547691X.2017.136Q968.
Gyllenhammar. I: Diderholm. B; Gustafsson. J: Berger. U; Ridefelt. P; Benskin. JP; Lignell. S;
Lampa. E; Glynn. A. (2018). Perfluoroalkyl acid levels in first-time mothers in relation to
offspring weight gain and growth. Environ Int 111: 191-199.
http://dx.doi.Org/10.1016/i.envint.2017.12.002.
Hack. M; Klein. INK; Taylor. HG. (1995). Long-term developmental outcomes of low birth weight
infants [Review]. Future Child 5: 176-196. http://dx.doi.org/10.2307/16Q2514.
Hall. AP; Elcombe. CR; Foster. JR; Harada. T; Kaufmann. W; Knippel. A: Kuttler. K; Malarkey. DE;
Maronpot. RR; Nishikawa. A: Nolte. T; Schulte. A: Strauss. V; York. MJ. (2012). Liver
hypertrophy: a review of adaptive (adverse and non-adverse) changes-conclusions from
the 3rd International ESTP Expert Workshop [Review]. Toxicol Pathol 40: 971-994.
http://dx.doi.org/10.1177/0192623312448935.
Han. CY. (2018). Update on FXR Biology: Promising Therapeutic Target? International Journal of
Molecular Sciences 19. http://dx.doi.org/10.3390/iimsl9072069.
Harris. MW; Birnbaum. LS. (1989). Developmental toxicity of perfluorodecanoic acid in
C57BL/6N mice. Fundam Appl Toxicol 12: 442-448. http://dx.doi.org/10.1016/Q272-
0590(89)90018-3.
Harrison. EH: Lane. JS; Luking. S; Van Rafelghem. MJ: Andersen. ME. (1988). Perfluoro-n-
decanoic acid: Induction of peroxisomal beta-oxidation by a fatty acid with dioxin-like
toxicity. Lipids 23: 115-119. http://dx.doi.org/10.1007/BF0253529Q.
Hu. J: Li. J: Wang. J: Zhang. A: Dai. J. (2014). Synergistic effects of perfluoroalkyl acids mixtures
with J-shaped concentration-responses on viability of a human liver cell line.
Chemosphere 96: 81-88. http://dx.doi.Org/10.1016/i.chemosphere.2013.07.033.
Huang. CY: Wilson. MW: Lav. LT; Chow. CK; Robertson. LW; Glauert. HP. (1994). Increased 8-
hydroxydeoxyguanosine in hepatic DNA of rats treated with the peroxisome
proliferators ciprofibrate and perfluorodecanoic acid. Cancer Lett 87: 223-228.
http://dx.doi.org/10.1016/0304-3835(94)90226-7.
Ikeda. T; Aiba. K; Fukuda. K; Tanaka. M. (1985). The induction of peroxisome proliferation in rat
liver by perfluorinated fatty acids, metabolically inert derivatives of fatty acids. J
Biochem 98: 475-482. http://dx.doi.org/10.1093/oxfordiournals.ibchem.al35302.
Intrasuksri, U; Feller, DR. (1991). Comparison of the effects of selected monocarboxylic,
dicarboxylic and perfluorinated fatty acids on peroxisome proliferation in primary
cultured rat hepatocytes. Biochem Pharmacol 42: 184-188.
http://dx.doi.org/10.1016/0006-2952(91)90698-5.
IPCS (International Programme on Chemical Safety). (2007). Harmonization project document
no. 4: Part 2: IPCS framework for analysing the relevance of a non-cancer mode of
action for humans. Geneva, Switzerland: World Health Organization.
http://www.who.int/ipcs/methods/harmonization/areas/cancer mode.pdf?ua=l.
This document is a draft for review purposes only and does not constitute Agency policy.
R-5 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Ishibashi. H; Hirano. M; Kim. EY; Iwata. H. (2019). In vitro and in silico evaluations of binding
affinities of perfluoroalkyl substances to baikal seal and human peroxisome proliferator-
activated receptor a. Environ Sci Technol 53: 2181-2188.
http://dx.doi.org/10.1021/acs.est.8b07273.
Jaeschke. H; Mcgill. MR: Ramachandran. A. (2012). Oxidant stress, mitochondria, and cell death
mechanisms in drug-induced liver injury: lessons learned from acetaminophen
hepatotoxicity [Review]. Drug Metab Rev 44: 88-106.
http://dx.doi.org/10.3109/03602532.2011.6Q2688.
Johnson. PR: Klaassen. CD. (2002). Regulation of rat multidrug resistance protein 2 by classes of
prototypical microsomal enzyme inducers that activate distinct transcription pathways.
Toxicol Sci 67: 182-189. http://dx.doi.Org/10.1093/toxsci/67.2.182.
