A EPA
EPA/635/R-18/211 Fb
www.epa.gov/iris
Toxicological Review of Hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
(CASRN 121-82-4]
Supplemental Information
August 2018
Integrated Risk Information System
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection Agency
policy and approved for publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
ii

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
CONTENTS
APPENDIX A. ASSESSMENTS BY OTHER NATIONAL AND INTERNATIONAL HEALTH AGENCIES	A-l
APPENDIX B. ADDITIONAL DETAILS OF LITERATURE SEARCH STRATEGY | STUDY SELECTION AND
EVALUATION	B-l
B.l.	DEFENSE TECHNICAL INFORMATION CENTER (DTIC) LITERATURE SEARCH AND SCREEN	B-l
APPENDIX C. INFORMATION IN SUPPORT OF HAZARD IDENTIFICATION AND DOSE-RESPONSE
ANALYSIS	C-l
C.l.TOXICOKINETIC	S	C-l
C.l.l. Absorption	C-l
C.l.2. Distribution	C-6
C.l.3. Metabolism	C-8
C.1.4. Excretion	C-10
C.l.5. Physiologically Based Pharmacokinetic (PBPK) Models	C-ll
C.2. HUMAN STUDIES	C-34
C.3.	OTHER PERTINENT TOXICITY INFORMATION	C-34
C.3.1. Mortality in Animals	C-34
C.3.2. Other Noncancer Effects	C-48
C.3.3.	Genotoxicity	C-81
APPENDIX D. DOSE-RESPONSE MODELING FOR THE DERIVATION OF REFERENCE VALUES FOR
EFFECTS OTHER THAN CANCER AND THE DERIVATION OF CANCER RISK ESTIMATES	D-l
D.l.BENCHMARK	DOSE MODELING SUMMARY FOR NONCANCER ENDPOINTS	D-l
D.l.l.	Evaluation of Model Fit and Model Selection	D-3
D.l.2. Modeling Results	D-4
D.2.BENCHMARK DOSE MODELING SUMMARY FOR CANCER ENDPOINTS	D-30
D.2.1. Evaluation of Model Fit and Model Selection for Mouse Tumor Data	D-30
D.2.2. Modeling Results for Female Mouse Tumor Data	D-31
D.2.3. Additional Dose-Response Analysis: Male Mice and Rats	D-39
D.2.4. Sensitivity Analysis on Dose-Response Modeling of Female Mouse Tumor Data	D-46
APPENDIX E. SUMMARY OF SCIENCE ADVISORY BOARD (SAB) PEER REVIEW COMMENTS AND
EPA'S DISPOSITION	E-l
REFERENCES FOR APPENDICES	R-l
iii

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
TABLES
Table A-l. Assessments by other national and international health agencies	A-l
Table B-l. Summary of detailed search strategies for hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX) (PubMed, Toxline, Toxcenter, TSCATS)	B-3
Table B-2. Summary of detailed search strategies for hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX) (DTIC online database)	B-16
Table B-3. Processes used to augment the search of core databases for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX)	B-18
Table C-l. Distribution of hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) or radiolabel from
administered RDXa	C-7
Table C-2. Principal urinary metabolites of hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) in
miniature swine 24 hours after dosing with RDX	C-9
Table C-3. Elimination ty2 values for hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) or
radiolabeled RDX	C-ll
Table C-4. Parameter values used in the Sweeney et al. (2012a) and Sweeney et al. (2012b)
physiologically based pharmacokinetic (PBPK) models for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) in rats, humans, and mice as reported by the study
authors	C-13
Table C-5. Parameters values used in the EPA application of the rat, human, and mouse
models	C-17
Table C-6. Doses, dosing formulations, and absorption rate constants in animal and human
studies	C-22
Table C-7. Sensitivity coefficients for rat and human hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX) physiologically based pharmacokinetic (PBPK) models	C-28
Table C-8. Summary of case reports of exposure to hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX)	C-3 5
Table C-9. Occupational epidemiologic studies of hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX): summary of methodologic features	C-40
Table C-10. Evidence pertaining to mortality in animals	C-43
Table C-ll. Evidence pertaining to ocular effects in animals	C-50
Table C-12. Evidence pertaining to cardiovascular effects in animals	C-52
Table C-13. Evidence pertaining to immune effects in animals	C-58
Table C-14. Evidence pertaining to gastrointestinal effects in animals	C-64
Table C-15. Evidence pertaining to other noncancer effects (hematological) in humans	C-67
Table C-16. Evidence pertaining to hematological effects in animals	C-69
Table C-17. Evidence pertaining to male reproductive effects in animals	C-76
Table C-18. Evidence pertaining to body-weight effects in animals	C-79
Table C-19. Summary of in vitro studies of the genotoxicity of hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX)	C-83
Table C-20. Summary of in vivo studies of the genotoxicity of hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX)	C-87
Table C-21. Summary of in vitro and in vivo studies of the genotoxicity of hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) metabolites	C-88
Table D-l. Noncancer endpoints selected for dose-response modeling for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX)	D-2
iv

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-2. Convulsion or mortality endpoints from Crouse et al. (2006) selected for dose-
response modeling for hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)	D-3
Table D-3. Model predictions for convulsions in male and female F344 rats exposed to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by gavage for 90 days (Crouse et
al., 2006); BMR = 5% ER	D-5
Table D-4. Model predictions for convulsions in female F344 rats exposed to hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX) by gavage on gestational days (GDs) 6-19
(Cholakis et al., 1980); BMR = 5% ER	D-8
Table D-5. Model predictions for combined incidence of convulsion and mortality in male
and female F344 rats exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by
gavage for 90 days (Crouse et al., 2006); BMR = 1% ER	D-ll
Table D-6. Model predictions for hemorrhagic/suppurative cystitis of the urinary bladder in
male F344 rats exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet
for 24 months (Levine et al., 1983); BMR = 10% ER	D-14
Table D-7. Model predictions for prostate suppurative inflammation in male F344 rats
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months
(Levine et al., 1983); BMR = 10% ER	D-16
Table D-8. Mortality data selected for dose-response modeling for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX)	D-19
Table D-9. Model predictions for combined mortality in male and female F344 rats exposed
to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 13 weeks (Levine et
al., 1981b); BMR = 1% ER	D-22
Table D-10. Model predictions for mortality (number found dead) in rats exposed to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) in the diet for 13 weeks (von
Oettingen et al., 1949); BMR = 1% ER	D-24
Table D-ll. Model predictions for combined mortality (number found dead) in male and
female F344 rats exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by
gavage for 90 days (Crouse et al., 2006); BMR = 1% ER	D-25
Table D-12. Model predictions for mortality in female Sprague-Dawley rats exposed to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by gavage on Gestation Days 6-15
(Angerhofer et al., 1986); BMR = 1% ER	D-28
Table D-13. Cancer endpoints selected for dose-response modeling for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX)	D-30
Table D-14. Model predictions for combined alveolar/bronchiolar adenoma and carcinoma
in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
by diet for 24 months (Lish et al., 1984b), with highest dose dropped; BMR =
10% ER	D-32
Table D-15. Model predictions for combined hepatocellular adenoma and carcinoma in
female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by
diet for 24 months (Parker et al., 2006; Lish et al., 1984b), with highest dose
dropped; BMR = 10% ER	D-34
Table D-16. Model predictions for combined hepatocellular adenoma and carcinoma in
female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by
diet for 24 months (Parker et al., 2006; Lish et al., 1984b), with highest dose
dropped and sample sizes from Lish et al. (1984b); BMR = 10% ER	D-37
Table D-17. Carcinoma data from Lish et al. (1984b) and Levine et al. (1983)	D-39
v

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-18. Model predictions and oral slope factor for alveolar/bronchiolar carcinomas in
male B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by
diet for 2 years (Lish et al., 1984b)	D-40
Table D-19. Model predictions for alveolar/bronchiolar carcinoma in male B6C3Fi mice
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 2 years (Lish
et al., 1984b), with highest dose dropped; BMR = 10% ER	D-40
Table D-20. Model predictions and oral slope factor for hepatocellular carcinomas in male
F344 rats administered hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) in the diet
for 2 years (Levine et al., 1983)	D-44
Table D-21. Model predictions for combined hepatocellular adenoma and carcinoma in F344
rats exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24
months (Levine et al., 1983); BMR = 5% ER	D-44
Table D-22. Model predictions for combined hepatocellular adenoma and carcinoma in
female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by
diet for 24 months (Parker et al., 2006; Lish et al., 1984b), with all doses
included; BMR = 10% ER	D-48
Table D-23. Model predictions for combined hepatocellular adenoma and carcinoma in
female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by
diet for 24 months (Parker et al., 2006; Lish et al., 1984b), with highest dose
dropped and multistage and nonmultistage models fitted; BMR = 10% ER	D-52
Table D-24. Model predictions for combined alveolar/bronchiolar adenoma and carcinoma
in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
by diet for 24 months (Lish et al., 1984b), with all doses included and multistage
and nonmultistage models fitted; BMR = 10% ER	D-55
Table D-25. Model predictions for combined alveolar/bronchiolar adenoma and carcinoma
in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
by diet for 24 months (Lish et al., 1984b), with highest dose dropped and
multistage and nonmultistage models fitted; BMR = 10% ER	D-58
Table D-26. Tumor data used for dose-response modeling with historical control	D-61
Table D-27. Model predictions for combined alveolar/bronchiolar adenoma and carcinoma
in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
by diet for 24 months (Lish et al., 1984b), using historical control and with
highest dose dropped; BMR = 10% ER	D-62
Table D-28. Model predictions for combined hepatocellular adenoma and carcinoma in
female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by
diet for 24 months (Parker et al., 2006; Lish et al., 1984b), using historical
control and with highest dose dropped; BMR = 10% ER	D-65
Table D-29. Comparison of model predictions, for lung and liver tumors in female B6C3Fi
mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24
months (Parker et al., 2006; Lish et al., 1984b), using concurrent and historical
control incidence (with highest dose dropped); BMR = 10% ER	D-68
vi

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
FIGURES
Figure C-l. Physiologically based pharmacokinetic (PBPK) model structure for hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX) in rats and humans	C-12
Figure C-2. EPA rat physiologically based pharmacokinetic (PBPK) model predictions fitted
to observed hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) blood concentrations
in male and female Sprague-Dawley rats following i.v. exposure	C-18
Figure C-3. EPA rat physiologically based pharmacokinetic (PBPK) model predictions fitted
to observed hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) blood concentrations
following oral exposure to RDX dissolved in water	C-20
Figure C-4. EPA rat model predictions fitted to observed hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) blood concentrations following oral exposure to RDX in dry
capsules	C-20
Figure C-5. Effect of varying oral absorption parameters on EPA rat model predictions fitted
to observed hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) blood concentrations
following oral exposure to coarse-grain RDX	C-21
Figure C-6. EPA rat model predictions fitted to observed hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) brain tissue concentrations following oral exposure to RDX	C-21
Figure C-l. EPA rat model predictions fitted to observed hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) blood concentrations following oral exposure to fine-grain RDX in
a saline slurry	C-23
Figure C-8. Comparison of EPA rat model predictions with data from Schneider et al. (1978)
for the subchronic gavage study	C-24
Figure C-9. Comparison of EPA rat model predictions with data from Schneider et al. (1978)
for the subchronic drinking water study	C-24
Figure C-10. EPA human model predictions fitted to observed hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) blood concentrations resulting from an accidental ingestion of
RDX by a 14.5-kg boy (Woody et al., 1986)	C-26
Figure C-ll. EPA human model predictions fitted to observed hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) blood concentrations resulting from accidental exposure to adults
assumed to be 70 kg (Ozhan et al., 2003)	C-26
Figure C-12. Comparison of EPA mouse physiologically based pharmacokinetic (PBPK) model
predictions with data from oral exposure to hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) dissolved in water	C-29
Figure D-l. Plot of incidence rate by dose, with the fitted curve of the Multistage 3° model,
for convulsions in male and female F344 rats exposed to hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) by gavage for 90 days (Crouse et al., 2006)	D-6
Figure D-2. Plot of incidence rate by dose, with fitted curve for Quantal-Linear model, for
convulsions in female F344 rats exposed to hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) by gavage on gestational days (GDs) 6-19 (Cholakis et al., 1980);
BMR = 5% ER; dose shown in mg/kg-day	D-9
Figure D-3. Plot of incidence rate by dose, with fitted curve for Multistage 3° model, for
combined incidence of convulsion and mortality in male and female F344 rats
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by gavage for 90 days
(Crouse et al., 2006); BMR = 1% ER; dose shown in mg/kg-day	D-12
Figure D-4. Plot of incidence rate by dose, with fitted curve for Multistage 3° model, for
hemorrhagic/suppurative cystitis of the urinary bladder in male F344 rats
vii

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months
(Levine et al., 1983); BMR = 10% ER; dose shown in mg/kg-day	D-14
Figure D-5. Plot of incidence rate by dose, with fitted curve for the Log-Probit model, for
prostate suppurative inflammation in male F344 rats exposed to hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months (Levine et al., 1983);
BMR = 10% ER; dose shown in mg/kg-day	D-17
Figure D-6. Plot of incidence rate by dose, with the fitted curve of the Multistage 2° model,
for combined mortality in male and female F344 rats exposed to hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 13 weeks (Levine et al., 1981b);
BMR = 1% ER; dose shown in mg/kg-day	D-22
Figure D-7. Plot of incidence rate by dose, with fitted curve for Dichotomous-Hill model, for
mortality (number found dead) in rats exposed to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) in the diet for 13 weeks (von Oettingen et al., 1949); BMR =
1% ER; dose shown in mg/kg-day	D-24
Figure D-8. Plot of incidence rate by dose, with the fitted curve of the Multistage 2° model,
for mortality in male and female F344 rats exposed to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) by gavage for 90 days (Crouse et al., 2006); BMR = 1% ER;
dose shown in mg/kg-day	D-26
Figure D-9. Plot of incidence rate by dose, with the fitted curve of the Multistage 3° model,
for mortality in female Sprague-Dawley rats exposed to hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) by gavage on Gestation Days 6-15 (Angerhofer et
al., 1986); BMR = 1% ER; dose shown in mg/kg-day	D-28
Figure D-10. Plot of incidence rate by dose, with fitted curve for Multistage-Cancer 1° model,
for combined alveolar/bronchiolar adenoma and carcinoma in female B6C3Fi
mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24
months (Lish et al., 1984b), with highest dose dropped; BMR = 10% ER; dose
shown in mg/kg-day	D-32
Figure D-ll. Plot of incidence rate by dose, with fitted curve for Multistage-Cancer 1° model,
for combined hepatocellular adenoma and carcinoma in female B6C3Fi mice
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months
(Parker et al., 2006; Lish et al., 1984b), with highest dose dropped; BMR = 10%
ER; dose shown in mg/kg-day	D-34
Figure D-12. Plot of incidence rate by dose, with fitted curve for Multistage-Cancer 1° model,
for combined hepatocellular adenoma and carcinoma in female B6C3F1 mice
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months
(Parker et al., 2006; Lish et al., 1984b), with highest dose dropped and sample
sizes from Lish et al. (1984b); BMR = 10% ER; dose shown in mg/kg-day	D-37
Figure D-13. Plot of incidence rate by dose, with fitted curve for Multistage-Cancer 3° model,
for alveolar/bronchiolar carcinoma in male B6C3Fi mice exposed to hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months (Lish et al., 1984b), with
highest dose dropped; BMR = 10% ER; dose shown in mg/kg-day	D-41
Figure D-14. Plot of incidence rate by dose, with fitted curve for Multistage 1° model, for
combined hepatocellular adenoma and carcinoma in F344 rats exposed to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months (Levine et al.,
1983); BMR = 5% ER; dose shown in mg/kg-day	D-45
Figure D-15. Plot of incidence rate by dose, with fitted curve for Multistage 1° model, for
combined hepatocellular adenoma and carcinoma in female B6C3Fi mice
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months
Vlll

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
(Parker et al., 2006; Lish et al., 1984b), with all doses included; BMR = 10% ER;
dose shown in mg/kg-day	D-49
Figure D-16. Plot of incidence rate by dose, with fitted curve for Log-Probit model, for
combined hepatocellular adenoma and carcinoma in female B6C3Fi mice
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months
(Parker et al., 2006; Lish et al., 1984b), with all doses included; BMR = 10% ER;
dose shown in mg/kg-day	D-50
Figure D-17. Plot of incidence rate by dose, with fitted curve for Multistage 1° model, for
combined hepatocellular adenoma and carcinoma in female B6C3Fi mice
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months
(Parker et al., 2006; Lish et al., 1984b), with highest dose dropped; BMR = 10%
ER; dose shown in mg/kg-day	D-53
Figure D-18. Plot of incidence rate by dose, with fitted curve for Multistage 1° model, for
combined alveolar/bronchiolar adenoma and carcinoma in female B6C3Fi mice
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months
(Lish et al., 1984b), with all doses included; BMR = 10% ER; dose shown in
mg/kg-day	D-56
Figure D-19. Plot of incidence rate by dose, with fitted curve for Multistage 1° model, for
combined alveolar/bronchiolar adenoma and carcinoma in female B6C3Fi mice
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months
(Lish et al., 1984b), with highest dose dropped; BMR = 10% ER; dose shown in
mg/kg-day	D-59
Figure D-20. Plot of incidence rate by dose, with fitted curve for Multistage-Cancer 1° model,
for combined alveolar/bronchiolar adenoma and carcinoma in female B6C3Fi
mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24
months (Lish et al., 1984b), using historical control and with highest dose
dropped; BMR = 10% ER; dose shown in mg/kg-day	D-63
Figure D-21. Plot of incidence rate by dose, with fitted curve for Multistage-Cancer 1° model,
for combined hepatocellular adenoma and carcinoma in female B6C3Fi mice
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months
(Parker et al., 2006; Lish et al., 1984b), using historical control and with highest
dose dropped; BMR = 10% ER; dose shown in mg/kg-day	D-65
IX

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
ABBREVIATIONS
AAP Army ammunition plant	FUDS
ACGIH American Conference of Governmental	GABA
Industrial Hygienists	GD
AChE acetylcholinesterase	GI
ADAF age-dependent adjustment factor	HED
AIC Akaike's information criterion	HERO
ALP alkaline phosphatase
ALT alanine aminotransferase	HI
AOP adverse outcome pathway	HMX
AST aspartate aminotransferase
atm atmosphere	i.p.
ATSDR Agency for Toxic Substances and	i.v.
Disease Registry	IRIS
AUC area under the curve	IUR
AUCiotai area under the curve for blood	KAD
concentration versus time from the
time of doxing to the time RDX is	KAS
completely eliminated
BDNF brain-derived neurotrophic factor	Kel
BMD benchmark dose	KfC
BMDL benchmark dose lower confidence limit	KQB
BMDS Benchmark Dose Software	KQC
BMDU benchmark dose upper bound	KQF
BMR benchmark response	KQL
BUN blood urea nitrogen	KQR
BW body weight
BW0-33 body weight scaling to the 0.33 power	KQS
BW2/3 body weight scaling to the 2/3 power
BW3/4 body weight scaling to the % power	KVB
BWa animal body weight	KVF
BWh human body weight	KVL
CAAC Chemical Assessment Advisory	KVR
Committee
CASRN Chemical Abstracts Service registry	KVS
number
CI	confidence interval	KW
Cmax peak concentration
CNS central nervous system	LDoi
CSF cerebrospinal fluid
CYP450 cytochrome P450	LDH
DAF dosimetric adjustment factor	LDLoi
d.f.	degrees of freedom	LOAEL
DMSO dimethylsulfoxide	LOD
DNA deoxyribonucleic acid	miRNA
DNX l-nitro-3,5-dinitroso-	MNX
1,3,5-triazacyclohexane
DTIC Defense Technical Information Center	MOA
EEG electroencephalogram	MRL
EPA Environmental Protection Agency	NA
ER	extra risk	NCE
FOB functional observational battery
Formerly Used Defense Sites
gamma-amino butyric acid
gestational day
gastrointestinal
human equivalent dose
Health and Environmental Research
Online
hazard index
octahydro-l,3,5,7-tetranitro-
1,3,5,7-tetrazocine
intraperitoneal
intravenous
Integrated Risk Information System
inhalation unit risk
rate constant for oral absorption,
compartment 2
rate constant for oral absorption,
compartment 1
terminal elimination rate constant
metabolic rate constant
fractional blood flow to brain
cardiac output
fractional blood flow to fat
fractional blood flow to liver
fractional blood flow to richly perfused
tissue
fractional blood flow to slowly perfused
tissue
fractional tissue volume of brain
fractional tissue volume of fat
fractional tissue volume of liver
fractional tissue volume of richly
perfused tissue
fractional tissue volume of slowly
perfused tissue
fractional tissue volume of blood
volume
the dose expected to be lethal to 1% of
the animals
lactate dehydrogenase
lower confidence limit on the LDoi
lowest-observed-adverse-effect level
limit of detection
microRNA
hexahydro-l-nitroso-3,5-dinitro-
1,3,5-triazine
mode of action
minimal risk level
not available
normochromatic erythrocyte
x

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
NCEA
ND
NOAEL
NPL
NTP
NZW
OR
ORD
OSF
OSHA
PB
PBPK
PCE
PECO
PEL
PF
PL
PND
POD
PR
PS
PWG
RBC
RDX
RfC
RfD
RNA
SAB
SD
SDMS
SDWA
SE
SGOT
SGPT
SLE
SS
m
TLV
TNT
TNX
TS
National Center for Environmental
Assessment
not determined
no-observed-adverse-effect level
National Priorities List
National Toxicology Program
New Zealand White
odds ratio
Office of Research and Development
oral slope factor
Occupational Safety and Health
Administration
tissue:blood partition coefficient for
brain
physiologically based pharmacokinetic
polychromatic erythrocyte
populations, exposures, comparitors,
and outcomes
permissible exposure limit
tissue:blood partition coefficient for fat
tissue:blood partition coefficient for
liver
postnatal day
point of departure
tissue:blood partition coefficient for
richly perfused tissue
tissue:blood partition coefficient for
slowly perfused tissue
Pathology Working Group
red blood cell
Royal Demolition eXplosive
(hexahydro-l,3,5-trinitro-
1,3,5-triazine)
inhalation reference concentration
oral reference dose
ribonucleic acid
Science Advisory Board
standard deviation
spontaneous death or moribund
sacrifice
Safe Drinking Water Act
standard error
glutamic oxaloacetic transaminase, also
known as AST
glutamic pyruvic transaminase, also
known as ALT
systemic lupus erythematosus
scheduled sacrifice
half life
threshold limit value
trinitrotoluene
hexahydro-l,3,5-trinitroso-
1,3,5-triazine
terminal sacrifice
TSCATS	Toxic Substances Control Act Test
Submissions
TWA	time-weighted average
U.S.	United States of America
UCM	Unregulated Contaminant Monitoring
UF	uncertainty factor
UFa	animal-to-human uncertainty factor
UFd	database deficiencies uncertainty factor
UFh	human variation uncertainty factor
UFl	LOAEL-to-NOAEL uncertain factor
UFs	subchronic-to-chronic uncertainty
factor
WBC	white blood cell
xi

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
APPENDIX A. ASSESSMENTS BY OTHER NATIONAL
AND INTERNATIONAL HEALTH AGENCIES
Table A-l. Assessments by other national and international health agencies
Organization
Toxicity value
Agency for Toxic Substances and
Disease Registry (ATSDR, 2012)
Acute oral MRL—0.2 mg/kg-d
Basis: Tremors and convulsions in rats (Crouse et al., 2006); application
of a composite uncertainty factor (UF) of 30 (3 for extrapolation from
animals to humans with dosimetric adjustments [PBPK modeling] and 10
for human variability)
Intermediate oral MRL—0.1 mg/kg-d
Basis: Convulsions in rats (Crouse et al.. 2006); application of a
composite UF of 30 (3 for extrapolation from animals to humans with
dosimetric adjustments [PBPK modeling] and 10 for human variability)
Chronic oral MRL—0.1 mg/kg-d
Basis: Tremors and convulsions in rats (Levine et al., 1983) application of
a UF of 30 (3 for extrapolation from animals to humans with dosimetric
adjustments [PBPK modeling] and 10 for human variability)
National Institute for Occupational
Safety and Health (NIOSH, 2012)
Recommended exposure limit—1.5 mg/m3 TWA for up to a 10-hr workday
during a 40-hr workweek; short-term (15-min) limit—3 mg/m3
Basis: Agreed with OSHA-proposed PEL in 1988 PEL hearings
Skin designation indicates potential for dermal absorption
Basis: Agreed with OSHA's proposal for skin notation in 1988 PEL
hearings
Occupational Safety and Health
Administration (OSHA, 2012)
PEL—1.5 mg/m3 TWA for an 8-hr workday in a 40-hr workweek
Basis: Adopted from the ACGIH TLV established in 1969
Skin designation indicates that cutaneous exposure may contribute to
overall exposure and measures should be taken to prevent skin
absorption
Basis: Adopted from ACGIH
Hazardous Chemical Information
System (Safe Work Australia, 2018)
Exposure standard—1.5 mg/m3 TWA for an 8-hr workday
Basis: Adopted from the ACGIH TLV established in 1991
Skin absorption notice indicates that absorption through the skin may be a
significant source of exposure
Basis: Adopted from ACGIH
ACGIH = American Conference of Governmental Industrial Hygienists; MRL = minimal risk level;
OSHA = Occupational Safety and Health Administration; PBPK = physiologically based pharmacokinetic;
PEL = permissible exposure limit; TLV = threshold limit value; TWA = time-weighted average; UF = uncertainty
factor.
A-l

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
APPENDIX B. ADDITIONAL DETAILS OF LITERATURE
SEARCH STRATEGY | STUDY SELECTION AND
EVALUATION
The literature search for hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) was conducted in
five online scientific databases through May 2016. The detailed search strategy used to search four
of these databases—PubMed, Toxline, Toxic Substances Control Act Test Submissions (TSCATS), and
Toxcenter—is provided in Table B-l. Toxcenter, a fee-based scientific database, was searched
outside of Health and Environmental Research Online (HERO).1 Toxcenter searches initially yield
titles only; obtaining complete citations and abstracts incurs additional costs. Thus, titles only were
initially screened; for titles identified as potentially relevant, complete citations with abstracts,
when available, were downloaded and rescreened. Of the rescreened citations, only those selected
for full text review were added to HERO and the RDX project page. The search strategy used to
search the Defense Technical Information Center (DTIC) database is described in Table B-2. The
computerized database searches were augmented by review of online regulatory sources, as well as
"forward" and "backward" Web of Science searches of two recent reviews (see Table B-3). Forward
searching was used to identify articles that cited the selected studies (i.e., the two reviews identified
in Table B-3), and backward searching was used to identify articles that the selected studies cited.
A post-peer-review literature search update was conducted in PubMed, Toxline, and
TSCATS for the period May 2016 to November 2017 and in DTIC for the period 2016 to February
2018 using a search strategy consistent with previous literature searches (see Tables B-l and B-2).
B.l. DEFENSE TECHNICAL INFORMATION CENTER (DTIC) LITERATURE
SEARCH AND SCREEN
Among the RDX-related citations thatwere identified in the January 2015 search ofthe
DTIC database, 826 (722 after duplicate removal within DTIC) were classified with the distribution
"approved for public release," 239 (217 after duplicate removal) were classified as "distribution
limited to U.S. government agencies and their contractors," and 199 (181 after duplicate removal)
were classified as "distribution limited to U.S. government agencies only." A preliminary screen of
the 1,120 unique citations was performed; 85 citations with unlimited distribution and 10 citations
'HERO is a database of scientific studies and other references used to develop EPA's assessments aimed at
understanding the health and environmental effects of pollutants and chemicals. It is developed and
managed in EPA's Office of Research and Development by the National Center for Environmental Assessment.
The database includes more than 1.6 million scientific references, including articles from the peer-reviewed
literature. New studies are added continuously to HERO.
B-l

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
with limited distribution were selected for further review as potential sources of health effects data
or supporting information. The remaining 1,025 unlimited- and limited-distribution DTIC
references not selected for further consideration either were not studies of RDX or did not contain
information pertinent to the assessment of the health effects of RDX (e.g., documents were related
to environmental properties such as leaching, explosive properties, fuel and propellant properties,
weapons systems, treatment of wastewater containing explosives, and disposal technologies). An
update of the DTIC search was performed in May 2016. The update search identified 21 items
classified as "approved for public release," 9 classified as "distribution limited to U.S. government
agencies and their contractors," and 9 classified as "distribution limited to U.S. government agencies
only"; none of these were selected for further review, as none met the inclusion criteria outlined in
Table LS-1 of the main document (i.e., none contained health effects data or supporting
information). A second update of the DTIC search was performed in February 2018, after peer
review of the RDX assessment. Included records were those with creation dates of January 2016 to
February 2018. The update search identified 46 items classified as "approved for public release," 9
classified as "distribution limited to U.S. government agencies and their contractors," and 13
classified as "distribution limited to U.S. government agencies only"; none of these were selected for
further review because none met the inclusion criteria outlined in Table LS-1 of the main document
or they were duplicates of references already cited in the RDX assessment.
The 85 unique selected citations with unlimited distribution from the January 2015 search
were uploaded to the HERO website fhttp://hero.epa.gov!. The 10 citations with limited
distribution were subject to a more in-depth screen to determine whether the references provided
additional primary health effects data and whether the U.S. Environmental Protection Agency (EPA)
should seek authorization for public distribution and upload to HERO. A review of the abstract or
full text of the documents associated with the limited-distribution citations resulted in the following
determinations:
•	One citation was excluded because it did not provide additional primary health effects data.
The citation reported data from a study that was subsequently published (Hathaway and
Buck. 1977) and had already been identified by the literature search strategy.
•	One citation (dated 1944) provided human and animal inhalation data and was considered
pertinent, but was not brought forward for further review because flaws in the design of
both the human and animal studies were such that results would not be considered
credible. Experimental animal study design issues included lack of a control group, small
numbers of animals, incomplete information on dosage or exposure levels, and inadequate
reporting. The human study described a case series and lacked a referent group and
measures of RDX exposure.
•	Eight citations were regulatory documents, reviews, or risk assessments that did not
specifically identify RDX and did not appear to contain primary health effects data.
B-2

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Based on these determinations, none of the 10 limited distribution citations that were
subject to further review were selected for further consideration or added to HERO.
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
Database
Terms
Hits
PubMed
Date: 4/2012
((((121-82-4) OR (Cyclonite[tw] OR Cyclotrimethylenetrinitramine[tw] OR
"cyclotrimethylene trinitramine"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazine"[tw] OR "Hexahydro-l,3,5-trinitro-s-triazine"[tw] OR
Hexogen[tw] OR "l,3,5-trinitro-l,3,5-triazine"[tw] OR "1,3,5-Triaza-
l,3,5-trinitrocyclohexane"[tw] OR "l,3,5-Trinitro-l,3,5-triazacyclohexane"[tw]
OR "l,3,5-Trinitrohexahydro-l,3,5-triazine"[tw] OR "1,3,5-Trinitrohexahydro-
s-triazine"[tw] OR "l,3,5-Trinitroperhydro-l,3,5-triazine"[tw] OR "Esaidro-
l,3,5-trinitro-l,3,5-triazina"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazin"[tw] OR "Perhydro-l,3,5-trinitro-l,3,5-triazine"[tw] OR
Cyclotrimethylenenitramine[tw] OR Trimethylenetrinitramine[tw] OR
"Trimethylene trinitramine"[tw] OR Trimethyleentrinitramine[tw] OR
"Trinitrocyclotrimethylene triamine"[tw] OR Trinitrotrimethylenetriamine[tw]
OR "CX 84A"[tw] OR Cyklonit[tw] OR Geksogen[tw] OR Heksogen[tw] OR
Hexogeen[tw] OR Hexolite[tw] OR "KHP 281"[tw] OR "PBX (af) 108"[tw] OR
"PBXW 108(E)"[tw] OR "Pbx(AF) 108"[tw]) OR (rdx[tw])) NOT medline[sb]) OR
(((121-82-4) OR (Cyclonite[tw] OR Cyclotrimethylenetrinitramine[tw] OR
"cyclotrimethylene trinitramine"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazine"[tw] OR "Hexahydro-l,3,5-trinitro-s-triazine"[tw] OR
Hexogen[tw] OR "l,3,5-trinitro-l,3,5-triazine"[tw] OR "1,3,5-Triaza-
l,3,5-trinitrocyclohexane"[tw] OR "l,3,5-Trinitro-l,3,5-triazacyclohexane"[tw]
OR "l,3,5-Trinitrohexahydro-l,3,5-triazine"[tw] OR "1,3,5-Trinitrohexahydro-
s-triazine"[tw] OR "l,3,5-Trinitroperhydro-l,3,5-triazine"[tw] OR "Esaidro-
l,3,5-trinitro-l,3,5-triazina"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazin"[tw] OR "Perhydro-l,3,5-trinitro-l,3,5-triazine"[tw] OR
Cyclotrimethylenenitramine[tw] OR Trimethylenetrinitramine[tw] OR
"Trimethylene trinitramine"[tw] ORTrimethyleentrinitramine[tw] OR
"Trinitrocyclotrimethylene triamine"[tw] ORTrinitrotrimethylenetriamine[tw]
OR "CX 84A"[tw] OR Cyklonit[tw] OR Geksogen[tw] OR Heksogen[tw] OR
Hexogeen[tw] OR Hexolite[tw] OR "KHP 281"[tw] OR "PBX (af) 108"[tw] OR
"PBXW 108(E)"[tw] OR "Pbx(AF) 108"[tw]) OR (rdx[tw]» AND (to[sh] OR po[sh]
OR ae[sh] OR pk[sh] OR (me[sh] AND (humans[mh] OR animals[mh])) OR ci[sh]
OR bl[sh] OR cf[sh] OR ur[sh] OR ((pharmacokinetics[mh] OR metabolism[mh])
AND (humans[mh] OR mammals[mh])) OR "dose-response relationship,
drug"[mh] OR risk[mh] OR "toxicity tests"[mh] OR noxae[mh] OR cancer[sb]
OR "endocrine system"[mh] OR "endocrine disruptors"[mh] OR "Hormones,
Hormone Substitutes, and Hormone Antagonists"[mh] OR triazines/ai OR
("Inhalation Exposure"[Mesh] OR "Maternal Exposure"[Mesh] OR "Maximum
Allowable Concentration"[Mesh] OR "Occupational Exposure"[Mesh] OR
"Paternal Exposure"[Mesh] OR "Environmental Exposure"[Mesh:noexp]))))
NOT (((((121-82-4) OR (Cyclonite[tw] OR Cyclotrimethylenetrinitramine[tw] OR
337
B-3

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
PubMed
Date: 4/2012
(continued)
"cyclotrimethylene trinitramine"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazine"[tw] OR "Hexahydro-l,3,5-trinitro-s-triazine"[tw] OR
Hexogen[tw] OR "l,3,5-trinitro-l,3,5-triazine"[tw] OR "1,3,5-Triaza-
l,3,5-trinitrocyclohexane"[tw] OR "l,3,5-Trinitro-l,3,5-triazacyclohexane"[tw]
OR "l,3,5-Trinitrohexahydro-l,3,5-triazine"[tw] OR "1,3,5-Trinitrohexahydro-
s-triazine"[tw] OR "l,3,5-Trinitroperhydro-l,3,5-triazine"[tw] OR "Esaidro-
l,3,5-trinitro-l,3,5-triazina"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazin"[tw] OR "Perhydro-l,3,5-trinitro-l,3,5-triazine"[tw] OR
Cyclotrimethylenenitramine[tw] OR Trimethylenetrinitramine[tw] OR
"Trimethylene trinitramine"[tw] OR Trimethyleentrinitramine[tw] OR
"Trinitrocyclotrimethylene triamine"[tw] OR Trinitrotrimethylenetriamine[tw]
OR "CX 84A"[tw] OR Cyklonit[tw] OR Geksogen[tw] OR Heksogen[tw] OR
Hexogeen[tw] OR Hexolite[tw] OR "KHP 281"[tw] OR "PBX (af) 108"[tw] OR
"PBXW 108(E)"[tw] OR "Pbx(AF) 108"[tw]) OR (rdx[tw])) NOT medline[sb]) OR
(((121-82-4) OR (Cyclonite[tw] OR Cyclotrimethylenetrinitramine[tw] OR
"cyclotrimethylene trinitramine"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazine"[tw] OR "Hexahydro-l,3,5-trinitro-s-triazine"[tw] OR
Hexogen[tw] OR "l,3,5-trinitro-l,3,5-triazine"[tw] OR "1,3,5-Triaza-
l,3,5-trinitrocyclohexane"[tw] OR "l,3,5-Trinitro-l,3,5-triazacyclohexane"[tw]
OR "l,3,5-Trinitrohexahydro-l,3,5-triazine"[tw] OR "1,3,5-Trinitrohexahydro-
s-triazine"[tw] OR "l,3,5-Trinitroperhydro-l,3,5-triazine"[tw] OR "Esaidro-
l,3,5-trinitro-l,3,5-triazina"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazin"[tw] OR "Perhydro-l,3,5-trinitro-l,3,5-triazine"[tw] OR
Cyclotrimethylenenitramine[tw] OR Trimethylenetrinitramine[tw] OR
"Trimethylene trinitramine"[tw] ORTrimethyleentrinitramine[tw] OR
"Trinitrocyclotrimethylene triamine"[tw] ORTrinitrotrimethylenetriamine[tw]
OR "CX 84A"[tw] OR Cyklonit[tw] OR Geksogen[tw] OR Heksogen[tw] OR
Hexogeen[tw] OR Hexolite[tw] OR "KHP 281"[tw] OR "PBX (af) 108"[tw] OR
"PBXW 108(E)"[tw] OR "Pbx(AF) 108"[tw]) OR (rdx[tw]» AND (to[sh] OR po[sh]
OR ae[sh] OR pk[sh] OR (me[sh] AND (humans[mh] OR animals[mh])) OR ci[sh]
OR bl[sh] OR cf[sh] OR ur[sh] OR ((pharmacokinetics[mh] OR metabolism[mh])
AND (humans[mh] OR mammals[mh])) OR "dose-response relationship,
drug"[mh] OR risk[mh] OR "toxicity tests"[mh] OR noxae[mh] OR cancer[sb]
OR "endocrine system"[mh] OR "endocrine disruptors"[mh] OR "Hormones,
Hormone Substitutes, and Hormone Antagonists"[mh] OR triazines/ai OR
("Inhalation Exposure"[Mesh] OR "Maternal Exposure"[Mesh] OR "Maximum
Allowable Concentration"[Mesh] OR "Occupational Exposure"[Mesh] OR
"Paternal Exposure"[Mesh] OR "Environmental Exposure"[Mesh:noexp]))))
AND (invertebrates OR aquatic organisms OR fish OR fishes OR amphibians OR
earthworm*))

B-4

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
PubMed
Date limit:
1/2012-
2/2013
(Cyclonite[tw] OR RDX[tw] OR Cyclotrimethylenetrinitramine[tw] OR
"cyclotrimethylene trinitramine"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazine"[tw] OR "Hexahydro-l,3,5-trinitro-s-triazine"[tw] OR
Hexogen[tw] OR "l,3,5-trinitro-l,3,5-triazine"[tw] OR "1,3,5-Triaza-
l,3,5-trinitrocyclohexane"[tw] OR "l,3,5-Trinitro-l,3,5-triazacyclohexane"[tw]
OR "l,3,5-Trinitrohexahydro-l,3,5-triazine"[tw] OR "1,3,5-Trinitrohexahydro-
s-triazine"[tw] OR "l,3,5-Trinitroperhydro-l,3,5-triazine"[tw] OR "Esaidro-
l,3,5-trinitro-l,3,5-triazina"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazin"[tw] OR "Perhydro-l,3,5-trinitro-l,3,5-triazine"[tw] OR
Cyclotrimethylenenitramine[tw] OR Trimethylenetrinitramine[tw] OR
"Trimethylene trinitramine"[tw] OR Trimethyleentrinitramine[tw] OR
"Trinitrocyclotrimethylene triamine"[tw] OR Trinitrotrimethylenetriamine[tw]
OR "CX 84A"[tw] OR Cyklonit[tw] OR Geksogen[tw] OR Heksogen[tw] OR
Hexogeen[tw] OR Hexolite[tw] OR "KHP 281"[tw] OR "PBX (af) 108"[tw] OR
"PBXW 108(E)"[tw] OR "Pbx(AF) 108"[tw]) AND (("2012/01/01"[Date - MeSH] :
"3000"[Date - MeSH]) OR ("2012/01/01"[Date - Entrez] : "3000"[Date - Entrez])
OR ("2012/01/01"[Date - Create] : "3000"[Date - Create]))
112
PubMed
Date limit:
11/2012-
1/2014
(Cyclonite[tw] OR RDX[tw] OR Cyclotrimethylenetrinitramine[tw] OR
"cyclotrimethylene trinitramine"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazine"[tw] OR "Hexahydro-l,3,5-trinitro-s-triazine"[tw] OR
Hexogen[tw] OR "l,3,5-trinitro-l,3,5-triazine"[tw] OR "1,3,5-Triaza-
l,3,5-trinitrocyclohexane"[tw] OR "l,3,5-Trinitro-l,3,5-triazacyclohexane"[tw]
OR "l,3,5-Trinitrohexahydro-l,3,5-triazine"[tw] OR "1,3,5-Trinitrohexahydro-
s-triazine"[tw] OR "l,3,5-Trinitroperhydro-l,3,5-triazine"[tw] OR "Esaidro-
l,3,5-trinitro-l,3,5-triazina"[tw] OR "Hexahydro-l,3,5-trinitro-
l,3,5-triazin"[tw] OR "Perhydro-l,3,5-trinitro-l,3,5-triazine"[tw] OR
Cyclotrimethylenenitramine[tw] OR Trimethylenetrinitramine[tw] OR
"Trimethylene trinitramine"[tw] OR Trimethyleentrinitramine[tw] OR
"Trinitrocyclotrimethylene triamine"[tw] OR Trinitrotrimethylenetriamine[tw]
OR "CX 84A"[tw] OR Cyklonit[tw] OR Geksogen[tw] OR Heksogen[tw] OR
Hexogeen[tw] OR Hexolite[tw] OR "KHP 281"[tw] OR "PBX (af) 108"[tw] OR
"PBXW 108(E)"[tw] OR "Pbx(AF) 108"[tw]) AND (("2012/ll/01"[Date - MeSH] :
"3000"[Date - MeSH]) OR ("2012/ll/01"[Date - Entrez] : "3000"[Date - Entrez])
OR ("2012/11/01"[Date - Create] : "3000"[Date - Create]))
138
B-5

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
PubMed
Date limit:
11/2013-
1/2015
("cyclonite"[nm] OR Cyclonite[tw] OR RDX[tw] OR
Cyclotrimethylenetrinitramine[tw] OR "cyclotrimethylene trinitramine"[tw] OR
"Hexahydro-l,3,5-trinitro-l,3,5-triazine"[tw] OR "Hexahydro-l,3,5-trinitro-
s-triazine"[tw] OR Hexogen[tw] OR "l,3,5-trinitro-l,3,5-triazine"[tw] OR
"l,3,5-Triaza-l,3,5-trinitrocyclohexane"[tw] OR "1,3,5-Trinitro-
l,3,5-triazacyclohexane"[tw] OR "l,3,5-Trinitrohexahydro-l,3,5-triazine"[tw]
OR "l,3,5-Trinitrohexahydro-s-triazine"[tw] OR "1,3,5-Trinitroperhydro-
l,3,5-triazine"[tw] OR "Esaidro-l,3,5-trinitro-l,3,5-triazina"[tw] OR
"Hexahydro-l,3,5-trinitro-l,3,5-triazin"[tw] OR "Perhydro-l,3,5-trinitro-
l,3,5-triazine"[tw] OR Cyclotrimethylenenitramine[tw] OR
Trimethylenetrinitramine[tw] OR "Trimethylene trinitramine"[tw] OR
Trimethyleentrinitramine[tw] OR "Trinitrocyclotrimethylene triamine"[tw] OR
Trinitrotrimethylenetriamine[tw] OR "CX 84A"[tw] OR Cyklonit[tw] OR
Geksogen[tw] OR Heksogen[tw] OR Hexogeen[tw] OR Hexolite[tw] OR "KHP
281"[tw] OR "PBX (af) 108"[tw] OR "PBXW 108(E)"[tw] OR "Pbx(AF) 108"[tw])
AND (2013/11/01: 3000[crdat] OR 2013/11/01: 3000[edat])
76
PubMed
Date limit:
11/2014-
5/2016
("cyclonite"[nm] OR Cyclonite[tw] OR RDX[tw] OR
Cyclotrimethylenetrinitramine[tw] OR "cyclotrimethylene trinitramine"[tw] OR
"Hexahydro-l,3,5-trinitro-l,3,5-triazine"[tw] OR "Hexahydro-l,3,5-trinitro-s-
triazine"[tw] OR Hexogen[tw] OR "l,3,5-trinitro-l,3,5-triazine"[tw] OR "1,3,5-
Triaza-l,3,5-trinitrocyclohexane"[tw] OR "l,3,5-Trinitro-l,3,5-
triazacyclohexane"[tw] OR "l,3,5-Trinitrohexahydro-l,3,5-triazine"[tw] OR
"l,3,5-Trinitrohexahydro-s-triazine"[tw] OR "l,3,5-Trinitroperhydro-l,3,5-
triazine"[tw] OR "Esaidro-l,3,5-trinitro-l,3,5-triazina"[tw] OR "Hexahydro-
l,3,5-trinitro-l,3,5-triazin"[tw] OR "Perhydro-l,3,5-trinitro-l,3,5-triazine"[tw]
OR Cyclotrimethylenenitramine[tw] OR Trimethylenetrinitramine[tw] OR
"Trimethylene trinitramine"[tw] OR Trimethyleentrinitramine[tw] OR
"Trinitrocyclotrimethylene triamine"[tw] OR Trinitrotrimethylenetriamine[tw]
OR "CX 84A"[tw] OR Cyklonit[tw] OR Geksogen[tw] OR Heksogen[tw] OR
Hexogeen[tw] OR Hexolite[tw] OR "KHP 281"[tw] OR "PBX (af) 108"[tw] OR
"PBXW 108(E)"[tw] OR "Pbx(AF) 108"[tw]) AND (2014/11/01: 3000[crdat] OR
2014/11/01: 3000[edat])
118
B-6

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
PubMed
Date limit:
5/2016-
11/2017
(post-peer
review)
("cyclonite"[nm] OR Cyclonite[tw] OR RDX[tw] OR
Cyclotrimethylenetrinitramine[tw] OR "cyclotrimethylene trinitramine"[tw] OR
"Hexahydro-l,3,5-trinitro-l,3,5-triazine"[tw] OR "Hexahydro-l,3,5-trinitro-s-
triazine"[tw] OR Hexogen[tw] OR "l,3,5-trinitro-l,3,5-triazine"[tw] OR "1,3,5-
Triaza-l,3,5-trinitrocyclohexane"[tw] OR "l,3,5-Trinitro-l,3,5-
triazacyclohexane"[tw] OR "l,3,5-Trinitrohexahydro-l,3,5-triazine"[tw] OR
"l,3,5-Trinitrohexahydro-s-triazine"[tw] OR "l,3,5-Trinitroperhydro-l,3,5-
triazine"[tw] OR "Esaidro-l,3,5-trinitro-l,3,5-triazina"[tw] OR "Hexahydro-
l,3,5-trinitro-l,3,5-triazin"[tw] OR "Perhydro-l,3,5-trinitro-l,3,5-triazine"[tw]
OR Cyclotrimethylenenitramine[tw] OR Trimethylenetrinitramine[tw] OR
"Trimethylene trinitramine"[tw] OR Trimethyleentrinitramine[tw] OR
"Trinitrocyclotrimethylene triamine"[tw] OR Trinitrotrimethylenetriamine[tw]
OR "CX 84A"[tw] OR Cyklonit[tw] OR Geksogen[tw] OR Heksogen[tw] OR
Hexogeen[tw] OR Hexolite[tw] OR "KHP 281"[tw] OR "PBX (af) 108"[tw] OR
"PBXW 108(E)"[tw] OR "Pbx(AF) 108"[tw]) AND (2016/05/01: 3000)
229
Toxline
Date: 4/2012
Notes: Searched CASRN or synonyms; removed invertebrates, aquatic
organisms, amphibians, earthworms.
507
Toxline
Date limit:
2011-2/2013
@OR+("Cyclonite"+"RDX"+"Cyclotrimethylenetrinitramine"+"cyclotrimethylen
e trinitramine"+"Hexahydro-l,3,5-trinitro-l,3,5-triazine"+"Hexahydro-
l,3,5-trinitro-s-triazine"+"Hexogen"+"l,3,5-trinitro-
l,3,5triazine"+"l,3,5-Triaza-l,3,5-trinitrocyclohexane"+"l,3,5-Trinitro-
l,3,5-triazacyclohexane"+"l,3,5-Trinitrohexahydro-l,3,5-triazine"+
"l,3,5-Trinitrohexahydro-s-triazine"+@term+@rn+121-82-4)+
@AND+@range+yr+2011+2013+@NOT+@org+pubmed+pubdart+crisp+tscats
5