Joshi-Barve. S; Kirpich. I: Cave. MC; Marsano. LS; McClain. G. (2015). Alcoholic, nonalcoholic,
and toxicant-associated steatohepatitis: Mechanistic similarities and differences
[Review], CMGH 1: 356-367. http://dx.doi.org/10.1016/Mcmgh.2015.05.006.
Judson. RS; Magpantay. FM; Chickarmane. V; Haskell. C; Tania. IN; Taylor. J: Xia. M; Huang. R;
Rotroff. DM: Filer. PL: Houck. KA; Martin. MT; Sipes. IN; Richard. AM: Mansouri. K;
Setzer. RW; Knudsen. TB; Crofton. KM; Thomas. RS. (2015). Integrated model of
chemical perturbations of a biological pathway using 18 in vitro high throughput
screening assays for the estrogen receptor. Toxicol Sci 148: 137-154.
http://dx.doi.org/10.1093/toxsci/kfvl68.
Kaiser. JP; Lipscomb. JC; Wesselkamper. SC. (2012). Putative mechanisms of environmental
chemical-induced steatosis. Int J Toxicol 31: 551-563.
http://dx.doi.org/10.1177/1091581812466418.
Kawashima. Y; Kobayashi. H; Miura. H; Kozuka. H. (1995). Characterization of hepatic responses
of rat to administration of perfluorooctanoic and perfluorodecanoic acids at low levels.
Toxicology 99:169-178. http://dx.doi.org/10.1016/0300-483X(95)03027-D.
Kelling, CK; Van Rafelghem, MJ; Drake, RL; Menahan, LA; Peterson, RE. (1986). Regulation of
hepatic malic enzyme by perfluorodecanoic acid. J Biochem Toxicol 1: 23-37.
http://dx.doi.org/10.1002/ibt.25700103Q4.
Kelling. CK; Van Rafelghem. MJ; Menahan. LA; Peterson. RE. (1987). Effects of
perfluorodecanoic acid on hepatic indices of thyroid status in the rat. Biochem
Pharmacol 36: 1337-1344. http://dx.doi.org/10.1016/0006-2952(87)90091-8.
Kim. SC; Hong. JT; Jang. SJ; Kang. WS; Yoo. HS; Yun. YP. (1998). Formation of 8-
oxodeoxyguanosine in liver DNA and hepatic injury by peroxisome proliferator clofibrate
and perfluorodecanoic acid in rats. J Toxicol Sci 23: 113-119.
http://dx.doi.Org/10.2131/its.23.2 113.
Kim. SJ; Choi. EJ; Choi. GW; Lee. YB; Cho. HY. (2019). Exploring sex differences in human health
risk assessment for PFNA and PFDA using a PBPK model. Arch Toxicol 93: 311-330.
http://dx.doi.org/10.1007/sQ0204-018-2365-v.
Kleinstreuer. NC; Ceger. P; Watt. ED; Martin. M; Houck. K; Browne. P; Thomas. RS; Casey. WM;
Dix. DJ; Allen. D; Sakamuru. S; Xia. M; Huang. R; Judson. R. (2017). Development and
validation of a computational model for androgen receptor activity. Chem Res Toxicol
30: 946-964. http://dx.doi.org/10.1021/acs.chemrestox.6b00347.
This document is a draft for review purposes only and does not constitute Agency policy.
R-6 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Kozuka, H; Watanabe, T; Horie, S; Yamada, J; Suga, T; Ikeda, T. (1991a). Characteristics of
peroxisome proliferation: co-induction of peroxisomal fatty acid oxidation-related
enzymes with microsomal laurate hydroxylase. Chem Pharm Bull (Tokyo) 39:1267-1271.
http://dx.doi.org/10.1248/cpb.39.1267.
Kozuka, H; Yamada, J; Horie, S; Watanabe, T; Suga, T; Ikeda, T. (1991b). Characteristics of
induction of peroxisomal fatty acid oxidation-related enzymes in rat liver by
drugs:Relationships between structure and inducing activity. Biochem Pharmacol 41:
617-623. http://dx.doi.ore/10.1016/0006-2952f9H90635-l.
Kruschke. JK. (2021). Bayesian analysis reporting guidelines. Nat Hum Behav 5: 1282-1291.
http://dx.doi.ore/10.1038/s41562-021-01177-7.
Kudo, N; Bandai, N; Suzuki, E; Katakura, M; Kawashima, Y. (2000). Induction by perfluorinated
fatty acids with different carbon chain length of peroxisomal beta-oxidation in the liver
of rats. Chem Biol Interact 124: 119-132. http://dx.doi.org/10.1016/S00Q9-
2797(99)00150-7.
Kudo. IN; Kawashima. Y. (2003). Induction of triglyceride accumulation in the liver of rats by
perfluorinated fatty acids with different carbon chain lengths: Comparison with
induction of peroxisomal beta-oxidation. Biol Pharm Bull 26: 47-51.
http://dx.doi.org/10.1248/bpb.26.47.
Langley. AE. (1990). Effects of perfluoro-n-decanoic acid on the respiratory activity of isolated
rat liver mitochondria. J Toxicol Environ Health 29: 329-336.
http://dx.doi.ore/10.1080/15287399009531395.
Lenters. V; Portengen. L; Rignell-Hydbom. A; Jonsson. BA; Lindh. CH; Piersma. AH: Toft. G;
Bonde. JP; Heederik. D; Rylander. L; Vermeulen. R. (2016). Prenatal phthalate,
perfluoroalkyl acid, and organochlorine exposures and term birth weight in three birth
cohorts: multi-pollutant models based on elastic net regression. Environ Health Perspect
124: 365-372. http://dx.doi.ore/10.1289/ehp.1408933.
Li. C; Ren. X: Cao. L; Qin. W; Guo. LH. (2019). Investigation of binding and activity of
perfluoroalkyl substances to the human peroxisome proliferator-activated receptor 3/6.
Environ Sci Process Impacts 21:1908-1914. http://dx.doi.org/10.1039/c9em00218a.
Li. K; Gao. P; Xiang. P; Zhang. X: Cui. X: Ma. LQ. (2017). Molecular mechanisms of PFOA-induced
toxicity in animals and humans: Implications for health risks [Review]. Environ Int 99: 43-
54. http://dx.doi.ore/10.1016/j.envint.2016.11.014.
Li. KM: Zhao. Q; Fan. ZY; Jia. SY; Liu. Q; Liu. FY: Liu. SL. (2022). The toxicity of perfluorodecanoic
acid is mainly manifested as a deflected immune function. Mol Biol Rep 49: 4365-4376.
http://dx.doi.ore/10.1007/sll033-022-07272-w.
Li. T; Yu. RT; Atkins. AR; Downes. M; Tukey. RH; Evans. RM. (2012). Targeting the pregnane X
receptor in liver injury [Review]. Expert Opin Ther Targets 16: 1075-1083.
http://dx.doi.org/10.1517/14728222.2012.715634.
Liang. JL; Tiwari. T; Moro. P; Messonnier. NE; Reingold. A: Sawyer. M; Clark. TA. (2018).
Prevention of pertussis, tetanus, and diphtheria with vaccines in the United States:
Recommendations of the Advisory Committee on Immunization Practices (ACIP).
MMWR Recomm Rep 67: 1-44. http://dx.doi.org/10.15585/mmwr.rr6702al.
This document is a draft for review purposes only and does not constitute Agency policy.
R-7 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Lim. JJ; Suh. Y; Faustman. EM; Cui. JY. (2021). Regulation of transporters by perfluorinated
carboxylic acids in HepaRG cells. Drug Metab Dispos 50: 1396-1413.
http://dx.doi.org/10.1124/dmd.121.000477.
Long. M; Ghisari. M; Bonefeld-J0rgensen, EC. (2013). Effects of perfluoroalkyl acids on the
function of the thyroid hormone and the aryl hydrocarbon receptor. Environ Sci Pollut
Res Int 20: 8045-8056. http://dx.doi.org/10.1007/sll356-013-1628-7.
Luo. D; Wu. WX; Pan. YA; Du. BB; Shen. MJ; Zeng. LX. (2021). Associations of prenatal exposure
to per- and polyfluoroalkyl substances with the neonatal birth size and hormones in the
growth hormone/insulin-like growth factor axis. Environ Sci Technol 55: 11859-11873.
http://dx.doi.org/10.1021/acs.est.lc02670.
Luo. M; Tan. Z; Dai. M; Song. D; Lin. J: Xie. M; Yang. J: Sun. L; Wei. D; Zhao. J: Gonzalez. FJ; Liu.
(2017). Dual action of peroxisome proliferator-activated receptor alpha in
perfluorodecanoic acid-induced hepatotoxicity. Arch Toxicol 91: 897-907.
http://dx.doi.org/10.1007/sQ0204-016-1779-7.
Ma. K; Saha. PK; Chan. L; Moore. DP. (2006). Farnesoid X receptor is essential for normal
glucose homeostasis. J Clin Invest 116:1102-1109. http://dx.doi.org/10.1172/JCI25604.
Ma. Y; Sachdeva. K; Liu. J: Song. X: Li. Y; Yang. D; Deng. R; Chichester. CO: Yan. B. (2005).
Clofibrate and perfluorodecanoate both upregulate the expression of the pregnane X
receptor but oppositely affect its ligand-dependent induction on cytochrome P450
3A23. Biochem Pharmacol 69: 1363-1371. http://dx.doi.Org/10.1016/i.bcp.2005.02.011.
Mackowiak. B; Hodge. J: Stern. S; Wang. H. (2018). The roles of xenobiotic receptors: Beyond
chemical disposition [Review]. Drug Metab Dispos 46:1361-1371.
http://dx.doi.org/10.1124/dmd.118.081042.