@OR+("l,3,5-Trinitroperhydro-l,3,5-triazine"+"Esaidro-l,3,5-trinitro-
l,3,5-triazina"+"Hexahydro-l,3,5-trinitro-l,3,5-triazin"+"Perhydro-
l,3,5-trinitro-l,3,5-triazine"+"Cyclotrimethylenenitramine"+
"Trimethylenetrinitramine"+"Trimethylene+trinitramine"+
"Trimethyleentrinitramine"+"Trinitrocyclotrimethylene+triamine"+
"Trinitrotrimethylenetriamine"+"CX+84A"+"Cyklonit"+"Geksogen"+
"Heksogen"+"Hexogeen"+"Hexolite"+"KHP+281")+@AND+@range+yr+2011+
2013+@NOT+@org+pubmed+pubdart+crisp+tscats
0
B-7

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
Toxline
Date limit:
2012-1/2014
@OR+("Cyclonite"+"RDX"+"Cyclotrimethylenetrinitramine"+"cyclotrimethylen
e trinitramine"+"Hexahydro-l,3,5-trinitro-l,3,5-triazine"+"Hexahydro-
l,3,5-trinitro-s-triazine"+"Hexogen"+"l,3,5-trinitro-l,3,5-triazine"+
"l,3,5-Triaza-l,3,5-trinitrocyclohexane"+"l,3,5-Trinitro-
l,3,5-triazacyclohexane"+"l,3,5-Trinitrohexahydro-l,3,5-triazine"+
"l,3,5-Trinitrohexahydro-s-triazine"+@term+@rn+121-82-4)+ @AND+
@range+yr+2012+2014+@NOT+@org+pubmed+pubdart+crisp+tscats
10

@OR+("l,3,5-Trinitroperhydro-l,3,5-triazine"+"Esaidro-l,3,5-trinitro-
l,3,5-triazina"+"Hexahydro-l,3,5-trinitro-l,3,5-triazin"+"Perhydro-
l,3,5-trinitro-l,3,5-triazine"+"Cyclotrimethylenenitramine"+
"Trimethylenetrinitramine"+"Trimethylene+trinitramine"+
"Trimethyleentrinitramine"+"Trinitrocyclotrimethylene+triamine"+
"Trinitrotrimethylenetriamine"+"CX+84A"+"Cyklonit"+"Geksogen"+
"Heksogen"+"Hexogeen"+"Hexolite"+"KHP+281")+@AND+@range+yr+2012+
2014+@NOT+@org+pubmed+pubdart+crisp+tscats
0
Toxline
Date limit:
2013-1/2015
@SYNO+@OR+(RDX+"cyclotrimethylene+trinitramine"+"l,3,5-trinitro-
l,3,5-triazine"+"l,3,5-Triaza-l,3,5-trinitrocyclohexane"+"l,3,5-Trinitro-
l,3,5-triazacyclohexane"+"l,3,5-Trinitrohexahydro-l,3,5-triazine"+
"l,3,5-Trinitrohexahydro-s-triazine"+"l,3,5-Trinitroperhydro-l,3,5-triazine"+
"CX+84A")+@AND+@range+yr+2012+2015+@NOT+@org+pubmed+pubdart+"
nih+reporter"+tscats+crisp
19

@SYNO+@OR+("Cyclonite"+"Cyclotrimethylenenitramine"+"Cyclotrimethylene
trinitramine"+"Hexahydro-l,3,5-trinitro-l,3,5-triazine"+"Hexahydro-
l,3,5-trinitro-s-triazine"+"Hexogen"+"Hexolite"+"KHP+281"+"PBX+(af)+108"+
"PBXW+108(E)"+"Pbx(AF)+108"+"Perhydro-l,3,5-trinitro-l,3,5-triazine")+
@AND+@range+yr+2012+2015+@NOT+@org+pubmed+pubdart+
"nih+reporter"+tscats+crisp
9

@SYNO+@OR+("Research+Development+Explosive"+"Royal+Demolition+eXpl
osive+"Trimethylenetrinitramine"+"Trinitrocyclotrimethylene+triamine"+"Trini
trotrimethylenetriamine"+"sym-Trimethylene+trinitramine"+@term+
@rn+121-82-4+@term+@rn+204655-61-8+@term+@rn+50579-23-
2+@term+@rn+53800-53-6+@term+@rn+57608-45-4+@term+@rn+82030-
42-0)+ @AND+@range+yr+2012+2015+@NOT+@org+pubmed+pubdart+
"nih+reporter"+tscats+crisp
0
B-8

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
Toxline
Date limit:
2014-5/2016
@SYNO+@OR+(RDX+"cyclotrimethylene+trinitramine"+"l,3,5-trinitro-l,3,5-
triazine"+"l,3,5-Triaza-l,3,5-trinitrocyclohexane"+"l,3,5-Trinitro-l,3,5-
triazacyclohexane"+"l,3,5-Trinitrohexahydro-l,3,5-triazine"+"l,3,5-
Trinitrohexahydro-s-triazine"+"l,3,5-Trinitroperhydro-l,3,5-
triazine"+"CX+84A")+@AND+@range+yr+2014+2016+@NOT+@org+pubmed+
pubdart+"nih+reporter"+tscats+crisp
1

@SYNO+@OR+("Cyclonite"+"Cyclotrimethylenenitramine"+"Cyclotrimethylene
trinitramine"+"Hexahydro-l,3,5-trinitro-l,3,5-triazine"+"Hexahydro-l,3,5-
trinitro-s-
triazine"+"Hexogen"+"Hexolite"+"KHP+281"+"PBX+(af)+108"+"PBXW+108(E)"
+"Pbx(AF)+108"+"Perhydro-l,3,5-trinitro-l,3,5-
triazine")+@AND+@range+yr+2014+2016+@NOT+@org+pubmed+pubdart+"n
ih+reporter"+tscats+crisp
0

@SYNO+@OR+("Research+Development+Explosive"+"Royal+Demolition+eXpl
osive+"Trimethylenetrinitramine"+"Trinitrocyclotrimethylene+triamine"+"Trini
trotrimethylenetriamine"+"sym-
Trimethylene+trinitramine"+@term+@rn+121-82-4+@term+@rn+204655-61-
8+@term+@rn+50579-23-2+@term+@rn+53800-53-6+@term+@rn+57608-
45-4+@term+@ rn+82030-42-
0)+@AND+@range+yr+2014+2016+@NOT+@org+pubmed+pubdart+"nih+rep
orter"+tscats+crisp
0
Toxline
Date limit:
5/2016-
11/2017
(post-peer
review)
@BOOL+("Esaidro-l,3,5-trinitro-l,3,5-triazina" or "Hexahydro-l,3,5-trinitro-
1,3,5-triazin" or "Perhydro-l,3,5-trinitro-l,3,5-triazine" or
"Cyclotrimethylenenitramine" or "Trimethylenetrinitramine" or
"Trimethylene+trinitramine" or "Trimethyleentrinitramine" or
"Trinitrocyclotrimethylene+triamine" or "Trinitrotrimethylenetriamine" or
"CX+84A" or "Cyklonit" or "Geksogen"or "Heksogen" or "Hexogeen" or
"Hexolite" or "KHP+281" or "PBX+af+108" or
"PBXW+108+E")+@RANGE+yr+2016+2017+@NOT+@org+pubmed+pubdart+cr
isp+tscats+nih
@BOOL+("cyclonite" or "Cyclonite" or "RDX" or
"Cyclotrimethylenetrinitramine" or "cyclotrimethylene trinitramine" or
"Hexahydro-l,3,5-trinitro-l,3,5-triazine" or "Hexahydro-l,3,5-trinitro-s-
triazine" or "Hexogenor"l,3,5-trinitro-l,3,5-triazine" or "1,3,5-Triaza-1,3,5-
trinitrocyclohexane" or "l,3,5-Trinitro-l,3,5-triazacyclohexane" or "1,3,5-
Trinitrohexahydro-l,3,5-triazine" or "1,3,5-Trinitrohexahydro-s-triazine" or
"l,3,5-Trinitroperhydro-l,3,5-
triazine")+@RANGE+yr+2016+2017+@NOT+@org+pubmed+pubdart+crisp+tsc
ats+nih
0
TSCATS
Date: 2/2013
@term+@rn+121-82-4+@AND+@org+tscats
4
B-9

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
TSCATS 2
Date: 5/2016
121-82-4 from EPA receipt date 01/01/2000
0
TSCATS
8e/FYI
Date: 5/2016
("121-82-4" OR "1,3,5-Triazine, hexahydro-l,3,5-trinitro-") tsca (8e OR FYI)
0
TSCATS
Date:
11/2017
(post-peer
review)
Not searched. TSCATS is no longer being updated.

B-10

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
Toxcenter
Date: 4/2012
((121-82-4 OR Cyclonite OR RDX OR Cyclotrimethylenetrinitramine OR
"cyclotrimethylene trinitramine" OR "Hexahydro-l,3,5-trinitro-l,3,5-triazine"
OR "Hexahydro-l,3,5-trinitro-s-triazine" OR Hexogen OR "1,3,5-trinitro-
1,3,5-triazine" OR "l,3,5-Triaza-l,3,5-trinitrocyclohexane" OR "1,3,5-Trinitro-
1,3,5-triazacyclohexane" OR "l,3,5-Trinitrohexahydro-l,3,5-triazine" OR
"1,3,5-Trinitrohexahydro-s-triazine" OR "l,3,5-Trinitroperhydro-l,3,5-triazine"
OR "Esaidro-l,3,5-trinitro-l,3,5-triazina" OR "Hexahydro-l,3,5-trinitro-
1,3,5-triazin" OR "Perhydro-l,3,5-trinitro-l,3,5-triazine" OR
Cyclotrimethylenenitramine ORTrimethylenetrinitramine OR "Trimethylene
trinitramine" ORTrimethyleentrinitramine OR "Trinitrocyclotrimethylene
triamine" OR Trinitrotrimethylenetriamine OR "CX 84A" OR Cyklonit OR
Geksogen OR Heksogen OR Hexogeen OR Hexolite OR "KHP 281" OR "PBX (af)
108" OR "PBXW 108(E)" OR "Pbx(AF) 108") NOT (patent/dt OR tscats/fs))AND
((chronic OR immunotox? OR neurotox? OR toxicokin? OR biomarker? OR
neurolog? OR pharmacokin? OR subchronic OR pbpk OR epidemiology/st,ct,it)
OR acute OR subacute OR Id50# OR Ic50# OR (toxicity OR adverse OR
poisoning)/st,ct,it OR inhal? OR pulmon? OR nasal? OR lung? OR respir? OR
occupation? OR workplace? OR worker? OR oral OR orally OR ingest? OR
gavage? OR diet OR diets OR dietary OR drinking(w)water OR (maximum and
concentration? and (allowable OR permissible)) OR (abort? OR abnormalit? OR
embryo? OR cleft? OR fetus? OR foetus? OR fetal? OR foetal? OR fertil? OR
malform? OR ovum OR ova OR ovary OR placenta? OR pregnan? OR prenatal
OR perinatal? OR postnatal? OR reproduc? OR steril? OR teratogen? OR sperm
OR spermac? OR spermag? OR spermati? OR spermas? OR spermatob? OR
spermatoc? OR spermatog? OR spermatoi? OR spermatol? OR spermator? OR
spermatox? OR spermatoz? OR spermatu? OR spermi? OR spermo? OR
neonat? OR newborn OR development OR developmental? OR zygote? OR
child OR children OR adolescen? OR infant OR wean? OR offspring OR
age(w)factor? OR dermal? OR dermis OR skin OR epiderm? OR cutaneous? OR
carcinog? OR cocarcinog? OR cancer? OR precancer? OR neoplas? OR tumor?
OR tumour? OR oncogen? OR lymphoma? OR carcinom? OR genetox? OR
genotox? OR mutagen? OR genetic(w)toxic? OR nephrotox? OR hepatotox? OR
endocrin? OR estrogen? OR androgen? OR hormon? OR rat OR rats OR mouse
OR mice OR muridae OR dog OR dogs OR rabbit? OR hamster? OR pig OR pigs
OR swine OR porcine OR goat OR goats OR sheep OR monkey? OR macaque?
OR marmoset? OR primate? OR mammal? OR ferret? OR gerbil? OR rodent?
OR lagomorpha OR baboon? OR bovine OR canine OR cat OR cats OR feline OR
pigeon? OR occupation? OR worker? OR epidem?) AND ((biosis/fs AND
py>1999) OR caplus/fs))
Notes: Duplicates were removed; Biosis subfile results were date limited to
avoid extensive overlap with Toxline.
337 titles
screened
(20 selected for
full records and
added to HERO)
B-ll

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
Toxcenter
Date limit:
1/2012-
2/2013
(((121-82-4 OR Cyclonite OR RDX OR Cyclotrimethylenetrinitramine OR
"cyclotrimethylene trinitramine" OR "Hexahydro-l,3,5-trinitro-l,3,5-triazine"
OR "Hexahydro-l,3,5-trinitro-s-triazine" OR Hexogen OR "1,3,5-trinitro-
1,3,5-triazine" OR "l,3,5-Triaza-l,3,5-trinitrocyclohexane" OR "1,3,5-Trinitro-
1,3,5-triazacyclohexane" OR "l,3,5-Trinitrohexahydro-l,3,5-triazine" OR
"1,3,5-Trinitrohexahydro-s-triazine" OR "l,3,5-Trinitroperhydro-l,3,5-triazine"
OR "Esaidro-l,3,5-trinitro-l,3,5-triazina" OR "Hexahydro-l,3,5-trinitro-
1,3,5-triazin" OR "Perhydro-l,3,5-trinitro-l,3,5-triazine" OR
Cyclotrimethylenenitramine ORTrimethylenetrinitramine OR "Trimethylene
trinitramine" ORTrimethyleentrinitramine OR "Trinitrocyclotrimethylene
triamine" OR Trinitrotrimethylenetriamine OR "CX 84A" OR Cyklonit OR
Geksogen OR Heksogen OR Hexogeen OR Hexolite OR "KHP 281" OR "PBX (af)
108" OR "PBXW 108(E)" OR "Pbx(AF) 108") NOT (patent/dt OR tscats/fs)) AND
(py>2012 OR ed>20120101)) AND (chronic OR immunotox? OR neurotox? OR
toxicokin? OR biomarker? OR neurolog? OR pharmacokin? OR subchronic OR
pbpk OR epidemiology/st,ct, it) OR acute OR subacute OR Id50# OR Ic50# OR
(toxicity OR adverse OR poisoning)/st,ct,it OR inhal? OR pulmon? OR nasal? OR
lung? OR respir? OR occupation? OR workplace? OR worker? OR oral OR orally
OR ingest? OR gavage? OR diet OR diets OR dietary OR drinking(w)water OR
(maximum and concentration? and (allowable OR permissible)) OR (abort? OR
abnormalit? OR embryo? OR cleft? OR fetus? OR foetus? OR fetal? OR foetal?
OR fertil? OR malform? OR ovum OR ova OR ovary OR placenta? OR pregnan?
OR prenatal OR perinatal? OR postnatal? OR reproduc? OR steril? OR
teratogen? OR sperm OR spermac? OR spermag? OR spermati? OR spermas?
OR spermatob? OR spermatoc? OR spermatog? OR spermatoi? OR spermatol?
OR spermator? OR spermatox? OR spermatoz? OR spermatu? OR spermi? OR
spermo? OR neonat? OR newborn OR development OR developmental? OR
zygote? OR child OR children OR adolescen? OR infant OR wean? OR offspring
OR age(w)factor? OR dermal? OR dermis OR skin OR epiderm? OR cutaneous?
OR carcinog? OR cocarcinog? OR cancer? OR precancer? OR neoplas? OR
tumor? OR tumour? OR oncogen? OR lymphoma? OR carcinom? OR genetox?
OR genotox? OR mutagen? OR genetic(w)toxic? OR nephrotox? OR hepatotox?
OR endocrin? OR estrogen? OR androgen? OR hormon? OR rat OR rats OR
mouse OR mice OR muridae OR dog OR dogs OR rabbit? OR hamster? OR pig
OR pigs OR swine OR porcine OR goat OR goats OR sheep OR monkey? OR
macaque? OR marmoset? OR primate? OR mammal? OR ferret? OR gerbil? OR
rodent? OR lagomorpha OR baboon? OR bovine OR canine OR cat OR cats OR
feline OR pigeon? OR occupation? OR worker? OR epidem?) AND (biosis/fs OR
caplus/fs))
Notes: Duplicates were removed.
26 titles screened
(6 selected for
full records and
added to HERO)
B-12

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
Toxcenter
Date limit:
11/2012-
1/2014
(((121-82-4 OR Cyclonite OR RDX OR Cyclotrimethylenetrinitramine OR
"cyclotrimethylene trinitramine" OR "Hexahydro-l,3,5-trinitro-l,3,5-triazine"
OR "Hexahydro-l,3,5-trinitro-s-triazine" OR Hexogen OR "1,3,5-trinitro-
1,3,5-triazine" OR "l,3,5-Triaza-l,3,5-trinitrocyclohexane" OR "1,3,5-Trinitro-
1,3,5-triazacyclohexane" OR "l,3,5-Trinitrohexahydro-l,3,5-triazine" OR
"1,3,5-Trinitrohexahydro-s-triazine" OR "l,3,5-Trinitroperhydro-l,3,5-triazine"
OR "Esaidro-l,3,5-trinitro-l,3,5-triazina" OR "Hexahydro-l,3,5-trinitro-
1,3,5-triazin" OR "Perhydro-l,3,5-trinitro-l,3,5-triazine" OR
Cyclotrimethylenenitramine ORTrimethylenetrinitramine OR "Trimethylene
trinitramine" ORTrimethyleentrinitramine OR "Trinitrocyclotrimethylene
triamine" OR Trinitrotrimethylenetriamine OR "CX 84A" OR Cyklonit OR
Geksogen OR Heksogen OR Hexogeen OR Hexolite OR "KHP 281" OR "PBX (af)
108" OR "PBXW 108(E)" OR "Pbx(AF) 108") NOT (patent/dt OR tscats/fs)) AND
(py>2012 OR ed>20121101)) AND (chronic OR immunotox? OR neurotox? OR
toxicokin? OR biomarker? OR neurolog? OR pharmacokin? OR subchronic OR
pbpk OR epidemiology/st,ct, it) OR acute OR subacute OR Id50# OR Ic50# OR
(toxicity OR adverse OR poisoning)/st,ct,it OR inhal? OR pulmon? OR nasal? OR
lung? OR respir? OR occupation? OR workplace? OR worker? OR oral OR orally
OR ingest? OR gavage? OR diet OR diets OR dietary OR drinking(w)water OR
(maximum and concentration? and (allowable OR permissible)) OR (abort? OR
abnormalit? OR embryo? OR cleft? OR fetus? OR foetus? OR fetal? OR foetal?
OR fertil? OR malform? OR ovum OR ova OR ovary OR placenta? OR pregnan?
OR prenatal OR perinatal? OR postnatal? OR reproduc? OR steril? OR
teratogen? OR sperm OR spermac? OR spermag? OR spermati? OR spermas?
OR spermatob? OR spermatoc? OR spermatog? OR spermatoi? OR spermatol?
OR spermator? OR spermatox? OR spermatoz? OR spermatu? OR spermi? OR
spermo? OR neonat? OR newborn OR development OR developmental? OR
zygote? OR child OR children OR adolescen? OR infant OR wean? OR offspring
OR age(w)factor? OR dermal? OR dermis OR skin OR epiderm? OR cutaneous?
OR carcinog? OR cocarcinog? OR cancer? OR precancer? OR neoplas? OR
tumor? OR tumour? OR oncogen? OR lymphoma? OR carcinom? OR genetox?
OR genotox? OR mutagen? OR genetic(w)toxic? OR nephrotox? OR hepatotox?
OR endocrin? OR estrogen? OR androgen? OR hormon? OR rat OR rats OR
mouse OR mice OR muridae OR dog OR dogs OR rabbit? OR hamster? OR pig
OR pigs OR swine OR porcine OR goat OR goats OR sheep OR monkey? OR
macaque? OR marmoset? OR primate? OR mammal? OR ferret? OR gerbil? OR
rodent? OR lagomorpha OR baboon? OR bovine OR canine OR cat OR cats OR
feline OR pigeon? OR occupation? OR worker? OR epidem?) AND (biosis/fs OR
caplus/fs))
Notes: Duplicates were removed.
20 titles screened
(0 selected for
full records; none
added to HERO)
B-13

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
Toxcenter
Date limit:
11/2013-
1/2015
(((121-82-4 OR Cyclonite OR RDX OR Cyclotrimethylenetrinitramine OR
"cyclotrimethylene trinitramine" OR "Hexahydro-l,3,5-trinitro-l,3,5-triazine"
OR "Hexahydro-l,3,5-trinitro-s-triazine" OR Hexogen OR "1,3,5-trinitro-
1,3,5-triazine" OR "l,3,5-Triaza-l,3,5-trinitrocyclohexane" OR "1,3,5-Trinitro-
1,3,5-triazacyclohexane" OR "l,3,5-Trinitrohexahydro-l,3,5-triazine" OR
"1,3,5-Trinitrohexahydro-s-triazine" OR "l,3,5-Trinitroperhydro-l,3,5-triazine"
OR "Esaidro-l,3,5-trinitro-l,3,5-triazina" OR "Hexahydro-l,3,5-trinitro-
1,3,5-triazin" OR "Perhydro-l,3,5-trinitro-l,3,5-triazine" OR
Cyclotrimethylenenitramine ORTrimethylenetrinitramine OR "Trimethylene
trinitramine" ORTrimethyleentrinitramine OR "Trinitrocyclotrimethylene
triamine" OR Trinitrotrimethylenetriamine OR "CX 84A" OR Cyklonit OR
Geksogen OR Heksogen OR Hexogeen OR Hexolite OR "KHP 281" OR "PBX (af)
108" OR "PBXW 108(E)" OR "Pbx(AF) 108") NOT (patent/dt OR tscats/fs)) AND
(py >2013 OR ed>20131101)) AND (chronic OR immunotox? OR neurotox? OR
toxicokin? OR biomarker? OR neurolog? OR pharmacokin? OR subchronic OR
pbpk OR epidemiology/st,ct, it) OR acute OR subacute OR Id50# OR Ic50# OR
(toxicity OR adverse OR poisoning)/st,ct,it OR inhal? OR pulmon? OR nasal? OR
lung? OR respir? OR occupation? OR workplace? OR worker? OR oral OR orally
OR ingest? OR gavage? OR diet OR diets OR dietary OR drinking(w)water OR
(maximum and concentration? and (allowable OR permissible)) OR (abort? OR
abnormalit? OR embryo? OR cleft? OR fetus? OR foetus? OR fetal? OR foetal?
OR fertil? OR malform? OR ovum OR ova OR ovary OR placenta? OR pregnan?
OR prenatal OR perinatal? OR postnatal? OR reproduc? OR steril? OR
teratogen? OR sperm OR spermac? OR spermag? OR spermati? OR spermas?
OR spermatob? OR spermatoc? OR spermatog? OR spermatoi? OR spermatol?
OR spermator? OR spermatox? OR spermatoz? OR spermatu? OR spermi? OR
spermo? OR neonat? OR newborn OR development OR developmental? OR
zygote? OR child OR children OR adolescen? OR infant OR wean? OR offspring
OR age(w)factor? OR dermal? OR dermis OR skin OR epiderm? OR cutaneous?
OR carcinog? OR cocarcinog? OR cancer? OR precancer? OR neoplas? OR
tumor? OR tumour? OR oncogen? OR lymphoma? OR carcinom? OR genetox?
OR genotox? OR mutagen? OR genetic(w)toxic? OR nephrotox? OR hepatotox?
OR endocrin? OR estrogen? OR androgen? OR hormon? OR rat OR rats OR
mouse OR mice OR muridae OR dog OR dogs OR rabbit? OR hamster? OR pig
OR pigs OR swine OR porcine OR goat OR goats OR sheep OR monkey? OR
macaque? OR marmoset? OR primate? OR mammal? OR ferret? OR gerbil? OR
rodent? OR lagomorpha OR baboon? OR bovine OR canine OR cat OR cats OR
feline OR pigeon? OR occupation? OR worker? OR epidem?) AND (biosis/fs OR
caplus/fs))
Note: Duplicates were removed.
80 titles screened
(3 selected for
full records and
added to HERO)
B-14

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-l. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (PubMed, Toxline, Toxcenter, TSCATS)
(continued)
Database
Terms
Hits
Toxcenter
Date limit:
11/2014-
5/2016
(((121-82-4 OR Cyclonite OR RDX OR Cyclotrimethylenetrinitramine OR
"cyclotrimethylene trinitramine" OR "Hexahydro-l,3,5-trinitro-l,3,5-triazine"
OR "Hexahydro-l,3,5-trinitro-s-triazine" OR Hexogen OR "l,3,5-trinitro-l,3,5
triazine" OR "l,3,5-Triaza-l,3,5-trinitrocyclohexane" OR "l,3,5-Trinitro-l,3,5-
triazacyclohexane" OR "l,3,5-Trinitrohexahydro-l,3,5-triazine" OR "1,3,5-
Trinitrohexahydro-s-triazine" OR "l,3,5-Trinitroperhydro-l,3,5-triazine" OR
"Esaidro-l,3,5-trinitro-l,3,5-triazina" OR "Hexahydro-l,3,5-trinitro-l,3,5
triazin" OR "Perhydro-l,3,5-trinitro-l,3,5-triazine" OR
Cyclotrimethylenenitramine ORTrimethylenetrinitramine OR "Trimethylene
trinitramine" ORTrimethyleentrinitramine OR "Trinitrocyclotrimethylene
triamine" OR Trinitrotrimethylenetriamine OR "CX 84A" OR Cyklonit OR
Geksogen OR Heksogen OR Hexogeen OR Hexolite OR "KHP 281" OR "PBX (af)
108" OR "PBXW 108(E)" OR "Pbx(AF) 108") NOT (patent/dt OR tscats/fs)) AND
(py >2013 OR ed>20131101)) AND (chronic OR immunotox? OR neurotox? OR
toxicokin? OR biomarker? OR neurolog? OR pharmacokin? OR subchronic OR
pbpk OR epidemiology/st,ct, it) OR acute OR subacute OR Id50# OR Ic50# OR
(toxicity OR adverse OR poisoning)/st,ct,it OR inhal? OR pulmon? OR nasal? OR
lung? OR respir? OR occupation? OR workplace? OR worker? OR oral OR orally
OR ingest? OR gavage? OR diet OR diets OR dietary OR drinking(w)water OR
(maximum and concentration? and (allowable OR permissible)) OR (abort? OR
abnormalit? OR embryo? OR cleft? OR fetus? OR foetus? OR fetal? OR foetal?
OR fertil? OR malform? OR ovum OR ova OR ovary OR placenta? OR pregnan?
OR prenatal OR perinatal? OR postnatal? OR reproduc? OR steril? OR
teratogen? OR sperm OR spermac? OR spermag? OR spermati? OR spermas?
OR spermatob? OR spermatoc? OR spermatog? OR spermatoi? OR spermatol?
OR spermator? OR spermatox? OR spermatoz? OR spermatu? OR spermi? OR
spermo? OR neonat? OR newborn OR development OR developmental? OR
zygote? OR child OR children OR adolescen? OR infant OR wean? OR offspring
OR age(w)factor? OR dermal? OR dermis OR skin OR epiderm? OR cutaneous?
OR carcinog? OR cocarcinog? OR cancer? OR precancer? OR neoplas? OR
tumor? OR tumour? OR oncogen? OR lymphoma? OR carcinom? OR genetox?
OR genotox? OR mutagen? OR genetic(w)toxic? OR nephrotox? OR hepatotox?
OR endocrin? OR estrogen? OR androgen? OR hormon? OR rat OR rats OR
mouse OR mice OR muridae OR dog OR dogs OR rabbit? OR hamster? OR pig
OR pigs OR swine OR porcine OR goat OR goats OR sheep OR monkey? OR
macaque? OR marmoset? OR primate? OR mammal? OR ferret? OR gerbil? OR
rodent? OR lagomorpha OR baboon? OR bovine OR canine OR cat OR cats OR
feline OR pigeon? OR occupation? OR worker? OR epidem?) AND (biosis/fs OR
caplus/fs))
Note: Duplicates were removed.
33 titles screened
(0 selected for
full records and
added to HERO)
CASRN = Chemical Abstracts Registry Number; CI = confidence interval.
B-15

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-2. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (DTIC online database)
Date
accessed
Terms
Hits
1/2015
Synonyms in all fields search box
("121-82-4" OR "RDX" OR "Cyclotrimethylenetrinitramine" OR "Cyclonite"
OR "cyclotrimethylene trinitramine" OR "Hexogen" OR "Hexahydro-1,3,5-
trinitro-l,3,5-triazine" OR "Hexahydro-l,3,5-trinitro-s-triazine" OR
"Trimethylene trinitramine" OR "Trimethylenetrinitramine" OR "Hexolite"
OR "Trinitrotrimethylenetriamine")
Keywords in citation box
("toxicity" OR "toxicology" OR "poisoning" OR "cancer" OR "carcinogens"
OR "carcinogen" OR "neoplasms" OR "neoplasm" OR "oncogenesis" OR
"teratogenic compounds" OR "lethality" OR "death" OR "body weight" OR
"immunology" OR "genotoxicity" OR "mutagenicity" OR "mutagens" OR
"mutations" OR "oral" OR "gavage" OR "inhalation" OR "dermal" OR
"metabolism" OR "pharmacokinetics" OR "pharmacokinetic" OR "PBPK"
OR "pharmacology" OR "organs" OR "skin" OR "tissues" OR "body fluids"
OR "toxic agents" OR "rats" OR "mice" OR "mouse" OR "rat")
Limited to content type: Documents


Distribution: Approved for public release
826 (85 selected
and added to HERO)

Distribution: U.S. government and contractors
239 (0 selected and
added to HERO)

Distribution: U.S. government only
199 (0 selected and
added to HERO)
5/2016
Synonyms in all fields search box
("121-82-4" OR "RDX" OR "Cyclotrimethylenetrinitramine" OR "Cyclonite"
OR "cyclotrimethylene trinitramine" OR "Hexogen" OR "Hexahydro-1,3,5-
trinitro-l,3,5-triazine" OR "Hexahydro-l,3,5-trinitro-s-triazine" OR
"Trimethylene trinitramine" OR "Trimethylenetrinitramine" OR "Hexolite"
OR "Trinitrotrimethylenetriamine")
Keywords in citation box
("toxicity" OR "toxicology" OR "poisoning" OR "cancer" OR "carcinogens"
OR "carcinogen" OR "neoplasms" OR "neoplasm" OR "oncogenesis" OR
"teratogenic compounds" OR "lethality" OR "death" OR "body weight" OR
"immunology" OR "genotoxicity" OR "mutagenicity" OR "mutagens" OR
"mutations" OR "oral" OR "gavage" OR "inhalation" OR "dermal" OR
"metabolism" OR "pharmacokinetics" OR "pharmacokinetic" OR "PBPK"
OR "pharmacology" OR "organs" OR "skin" OR "tissues" OR "body fluids"
OR "toxic agents" OR "rats" OR "mice" OR "mouse" OR "rat")
Limited to content type: Documents


Distribution: Approved for public release
21 (0 selected and
added to HERO)

Distribution: U.S. government and contractors
9 (0 selected and
added to HERO)
B-16

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-2. Summary of detailed search strategies for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (DTIC) (continued)
Database
Terms
Hits
5/2016
(continued)
Distribution: U.S. government only
9 (0 selected and
added to HERO)
2/16/2018
(post-peer
review)
"cyclonite" OR Cyclonite OR RDX OR Cyclotrimethylenetrinitramine OR
"cyclotrimethylene trinitramine" OR "Hexahydro-l,3,5-trinitro-l,3,5-triazine"
OR "Hexahydro-l,3,5-trinitro-s-triazine" OR Hexogen OR "l,3,5-trinitro-l,3,5-
triazine OR "l,3,5-Triaza-l,3,5-trinitrocyclohexane" OR "l,3,5-Trinitro-l,3,5-
triazacyclohexane" OR "l,3,5-Trinitrohexahydro-l,3,5-triazine" OR "1,3,5-
Trinitrohexahydro-s-triazine" OR "l,3,5-Trinitroperhydro-l,3,5-triazine" OR
"Esaidro-l,3,5-trinitro-l,3,5-triazina" OR Hexahydro-l,3,5-trinitro-l,3,5-
triazin" OR "Perhydro-l,3,5-trinitro-l,3,5-triazine" OR
Cyclotrimethylenenitramine ORTrimethylenetrinitramine OR "Trimethylene
trinitramine" ORTrimethyleentrinitramine OR "Trinitrocyclotrimethylene
triamine" OR 'Trinitrotrimethylenetriamine" OR "CX 84A" OR Cyklonit OR
Geksogen OR Heksogen OR Hexogeen OR Hexolite OR "KHP 281" OR "PBX af
108" OR "PBXW 108E" OR "PbxAF 108"
Limited to content type: Documents
Includes records with creation dates of January 2016 to February 2018.


Distribution: Approved for public release
46 (0 selected and
added to HERO)

Distribution: U.S. government and contractors
9 (0 selected and
added to HERO)

Distribution: U.S. government only
13 (0 selected and
added to HERO)
B-17

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-3. Processes used to augment the search of core databases for
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
Selected key reference(s) or sources
Date
Additional
references
identified
"Forward" and "backward" Web of Science searches3
Sweeney et al. (2012a). Assessing the noncancer risk for RDX (hexahydro-1,3,5-
trinitro-l,3,5-triazine) using physiologically based pharmacokinetic (PBPK) modeling.
Regul Toxicol Pharmacol 62(1):107-114. (forwardsearch)
1 search result
3/2013
0 citations added
Sweenev et al. (2012b). Cancer mode of action, weight of evidence, and proposed
cancer reference value for hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX). Regul
Toxicol Pharmacol 64(2):205-224 (backwards search)
0 search results
3/2013
0 citations added
Sweenev et al. (2012a). Assessing the noncancer risk for RDX (hexahvdro-1,3,5-
trinitro-l,3,5-triazine) using physiologically based pharmacokinetic (PBPK) modeling.
Regul Toxicol Pharmacol 62(1):107-114.
(review of 35 references cited in this paper)
3/2013
0 citations added
Sweenev et al. (2012b). Cancer mode of action, weight of evidence, and proposed
cancer reference value for hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX). Regul
Toxicol Pharmacol 64(2):205-224
(review of 69 references cited in this paper)
3/2013
3 citations added
B-18

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table B-3. Processes used to augment the search of core databases for
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) (continued)
Selected key reference(s) or sources
Date
Additional
references
identified
Online regulatory sources
Combination of CASRN and synonyms searched on the following websites:
ATSDR (http://www.atsdr.cdc.gov/substances/index.asp)
(Note: the reference list for the ATSDR toxicological profile for RDX was compared to
the search results and relevant references were added)
California Environmental Protection Agency (Office of Environmental Health Hazard
Assessment) (http://www.oehha.ca.gov/risk.html)
eChemPortal (http://www.echemDortal.org/echemDortal/DarticiDant/Dage.action?
oagelD=9)
EPA Acute Exposure Guideline Levels
(httD://www. eDa.gov/oDDt/aegl/Dubs/chemlist. htm)
(httD://www.eDa.gov/ncea/iris/index.html) to find data
EPA National Service Center for Environmental Publications
(htto://www. eoa.gov/nceDihom/)
EPA Science Inventory (htto://cfoub.eoa.gov/si/)
Federal Docket (www.regulations.gov)
Health Canada First Priority List Assessments (htto://www.hc-sc.gc.ca/ewh-
semt/Dubs/contaminants/osll-lsDl/index-eng.ohD)Health Canada Second Priority List
Assessments (htto://www.hc-sc.gc.ca/ewh-semt/Dubs/contaminants/osl2-lsD2/index-
4/2012
1/2014
1/2015
5/2016
15 citations
added
1 citation added
0 citations added
0 citations added
eng.oho)
International Agency for Research on Cancer
(httoV/monograohs. iarc.fr/htdig/search. html)
International Programme on Chemical Safetv INCHEM (htto://www.inchem.org/)
National Academy of Science via the National Academies Press
(htto://www. nao.edu/)
National Cancer Institute (htto://www.cancer.gov)
National Center for Toxicological Research
(htto://www.fda.gov/AboutFDA/CentersOffices/OC/OfficeofScientificandMedicalProg


rams/NCTR/default.htm)
National Institute of Environmental Health Sciences (NIEHS)
(htto://www. niehs.nih.gov/)
National Institute for Occupational Safety and Health NIOSHTIC 2
(httD://www2a. cdc.gov/nioshtic-2/)
NTP—Report on Carcinogens, status, results, and management reports
(httD://ntDsearch. niehs.nih.gov/auerv. html)
World Health Organization assessments—Concise International Chemical Assessment
Documents, Environmental Health Criteria
(htto://www.who.int/iDcs/assessment/en/)


ATSDR = Agency for Toxic Substances and Disease Registry; CASRN = Chemical Abstracts Registry Number;
NTP = National Toxicology Program; PBPK = physiologically based pharmacokinetic.
aSweeney et al. (2012a) and Sweeney et al. (2012b) were selected for "forward" and "backward" searching in the
Web of Science as the two more recent reviews of the health effects of RDX toxicity in the published literature.
B-19

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
APPENDIX C. INFORMATION IN SUPPORT OF
HAZARD IDENTIFICATION AND DOSE-RESPONSE
ANALYSIS
C.l. TOXICOKINETICS
Hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) is absorbed following exposure by inhalation
and oral routes. The rate and extent of absorption are dependent upon the dosing preparation.
RDX is systemically distributed, can be transferred from mother to fetus, and can transfer in
maternal milk. Metabolism of RDX is extensive and includes denitration, ring cleavage, and
generation of C02 possibly through cytochrome P450 (CYP450). RDX metabolites are eliminated
primarily via urinary excretion and exhalation of C02.
C.l.l. Absorption
Absorption of RDX following oral exposure has been demonstrated in humans and
laboratory animals (rats, mice, swine, and voles) by measuring radiolabeled RDX and/or
metabolites in excreta (urine and expired air) and tissues (including blood). Quantitative estimates
of oral absorption (e.g., oral bioavailability or fractional absorption) are not available in humans.
Results of animal studies indicate that oral bioavailability ranges from approximately 50 to 90%
and may vary based on the physical form of RDX and the matrix (e.g., soil, plants) in which it is
administered. No studies investigating absorption of RDX following inhalation exposure have been
identified. Results of an intratracheal administration study in rats provide limited evidence of
absorption of RDX from the respiratory tract. The only data evaluating dermal absorption of RDX is
provided by in vitro studies showing that RDX can be absorbed through excised skin of humans and
animals.
Oral Absorption
Quantitative information on blood levels following accidental exposure to RDX is limited to
two studies of accidental oral exposures (Kuctikardali etal.. 2003: Woody etal.. 1986) and one
study of mixed dermal and inhalation exposure (Ozhan etal.. 2003). A number of qualitative case
studies of accidental exposures with similar toxic effects provide additional support that RDX is
absorbed into the body fHettand Fichtner. 2002: Harrell-Bruder and Hutchins. 1995: Goldberg et
al.. 1992: Ketel and Hughes. 1972: Hollander and Colbach. 1969: Stone etal.. 1969). The oral
absorption of RDX in humans was demonstrated in a case report of a 3-year-old male child who
ingested plasticized RDX material that adhered to his mother's work boots and clothing (Woody et
C-l

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
al.. 19861. RDX was measured in serum, urine, cerebrospinal fluid, and feces. Based on a kinetic
analysis of serum RDX concentrations, the dose was estimated to be 85 mg/kg and the first-order
absorption rate constants were estimated to be 0.34-2.20/hour (Woody et al.. 19861.2 Sweeney et
al. (2012a) estimated the absorption rate constant for this same subject to be 0.060/hour. The
large range in the calculated absorption rate constants resulted from uncertainty in the dose and
time to peak serum RDX levels, and the models that were used to simulate the RDX toxicokinetics.
Ozhan etal. (20031 summarized plasma RDX levels in five military personnel who were accidentally
exposed to toxic levels of RDX. Although Ozhan etal. f20031 reported that personnel were exposed
through dermal contact and inhalation, secondary oral exposure may have occurred. Based on
physiologically based pharmacokinetic (PBPK) model fits to the plasma RDX concentration data,
Sweeney etal. (2012a) estimated a first-order absorption rate constant of 0.033/hour. Kuctikardali
etal. (20031 summarized plasma RDX levels in another group of five military personnel who
ingested toxic levels of RDX (doses were not reported). RDX was detected in plasma of all patients
within 3 hours after ingestion.
Quantitative data to directly support estimates of oral bioavailability are available from
studies in rats and mice fGuo etal.. 1985: Schneider etal.. 1978.19771. Results of single and
repeated oral dose studies in adult Sprague-Dawley rats indicate that approximately 83-87% of the
administered dose is absorbed from the gastrointestinal (GI) tract Following gavage
administration of 50 mg/kg [14C]-RDX dissolved in dimethylsulfoxide (DMSO), approximately 90%
of the administered carbon-14 was recovered 4 days after dosing, with ~3% in feces, 34% in urine,
43% in expired air, and 10% in the carcass (Schneider etal.. 19771. It is unclear if the carcass
included the GI tract, which may have included unabsorbed RDX. Assuming that all of the
carbon-14 in feces represented unabsorbed RDX (rather than RDX that was absorbed and
subsequently secreted into the intestine), results of this study indicate that at least 87% of the
administered dose was absorbed from the GI tract. Similar results were observed following
repeated daily oral exposure of Sprague-Dawley rats to [14C]-RDX by gavage (in DMSO) or drinking
water for 1 week. Based on recovery of carbon-14 in urine and expired air and the carbon-14
retained in carcass, approximately 83% (drinking water) to 85% (gavage) of the administered dose
was absorbed fSchneider etal.. 19781.
An estimate of oral bioavailability in rats can also be obtained from data on blood RDX
concentrations reported in Krishnan et al. (2009). Male Sprague-Dawley rats received a single
intravenous (i.v.) (0.77 or 1.04 mg/kg) or oral (1.53 or 2.07 mg/kg, dissolved in water) dose of RDX.
Estimates of bioavailability were obtained based on the reported blood RDX concentrations,
terminal elimination rate constants (estimated for this review by fitting the serum RDX data with a
first-order exponential function, see Table C-5 in Section C.1.4, Excretion) and the blood area under
the curve (AUC) values (calculated for this review using the trapezoid rule extrapolated to infinite
2Woodv et al. (19861 reported the absorption rate constants in units of L/hour; however, this appears to have
been a typographical error for l/hour or hour-1.
C-2

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
time). Calculated dose-adjusted AUC values were 9.6 and 8.4 hours-kg/L for the i.v. studies and
4.7 and 6.0 hours-kg/L for the oral dosing studies. These AUC values correspond to estimated oral
bioavailability ranging from 50 to 70%. Recovery of the administered radiolabel was incomplete
(~90% of the administered carbon-14) in the studies (Schneider et al.. 1978.1977): therefore, it is
possible that oral bioavailability is actually higher than 83-87%. Guo etal. (1985) reported data on
blood tritium kinetics in mice that received i.v. (0.055 mg RDX or ~2.5 mg/kg body weight) or oral
(50 mg/kg) doses of [3H]-RDX. Based on the reported blood tritium concentrations (% of dose/g)
and terminal values for concentrations of tritium in blood (1.1 days for i.v. and 2.2 days for oral),
the corresponding AUCs of the blood concentration versus time curves were calculated (calculated
for this review using the trapezoid rule extrapolated to infinite time) to be 30 and
16 hour-% dose/g for i.v. and oral dosing, respectively. This corresponds to an oral bioavailability
of RDX-derived tritium concentration of approximately 50% (i.e., 16/30).
In Yucatan miniature pigs (minipigs) administered a single dose of [14C]-RDX (43-45 mg/kg
as a suspension in carboxymethylcellulose), approximately 0.8-6% of the administered carbon-14
was eliminated in feces 24 hours after dosing fMusick etal.. 2010: Maior etal.. 20071. Although
results of swine studies suggest that GI absorption of RDX was nearly complete, data cannot be used
to determine a quantitative estimate of oral bioavailability because it is unlikely that fecal excretion
ofunabsorbed RDX was complete 24 hours after dosing (Snoeck etal.. 2004).
Oral bioavailability of RDX appears to vary depending on the physical form of RDX and the
matrix (e.g., soil, vegetation) in which it is administered. Schneider et al. T19771 compared the oral
absorption of a single 100 mg/kg gavage dose of coarse granular [14C]-RDX as a slurry in isotonic
saline with a single 50 mg/kg gavage dose of a finely powdered [14C]-RDX solution in saline in
Sprague-Dawley rats. Plasma carbon-14 levels were measured for 24 hours after dosing. For both
[14C]-RDX preparations, peak plasma levels of carbon-14 were observed 24 hours after oral
administration, with a higher 24-hour plasma concentration for the 50 mg/kg dose (~4.7 ng/mL)
compared to the 100 mg/kg dose (3.12 ng/mL). Results of this study indicate that the oral
bioavailability of RDX may be greater for the finely powdered preparation than for the coarse
granular preparation, a result consistent with the greater surface area available for absorption with
the finely powdered RDX. However, blood levels were only measured 24 hours after dosing, and
the lower 2-hour carbon-14 plasma concentration for the coarse granular preparation could be due
to slower absorption of the coarse RDX granules compared with fine RDX powder, rather than
lower overall bioavailability.
Oral bioavailability of RDX is lower when administered as RDX-contaminated soil or when
RDX is in plant materials that were grown on RDX-contaminated soils. Crouse etal. f20081
investigated the oral bioavailability of RDX in contaminated soils relative to pure RDX by comparing
the AUC for the RDX blood concentration versus time curves. Adult male Sprague-Dawley rats were
administered oral doses (in gelatin capsules) of pure RDX (99.9% purity; neat) or an equivalent
amount of RDX in contaminated soils from the Louisiana Army Ammunition Plant (AAP) or Fort
C-3