Maher. JM; Aleksunes. LM; Dieter. MZ; Tanaka. Y; Peters. JM; Manautou. JE; Klaassen. CD.
(2008). Nrf2- and PPAR alpha-mediated regulation of hepatic Mrp transporters after
exposure to perfluorooctanoic acid and perfluorodecanoic acid. Toxicol Sci 106: 319-
328. http://dx.doi.org/10.1093/toxsci/kfnl77.
Malhi. H; Gores. GJ. (2008). Cellular and molecular mechanisms of liver injury [Review].
Gastroenterology 134:1641-1654. http://dx.doi.Org/10.1053/i.gastro.2008.03.002.
Manzano-Salgado. CB; Casas. M; Lopez-Espinosa. MJ: Ballester. F; Iniguez. C; Martinez. D; Costa.
O: Santa-Marina. L; Pereda-Pereda. E; Schettgen. T; Sunyer. J: Vrijheid. M. (2017).
Prenatal exposure to perfluoroalkyl substances and birth outcomes in a Spanish birth
cohort. Environ Int 108: 278-284. http://dx.doi.Org/10.1016/i.envint.2017.09.006.
Meek. ME: Boobis. A: Cote. I: Dellarco. V; Fotakis. G; Munn. S; Seed. J: Vickers. C. (2014). New
developments in the evolution and application of the WHO/IPCS framework on mode of
action/species concordance analysis [Review]. J Appl Toxicol 34:1-18.
http://dx.doi.org/10.10Q2/iat.2949.
Mellor. CL; Steinmetz. FP; Cronin. MT. (2016). The identification of nuclear receptors associated
with hepatic steatosis to develop and extend adverse outcome pathways [Review]. Crit
Rev Toxicol 46: 138-152. http://dx.doi.org/10.3109/10408444.2Q15.1089471.
Meng. Q; Inoue. K; Ritz. B; Olsen. J: Liew. Z. (2018). Prenatal exposure to perfluoroalkyl
substances and birth outcomes; an updated analysis from the danish national birth
This document is a draft for review purposes only and does not constitute Agency policy.
R-8 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
cohort. Int J Environ Res Public Health 15: 1832.
http://dx.doi.org/10.3390/iierphl5Q91832.
Morken. NH; Travlos. GS; Wilson. RE: Eggesb0. M; Longnecker. MP. (2014). Maternal glomerular
filtration rate in pregnancy and fetal size. PLoS ONE 9: el01897.
http://dx.doi.org/10.1371/iournal.pone.0101897.
NTP (National Toxicology Program). (2018). 28-day evaluation of the toxicity (C20615) of
perfluorodecanoic acid (PFDA) (335-76-2) on Harlan Sprague-Dawley rats exposed via
gavage [NTP]. U.S. Department of Health and Human Services.
http://dx.doi.org/10.22427/NTP-DATA-002-02652-0004-000Q-l.
Oguro. T; Hayashi. M; Nakajo. S; Numazawa. S; Yoshida. T. (1998). The expression of heme
oxygenase-1 gene responded to oxidative stress produced by phorone, a glutathione
depletor, in the rat liver; the relevance to activation of c-jun n-terminal kinase. J
Pharmacol Exp Ther 287: 773-778.
Ohmori. K; Kudo. N; Katayama. K; Kawashima. Y. (2003). Comparison of the toxicokinetics
between perfluorocarboxylic acids with different carbon chain length. Toxicology 184:
135-140. http://dx.doi.org/10.1016/S0300-483X(02)00573-5.
Ojo. AF; Peng. C; Ng. JC. (2020). Combined effects and toxicological interactions of
perfluoroalkyl and polyfluoroalkyl substances mixtures in human liver cells (HepG2).
Environ Pollut 263:114182. http://dx.doi.Org/10.1016/i.envpol.2020.114182.
Ojo. AF: Xia. Q; Peng. C; Ng. JC. (2021). Evaluation of the individual and combined toxicity of
perfluoroalkyl substances to human liver cells using biomarkers of oxidative stress.
Chemosphere 281: 130808. http://dx.doi.Org/10.1016/i.chemosphere.2021.130808.
Olson. CT; Andersen. ME. (1983). The acute toxicity of perfluorooctanoic and perfluorodecanoic
acids in male rats and effects on tissue fatty acids. Toxicol Appl Pharmacol 70: 362-372.
http://dx.doi.org/10.1016/0041-008x(83)90154-0.
Passen. EL: Andersen. BR. (1986). Clinical tetanus despite a protective level of toxin-neutralizing
antibody [Case Report]. JAMA 255: 1171-1173.
http://dx.doi.org/10.1001/iama.1986.03370090093Q29.
Patel. JC: Mehta. BC. (1999). Tetanus: Study of 8,697 cases. Indian J Med Sci 53: 393-401.