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Meade. Blood concentrations for rats dosed with Louisiana AAP soil (1.24 mg/kg) and neat RDX
(1.24 mg/kg) peaked at approximately 6 hours. The AUC and 6-hour RDX blood concentration
were both approximately 25% lower for Louisiana AAP soil than for neat RDX (p < 0.003 for AUC),
suggesting that oral bioavailability of RDX from Louisiana AAP soil was 25% lower than neat RDX.
For Fort Meade soil (0.2 mg/kg), RDX blood concentrations peaked at 6 hours compared to 4 hours
for neat RDX (0.2 mg/kg). The 4-hour blood concentration for Fort Meade soil was approximately
15% lower than for neat RDX, although the AUC for Fort Meade soil was only 5% lower than for
neat RDX (not statistically significant). Collectively, these results suggest that RDX in soil is
absorbed following oral exposure and that it has a lower bioavailability than neat RDX.
Fellows etal. (2006) showed that plants (alfalfa shoots and corn leaves) incorporated
[14C]-RDX grown on [14C]-RDX-amended soils. [14C]-RDX and plant metabolites of [14C]-RDX were
absorbed by voles following oral administration (Fellows etal.. 20061. In adult male prairie voles
[Microtus ochrogaster) fed diets containing RDX incorporated in plants for 5 or 7 days (average RDX
dose per animal of 2.3 mg/kg-day), 94.8, and 96.6%, respectively, of the administered carbon-14
was eliminated in excreta (combined feces, urine, and CO2) and 3-5% was retained in the carcass.
Feces, urine, and CO2 contained 74-79,13-14, and 8-12% of the total carbon-14 in excreta,
respectively. Based on carbon-14 elimination in urine and CO2 plus that retained by the carcass, the
study authors estimated the oral bioavailability of plant-derived RDX to be >20%. However, if
biliary excretion of RDX and/or RDX metabolites is a major excretory pathway in voles (as is the
case with mice), estimates of bioavailability of plant-derived RDX could be substantially higher.
In Yorkshire piglets administered single doses of 5 or 10 mg/kg in gelatin capsules, peak
plasma concentrations were proportional to the administered dose fBannon. 20061. However, the
potential for dose-dependence has not been evaluated over a wide range of doses.
RDX appears in blood within 1 hour following oral dosing; however, the rate of absorption
may depend upon the physical form or dose of RDX (Bannon et al.. 2009: Crouse etal.. 2008:
Bannon. 2006: Guo etal.. 1985: MacPhail etal.. 1985: Schneider et al.. 19771. Oral absorption of
RDX was rapid in LACA mice following stomach perfusion with [3H]-RDX (50 mg/kg in methyl
cellulose) (Guo etal.. 19851. The tritium radiolabel was detected in blood 15 minutes following
dosing, and the highest concentrations in blood were observed 30 minutes after dosing. Based on
an analysis of the blood tritium concentration kinetics, the study authors estimated an absorption
rate constant of 8.7/hour. In Sprague-Dawley rats administered single doses (0.2-18.0 mg/kg) of
RDX in gelatin capsules, peak blood RDX concentrations were observed between 2.5 and 6 hours
(Bannon etal.. 2009: Krishnan etal.. 2009: Crouse etal.. 20081. Peak blood concentrations
occurred at 24 hours after Sprague-Dawley rats were administered a single oral dose (100 mg/kg)
of coarse granular RDX in saline (Schneider etal.. 19771. Similarly, peak RDX blood concentrations
in Yorkshire piglets administered single doses (5-10 mg/kg) of finely powdered (>98% pure) RDX
in gelatin capsules occurred at 3-8 hours after dosing (Bannon. 20061 compared to 24 hours after
a single dose (100 mg/kg) of RDX administered as a fine powder in saline (Schneider etal.. 19771.
C-4

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Peak plasma concentrations in minipigs administered a single dose of [14C]-RDX (45 mg/kg as a
suspension in carboxymethylcellulose) were reached within 6-12 hours after dosing fMusicketal..
20101. Krishnan et al. (20091 and Sweeney etal. (2012a) estimated absorption rates in rats dosed
with higher doses of coarse granular RDX to be approximately 100 times slower than absorption
rates in rats dosed with lower doses of finely powdered neat RDX preparations or neat RDX
dissolved in aqueous vehicles. For example, Krishnan et al. (20091 estimated the absorption rate
constant to be 0.75/hour for rats dosed with neat RDX dissolved in water (1.53 or 2.07 mg/kg) or
neat RDX in a gelatin capsule (0.2 or 1.24 mg/kg) fCrouse etal.. 20081. compared to 0.0075/hour
for rats dosed with coarse granular RDX (100 mg/kg) f Schneider etal.. 19771.
Inhalation Absorption
Studies evaluating absorption of RDX in humans following inhalation exposure were not
identified. Several case reports have documented seizures and other neurological effects in
individuals exposed to RDX either in a manufacturing setting or in the course of using RDX as a
cooking fuel fTestud etal.. 1996a: Testud etal.. 1996b: Ketel and Hughes. 1972: Hollander and
Colbach. 1969: Kaplan etal.. 1965: Barsotti and Crotti. 19491. These reports suggest that RDX may
be absorbed by the respiratory system. However, in several cases, the study authors were unable
to clearly identify the primary route of exposure. In some cases, incidental oral exposure was
suggested. Studies in laboratory animals have not investigated the absorption of RDX following
inhalation exposure.
Dermal Absorption
In vitro studies have demonstrated the dermal absorption of RDX in human and pig skin
(Reddv etal.. 2008: Reifenrath etal.. 20081. Reddv etal. (20081 reported that 5.7% of the applied
RDX dose (in acetone) was absorbed into excised human skin in 24 hours. Dermal absorption of
[14C]-RDX from both a low-carbon (1.9%) and a high-carbon (9.5%) soil was also assessed in this
system. Approximately 2.6% of the RDX applied in the low-carbon soil and 1.4% applied in the
high-carbon soil was absorbed after 24 hours. Thus, the dermal absorption of RDX from soils was
reduced when compared with absorption from acetone, and absorption was lower in the
high-carbon soil than in the low-carbon soil.
Reifenrath etal. (20081 investigated the influence of skin surface moisture conditions, soil
carbon content, and soil aging on the in vitro percutaneous penetration of [14C]-labeled RDX in
excised pig skin. Mean skin absorption of RDX was higher for low-carbon soil (1.2%), regardless of
soil age and skin surface moisture. Absorption and evaporation were <1% for RDX regardless of
soil type and age. While dermal absorption of certain munitions (e.g., 2,6-dinitrotoluene) is greatly
enhanced by hydration of the skin surface, hydration had minimal effect on RDX, mostly due to the
lack of RDX volatility (e.g., <0.5% evaporation).
C-5

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
C.1.2. Distribution
Information on the distribution of absorbed RDX in humans is limited to a few cases of
accidental exposures to RDX that provide data on the kinetics of RDX in blood and cerebrospinal
fluid (Kiiciikardali etal.. 2003: Ozhan etal.. 2003: Woody etal.. 19861. Concentrations of RDX in
serum and cerebrospinal fluid were similar (11 and 9 mg/L, respectively) in a child 24 hours after
ingesting an estimated dose of 85 mg/kg RDX (Woody etal.. 19861. More extensive information on
tissue distribution is available for animals, including mice, rats, and swine fPan etal.. 2013: Musick
etal.. 2010: Bannon. 2006: Reddvetal.. 1989: Guo etal.. 1985: MacPhail etal.. 1985: Schneider et
al.. 19771. In these studies, RDX or radiolabeled RDX ([14C] or [3H]) was administered by the oral,
intraperitoneal (i.p.), i.v., or intratracheal route and the distribution of the RDX or radiolabel was
measured. Because metabolism of RDX can result in loss of carbon-14 or tritium from the parent
compound, the distribution of radiolabel will not necessarily reflect the distribution of RDX
(Schneider etal.. 19771. To compare tissue distributions in studies in which animals received
different doses by different routes of administration, distribution data are expressed as ratios of
tissue RDX or radiolabel to that of either whole blood or plasma, whichever was reported. RDX in
blood distributes into red blood cells (RBCs) and plasma to achieve concentration ratios that are
close to unity. The plasma:whole blood carbon-14 ratio in swine that received a single oral dose of
[14C]-RDX (45 mg/kg) was approximately 1.3 (Musick et al.. 20101. and whole rat blood incubated
in vitro with RDX had a plasma:RBC RDX ratio of approximately 1.0 fKrishnan et al.. 20091. As a
result of the similarity between plasma and whole blood concentrations, tissue distribution is
approximately equivalent when expressed as ratios of blood or plasma.
Studies conducted in rats, mice, and swine indicate that absorbed RDX distributes to many
different tissues. Schneider et al. (1977) estimated the volume of distribution of RDX to be
approximately 2.18 L/kg in rats, based on plasma RDX kinetics in rats that received a single i.p.
dose of RDX (5-6 mg/kg). Consistent with this estimate are observations of tissue :blood (or
plasma) concentration ratios that exceed 1 in various tissues, including brain (showing that RDX
can cross the blood-brain barrier), heart, kidney, and liver (Musick etal.. 2010: Bannon etal.. 2006:
MacPhail etal.. 1985: Schneider etal.. 19771. Distribution within the brain may not be uniform.
However, Bannon etal. (20061 observed tissue:blood concentrations for RDX of approximately 4 in
brain hippocampus and 3.5 in brain cortex of swine that received a single oral dose of 10 mg/kg
[14C]-RDX, although this is the only study that reported distribution for brain regions. Reported
tissue:blood (or plasma) concentration ratios of RDX 24 hours following a single dose (oral or i.p.)
were 1-9 for kidney, 1-7 for liver, and 1-3 for heart (see Table C-l) fBannon. 2006: Schneider et
al.. 19771. With repeated oral dosing (e.g., 28-90 days), tissue:blood ratios of RDX for these tissues
were consistently greater than unity fPan etal.. 2013: Schneider etal.. 19781. There is no
consistent evidence that RDX accumulates in fat, although estimates of the fat:blood partition
coefficient range from 6 to 8 and exceed that of other tissues (Sweeney etal.. 2012a: Krishnan etal..
20091.
C-6

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Plasma protein binding may be a factor impacting the concentration of RDX available to
diffuse across the blood-brain barrier and to be metabolized. While empirical measurements of
protein binding are not available, some of the toxicokinetic studies can be informative. The
plasma:whole blood ratios measured in Musicketal. (20101 and Krishnan etal. (20091 are
approximately 1.3 and 1.0, which are not suggestive of a high affinity for protein binding. The
volume of distribution estimated in Schneider etal. (19771 is 2.18 L/kg, approximately twice body
size, which is again not suggestive of a high affinity for protein binding. In the absence of evidence
of protein binding, the total amount of RDX is considered potentially available for diffusion across
the blood-brain barrier and for metabolism.
Table C-l. Distribution of hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) or
radiolabel from administered RDXa
Animal
Route
Dose
(mg/kg)
Time
(hrs)
Brain
Heart
Kidney
Liver
Fat
Source
Swine
Oral
45b
24
0.6°
0.7
2.4
7.3
0.4
Musick et al. (2010)
Swine
Oral
10d
3
3.5-4.0d
2
<1
<1
NAe
Bannon et al. (2006)
Swine
Oral
100d
24
1.5°
1.1
1.2-1.9
0.9
1.8
Schneider et al. (1977)
Rat
Oral
100d
24
3.4°
2.9
6.6
0.7
NA
Schneider et al. (1977)
Rat
i.p.
50d
2
3.4°
2.6
8.8
5.7
NA
Schneider et al. (1977)
Rat
i.p.
500d
<6.5
2.5°
2.1
4.8
3.3
NA
Schneider et al. (1977)
Mouse
Oral
50f
24
lc
0.8
1
1.4
0.8
Guoetal. (1985)
Mouse
i.v.
2.5f
24
0.6g
0.8
0.7
1.6
0.4
Guoetal. (1985)
NA = not available.
aValues are tissue:blood or tissue:plasma ratios following a single dose of either RDX, [14C]-RDX, or [3H]-RDX.
bCarbon-14
cTissue:plasma
dRDX
eNot available
Tritium
gTissue:blood
In rats, RDX can cross the placental-blood barrier resulting in exposure to the fetus, and can
also be transported into maternal milk. Hess-Ruth et al. (20071 detected RDX in the brain tissue of
Postnatal Day (PND) 1 rat pups (concentrations ranged from 0.64 to 7.6 ng/g brain tissue, with no
differences between males and females) after maternal exposure to 6 mg/kg RDX via gavage from
Gestational Day (GD) 6 to PND 10. RDX was also detected in maternal milk (concentrations ranged
from 3 to 5.7 ng/mL on PND 1 and from 0.7 to 3.1 ng/mL on PND 10).
C-7

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
C.1.3. Metabolism
The metabolism of RDX is not well characterized. No studies investigating the metabolism
of RDX in humans were identified. Studies in animals indicate that RDX undergoes extensive
metabolism, including denitration, ring cleavage, and generation of CO2. Predominant metabolic
pathways and major organs involved in RDX metabolism have not been identified, although results
of in vitro studies suggest a role for CYP450.
RDX undergoes metabolism through processes that generate the nitrosamine RDX
metabolites hexahydro-l-nitroso-3,5-dinitro-l,3,5-triazine (MNX), hexahydro-l,3-dinitroso-
5-nitro-l,3,5-triazine (DNX), andhexahydro-l,3,5-trinitroso-l,3,5-triazine (TNX), and further
metabolism to CO2. In mice at the end of a 28-day exposure to RDX in feed (ad libitum), the
nitrosamine RDX metabolites (MNX, DNX, and TNX) were measured (Pan etal.. 2013). RDX and the
metabolites MNX, DNX, and TNX were detected in the brain and other tissues. In Sprague-Dawley
rats administered a single 50 mg/kg gavage dose of [14C]-RDX, 43% was recovered as exhaled
[14C02] after 4 days f Schneider etal.. 19771. Similarly, approximately 30-50% of the radioactivity
was recovered as exhaled [14CC>2] in rats administered [14C]-RDX in saturated drinking water or
daily gavage for up to 3 months (Schneider etal.. 1978). Metabolism of RDX to CO2 was also
observed in prairie voles following dietary exposure (average RDX dose per animal of
2.3 mg/kg-day) to [14C]-RDX incorporated plant materials for 5-7 days, with approximately 9% of
the administered [14C]-RDX dose eliminated as exhaled [14CC>2] fFellows etal.. 20061. Terminal
metabolites of RDX have been identified in the urine of rats and swine, with very little urinary
excretion of the parent compound, indicating extensive metabolism of RDX. Following oral
administration of a single 50 mg/kg gavage dose of [14C]-RDX, 3.6% of the urinary radioactivity was
identified as unmetabolized RDX (Schneider et al.. 1977). Total urinary radiolabel accounted for
about one-third of the administered label and unmetabolized RDX contributed 3-5% of total
urinary radioactivity in rats exposed to [14C]-RDX-saturated drinking water for 1 or 13 weeks
f Schneider etal.. 19781. Similar results were observed in Yucatan minipigs administered a single
45 mg/kg oral dose of [14C]-RDX, with approximately 1-3.5% of the urinary radioactivity as parent
RDX fMaior etal.. 20071. Urinary metabolites were not characterized in these studies fSchneider et
al.. 1978.1977). However, Schneider et al. (1978) cited unpublished findings in their laboratory
that, in addition to carbon dioxide, other one-carbon intermediates were produced, including
bicarbonate and formic acid.
In the environment, the predominant breakdown products of RDX are MNX, DNX, and TNX;
methylene dinitramine and 4-nitro-2-diazbutanal have also been detected flaligama etal.. 2013:
Halasz etal.. 2012: Sweeney etal.. 2012b: Paquet etal.. 2011: Fuller etal.. 2010: Smith etal.. 2006:
Meyer etal.. 2005: Beller and Tiemeier. 20021 (see additional discussion in Section 1.1.1 of the
Toxicological Review). The toxicity of these environmental breakdown products has received little
investigation [e.g., see Meyer etal. (2005) and Smith etal. (2006)].
C-8

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
RDX metabolism in animals is less well understood. iV-nitroso RDX metabolites have been
identified as derived through anaerobic metabolism fATSDR. 2012: Pan etal.. 2007bl. Based on
characterization of RDX metabolites in urine and plasma of minipigs, metabolism of RDX appears to
involve loss of nitro groups and ring cleavage (Musick etal.. 2010: Major etal.. 20071. The two
principal urinary metabolites identified in minipigs following a single oral dose of 43 or 45 mg/kg
were 4-nitro-2,4-diazabutanal and 4-nitro-2,4-diaza-butanamide (see Table C-2). Bhushan et al.
(20031 suggested that the formation of the 4-nitro-2,4-diazabutanal metabolite occurred via
denitration followed by hydroxylation and spontaneous hydrolytic decomposition resulting in ring
cleavage and aldehyde formation. In gavage studies with Yucatan minipigs, only trace amounts of
the nitrosamine RDX metabolites MNX and DNX were found in urine (Musick etal.. 2010: Major et
al.. 20071. In plasma, most of the radioactivity existed as RDX, with trace levels of MNX, DNX, and
TNX. The study authors suggested that the trace levels of these metabolites in plasma may have
been formed within the GI tract via sequential nitrogen reduction by intestinal bacteria fMaior et
al.. 20071. The low levels of these compounds in urine and plasma were attributed to the nearly
complete absorption of RDX from the GI tract, leaving little parent compound available for bacterial
metabolism within the GI tract In a study of female deer mice (Peromyscus maniculatus) fed diets
containing RDX at concentrations of 12 and 120 mg/kg for 9 days, MNX and DNX were identified in
the stomach, but TNX was not detected (Pan etal.. 2007b). MNX and DNX were also measured in
various organs of female B6C3Fi mice provided RDX in feed at doses of 0.38-522 mg/kg; TNX was
also detected in some organ compartments, but not in the liver. The study authors concluded that
RDX can be metabolized into its iV-nitroso compounds in mice, but did not identify a mechanism for
the formation of the metabolites. Comparing RDX with MNX and TNX, RDX was the most potent
compound at causing overt signs of toxicity (seizures and mortality) as determined through
identification of the median lethal dose using the U.S. Environmental Protection Agency (EPA)
up-and-down procedure in deer mice of varying ages (Smith etal.. 2009: Rispin etal.. 20021.
Table C-2. Principal urinary metabolites of hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) in minipigs 24 hours after dosing with RDX
Sample origin
Metabolite name
Metabolite structure
Urine peak 1 Ml
4-Nitro-2,4-diazabutanal
H H
N' „N^
02N "If
6
Urine peak 2 M2
4-Nitro-2,4-diaza-butanamide
H H
.,-N, NH2
o2n"" "y
O
Sources: Major et al. (2007): Musick et al. (2010).
C-9

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Although the metabolic pathways and major tissues involved in RDX metabolism have not
been identified, there is some evidence for the involvement of the liver and CYP450 enzymes.
Comparison of hepatic radioactivity to liver concentrations of RDX after a single gavage dose to rats
suggested the presence of RDX metabolites and a possible role for hepatic metabolism of RDX
(Schneider etal.. 19771. In vitro data indicated that CYP450 may be involved in the metabolism of
RDX fBhushan et al.. 20031. Incubation of RDX with nicotinamide adenine dinucleotide phosphate
and rabbit liver CYP450 2B4 under anaerobic conditions produced nitrite, 4-nitro-2,4-diazabutanal,
formaldehyde, and ammonium ions fBhushan et al.. 20031. The reaction rate under aerobic
conditions was approximately one-third of that observed under anaerobic conditions. Several
CYP450 inhibitors (ellipticine, metyrapone, phenylhydrazine, 1-aminobenzotriazole, and carbon
monoxide) decreased the formation of RDX metabolites (55-82% inhibition), providing support for
the role of CYP450 in RDX metabolism.
C.1.4. Excretion
The primary routes of elimination of absorbed RDX are excretion of RDX and metabolites in
urine, and exhalation of CO2 liberated from metabolism of RDX fSweenev etal.. 2012a: Musick etal..
2010: Krishnan et al.. 2009: Major etal.. 2007: Schneider et al.. 19771. Tritium derived from
administered [3H]-RDX has been detected in mouse gall bladder contents, suggesting biliary
secretion in this species (Guo etal.. 19851: however, biliary secretion of RDX or metabolites has not
been confirmed in other animal species. Studies conducted in the rat and swine suggest that
metabolism is the dominant mechanism of elimination of absorbed RDX. In both species,
metabolites dominated the carbon-14 distribution in urine of animals that received doses of
[14C]-RDX, with RDX accounting for <5% of the urinary carbon-14 f Musick etal.. 2010: Schneider et
al.. 19771.
Data on kinetics of elimination of absorbed RDX from blood are available from reports of
accidental exposures of humans to RDX (see Table C-3). Woody etal. (1986) estimated the
elimination ty2 to be approximately 15 hours in a child who ingested approximately 85 mg of RDX
per kg of body weight The ty2 estimate was based on measured serum concentrations of RDX made
between 24 and 120 hours following ingestion for RDX. Based on plasma RDX concentration data
from five adults exposed to RDX (measurements made between 24 and 96 hours following
exposure) (Ozhan etal.. 2003). a first-order elimination ty2 of 20-30 hours was derived (calculated
for this review by fitting the serum RDX data to a first-order exponential function). It is not
possible to draw reliable inferences from these values because they are based on accidental, acute
exposures, and in particular, the data for the child are based on a single set of measurements of one
individual.
C-10

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-3. Elimination £% values for hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX) or radiolabeled RDX
Animal
Route
Dose (mg/kg)
Time3
t/2 (hrs)
Source
Human (child)
Oral
_Q
LO
00
24-120 hrs
15.0°
Woodv et al. (1986)
Human (adult)
Oral
NA
24-96 hrs
21-29cd
Ozhan et al. (2003)
Rat
i.v.
5-6
0.5 min-6 hrs
_Q
o
1
Schneider et al. (1977)
Rat
i.v.
0.8-1.0
30 min-10 hrs
4.6c,d
Krishnan et al. (2009)
Rat
Oral
1.53-2.07
1-10 hrs
6.9c,d
Krishnan et al. (2009)
Mouse
Oral
35, 60, 80
45 min-4 hrs
1.2d
Sweenev et al. (2012b)
NA = not available.
Observation period following exposure on which the ty, values were based.
bReported estimate of dose based on blood kinetics.
cValue for blood RDX.
Calculated for this review based on reported blood RDX concentrations.
The kinetics of elimination of absorbed RDX from blood has been evaluated in rats and
mice. In rats, elimination kinetics were biphasic (Krishnan et al.. 2009: Guo etal.. 1985: Schneider
etal.. 19771. As shown in Table C-3, estimated ty2 values for the terminal elimination phase in rats
range from 5 to 10 hours (Krishnan etal.. 2009: Schneider et al.. 19771. Blood concentration time
course measurements of RDX can be used to estimate an apparent metabolism and elimination of
RDX from blood. The RDX blood concentrations reported in Sweeney etal. f2012bl after gavage
dosing of 35, 60, and 80 mg/kg RDX found a consistent terminal elimination ty2 of approximately
1.2 hours. The elimination ty2 estimated for rats (Krishnan et al.. 2009: Schneider et al.. 19771 is as
much as an order of magnitude longer than in mice (Sweeney etal.. 2012bl.
C.1.5. Physiologically Based Pharmacokinetic (PBPK) Models
Overview of Available Physiologically Based Pharmacokinetic (PBPK) Models
A PBPK model to simulate the pharmacokinetics of RDX in rats was first developed by
Krishnan et al. (20091 and improved upon to extend the model to humans and mice (Sweeney etal..
2012a: Sweeney etal.. 2012bl. The Sweeney etal. (2012al model consists of six main
compartments: blood, brain, fat, liver, and lumped compartments for rapidly perfused tissues and
slowly perfused tissues (see Figure C-l). The model can simulate RDX exposures via the i.v. or oral
route. Distribution of RDX to tissues is assumed to be flow limited. Oral absorption is represented
in this model as first-order uptake from the GI tract into the liver, with 100% of the dose absorbed.
RDX is assumed to be cleared by first-order metabolism in the liver. However, there is no
representation of the kinetics of any RDX metabolites. The acslX model code (Advanced Continuous
Simulation Language, Aegis, Inc., Huntsville, AL) was obtained from the study authors of Sweeney et
al. f~2012al.
C-ll