Permadi. H; Lundgren. B; Andersson. K; Sundberg. C; Depierre. JW. (1993). Effects of perfluoro
fatty acids on peroxisome proliferation and mitochondrial size in mouse liver: dose and
time factors and effect of chain length. Xenobiotica 23: 761-770.
http://dx.doi.org/10.3109/004982593Q9166782.
Powers, RH; Aust, SD. (1986). The effects of nonadecafluoro-n-decanoic acid on serum retinol
and hepatic retinyl palmitate hydrolase activity in male Sprague-Dawley rats. J Biochem
Toxicol 1: 27-42. http://dx.doi.org/10.1002/ibt.25700102Q4.
Reo. NV; Goecke. CM: Narayanan. L; Jarnot. BM. (1994). Effects of perfluoro-n-octanoic acid,
perfluoro-n-decanoic acid, and clofibrate on hepatic phosphorus metabolism in rats and
guinea pigs in vivo. Toxicol Appl Pharmacol 124:165-173.
http://dx.doi.org/10.1006/taap.1994.102Q.
Reyes. L; Manalich. R. (2005). Long-term consequences of low birth weight [Review]. Kidney Int
Suppl 68: S107-S111. http://dx.doi.Org/10.llll/i.1523-1755.2005.09718.x.
This document is a draft for review purposes only and does not constitute Agency policy.
R-9 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Robledo. CA; Yeung. E; Mendola. P; Sundaram. R; Maisog. J; Sweeney. AM; Barr. DB; Louis. GM.
(2015). Preconception maternal and paternal exposure to persistent organic pollutants
and birth size: the LIFE study. Environ Health Perspect 123: 88-94.
http://dx.doi.org/10.1289/ehp.1308Q16.
Rosen. MB: Das. KP; Wood. CR; Wolf. CJ; Abbott. BP: Lau. C. (2013). Evaluation of perfluoroalkyl
acid activity using primary mouse and human hepatocytes. Toxicology 308: 129-137.
http://dx.doi.Org/10.1016/i.tox.2013.03.011.
Rosenmai. AK; Ahrens. L; le Godec. T; Lundqvist. J: Oskarsson. A. (2018). Relationship between
peroxisome proliferator-activated receptor alpha activity and cellular concentration of
14 perfluoroalkyl substances in HepG2 cells. J Appl Toxicol 38: 219-226.
http://dx.doi.org/10.10Q2/iat.3515.
Roth. RA; Jaeschke. H; Luyendyk. JP. (2019). Chapter 13: Toxic responses of the liver. In CD
Klaassen (Ed.), Casarett & Doull's toxicology: The basic science of poisons (9th ed., pp.
719-766). New York, NY: McGraw Hill.
Routti. H; Berg. MK; Lille-Lang0v. R; 0ygarden. L; Harju. M; Dietz. R; Sonne. C; Goks0yr. A.
(2019). Environmental contaminants modulate the transcriptional activity of polar bear
(Ursus maritimus) and human peroxisome proliferator-activated receptor alpha
(PPARA). Sci Rep 9: 6918. http://dx.doi.org/10.1038/s41598-Q19-43337-w.
Russell. DW. (2003). The enzymes, regulation, and genetics of bile acid synthesis [Review]. Annu
Rev Biochem 72: 137-174.
http://dx.doi.org/10.1146/annurev.biochem.72.121801.161712.
Sagiv. SK; Rifas-Shiman. SL; Fleisch. AF; Webster. TF; Calafat. AM: Ye. X: Gillman. MW; Oken. E.
(2018). Early Pregnancy Perfluoroalkyl Substance Plasma Concentrations and Birth
Outcomes in Project Viva: Confounded by Pregnancy Hemodynamics? Am J Epidemiol
187: 793-802. http://dx.doi.org/10.1093/aie/kwx332.
Salvatier. J: Wiecki. TV: Fonnesbeck. C. (2016). Probabilistic programming in Python using
PyMC3. PeerJ Computer Science 2: e55. http://dx.doi.org/10.7717/peeri-cs.55.
Sanghavi. M; Rutherford. JD. (2014). Cardiovascular physiology of pregnancy. Circulation 130:
1003-1008. http://dx.doi.org/10.1161/CIRCULATIQNAHA.114.009029.
Schramm. H; Friedberg. T; Robertson. LW; Oesch. F; Kissel. W. (1989). Perfluorodecanoic acid
decreases the enzyme activity and the amount of glutathione S-transferases proteins
and mRNAs in vivo. Chem Biol Interact 70: 127-143. http://dx.doi.org/10.1016/00Q9-
2797(89)90068-9.
Selgrade. MK. (2007). Immunotoxicity: The risk is real [Review]. Toxicol Sci 100: 328-332.
http://dx.doi.org/10.1093/toxsci/kfm244.