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
IV dose
KQC
KQB
KQF
KQR
KQS
KQL
KAD
KfC
v
Metabolism
KfC
KAS
KAS
Metabolism
Stomach
Duodenum
Fat
Blood
Liver
Liver
Richly Perfused
Brain
Oral dose	Oral dose
Figure C-l. Physiologically based pharmacokinetic (PBPK) model structure
for hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) in rats and humans.3
aExposure to RDX is by the i.v. or oral route, and clearance occurs by metabolism in the liver. See Table C-4 for
definitions of parameter abbreviations. The Gl tract is represented as one compartment in Krishnan et al. (2009)
(on the left) and two compartments in Sweeney et al. (2012a) (on the right).
The parameter values used in the Sweeney etal. (2012a) rat model are listed in Table C-4.
The physiological model parameter values for cardiac output, tissue volumes, and blood perfusion
of tissues were obtained from the literature fT imchalk et al.. 2 0 0 2: Brown etal.. 19971. RDX
tissue:blood partition coefficients for liver (PL), brain (PB), and richly perfused tissues (PR) were
estimated with an algorithm that relates the measured n-octanol:water partition coefficient for RDX
to reported compositions of water and lipids in specific rat tissues (Poulin and Theil. 2000: Poulin
and Krishnan. 1995). Tissue:blood partition coefficients for fat (PF) and slowly perfused tissues
(PS), as well as the metabolic rate constant (KfC), were simultaneously optimized to fit rat blood
RDX concentrations following i.v. doses of 0.77 or 1.04 mg/kg RDX f Krishnan etal.. 20091.
producing values of 5.57, 0.15, and 2.6 kg0 33/hour for PF, PS, and KfC, respectively. While the
optimized value for PS is much smaller than that used by Krishnan et al. f20091 fl.O kg°-33/hour),
the optimized values for PF and KfC were fairly similar to those used by Krishnan et al. (2009)
(7.55, and 2.2 kg0 33/hour). The rat model with these parameter values also had good agreement
with blood RDX concentrations after a 5-6 mg/kg i.v. exposure (Schneider etal.. 1977).
C-12

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-4. Parameter values used in the Sweeney etal. (2012a) and Sweeney
etal. (2012b) physiologically based pharmacokinetic (PBPK) models for
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) in rats, humans, and mice as
reported by the study authors
Parameter (abbreviation; units)
Rat
Human
Mouse
Source
BW (kg)
0.3
70
0.0206
Default values; study-specific values used
if available
Cardiac output (KQC, L/hr/kg074)
15
14
15
Timchalk et al. (2002); Brown et al. (1997)
Tissue volumes (fraction of BW)
Liver (KVL)
0.04
0.026
0.04
Timchalk et al. (2002); Brown et al. (1997)
Brain (KVB)
0.012
0.02
0.012
Timchalk et al. (2002); Brown et al. (1997)
Fat (KVF)
0.07
0.21
0.07
Timchalk et al. (2002); Brown et al. (1997)
Richly perfused tissues (KVR)
0.04
0.052
0.04
Timchalk et al. (2002); Brown et al. (1997)
Blood (KVV)
0.06
0.079
0.06
Timchalk et al. (2002); Brown et al. (1997)
Slowly perfused tissues (KVS)
0.688
0.523
0.688
0.91-(KVL + KVB + KVF + KVR + KVV)
Blood flows (fraction of cardiac output)
Liver (KQL)
0.25
0.175
0.25
Timchalk et al. (2002); Brown et al. (1997)
Brain (KQB)
0.03
0.114
0.03
Timchalk et al. (2002); Brown et al. (1997)
Fat (KQF)
0.09
0.085
0.09
Timchalk et al. (2002); Brown et al. (1997)
Slowly perfused tissues (KQS)
0.2
0.2449
0.2
Timchalk et al. (2002); Brown et al. (1997)
Richly perfused tissues (KQR)
0.43
0.3811
0.43
1-(KQL + KQB + KQF + KQS)
Tissue:blood partition coefficients
Liver (PL)
1.2
1.3
1.3
Krishnan et al. (2009)a
Brain (PB)
1.4
1.6
1.6
Krishnan et al. (2009)a
Richly perfused tissues (PR)
1.4
1.6
1.6
Krishnan et al. (2009)a
Fat:blood (PF)
5.57
5.57
5.57
Sweenev et al. (2012a)b
Slowly perfused tissues (PS)
0.15
0.15
0.15
Sweeney et al. (2012a)b
Metabolism
First-order metabolic rate constant
(KfC; kg033/hr)
2.6
9.87 (child);
11.2 (adult)
102
Sweenev et al. (2012a)b c;
Sweenev et al. (2012b)d
C-13

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-4. Parameter values used in the Sweeney etal. f2012al and Sweeney
etal. f2012bl physiologically based pharmacokinetic (PBPK) models for RDX
in rats, humans, and mice as reported by the study authors (continued)
Parameter (abbreviation; units)
Rat
Human
Mouse
Source
GI absorption
Dosing via gavage
Absorption from compartment 1
(KAS, /hr)
0.83
0.033
0.51
Sweeney et al. (2012a); Sweeney et al.
(2012b)cde
Transfer from compartment 1 to
compartment 2 (KT, /hr)
1.37
0
0
Sweeney et al. (2012a)cd
Absorption from compartment 2
(KAD, /hr)
0.0258
0
0
Sweeney et al. (2012a)cd
Dosing via capsule (KAS, /hr)
0.12
NA
NA
Sweeney et al. (2012a)e
"coarse" RDX formulation (KAS, /hr)
0.005
NA
NA
Sweeney et al. (2012a)e
BW = body weight: KAD = rate constant for oral absorption, compartment 2; KAS = rate constant for oral
absorption, compartment 1; KQB = fractional blood flow to brain; KQC = cardiac output; KQF = fractional blood
flow to fat; KQL = fractional blood flow to liver; KQR = fractional blood flow to richly perfused tissue;
KQS = fractional blood flow to slowly perfused tissue; KVB = fractional tissue volume of brain; KVF = fractional
tissue volume of fat; KVL = fractional tissue volume of liver; KVR = fractional tissue volume of richly perfused
tissue; KVS = fractional tissue volume of slowly perfused tissue; KVV = fractional tissue volume of blood volume;
NA = not available.
Predicted from n-octanol:water partition coefficient.
bOptimized from rat i.v. data.
cOptimized from human data of Ozhan et al. (2003) and Woody et al. (1986).
dOptimized from mouse oral data.
eOptimized from rat oral data of Bannon et al. (2009), Crouse et al. (2008), Krishnan et al. (2009), and Schneider et
al. (1977).
Note: Parameter values used in the Sweeney et al. (2012a) and Sweeney et al. (2012b) PBPK models for RDX in
rats, humans, and mice.
The GI tract oral absorption rate constant (KAS) was optimized to fit the time-course
concentration data for rat oral dosing studies. The Krishnan et al. (2009) model used a
one-compartment GI tract KAS was fit to the RDX blood concentrations in Krishnan et al. (2009).
and the model with this parameter value had good agreement with the blood RDX concentrations
after 0.2 and 1.24 mg/kg oral exposures (Crouse etal.. 2008). The value of KAS was adjusted to fit
the RDX blood concentrations in the Schneider et al. f 19771 study. Sweeney etal. f2012al modified
the GI tract description by adding a second GI compartment and corresponding oral absorption
parameters (KAS, KAD, and KT) to fit the blood concentrations from Krishnan et al. f20091. For the
other oral dosing studies, the two-compartment GI model did not improve the model fit to the data,
so KT was set equal to zero making the GI submodel equivalent to a one-compartment model. The
C-14

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
value of KAS was adjusted separately to fit the oral studies with the RDX in capsules fBannon etal..
2009: Crouse etal.. 20081 and coarse-grain RDX in a saline slurry fSchneider etal.. 19771.
The Sweeney etal. (2012a) model fits to blood and brain RDX concentrations for rats were
mostly within a factor of 1.5 of the experimentally measured values indicating a tightly calibrated
model.
Human RDX toxicokinetics were modeled with the same model structure as for rats. Values
for the human physiological parameters such as tissue volumes and blood perfusion of tissues were
obtained from the literature fBrown et al.. 19971. Human absorption and metabolic clearance rate
constants were optimized to fit observed RDX blood concentrations from a case study of ingestion
by a 3-year-old boy (Woody etal.. 19861. and a study where five soldiers were intentionally or
accidentally exposed to RDX powder via inhalation or dermal contact (Ozhan etal.. 20031. The
amounts of RDX ingested in both studies were unknown, so Sweeney etal. (2012a) estimated the
dose amount by optimizing this parameter to fit the data (see Table C-4). Sweeney etal. f2012al
initially simulated each individual soldier's blood level data separately. The resulting parameter
values were similar, so data from the five soldiers were combined and the rate constants were
re-estimated using the combined data. For comparison, the rat metabolic rate constant (KfC) was
scaled to humans; the rat KfC (from fitting to in vivo data) was multiplied by the ratio of the human
to rat metabolic rate constants measured in vitro and by the ratio of human-to- rat microsomal
protein levels (Cao etal.. 2008: Lipscomb and Poet. 20081. The scaling from rats yielded a human in
vivo metabolic rate constant of 12.4 kg body weight scaling to the 0.33 power (BW0-33)/ hour, which
is similar to the values that Sweeney etal. (2012a) derived by fitting the combined Ozhan et al.
f20031 adult data (11.2 kg BW°-33/hour) and the Woody etal. f 19861 child data
(9.87 kg BW°-33/hour).
Mouse RDX toxicokinetics were also modeled by Sweeney etal. (2012b) using the same
model structure as for rats. Values for the mouse physiological parameters such as tissue volumes
and blood perfusion of tissues were assumed to be the same as the body weight normalized
parameter values in the rat model. RDX tissue:blood partition coefficients for liver (PL), brain (PB),
and richly perfused tissues (PR) were estimated with an algorithm that relates the measured
n-octanol:water partition coefficient for RDX to reported compositions of water and lipids in
specific mouse tissues fPoulin and Theil. 2000: Poulin and Krishnan. 19951. The KfC and KAS were
optimized to fit measured mouse RDX blood concentrations (Sweeney et al.. 2012b). The KfC value
estimated for the mouse (102 kg0 33/hour) is much higher than those estimated for rats and humans
(2.6 and 11.2 kg0 33/hour, respectively); however, the KAS value (0.51/hour) fit to mouse data is
similar to the value (0.83/hour) used in the RDX rat model for gavage in water. The Sweeney et al.
(2012b) model predictions of blood RDX concentrations were in good agreement with the
experimental mouse gavage data reported in the same study.
The above PBPK model was evaluated and subsequently modified by EPA for use in
dose-response modeling in this assessment This is detailed in the following section.
C-15

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Physiologically Based Pharmacokinetic (PBPK) Model Evaluation and Further Development of
the Sweeney et al. (2012a) and Sweeney et al. (2012b) Models
EPA evaluated and performed a quality control check of the PBPK models for RDX in rats,
humans, and mice published by Sweeney and colleagues (Sweeney et al.. 2012a: Sweeney etal..
2012b). The conclusions from these analyses are summarized below and then discussed in more
detail:
1)	The model code and the parameter values matched the published reports. Minor
discrepancies in physiological parameters (fractional tissue volume of richly perfused tissue
[KVR] and fractional blood flow to slowly perfused tissue [KQS]) were identified and
updated in the model by EPA.
2)	The absorption of RDX from the GI tract did not use a consistent structure; for gavage doses,
the model used a two-compartment GI submodel and for other oral exposures (e.g., gelatin
capsule), the model used a one-compartment GI submodel. The model was revised to have
a one-compartment GI submodel to simulate all oral exposures with a consistent set of
absorption parameters for each dosage formulation of administered RDX.
3)	Additional oral rat data were identified from single-dose studies (MacPhail etal.. 1985:
Schneider et al.. 1977) and subchronic studies (Schneider et al.. 1978) and were used for
model calibration as well as for independent comparison against model predictions.
4)	In addition to the sensitivity analysis conducted by Sweeney etal. f2012bl on the mouse
model, a sensitivity analysis in the rat and human models was performed.
5)	The Sweeney etal. (2012b) mouse model used the same physiological parameters scaled to
body weight as the rat model. This mouse model was revised to use mouse-specific
physiological parameters.
The Sweeney etal. (2012a) model for rats was modified by changing the oral absorption
rate constants (as discussed below) and the PF and PS as shown in Table C-5. The partition
coefficients for the fat and slowly perfused tissues were set to the values calculated by Krishnan et
al. f20091 relating the measured n-octanol:water partition coefficient for RDX to reported
compositions of water and lipids in those tissues. The fits to RDX blood time-course data after i.v.
exposure (see Figure C-2) are slightly worse than the Sweeney etal. f2012al rat model because the
Sweeney etal. (2012a) rat model optimized the fat:blood and slowly perfused tissue partition
coefficients to fit the data.
C-16

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-5. Parameters values used in the EPA application of the rat, human,
and mouse models
Parameter (abbreviation; units)
Rat
Human
Mouse
Source
BW (kg)
0.3
70
0.0206
Default values shown; study-specific
values used if available
Cardiac output (KQC; L/hr/kg074)
15
14
15
Timchalk et al. (2002); Brown et al. (1997)
Tissue volumes (fraction of BW)
Liver (KVL)
0.04
0.026
0.055
Timchalk et al. (2002); Brown et al. (1997)
Brain (KVB)
0.012
0.02
0.017
Timchalk et al. (2002); Brown et al. (1997)
Fat (KVF)
0.07
0.21
0.07
Timchalk et al. (2002); Brown et al. (1997)
Richly perfused tissues (KVR)
0.04
0.054
0.071
Timchalk et al. (2002); Brown et al. (1997)
Blood (KVV)
0.06
0.079
0.049
Timchalk et al. (2002); Brown et al. (1997)
Slowly perfused tissues (KVS)
0.688
0.523
0.648
0.91-(KVL + KVB + KVF + KVR + KVV)
Blood flows (fraction of cardiac output)
Liver (KQL)
0.25
0.175
0.25
Timchalk et al. (2002); Brown et al. (1997)
Brain (KQB)
0.03
0.114
0.03
Timchalk et al. (2002); Brown et al. (1997)
Fat (KQF)
0.09
0.085
0.09
Timchalk et al. (2002); Brown et al. (1997)
Slowly perfused tissues (KQS)
0.2
0.249
0.2
Timchalk et al. (2002); Brown et al. (1997)
Richly perfused tissues (KQR)
0.43
0.377
0.43
1-(KQL + KQB + KQF + KQS)
Tissue:blood partition coefficients and metabolism
Liver (PL)
1.2
1.3
1.3
Krishnan et al. (2009)a
Brain (PB)
1.4
1.6
1.6
Krishnan et al. (2009)a
Richly perfused tissues (PR)
1.4
1.6
1.6
Krishnan et al. (2009)a
Fat:blood PC (PF)
7.55
7.55
7.55
Krishnan et al. (2009)a
Slowly perfused tissues (PS)
1.0
1.0
0.9
Krishnan et al. (2009)a
First-order metabolic rate constant
(KfC, kg0 33/hr)
2.6
9.87 (small
boy); 11.2
(soldiers)
77
Sweenev et al. (2012a)bc; Sweenev et al.
(2012b)d
C-17

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-5. Parameters values used in the EPA application of the rat, human,
and mouse models (continued)
Parameter (abbreviation; units)
Rat
Human
Mouse
Source
Absorption
Absorption from Gl to liver (KAS,
/hr)
Table C-6
1.75
0.6
Fit to rat, human, and mouse oral data
Absorption from lung to blood
(Klung, /hr)

0.75

Fit to human data
BW = body weight; KQB = fractional blood flow to brain; KQC = cardiac output; KQF = fractional blood flow to fat;
KQL = fractional blood flow to liver; KQR = fractional blood flow to richly perfused tissue; KVB = fractional tissue
volume of brain; KVF = fractional tissue volume of fat; KVL = fractional tissue volume of liver; KVS = fractional
tissue volume of slowly perfused tissue; KVV = fractional tissue volume of blood volume.
Predicted from n-octanol:water partition coefficient.
bOptimized from rat i.v. data.
cOptimized from human data of Ozhan et al. (2003) and Woody et al. (1986).
dOptimized from mouse oral data, and differs from that obtained by Sweeney et al. (2012b).
A)
2.5
2
£. 1.5
1 -
0.5

¦ 1.038 mg/kg data
	1.038 mg/kg fit
A
A 0.767 mg/kg data
	0.767 mg/kg fit
3 4 5 6 7
time (hours)
10
5.5 mg/kg data 	5.5 rrig/kg fit
2	3	4
time (hours)
Figure C-2. EPA rat physiologically based pharmacokinetic (PBPK) model
predictions fitted to observed hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
blood concentrations in male and female Sprague-Dawley rats following i.v.
exposure.3
aA) Data from Krishnan et al. (2009) (0.4 kg rats) and B) data from Schneider et al. (1977) (simulation of 0.25 kg rats
and 5.5 mg/kg dose for 0.2-0.25 kg rats and 5-6 mg/kg dose).
C-18

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Absorption ofhexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) from the gastrointestinal (Gl)
Tract
As discussed above in the oral absorption section under toxicokinetics (Section C.l.l), the
rate of oral absorption depends on the physical form of RDX. This was demonstrated by comparing
the Schneider etal. (19771 studies, which used gavage doses of 100 mg/kg of course, granular RDX
and 50 mg/kg finely powdered RDX, and observing that the 50 mg/kg finely powdered RDX had a
higher peak plasma level. These results are likely explained by the smaller surface area-to-mass
ratio of the coarse-grain RDX, leading to slower dissolution and absorption.
To follow the rule of model parsimony (i.e., use no more parameters than needed for the
best fit to all of the data), oral absorption was modeled with a one-compartment GI tract submodel
for all simulations. To account for the differences in absorption due to the physical form of RDX,
separate rate constants for RDX oral absorption were optimized to fit measured blood
concentrations of RDX according to the type of dosing formulation; the model fits obtained with the
EPA's revised parameters for rats are shown in Figures C-3 to C-5. The oral dosing formulations
were grouped into four categories: RDX dissolved in water, RDX in capsules, fine-grain RDX, and
coarse-grain RDX. The absorption rate constant for RDX dissolved in water was optimized to the
data in the Krishnan et al. (2009) study (see Figure C-3). The absorption rate constant for RDX in
capsules was optimized to the data in the Crouse etal. (2008) and Bannonetal. (2009) studies (see
Figure C-4). The absorption rate constant for fine-grain RDX was optimized to the data described
below (Additional RDX Time-Course Data) in the MacPhail etal. f!9851 and Schneider etal. f 19771
studies (see Figure C-7). The Schneider et al. f 19771 study was used to estimate the absorption rate
constants for coarse-grain RDX (see Figure C-5; as represented by the fit to the data obtained by the
solid curve at 100% bioavailability). Overall, the fits of the EPA revised model to the blood time-
course data of these studies are similar to the fits of the Sweeney etal. f2012al rat model. The fits
to RDX brain time-course data after oral exposure to RDX in capsules (see Figure C-6A) are similar
to the fits of the Sweeney etal. (2012a) rat model. The absorption rate constants for each dosing
formulation are listed in Table C-6. As discussed in Section C.1.2, Distribution, plasma protein
binding could be a factor impacting RDX diffusion across the blood-brain barrier. However, in the
absence of empirical values for estimating protein binding, the total RDX blood concentration is
assumed available for diffusion across the blood-brain barrier.
C-19

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
2.25
¦ 1.53 mg/kg data 	1.53mg/kgfit
2
A 2.07 mg/kg data 	2.07 mg/kg fit
1.75
1.5
1.25
1
£0.75
0.5
0.25
0
0
1
2
3
4
5
e
7
8
9 10
time (hours)
Figure C-3. EPA rat physiologically based pharmacokinetic (PBPK) model
predictions fitted to observed hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
blood concentrations following oral exposure to RDX dissolved in water.3
aMale and female Sprague-Dawley rats (0.4 kg) were dosed by gavage (Krishnan et al., 2009).
A)
0.55
0.5
0.45
5 0.4
DjD
_E 0.35
1 0.3
| 0.25
| 02
g 0.15
0.1
0.05
0
B) 8
7
	18 mg/kg fit
¦ 18 mg/kg data
	3 mg/kg fit
A 3 mg/kg data
	1.24 mg/kg fit ¦ 1.24 mg/kg data
0123456789 10 11 12
time (hours)
0 4 8 12 16 20 24 28 32 36 40 44 48
time (hours)
Figure C-4. EPA rat model predictions fitted to observed hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) blood concentrations following oral exposure to
RDX in dry capsules.3
aThe ingested RDX doses were: A) 0.2 and 1.24 mg/kg RDX in male Sprague-Dawley rats [0.4 kg, data from Crouse
et al. (2008)1 and B) 3 and 18 mg/kg RDX in male and female Sprague-Dawley rats [0.35 kg, data from Bannon et
al. (2009)1 for KAS = 0.35/hour.
C-20

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
¦ 100 mg/kg data
	35%bioavail KAS=0,019/hr fit
	40%bioavail KAS=Q.019/hr fit
	100% bioavail KAS=0.00497/hr fit
£1.25
0 2 4 6 8 10 12 14 16 18 20 22 24
time (hours)
Figure C-5. Effect of varying oral absorption parameters on EPA rat model
predictions fitted to observed hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
blood concentrations following oral exposure to coarse-grain RDX.a
aSymbols denote observed RDX blood concentrations measured in male Sprague-Dawley rats (0.225 kg) resulting
from oral doses of 100 mg/kg RDX (Schneider et al.. 1977). The KAS fit to these data, assuming 100%
bioavailability, resulted in the same estimate (KAS = 0.00497/hour) as obtained by Sweeney et al. (2012a).
Alternatively, for KAS fixed at the value fit to fine-grain RDX in a saline slurry (KAS = 0.019/hour fit to data from
Schneider et al. (1977) and MacPhail et al. (1985); see Figure C-7), the estimate of oral bioavailability fit to the
RDX blood concentrations was 35%. A bioavailability of 40% and KAS = 0.019/hour is also shown for comparison
	18 mg/kg fit
¦ 18 mg/kg data
	3 mg/kg fit
A 3 mg/kg data
16 20 24 28 32 36 40 44 48
time (hours)
	50 mg/kg fit
¦ female data
A male data
Figure C-6. EPA rat model predictions fitted to observed hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) brain tissue concentrations following oral
exposure to RDX.a
8 12 16 20 24 28 32 36 40 44 48
time (hours)
13
12
___ 11
Is 10
i. 9
= s
" 7
.£ 6
g 5
" 4
3
2
1
0
aA) 3 and 18 mg/kg RDX in dry capsules [0.35 kg male and female rat data from Bannon et al. (2009)1; best-fit
KAS = 0.35/hour. B) 50 mg/kg fine-grain RDX in a saline slurry [0.25 kg male and female rat data from MacPhail et
al. (1985)1; best-fit KAS = 0.019/hour.
C-21

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-6. Doses, dosing formulations, and absorption rate constants in
animal and human studies
Formulation
Study
Dose
Estimated KA (/hr)
RDX dissolved in water
Krishnan et al. (2009)
1.53, 2.07 mg/kg, single gavage
1.75
Schneider et al. (1978)
~5-8 mg/kg-d, drinking water
90 d
Dry RDX in capsules3
Crouse et al. (2008)
0.2, 1.24 mg/kg, single dose
0.35
Bannon et al. (2009)
3,18 mg/kg, single dose
Fine-grain RDX in saline slurry
Schneider et al. (1977)
50 mg/kg, single gavage
0.19
MacPhail et al. (1985)b
50 mg/kg, single gavage
Coarse-grain RDX in saline slurry
Schneider et al. (1977)
100 mg/kg, single gavage
0.00497
aCapsules were filled with dry RDX from stock solution of acetone, and acetone was evaporated off.
bRDX particle size was <66 urn in diameter suspended in a 2% solution of carboxymethylcellulose.
An alternative to varying the KAS for each RDX formulation would be to vary the oral
bioavailability, in effect modifying the administered exposure concentration. Therefore, the
sensitivity of the model fit to variations in oral bioavailability was examined in Figure C-5 and an
analysis of model sensitivity to oral bioavailability was conducted as discussed further in the
section, Sensitivity Analysis of the Rat and Human PBPK Model.
Additional Hexahydro-1,3,5-trinitro-l,3,5-triazine (RDX) Time-Course Data
The EPA revised models were simultaneously fitted against additional RDX time-course
data [not used in the original Sweeney etal. f2012al model calibration]. These data came from
(1) two studies in which animals received oral doses of fine-grain RDX fMacPhail etal.. 1985:
Schneider et al.. 19771 (see Figure C-7) and (2) RDX brain time-course data from a study in which
animals received oral doses of fine-grain RDX fMacPhail etal.. 19851 (see Figure C-6B). Overall, the
calibrated EPA rat model predictions are within a factor of 1.5 of the measured values from
different data sets, and are therefore likely to provide a more robust estimated parameter.
C-22

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
A) 5
	50 mg/kg fit
¦ 50 mg/kg data
B) 4.5
	50 mg/kg fit
4.5
4
¦ female data
0
0 24 48 72 96 120 144 168 192 216 240 264
time (hours)
0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96
time (hours)
Figure C-7. EPA rat model predictions fitted to observed hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) blood concentrations following oral exposure to
fine-grain RDX in a saline slurry.3
aOral doses of 50 mg/kg RDX were administered to: A) male Sprague-Dawley rats (0.225 kg) (Schneider et al., 1977)
and B) male and female Sprague-Dawley rats (0.25 kg) data (MacPhail et al., 1985). Best-fit KAS = 0.019/hour.
Following calibration, the EPA model was further tested by comparison with results from
two other subchronic oral studies in male and female rats (Schneider etal.. 19781. These were a
gavage study in which 20 mg/kg RDX was administered in saline slurry and a drinking water study
in which rats were provided with RDX-saturated drinking water (50-70 |ig/mL] ad libitum for
which the study authors estimated a daily dose between 5 and 8 mg RDX/kg body weight. It is
striking that the observed RDX blood concentrations in the gavage study (see Figure C-8, symbols)
were virtually the same, or only slightly elevated, as compared to the blood concentrations reported
in the drinking water exposures, with an approximately threefold lower daily administered dose in
the drinking water study (see Figure C-9, symbols). This is counter to the expectation that higher
doses cause higher blood levels and is discussed further below.
C-23

-------
Supplemental Information—Toxicological Review of Hexahydro-l,3,5-trinitro-l,3,5-triazine
CLO
E
x
o
0£
i
i t
-20 mg/kg
model
20 mg/kg
data
0 10 20 30 40 50 60 70 80 90 100
time (days)
Figure C-8. Comparison of EPA rat model predictions with data from
Schneider et al. f19781 for the subchronic gavage study.3
aModel fits and mean observed RDX blood concentrations resulting from daily gavage doses of 20 mg/kg RDX for
90 days to male and female Sprague-Dawley rats (0.225 kg). The RDX in saline slurry was assumed to be
coarse-grain with an oral absorption rate constant KAS = 0.00497/hour,
model
time (days)
Figure C-9. Comparison of EPA rat model predictions with data from
Schneider et al. (19781 for the subchronic drinking water study.3
aModel fits and mean observed RDX blood concentrations resulting from a daily estimated dose of 6.5 mg
RDX/kg-day for 90 days to male and female Sprague-Dawley rats (0.225 kg). The large peak-to-trough change in
the simulation results from model representation of the daily oral ingestion of drinking water primarily during the
waking state. The oral absorption rate constant for RDX dissolved in water was used (KAS = 1.75/hour),
C-24

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
EPA's modified PBPK model was set up to simulate drinking water exposures with a
noncontinuous sipping pattern based on Spiteri f!9821. who assumed 80% of the consumption to
occur episodically at night when the rats were awake.3 The model predicts blood concentrations to
increase in proportion to the total dose; for the gavage study, the model predictions yielded an RDX
blood concentration approximately threefold higher than the reported mean blood concentrations
(see Figure C-8), while for the subchronic drinking water study, the model fit the data reasonably
well (see Figure C-9).
It is possible that multiple mechanisms such as elimination of unabsorbed RDX or metabolic
induction may explain why the observed RDX blood concentrations did not increase in proportion
to the higher administered dose in the gavage studies compared to the drinking water study.
Elimination of unmetabolized RDX may be an insignificant factor in the single-dose studies used for
calibration of the absorption constant for the RDX in saline slurry, but for repeated, higher doses
this elimination route could be significant. Schneider et al. T19781 found similar RDX
concentrations in the feces of rats in the gavage and drinking water studies (3.1 ± 2.0 and
2.7 ± 1.3 ng RDX per g dry weight feces, respectively). The total recovery of radioactivity in feces
was also similar in the gavage study (4.8 ± 0.8%, Week 1 only) and drinking water study
(4.4 ± 0.6%, measured over the course of the study). Thus, the difference in fecal elimination for
the two routes does not appear significant
It is also possible that metabolic induction occurred during the repeated dosing of RDX in
the gavage study leading to the lower observed RDX blood concentrations. The reasonably good fits
of the model to the drinking water data set demonstrated achievement of regular periodic levels,
and indicate a lack or much lower extent of metabolic induction over time from those repeated
doses, possibly because the dose rate was lower: 5-8 versus 20 mg/kg-day in the gavage study.
Overall, the reasonable agreement of the modified EPA RDX rat model with the subchronic drinking
water data support the use of the model in estimating and extrapolating blood levels following
chronic exposure at or below this exposure range (5-8 mg/kg-day), particularly in drinking water.
Simulating Exposures in Humans
The Sweeney etal. f2012al model for humans was modified in the same ways as the rats, by
changing the PF and PS as shown in Table C-5 and fitting the rate constants for oral absorption and
metabolism to RDX blood concentration data. In the studies of humans with measured RDX blood
concentrations by Woody etal. (1986) and Ozhan etal. (2003). the RDX doses were unknown and
the doses were therefore also optimized to fit the data. The model predictions for the Woody etal.
(1986) data using the best-fit values of dose = 58.9 mg/kg, KAS = 1.75/hour, and
KfC = 9.87 kg0 33/hour are shown in Figure C-10. The model predictions for the Ozhan etal. f20031
3A constant drinking water ingestion rate interspaced between episodes of no ingestion was assumed. Each
12-hour awake period consisted of eight cycles that alternated between 1.5-hour cycles of frequent sipping
(continuous ingestion) and zero ingestion for 45 minutes each. Each 12-hour sleeping period consisted of
four cycles with regular sipping periods of 30 minutes followed by 2.5 hours of no ingestion.
C-25

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
data using the best-fit values of an oral dose = 3.5 mg/kg, KAS = 1.75/hour, and
KfC = 9.87 kg0 33/hour are shown in Figure C-ll.
40
	58.9 mg/kg fit
~ 58.9 mi
35
data
E 25
20
.- 15
0
12
24
36
48
60
72
84
96 108 120
time (hours)
Figure C-10. EPA human model predictions fitted to observed hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX) blood concentrations resulting from an
accidental ingestion of RDX by a 14.5-kg boy (Woody et al.. 1986).a
aThe best-fit values were KAS = 1.75/hour, dose = 58.9 mg/kg, and KfC = 9.87 kg033/hour.
3-i	¦ data from accidental exposure
	inhalation fit assuming 3.5 mg/kg
¦5 T.		oral fit assuming 3.5 mg/kg
"O
§ 1.5
Si
c
X
Q
C£.
0.5
0
12
24
36
48
60
72
84
96
time (hours)
Figure C-ll. EPA human model predictions fitted to observed hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX) blood concentrations resulting from
accidental exposure to adults assumed to be 70 kg (Ozhan et al.. 20031.
For an assumed oral exposure, the best-fit values were KAS = 1.75/hour, dose = 3.5 mg/kg, and
KfC = 9.87 kg0 33/hour. For the same 3.5 mg/kg dose and metabolism rate constant, an inhalation exposure found
a best-fit value for Klung of 0.75/hour.
C-26

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
EPA's calibration of the model differed in another important respect from that carried out
by Sweeney etal. f2012al. As previously mentioned, Sweeney etal. f2012al simulated the soldiers'
exposure from the Ozhan etal. (20031 study as an oral exposure, although the study report states
that the exposure was via inhalation and dermal routes. An inhalation or dermal exposure could
change the amount of RDX reaching the blood compared with an oral exposure due to first-pass
metabolism in the liver after oral absorption. Dermal absorption was not considered by EPA to be a
significant route of RDX exposure and was therefore not modeled. This decision is supported by a
study that used excised human skin and reported that only 5.7% of the applied dose was absorbed
into the skin by 24 hours post dosing fReddv et al.. 20081. The model was modified to simulate an
inhalation exposure and compared with the data from Ozhan etal. (20031. There are insufficient
data on blood:air partitioning to modify the Sweeney etal. (2012a) model with a lung
compartment; therefore, inhalation exposure was modeled in an approximate manner as a direct
input to the blood with an optimized absorption rate to represent absorption from air containing
RDX into the blood. The soldiers' inhalation exposure was simulated as a continuous 8-hour
exposure (i.e., assuming that the soldiers were exposed occupationally during an 8-hour workday),
and for the same dose of 3.5 mg RDX/kg that was estimated by Sweeney etal. f2012al. The model
assumed that 100% of the inhaled dose was absorbed and that the absorption rate constant was
optimized to fit the measured blood concentrations of RDX. The model predictions were in good
agreement with the RDX blood concentrations reported by Ozhan etal. (20031 as shown in
Figure C-ll.
Sensitivity Analysis of the Rat and Human Physiologically Based Pharmacokinetic (PBPK)
Models
A sensitivity analysis was performed to see how each model parameter affects the model
output. A sensitivity coefficient, defined as the change in a specified dose metric due to a 1%
increase in the value of a parameter, was calculated for each parameter in the rat and human
models. This analysis was carried out for both short-term (24 hours following a single oral dose of
1.5 mg/kg RDX) and longer term (90 days of repeated oral dosing with 1.5 mg/kg RDX) exposures
for the dose metric of blood AUC. Parameters with sensitivity coefficients >0.1 in absolute value
(i.e., considered sensitive) are presented in Table C-7. For the blood AUC dose metric, the only
sensitive RDX-specific parameter is the KfC. This sensitivity is likely because bioavailability was
assumed to be 100% and metabolism is the only route of elimination in the model. These
assumptions mean that all administered RDX will be absorbed and metabolized; in other words, the
blood AUC is proportional to the dose and inversely proportional to the metabolic clearance rate
constant. For the parameter values in this model, the rate of metabolism is relatively slow
compared to the transport of RDX between other tissues and the site of metabolism in the liver, so
that the blood AUC is not sensitive to parameters that impact transport such as cardiac output
(KQC) and fractional blood flow to liver (KQL). Because the metabolic clearance rate constant is
scaled to body weight and by liver volume, the blood AUC is also sensitive to these parameters. The
C-27

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
sensitivity analysis by Sweeney etal. f2012bl for the AUC of RDX in the liver found the model was
sensitive to the PL in addition to the same parameters (KfC, fractional tissue volume of liver [KVL],
and body weight [BW]) found for the blood AUC.
Table C-7. Sensitivity coefficients for rat and human hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) physiologically based pharmacokinetic (PBPK) models
Parameter
Rat sensitivity coefficient
Human sensitivity coefficient
Fractional liver volume (KVL)
-1
-1
Body weight (BW)
0.3
0.3
Metabolic rate constant (KfC)
-1
-1
KQB = fractional blood flow to brain; KQF = fractional blood flow to fat; KVB = fractional tissue volume of brain;
KVF = fractional tissue volume of fat; KVV = fractional tissue volume of blood volume.
Parameters with sensitivity coefficients <0.1 in absolute value are considered not sensitive, and are listed below:
•	KQC
•	Fractional blood flow to all tissues (KQL; KQF; KQS; KQB)
•	Fractional tissue volume of KVF, KVB, and KVV
•	Blood partition coefficients to all tissues (PL; PF; rapidly PR; PS; PB)
•	Absorption rates from Gl (KAS, KT, KAD)
The model is also very sensitive to oral bioavailability, with a sensitivity coefficient of 0.8 in
the case of the rat model. As discussed above in the oral absorption section of toxicokinetics (see
Section C.l.l), estimates of the bioavailability of RDX range from 50 to 87% or greater and may
depend on the physical form of RDX fKrishnan et al.. 2009: Schneider etal.. 1978.19771. However,
as seen in Figure C-5, it was not possible to identify the bioavailability and KAS as separate
parameters by fitting to the available RDX blood concentration time-course. Introducing oral
bioavailability as an additional unknown parameter and recalibrating the model did not appear to
provide an advantage. Therefore, 100% bioavailability was assumed in the model and
acknowledged as an uncertainty.
Simulating Exposures in Mice
Physiological parameters specific to mice were obtained from the literature (Brown etal..
19971 and are shown in Table C-5. The partition coefficients calculated for mice by Sweeney et al.
(2012b) were used, and include the liver, brain, and richly perfused tissues. The partition
coefficients for the fat and slowly perfused tissues from the Sweeney etal. (2012b) mouse model
were not used because they were estimated via optimization of fits to rat i.v. data. Instead, the
partition coefficient for fat tissues was set equal to the value calculated by Krishnan et al. (20091 for
rat fat tissue, 7.55. The partition coefficient for slowly perfused tissues (0.9) was calculated for
mouse tissues using the same methodology as Krishnan et al. (20091. The rate constants for oral
absorption and metabolism were optimized to fit the data from Sweeney etal. f2012bl for mouse
C-28

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
blood RDX concentrations. The model predictions were in good agreement with the RDX blood
concentrations reported by Sweeney etal. f2012bl. as shown in Figure C-12.
A 80mg/kgdata 	80mg/kgfit
time hours
60mg/kgdata 	60mg/kgfit
time (hours)
~ 35mg/kgdata 	35mg/kgfit
time (hours)
Figure C-12. Comparison of EPA mouse physiologically based
pharmacokinetic (PBPK) model predictions with data from oral exposure to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) dissolved in water.3
aModel fits and mean and standard deviation of observed RDX blood concentrations in female B6C3Fi mice
(0.0205 kg) for doses of 35, 60, and 80 mg/kg with KAS = 0.6/hour and KfC = 77 kg0 33/hour. Experimental data
from Sweeney et al. (2012b).
The mouse RDX blood concentrations reported by Sweeney etal. f2012bl. as shown in
Figure C-12 were evaluated with a noncompartmental analysis and compared with the rat data.
The estimate of the area under the curve for blood concentration versus time from the time of
dosing to the time RDX is completely eliminated (AUCtotai) was calculated with a linear trapezoidal
C-29

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
sum plus an extrapolation of the blood concentration at the last time point divided by the terminal
elimination rate constant as shown in the following equation:
AUCtotai = 2(Ablood concentrations) At/2 + blood concentration at last time point/Kei
(C-l)
where Ablood concentrations are the successive blood concentrations, At is the time
between measured concentrations, and Kei is the terminal elimination rate constant
(calculated from the slope of the linear regression line to the log of blood concentrations).
For the mouse data from Sweeney etal. (2012b) for the doses of 35, 60, and 80 mg/kg, the
results of the AUCtotai calculation are 3.35, 3.70, and 4.75 mg/L hour; normalized to the
administered dose, these are 0.096, 0.062, and 0.059 mg/L hour per mg/kg. For the blood
concentrations measured in rats in the Krishnan et al. f20091 study (see Figure C-3), the animals
received a single oral (gavage) dose of RDX dissolved in water similar to the Sweeney etal. (2012b)
study. The Krishnan et al. f20091 study used doses of 1.53 and 2.07 mg/kg and the results of the
AUCtotai calculation are 6.1 and 11.9 mg/L hour. Including the extrapolation of the blood
concentration from the last time point with Kei had a major contribution to the AUCtotai
(approximately one-third), which adds uncertainty to the result, so the AUCtotai was also calculated
without this term and the results are 4.1 and 7.5 mg/L hour. The AUCtotai values normalized to the
administered doses are 4.0 and 5.8 mg/L hour per mg/kg (including the extrapolation from the last
time point) or 2.7 and 3.6 mg/L hour per mg/kg (excluding the extrapolation from the last time
point). Overall, the AUCtotai normalized to the administered doses for the rat are of the order
10-100 times greater than for the mouse. This noncompartmental analysis of the data is
independent of the PBPK modeling and shows the extent of the toxicokinetic differences for RDX
between the mouse and rat
The only additional information on RDX metabolism in the mouse comes from a study by
Pan etal. f20131. Pan etal. f20131 measured nitrosamine RDX metabolites of RDX (MNX, DNX, and
TNX, the latter representing a minor metabolic pathway) in mice at the end of a 28-day exposure to
RDX in feed (ad libitum). These measurements were a single time point without controlling the
time between the last RDX ingestion and measurement, and were therefore judged not to be
sufficient for use in parameterizing a PBPK model of the nitrosamine metabolites.
Rat to Human Extrapolations
The rat and human PBPK models as described above were applied to derive human
equivalent doses (HEDs) for candidate points of departure (PODs) for endpoints selected from rat
bioassays. The rat and human PBPK models were used to estimate two dose metrics—the AUC and
the peak concentration (Cmax) for RDX concentration in arterial blood. Tissue-specific dose metrics
for kidney, bladder, and prostate were not available in the PBPK model. The PBPK model estimates
RDX concentrations in the brain and these were considered as tissue-specific dose metrics for
C-30

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
neurotoxicity. The brain RDX concentrations were not used because the brain concentration data
are limited and the data are only moderately well fit. Based on the limited number of observations
of brain concentrations in toxicokinetic studies, the brain concentrations correlate with plasma
concentrations. The plasma concentrations were calibrated by fitting to multiple data set and
predictions from the rat model compared well with data not used in model calibration, supporting
the use of plasma concentrations.
The relationships between administered dose and both internal metrics (AUC and Cmax)
were evaluated with the rat PBPK model over the range of 1 ng/kg-day to 100 mg/kg-day and with
the human PBPK model over the range of 0.05 ng/kg-day to 200 mg/kg-day, ranges that encompass
the PODs. The times to reach steady state for the dose metrics were shorter than the duration of
the toxicity studies, so the steady state values were considered representative of the study and
were used. To calculate steady-state values for daily exposure, the simulations were run until the
daily average had a <1% change between consecutive days. For both the rat and human PBPK
models, both dose metrics correlated linearly with the administered dose. For rats dosed via
gavage, the slope of administered dose versus AUC was 6.800 mg/L-day / mg/kg-day and that for
Cmax was 0.4718 mg/L / mg/kg-day. For a continuous dose, the slope of dose versus AUC was the
same (6.800 mg/L-day / mg/kg-day) and for Cmax was 0.3951 mg/L / mg/kg-day. For humans,
assuming a drinking water dose sipping pattern, the slope of administered dose versus AUC was
13.95 mg/L-day / mg/kg-day and that for Cmax was 0.7316 mg/L / mg/kg-day. Given this linearity
in internal metrics and assuming that equal internal metrics in rats and humans are associated with
the same degree of response, the HEDs could then be directly determined by multiplying the
benchmark dose lower confidence limit (BMDL) in rats by the ratio of these slopes as follows:
•	For a gavage dose in rats converted to a human drinking water dose (i.e., applied to a
study that used gavage dosing, specifically Crouse etal. f20061 and Cholakis et al.
£1980}):
o The ratio for AUC was 6.800 / 13.95 = 0.487
o The ratio for Cmax was 0.4718 / 0.7316 = 0.645
•	For a continuous dose in rats converted to a human drinking water dose (i.e., applied to
a study that used dietary dosing, specifically Levine etal. (1983)):
o The ratio for AUC was 6.800 / 13.95 = 0.487
o The ratio for Cmax was 0.3951 / 0.7316 = 0.540
These ratios were applied in Table 2-2 to calculate the PODhed from the rat BMDLs and no-
observed-adverse-effect levels (NOAELs) for each endpoint
Mouse to Human Extrapolations
The mouse and human PBPK models as described above were applied to derive HEDs for
candidate PODs for endpoints selected from mouse bioassays. The mouse and human PBPK models
were used to estimate two dose metrics—the AUC and Cmax for RDX concentration in arterial
C-31

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
blood. The relationships between administered dose and both internal metrics (AUC and Cmax)
were evaluated with the mouse PBPK model over the range 10 ng/kg-day to 100 mg/kg-day and
with the human PBPK model over the range 0.05 ng/kg-day to 200 mg/kg-day, ranges that
encompass the PODs. The times to reach steady state for the dose metrics were shorter than the
duration of the toxicity studies, so the steady-state values were considered representative of the
study and were used. To calculate steady-state values for daily exposure, the simulations were run
until the daily average had a <1% change between consecutive days. For both the mouse and
human PBPK models, both dose metrics correlated linearly with the administered dose. For mouse
dosed via gavage, the slope of administered dose versus AUC was 0.0656 mg/L-day / mg/kg-day
and that for Cmax was 0.0273 mg/L / mg/kg-day. For a continuous dose, the slope of dose versus
AUC was the same 0.0656 mg/L-day / mg/kg-day and for Cmax was 0.0081 mg/L / mg/kg-day. For
humans, assuming a drinking water dose sipping pattern, the slope of administered dose versus
AUC was 13.95 mg/L-day / mg/kg-day and that for Cmax was 0.7316 mg/L / mg/kg-day. Given this
linearity in internal metrics and assuming that equal internal metrics in mice and humans are
associated with the same degree of response, the HEDs could then be directly determined by
multiplying the BMDL in mice by the ratio of these slopes as follows:
•	For a gavage dose in mice converted to a human drinking water dose (i.e., applied to a
study that used gavage dosing):
o The ratio for AUC was 0.0656 / 13.95 = 0.0047
o The ratio for Cmax was 0.0273 / 0.7316 = 0.373, respectively.
•	For a continuous dose in mice converted to a human drinking water dose (i.e., applied to
a study that used dietary dosing):
o The ratio for AUC was 0.0656 / 13.95 = 0.0047
o The ratio for Cmax was 0.0081 / 0.7316 = 0.011
Summary of Confidence in Physiologically Based Pharmacokinetic (PBPK) Models for
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
Overall, good fits to the rat, mouse, and human time-course data for RDX internal
concentrations were obtained. For the rat and human models, calibration was based generally on
fitting to more than one data set obtained from different studies originating in different
laboratories or accidental exposure settings. Predictions from the rat model compared well with
data from a subchronic drinking water study that was not used in model calibration.
The metabolic rate constant used in the human model was fit to limited data from
accidentally exposed humans; however, the value of the metabolic rate constant has additional
support from in vitro experimental data. The rat metabolic rate constant, fit to multiple
experimental data sets, was scaled to humans using the ratio of human-to-rat rate constants
measured with in vitro methods. This scaled value of the human metabolic rate constant was very
similar to the rate constant estimated by fitting the model to the human data. The congruence in
C-32

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
values increases the confidence in using the EPA-modified PBPK model for predicting human blood
RDX concentrations.
There are several uncertainties in these models (listed below), the most significant of which
pertain to the mouse PBPK model. The mouse model was based on a single data set, which used
high RDX doses to obtain detectable RDX blood concentrations, and the types of additional data that
increased the confidence in the rat and human models are not available for mice. The additional
data not available for mice are in vitro measurements of RDX metabolism by mouse cells and
quantification of potential routes of RDX elimination in mice. Overall, these uncertainties result in
lower confidence in the mouse model than in the rat and human models.
1)	RDX is readily metabolized in several species, yet there are no data on the toxicokinetics of
RDX metabolites in the rat and human. Some data are available for the iV-nitrosoamine
metabolites (a minor metabolic pathway) in mice, but the data are too sparse to help better
parameterize a PBPK model. Consequently, the PBPK models used in this assessment do
not incorporate the kinetics of RDX metabolites.
2)	The available toxicokinetic data are not sufficient to uniquely identify a parameter value for
RDX oral bioavailability. Consequently, the model assumes 100% bioavailability even
though some studies in rats suggest that a lower bioavailability is likely.
3)	The human model is based on single accidental exposures, and the exposure concentrations
are not known.
4)	The only route of clearance of RDX used in the models is that of total metabolism, which
appears reasonable for the rat for which only roughly 5% of the RDX was detected
unmetabolized in urine and feces. However, no data on the excretion of RDX are available
for the mouse. This inability to properly characterize the fraction of RDX that is
metabolized in the mouse is problematic considering some evidence to indicate that the role
of metabolism in RDX toxicity may be different across species. This uncertainty decreases
the confidence in the mouse PBPK model.
5)	The PBPK model for the mouse is based on a single data set This single data set is used to
fit both the absorption and metabolic rate constants. There are no in vitro data to
independently estimate the metabolic rate constant for the mouse. Consequently, the
confidence in the mouse model parameter values is low.
6)	The analytical detection limit in the mouse pharmacokinetic study is too high to enable
detection at the lower doses. The lowest dose that resulted in a detectable level of RDX in
blood was 35 mg/kg; this dose was high enough to manifest some toxicity in the chronic
mouse bioassay. The measured blood concentration at the final 4-hour time point at the
35 mg/kg dose was based on the level measured from one animal only (in the other five
animals exposed at this dose, three were nondetects, one was excluded as an outlier, and
one animal died). Data from a single animal decreases the confidence in the calibration of
the mouse PBPK model.
C-33