Sovadinova. I: Babica. P; Boke. H; Kumar. E; Wilke. A: Park. JS; Trosko. JE; Upham. BL. (2015).
Phosphatidylcholine Specific PLC-lnduced Dysregulation of Gap Junctions, a Robust
Cellular Response to Environmental Toxicants, and Prevention by Resveratrol in a Rat
Liver Cell Model. PLoS ONE 10: e0124454.
http://dx.doi.org/10.1371/iournal.pone.0124454.
Starling. AP; Adgate. JL; Hamman. RF; Kechris. K; Calafat. AM: Ye. X: Dabelea. D. (2017).
Perfluoroalkyl substances during pregnancy and offspring weight and adiposity at birth:
This document is a draft for review purposes only and does not constitute Agency policy.
R-10 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Examining mediation by maternal fasting glucose in the healthy start study. Environ
Health Perspect 125: 067016. http://dx.doi.org/10.1289/EHP641.
Steenland. K; Barry. V; Savitz. D. (2018). Serum perfluorooctanoic acid and birthweight: an
updated meta-analysis with bias analysis. Epidemiology 29: 765-776.
http://dx.doi.org/10.1097/EDE.00000000000009Q3.
Sterchele. PF; Sun. H; Peterson. RE: Vanden Heuvel. JP. (1996). Regulation of peroxisome
proliferator-activated receptor-alpha mRNA in rat liver. Arch Biochem Biophys 326: 281-
289. http://dx.doi.org/10.1006/abbi.1996.0Q77.
Sterchele. PF: Vanden Heuvel. JP: Davis. JW; Shrago. E; Knudsen. J: Peterson. RE. (1994).
Induction of hepatic acyl-CoA-binding protein and liver fatty acid-binding protein by
perfluorodecanoic acid in rats. Lack of correlation with hepatic long-chain acyl-CoA
levels. Biochem Pharmacol 48: 955-966. http://dx.doi.org/10.1016/00Q6-
2952(94)90366-2.
Takacs. ML: Abbott. BP. (2007). Activation of mouse and human peroxisome proliferator-
activated receptors (alpha, beta/delta, gamma) by perfluorooctanoic acid and
perfluorooctane sulfonate. Toxicol Sci 95:108-117.
http://dx.doi.org/10.1093/toxsci/kfll35.
Takagi. A: Sai. K; Umemura. T; Hasegawa. R; Kurokawa. Y. (1991). Short-term exposure to the
peroxisome proliferators, perfluorooctanoic acid and perfluorodecanoic acid, causes
significant increase of 8-hydroxydeoxyguanosine in liver DNA of rats. Cancer Lett 57: 55-
60. http://dx.doi.org/10.1016/0304-3835(91)90063-N.
Takahashi. S; Tanaka. IN; Golla. S; Fukami. T; Krausz. KW; Polunas. MA: Weig. BC; Masuo. Y; Xie.
C; Jiang. C; Gonzalez. FJ. (2017). Editor's highlight: farnesoid X receptor protects against
low-dose carbon tetrachloride-induced liver injury through the taurocholate-
JNKpathway. Toxicol Sci 158: 334-346. http://dx.doi.org/10.1093/toxsci/kfx094.
Tian. M. .; Reichetzeder. C. .; Li. J. .; Hocher. B. . (2019). Low birth weight, a risk factor for
diseases in later life, is a surrogate of insulin resistance at birth. J Hypertens 37: 2123-
2134. http://dx.doi.org/10.1097/HJH.00000000000Q2156.
U.S. EPA (U.S. Environmental Protection Agency). (2012). Benchmark dose technical guidance
[EPA Report]. (EPA100R12001). Washington, DC: U.S. Environmental Protection Agency,
Risk Assessment Forum, https://www.epa.gov/risk/benchmark-dose-technical-guidance.
U.S. EPA (U.S. Environmental Protection Agency). (2016a). Health effects support document for
perfluorooctane sulfonate (PFOS) [EPA Report]. (EPA 822-R-16-002). Washington, DC:
U.S. Environmental Protection Agency, Office of Water, Health and Ecological Criteria
Division, https://www.epa.gov/sites/production/files/2016-
05/documents/pfos hesd final 508.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2016b). Health effects support document for
perfluorooctanoic acid (PFOA) [EPA Report]. (EPA 822-R-16-003). Washington, DC: U.S.
Environmental Protection Agency, Office of Water, Health and Ecological Criteria
Division, https://www.epa.gov/sites/production/files/2016-
05/documents/pfoa hesd final-plain.pdf.
This document is a draft for review purposes only and does not constitute Agency policy.
R-ll DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
U.S. EPA (U.S. Environmental Protection Agency). (2019). CompTox Chemicals Dashboard
[Database]. Research Triangle Park, NC. Retrieved from
https://comptox.epa.gov/dashboard
Valvi. D; Oulhote. Y; Weihe. P; Dalgard. C; Bjerve. KS; Steuerwald. U; Grandjean. P. (2017).
Gestational diabetes and offspring birth size at elevated environmental pollutant
exposures. Environ Int 107: 205-215. http://dx.doi.Org/10.1016/i.envint.2017.07.016.
van Otterdijk. FM. (2007). Repeated dose 90-day oral toxicity study with MTDID 8391 by daily
gavage in the rat followed by a 3-week recovery period. (Study Number 06-398).
Maplewood, MN: 3M.
Van Rafelghem, MJ; Andersen, ME. (1988). The effects of perfluorodecanoic acid on hepatic
stearoyl-coenzyme A desaturase and mixed function oxidase activities in rats. Fundam
Appl Toxicol 11: 503-510. http://dx.doi.org/10.1016/0272-0590(88)90114-5.
Van Rafelghem. MJ: Mattie. PR: Bruner. RH; Andersen. ME. (1987). Pathological and hepatic
ultrastructural effects of a single dose of perfluoro-n-decanoic acid in the rat, hamster,
mouse, and guinea pig. Fundam Appl Toxicol 9: 522-540.
http://dx.doi.org/10.1016/0272-0590(87)90034-0.
Van Rafelghem, MJ; Vanden Heuvel, JP; Menahan, LA; Peterson, RE. (1988). Perfluorodecanoic
acid and lipid metabolism in the rat. Lipids 23: 671-678.
http://dx.doi.org/10.1007/BF02535666.
Vanden Heuvel. JP: Kuslikis. Bl; Shrago. E; Peterson. RE. (1991). Inhibition of long-chain acyl-CoA
synthetase by the peroxisome proliferator perfluorodecanoic acid in rat hepatocytes.
Biochem Pharmacol 42: 295-302. http://dx.doi.org/10.1016/0006-2952(91)90716-l.
Vanden Heuvel. JP: Sterchele. PF; Nesbit. DJ; Peterson. RE. (1993). Coordinate induction of acyl-
CoA binding protein, fatty acid binding protein and peroxisomal beta-oxidation by
peroxisome proliferators. Biochim Biophys Acta 1177: 183-190.
http://dx.doi.org/10.1016/0167-4889(93)90039-R.
Wahlang. B; Beier. Jl; Clair. HB; Bellis-Jones. HJ; Falkner. K; Mcclain. CJ; Cave. MC. (2013).
Toxicant-associated steatohepatitis [Review]. Toxicol Pathol 41: 343-360.
http://dx.doi.org/10.1177/0192623312468517.
Wallace. KB: Kissling. GE; Melnick. RL; Blystone. CR. (2013). Structure-activity relationships for
perfluoroalkane-induced in vitro interference with rat liver mitochondrial respiration.
Toxicol Lett 222: 257-264. http://dx.doi.Org/10.1016/i.toxlet.2013.07.025.
Wang. D; Gao. Q; Wang. T; Kan. Z; Li. X: Hu. L; Peng. CY; Qian. F; Wang. Y; Granato. D. (2020).
Green tea polyphenols and epigallocatechin-3-gallate protect against perfluorodecanoic
acid induced liver damage and inflammation in mice by inhibiting NLRP3 inflammasome
activation. Food Res Int 127: 108628. http://dx.doi.Org/10.1016/i.foodres.2019.108628.
Wang. YM; Chai. SC; Brewer. CT; Chen. T. (2014). Pregnane X receptor and drug-induced liver
injury [Review]. Expert Opin Drug Metab Toxicol 10: 1521-1532.
http://dx.doi.org/10.1517/17425255.2014.963555.
Watt. ED: Judson. RS. (2018). Uncertainty quantification in ToxCast high throughput screening.
PLoS ONE 13: e0196963. http://dx.doi.org/10.1371/iournal.pone.0196963.
Weisskopf. MG; Seals. RM; Webster. TF. (2018). Bias amplification in epidemiologic analysis of
exposure to mixtures. Environ Health Perspect 126. http://dx.doi.org/10.1289/EHP2450.
This document is a draft for review purposes only and does not constitute Agency policy.
R-12 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Weisskopf. MG; Webster. TF. (2017). Trade-offs of personal versus more proxy exposure
measures in environmental epidemiology. Epidemiology 28: 635-643.
http://dx.doi.org/10.1097/EDE.000000000000Q686.
Whitworth. KW; Haug. LS; Baird. DP: Becher. G; Hoppin. JA; Skjaerven. R; Thomsen. C; Eggesbo.
M; Travlos. G; Wilson. R; Longnecker. MP. (2012). Perfluorinated compounds and
subfecundity in pregnant women. Epidemiology 23: 257-263.
http://dx.doi.org/10.1097/EPE.0b013e31823b5031.
Wiels0e. M; Long. M; Ghisari. M; Bonefeld-J0rgensen, EC. (2015). Perfluoroalkylated substances
(PFAS) affect oxidative stress biomarkers in vitro. Chemosphere 129: 239-245.
http://dx.doi.Org/10.1016/i.chemosphere.2014.10.014.
Wikstrom. S; Lin. PI: Lindh. CH; Shu. H; Bornehag. CG. (2020). Maternal serum levels of
perfluoroalkyl substances in early pregnancy and offspring birth weight. Pediatr Res 87:
1093-1099. http://dx.doi.org/10.1038/s41390-019-072Q-l.
Witzmann. F; Coughtrie. M; Fultz. C; Lipscomb. J. (1996). Effect of structurally diverse
peroxisome proliferators on rat hepatic sulfotransferase. Chem Biol Interact 99: 73-84.
http://dx.doi.org/10.1016/0009-2797(95)03661-X.
Witzmann. FA: Parker. DIN. (1991). Hepatic protein pattern alterations following
perfluorodecanoic acid exposure in rats. Toxicol Lett 57: 29-36.
http://dx.doi.org/10.1016/0378-4274(91)90116-N .
Wolf. G; Schmid. JE; Lau. C; Abbott. BP. (2012). Activation of mouse and human peroxisome
proliferator-activated receptor-alpha (PPARa) by perfluoroalkyl acids (PFAAs): further
investigation of C4-C12 compounds. Reprod Toxicol 33: 546-551.
http://dx.doi.Org/10.1016/i.reprotox.2011.09.009.
Wolf. G; Takacs. ML: Schmid. JE: Lau. C; Abbott. BP. (2008). Activation of mouse and human
peroxisome proliferator-activated receptor alpha by perfluoroalkyl acids of different
functional groups and chain lengths. Toxicol Sci 106:162-171.
http://dx.doi.org/10.1093/toxsci/kfnl66.
Woods. MM: Lanphear. BP: Braun. JM; McCandless. LC. (2017). Gestational exposure to
endocrine disrupting chemicals in relation to infant birth weight: A Bayesian analysis of
the HOME Study. Environ Health 16: 115. http://dx.doi.org/10.1186/sl2940-017-Q332-
3.
Xu, K; Guidez, F; Glasow, A; Chung, P; Petrie, K; Stegmaier, K; Wang, KK; Zhang, J; Jing, Y; Zelent,
A; Waxman, S. (2005). Benzodithiophenes potentiate differentiation of acute
promyelocytic leukemia cells by lowering the threshold for ligand-mediated
corepressor/coactivator exchange with retinoic acid receptor alpha and enhancing
changes in all-trans-retinoic acid-regulated gene expression. Cancer Res 65: 7856-7865.
http://dx.doi.org/10.1158/0008-5472.CAN-Q5-1056.
Yamamoto. A: Kawashima. Y. (1997). Perfluorodecanoic acid enhances the formation of oleic
acid in rat liver. Biochem J 325 ( Pt 2): 429-434. http://dx.doi.org/10.1042/bi325Q429.
Yang. X: Schnakenberg. LK; Shi. Q; Salminen. WF. (2014). Hepatic toxicity biomarkers. In RC
Gupta (Ed.), Biomarkers in Toxicology (pp. 241-259). New York, NY: Academic Press.
http://dx.doi.org/10.1016/B978-0-12-404630-6.00013-Q.
This document is a draft for review purposes only and does not constitute Agency policy.
R-13 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Supplemental Information for the Toxicological Review ofPFDA and Related Salts
Yao. Q; Gao. Y; Zhang. Y; Qin. K; Liew. Z; Tian. Y. (2021). Associations of paternal and maternal
per- and polyfluoroalkyl substances exposure with cord serum reproductive hormones,
placental steroidogenic enzyme and birth weight. Chemosphere 285: 131521.
http://dx.doi.Org/10.1016/i.chemosphere.2021.131521.
Zhang, L; Ren, XM; Guo, LH. (2013). Structure-based investigation on the interaction of
perfluorinated compounds with human liver fatty acid binding protein. Environ Sci
Technol 47: 11293-11301. http://dx.doi.org/10.1021/es4Q26722.
Zhang. LY; Ren. XM: Wan. B; Guo. LH. (2014). Structure-dependent binding and activation of
perfluorinated compounds on human peroxisome proliferator-activated receptor y.
Toxicol Appl Pharmacol 279: 275-283. http://dx.doi.Org/10.1016/i.taap.2014.06.020.
Zhang. YM; Dong. XY; Fan. LJ; Zhang. ZL; Wang. Q; Jiang. IN; Yang. XS. (2017). Poly- and
perfluorinated compounds activate human pregnane X receptor. Toxicology 380: 23-29.
http://dx.doi.Org/10.1016/i.tox.2017.01.012.
This document is a draft for review purposes only and does not constitute Agency policy.
R-14 DRAFT-DO NOT CITE OR QUOTE
------- |