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
7) The metabolic rate constant as estimated by the PBPK model for mice was 30-fold higher
than the rat (after accounting for body-weight differences), suggesting that the
toxicokinetics of RDX could be significantly different in the mouse than in the rat Mice may
have more efficient or higher expression of the CYP450 enzymes. Alternatively, mice may
have other unknown metabolic pathways responsible for metabolizing RDX. Identifying the
specific CYP450 enzymes and measuring expression levels and in vitro metabolic rate
constants in mice would allow for in vitro scaling from rats to mice, which could be used to
independently evaluate the mouse metabolic rate constant. Given the high sensitivity of the
model to the metabolic rate constant, this uncertainty in the mouse toxicokinetics
significantly decreases the confidence in using the mouse PBPK model for predicting mouse
blood RDX concentrations.
Model Code for hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) Physiologically Based
Pharmacokinetic (PBPK) Model Used in the Assessment
The PBPK acslX model code is made available electronically through the Health and
Environmental Research Online (HERO) database. All model files may be downloaded in a zipped
workspace from HERO (U.S. EPA. 2014).
C.2. HUMAN STUDIES
Table C-8 presents a summary of case reports of humans acutely exposed to RDX. Table C-9
provides a chronological summary of the methodologic features of the available epidemiology
studies of RDX.
C.3. OTHER PERTINENT TOXICITY INFORMATION
C.3.1. Mortality in Animals
Evaluations of the evidence for specific health effects associated with RDX exposure are
provided in Sections 1.2.1 to 1.2.7. In addition to these specific organ/system health effects,
reduced survival associated with RDX exposure has been observed in experimental animals across
multiple studies of varying exposure duration and study design (see Table C-10). Evidence
pertaining to mortality in experimental animals exposed to RDX is summarized in Table C-10;
studies are ordered in the evidence table by duration of exposure and then species.
C-34

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-8. Summary of case reports of exposure to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX)
Reference,
number of cases,
exposure setting
Exposure
Effects observed
Comments
Barsotti and Crotti
(1949)
17 males among
20 male Italian
workers
(1939-1942)
Inhalation of RDX powder
during the drying, cooling,
sieving, and packing
processes of its manufacture
Generalized convulsions of a
tonic-clonic (epileptic) type
followed by postictal coma; loss
of consciousness without
convulsions; vertigo; vomiting and
confusion; transient arterial
hypertension
Tobacco and alcohol use
were considered by the
study authors to be
aggravating factors
Manufacturing

Symptoms occurred either
without prodromal symptoms or
were preceded by several days of
insomnia, restlessness, irritability,
or anxiety

Kaplan et al. (1965)
5 males among
26 workers
(April-July 1962)
Manufacturing
Inhalation, ingestion, and
possible skin absorption as a
result of the release of RDX
dust into the workroom air
during the dumping of dried
RDX powder, screening and
blending, and cleanup of
spilled material without
adequate ventilation
Sudden convulsions or loss of
consciousness without
convulsions; few or no
premonitory symptoms (e.g.,
headache, dizziness, nausea,
vomiting); stupor, disorientation,
nausea, vomiting, and weakness;
no changes in complete blood
counts or urinalysis
Mild cases of RDX
intoxication may have
been masked by viral
illness with nonspecific
symptoms (e.g.,
headache, weakness,
upset Gl tract); no
method was available
for determining RDX
concentrations in air;
recovery was complete
without sequelae
Merrill (1968)
2 males
Wartime, Vietnam
Ingestion of unknown
quantity of C-4 with
moderate amounts of alcohol
Coma, vomiting, hyperirritability,
muscle twitching, convulsions,
mental confusion, and amnesia;
kidney damage (oliguria, gross
hematuria, proteinuria, elevated
BUN); liver or muscle damage
(high AST); leukocytosis
Confounding factors
included ingestion of C-4
while intoxicated with
ethanol (vodka), which
may have caused Gl
symptoms, and smoking
(1-1.5 packs of
cigarettes per day)
C-35

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-8. Summary of case reports of exposure to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) (continued)
Reference,
number of cases,
exposure setting
Exposure
Effects observed
Comments
Stone et al. (1969)
4 males
(March-December
1968)
Wartime, Vietnam
Ingestion of 180 g (patient 1),
or 25 g (patients 2, 3) of C-4
(91% RDX)
Generalized seizures, lethargy,
nausea, vomiting, fever, muscle
soreness, headaches, twitching,
semicomatose, headaches,
hematuria, abnormal laboratory
findings, muscle injury, elevated
AST; no kidney damage
For the patient who ingested the
highest dose, anemia and loss of
recent memory was present after
30 d
Troops ingested small
quantities of RDX to get
a feeling of inebriation
similar to that induced
by ethanol
Hollander and
Colbach (1969)
5 males (June
1968-January 1969)
Wartime, Vietnam
Inhalation (all five cases) and
ingestion of unknown
quantity of C-4 (two cases)
Tonic-clonic seizures; nausea and
vomiting occurred before and
after admission; hyperirritability,
muscle twitching, convulsions,
mental confusion, and amnesia;
kidney damage (oliguria, gross
hematuria, proteinuria, elevated
BUN); liver or muscle damage
(high AST); leukocytosis;
symptoms cleared by the next day
except for amnesia (in case 2),
oliguria (lasted for 4 days), and
gross hematuria (decreased by
9th hospital day)

Knepshield and
Stone (1972)
6 males
Wartime, Vietnam
Ingestion of C-4, range
25-180 g, average 77 g
Generalized seizures, coma,
lethargy, severe neuromuscular
irritability with twitching and
hyperactive reflexes, myalgia,
headache, nausea, vomiting,
oliguria, gross hematuria, low-
grade fever; abnormal laboratory
findings (neutrophilic
leukocytosis, azotemia, elevated
AST)
Includes data on
two patients from
Merrill (1968)
C-36

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-8. Summary of case reports of exposure to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) (continued)
Reference,
number of cases,
exposure setting
Exposure
Effects observed
Comments
Ketel and Hughes
(1972)
18 males
(December
1968-December
1969)
Wartime, Vietnam
Inhalation while cooking with
C-4 and possible ingestion
Central nervous system signs
(confusion, marked
hyperirritability, involuntary
twitching of the extremities,
severe prolonged generalized
seizures, prolonged postictal
mental confusion, amnesia); renal
effects (oliguria and proteinuria,
one case of acute renal failure
requiring hemodialysis); Gl
toxicity (nausea, vomiting)
C-4 was cut with the
same knife used to
stir/prepare food
Woodv et al. (1986)
1 male child (August
1984)
Manufacturing
Ingestion of plasticized RDX
from mother's clothing
and/or boots; estimated
ingested dose of 1.23 g RDX
was normalized to the
patient's body weight
(84.82 mg/kg)
Seizures before and after
admission; EEG revealed
prominent diffuse slowing that
was greatest in the occipital
regions with no evidence of
epileptiform activity; elevated AST
on admission and after 24 hrs;
within 24 hrs, the child was
extubated and intensive care
withdrawn; normal mental status
and normal neurological
examination at discharge
Mother worked at an
explosive plant in which
RDX was manufactured
in a plasticized form
Goldberg et al.
(1992)
1 male
Nonwartime
Ingestion after chewing a
piece (unknown size) of
"Semtex" plastic explosive
4 hrs before first seizure
Frontal headache and two tonic-
clonic seizures; progressively
disseminating petechial rash
suggestive of meningococcal
infection apyrexial; normotensive;
no photophobia; no neurological
abnormalities; florid petechial
rash over the face and trunk;
lacerated tongue
Initial results included leukocyte
count of 10.8 x 109/dL (87%
neutrophils); hemoglobin, platelet
count, coagulation screen, serum
and CSF biochemistry all within
normal limits; CSF and blood
bacteriologically unremarkable
Shortly following admission,
headache and rash disappeared;
no further seizures

C-37

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-8. Summary of case reports of exposure to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) (continued)
Reference,
number of cases,
exposure setting
Exposure
Effects observed
Comments
Harrell-Bruder and
Hutchins (1995)
1 male
Ingestion of C-4 (chewing on
a piece of undetermined size)
Tonic-clonic seizures; postictal
state; EEGs were normal; brisk
deep tendon reflexes

Nonwartime



Testud et al.
(1996a)
1 male
Manufacturing
Inhalation and possible
dermal exposure during the
RDX manufacturing process
Malaise with dizziness, headache,
and nausea progressing to
unconsciousness and generalized
seizures without involuntary
urination or biting of the tongue;
blood chemistries were in the
normal range and blood alcohol
content was null

Testud et al.
(1996b)
2 males
Inhalation and possible
dermal exposure during the
RDX manufacturing process
Sudden loss of consciousness and
generalized seizures; blood serum
level of 2 mg/L RDX measured
Smoker and alcohol
drinker
Manufacturing



Hett and Fichtner
(2002)
1 male
Nonwartime
Ingestion of a cube (1 cm
across) of C-4
Nausea and vomiting; tonic-clonic
seizure lasting 2 min, followed by
two seizures of about 30 sec each;
myoclonic jerks in all limbs;
petechial hemorrhages around
face and trunk after seizures

Kucukardali et al.
(2003)
5 males
Nonwartime
Ingestion (accidental) of
37-250 mg/kg body weight
RDX during military training
via food contaminated with
RDX
Abdominal pain, nausea,
vomiting, myalgia, headache,
generalized weakness, repetitive
tonic-clonic convulsions, lethargic
or comatose between seizures,
hyperactive deep tendon reflexes,
sinusal tachycardia; elevated
serum levels of AST and alanine
aminotransferase; kidney
damage; plasma RDX levels 3 hrs
after ingestion ranged from 268
to 969 pg/mL

C-38

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-8. Summary of case reports of exposure to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) (continued)
Reference,
number of cases,
exposure setting
Exposure
Effects observed
Comments
Davies et al. (2007)
17 males
Nonwartime
Ingestion of an unknown
quantity of C-4 under unclear
circumstances, but unrelated
to recreational abuse
Seizures, headache, nausea, and
vomiting; hypokalemia and
elevated creatine kinase, lactate
dehydrogenase, and phosphate
noted in all but two patients;
metabolic acidosis only occurred
in two patients directly following
seizures
Patient histories may
have been affected by
the fact that the
incident was the focus
of a military police
investigation
Kasuske et al.
(2009)
2 males
Nonwartime
Ingestion of C-4 after
handling explosive ordnance
Seizures, postictal state,
confusion, drowsiness, headache,
nausea, and vomiting; blood work
revealed high WBC count and
elevated creatine phosphokinase;
proteinuria and gross hematuria
observed

AST = aspartate aminotransferase; BUN = blood urea nitrogen; CSF = cerebrospinal fluid; EEG =
electroencephalogram; WBC = white blood cell.
C-39

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-9. Occupational epidemiologic studies of hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX): summary of
methodologic features
Reference,
Participants,
Consideration


Consideration


setting and
selection,
of likely
Exposure measure and
Outcome
of likely
Analysis and

design
follow-up
selection bias
range
measure
confounding
results details
Sample size
Ma and Li
Details of industrial
Sparse reporting
Details of exposure
Neurobehavioral
No adjustment
Comparisons of
60 exposed;
(1993)
process and subject
of details on
monitoring not reported.
battery
for other risk
mean scaled score
Group A
(China)3
selection framework
subject
Two groups of exposed
administered by
factors (e.g.,
on memory
(n = 30;
Industrial
not reported;
recruitment and
subjects:
trained personnel:
alcohol
retention, letter
26 males,
workers
referents chosen
participation
Group A, intensity,
memory
consumption);
cancellation, or
4 females);
(translated
from same plant;

0.407 (0.332) mg/m3
retention, simple
no consideration
block design test;
Group B
article)
age, employment

[mean (standard
reaction time,
or attempt to
mean time on
(n = 30);

duration, and

deviation)], daily
choice reaction
distinguish TNT
reaction tests;
32 referents

education similar

cumulative,
time, letter

and total
(27 males,

across groups;

2.66 (1.89) mg/m3.
cancellation, and

behavioral score;
5 females)

participation rate

Group B, intensity,
block design

variance (F-test),


not reported

0.672 (0.556) mg/m3; daily
cumulative,
2.53 (8.40) mg/m3.


linear and
multiple
regression, and
correlation
analysis

C-40

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-9. Occupational epidemiologic studies of hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX): summary of
methodologic features (continued)
Reference,
Participants,
Consideration


Consideration


setting and
selection,
of likely
Exposure measure and
Outcome
of likely
Analysis and

design
follow-up
selection bias
range
measure
confounding
results details
Sample size
Hathawav
2,022 workers
Potential healthy
Atmospheric samples of all
Liver function,
Workers with
Comparison of
69 RDX
and Buck
[1,017 exposed to
worker effect
operations with potential
renal function,
TNT exposure
mean value;
exposed
(1977)
open explosives

exposure to open
and hematology
excluded from
prevalence of
(43 males,
(United
(TNT, RDX, HMX),

explosives taken in 1975.
tests (blood)
exposed groups
elevated value on
26 females),
States)
1,005 referents] at

Range: not detected to


an individual test
24 RDX/HMX
Ammunition
five U.S. Army

1.57 mg/m3. Seventy



exposed (all
workers
ammunition plants
(Iowa, Illinois,
Tennessee);
participation rate:
76% exposed, 71%
referents

exposed workers with RDX
at >0.01 mg/m3 (the LOD);
mean: 0.28 mg/m3
(standard deviation not
presented). Job title used
to initially identify exposed
or unexposed status and
reassigned to one of two
exposed groups
(nondetected,
>0.01 mg/m3) based on
subject's industrial hygiene
monitoring results.



males),
338 referents
(237 males,
101 females)
C-41

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-9. Occupational epidemiologic studies of hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX): summary of
methodologic features (continued)
Reference,
Participants,
Consideration


Consideration


setting and
selection,
of likely
Exposure measure and
Outcome
of likely
Analysis and

design
follow-up
selection bias
range
measure
confounding
results details
Sample size
West and
378 of 404 subjects
Former
Semiquantitative
Abnormal
Cases and
Unadjusted
32 cases
Stafford
(excluded 3 deaths
employees who
assessment; source of
hematology value
controls were
prevalence odds
(29 males,
(1997)
and 23 subjects with
were unable to
industrial hygiene data not
in 1991 survey
not matched and
ratios and 95%
3 females) and
(United
unknown addresses)
work due to
reported. Interviews with
indicating possible
statistical
CIs; analyses
322 controls
Kingdom)
previously studied in
adverse health
current and past
myelodysplasia:
analyses were
limited to males
(282 males,
Ammunition
1991, 32 cases with
outcome were
employees and job title
neutropenia
not adjusted for
because of low
12 females)
workers
abnormal
not included in
analysis were conducted to
(2.0 x 109/L), low
other risk factor
frequency of

(case-
hematology test and
the 1991
identify potential exposure
platelet count
or occupational
exposure to

control
322 controls with
prevalence study
to 100 agents, including
(<150 x 109/L), or
exposures; no
females

study)
normal hematology

RDX. Exposure surrogate
macrocytosis
consideration or



test; participation

was >50 hrs in duration
(mean
attempt to



rate among eligible

and intensity was low
corpuscular
distinguish TNT



subjects: 97% cases,

(1-10 ppm), moderate
volume = 99 fL or




93% controls

(10-100 ppm), or high
>6% macrocytes)






(100-1,000 ppm). RDX







exposure prevalence







(males) was 83%.




HMX = octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine; LOD = limit of detection; TNT = trinitrotoluene.
aMa and Li (1993) describe symptoms reported by subjects during a physical examination, but these are not included in the evidence table because responses
for individual symptoms were not identified.
C-42

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-10. Evidence pertaining to mortality in animals3
Reference and study design
Results
Lish et al. (1984b)
Mice, B6C3Fi, 85/sex/group; interim
sacrifices (10/sex/group) at 6 and 12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles <66 urn
0,1.5, 7.0, 35, or 175/100 mg/kg-d (high
dose reduced to 100 mg/kg-d in Week 11
due to excessive mortality)
Diet
2 yr (mortality incidence also provided for
mice through Week 11 when the high dose
was dropped because of high mortality at
that dose, and from the report of the 6-mo
interim sacrifice)
Doses
0
1.5
7.0
35
175/100
Mortality at 11 wk (incidence)
M
1/85
0/85
0/85
0/85
30/85
F
0/85
0/85
0/85
0/85
36/85
Mortality at 6 mos (incidence)
M
1/85
2/85
3/85
3/85
34/85
F
0/85
1/85
0/85
0/85
36/85
Mortality at 2 yr (incidence)
M
20/65
23/65
25/65
29/65
41/65
F
16/65
21/65
14/65
21/65
42/65
After the high dose was reduced to 100 mg/kg-d, survival was
similar to controls.
Levine et al. (1983)
Rats, F344, 75/sex/group; interim sacrifices
(10/sex/group) at 6 and 12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant: 83-89% of particles <66 nm
0, 0.3,1.5, 8.0, or 40 mg/kg-d
Diet
2 yr (mortality incidence also provided for
mice through wk 13, and from the report of
the 6-mo scheduled sacrifice)
Doses
0
0.3
1.5
8.0
40
Mortality at 13 wk (incidence)
M
0/75
0/75
0/75
0/75
0/75
F
0/75
0/75
0/75
0/75
0/75
Mortality in 6-mos interim sacrifice animals (incidence)*
"includes spontaneous death and moribund sacrifice
M
0/75
0/75
0/75
0/75
5/75
F
0/75
0/75
0/75
0/75
0/75
Mortality at 2 yr (incidence)
M
17/55
19/55
30/55*
26/55
51/55*
F
12/55
10/55
13/55
14/55
27/55*
C-43

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-10. Evidence pertaining to mortality in animals3 (continued)
Reference and study design
Results
Cholakis et al. (1980)
Mice, B6C3Fi, 10-12/sex/group
88.6% pure, with 9% HMX and 2.2% water
as contaminants; ~200 urn particle size
0, 40, 60, or 80 mg/kg-d for 2 wk followed
by 0, 320,160, or 80 mg/kg-d (TWA doses
of 0, 79.6,147.8, or 256.7 mg/kg-d for
males and 0, 82.4,136.3, or 276.4 mg/kg-d
for females)b
Diet
13 wk
Doses
0
79.6
147.8
256.7
Mortality (incidence)
M
0/10
0/10
0/10
4/10*
F
0/11
0/12
0/10
2/12
Cholakis et al. (1980)
Rats, F344,10/sex/group
88.6% pure, with 9% HMX and 2.2% water
as contaminants; ~200 nm particle size
0,10,14, 20, 28, or 40 mg/kg-d
Diet
13 wk
Doses
0
10
14
20
28
40
Mortality (incidence)
M
0/10
0/10
0/10
0/10
0/10
0/10
F
1/10
(accidental
death)
0/10
0/10
0/10
0/10
0/10
Cholakis et al. (1980)
Rats, CD, two-generation study;
F0: 22/sex/group; Fl: 26/sex/group;
F2: 10/sex/group
88.6% pure, with 9% HMX and 2.2% water
as contaminants; ~200 nm particle size
F0 and Fl parental animals: 0, 5,16, or
50 mg/kg-d
Diet
F0 exposure: 13 wk premating, and during
mating, gestation, and lactation of Fl;
Fl exposure: 13 wk after weaning, and
during mating, gestation, and lactation of
F2; F2 exposure: until weaning
Doses
0
5
16
50
Mortality in F0 adults (incidence^
M (F0)
0/22
0/22
0/22
2/22
F (F0)
0/22
0/22
0/22
6/22
M + F
(F0)
0/44
0/44
0/44
8/44*
Crouse et al. (2006)
Rats, F344,10/sex/group
99.99% pure
0, 4, 8,10,12, or 15 mg/kg-d
Gavage
13 wk
Doses
0
4
8
10
12
15
Mortality (incidence)
M
0/10
0/10
1/10
3/10
2/10
3/10
F
0/10
0/10
1/10
2/10
5/10
4/10
C-44

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-10. Evidence pertaining to mortality in animals3 (continued)
Reference and study design
Results
Levine et al. (1990); Levine et al. (1981a);
Levine et al. (1981b)d
Rats, F344,10/sex/group; 30/sex for
control
84.7 ± 4.7% purity, ~10% HMX, median
particle diameter 20 urn, ~90% of particles
<66 urn
0,10, 30,100, 300, or 600 mg/kg-d
Diet
13 wk
Doses
0
10
30
100
300
600
Mortality (incidence)e
M
0/30
0/10
0/10
8/10
10/10
10/10
F
0/30
1/10
0/10
5/10
10/10
10/10
von Oettingen et al. (1949)
Rats, sex/strain not specified, 20/group
90-97% pure, with 3-10% HMX; particle
size not specified
0,15, 25, or 50 mg/kg-d
Diet
13 wk
Doses
0
15
25
50
Mortality (incidence)

0/20
1/20
(probably
not related
to RDX)
8/20
8/20
Hart (1974)
Dogs, beagle, 3/sex/group
Premix with ground dog chow containing
20 mg RDX/g-chow, 60 g of dog food; purity
and particle size not specified
0, 0.1,1, or 10 mg/kg-d
Diet
13 wk
Doses
0
0.1
1
10
Mortality (incidence)
M
0/3
0/3
1/3
(not related to
RDX)
0/3
F
0/3
0/3
0/3
0/3
Martin and Hart (1974)
Monkeys, cynomolgus or rhesus/
3/sex/group
Purity and particle size not specified
0, 0.1,1, or 10 mg/kg-d
Gavage
13 wk
Doses
0
0.1
1
10
Mortality (incidence)
M
0/3
0/3
0/3
0/3
F
0/3
0/3
0/3
1/3
(animal exhibited
neurological
effects; euthanized)
von Oettingen et al. (1949)
Dogs, breed not specified, 5 females/group
(control); 7 females/group (exposed)
90-97% pure, with 3-10% HMX; particle
not specified
0 or 50 mg/kg-d
Diet
6 d/wk for 6 wk
Doses
0
50
Mortality (incidence)
F
0/5
1/7
C-45

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-10. Evidence pertaining to mortality in animals3 (continued)
Reference and study design
Results
MacPhail etal. (1985)
Rats, Sprague-Dawley derived CD,
8-10 males or females/group
Purity 84 ± 4.7%; <66 urn particle size
0,1, 3, or 10 mg/kg-d
Gavage
30 d
No mortality was reported (incidence data were not provided).
Cholakis et al. (1980)
Rabbits, New Zealand White,
11-12 pregnant females/group
88.6% pure, with 9% HMX and 2.2% water
as contaminants; ~200 urn particle size
0, 0.2, 2.0, or 20 mg/kg-d
Diet
GDs 7-29
Doses
0
0.2
2.0
20
Mortality (incidence)
F
0/11
0/11
0/11
0/12
Cholakis et al. (1980)
Rats, F344, 24-25 females/group
88.6% pure, with 9% HMX and 2.2% water
as contaminants
0, 0.2, 2.0, or 20 mg/kg-d
Gavage
GDs 6-19
Doses
0
0.2
2.0
20
Mortality (incidence)
F
0/24
0/24
0/24
5/24
(1 rat accidentally
killed; removed from
analysis)
Angerhofer et al. (1986)
(Range-finding study)
Rats, Sprague-Dawley, 6 pregnant
females/group
Purity 90%; 10% HMX and 0.3% acetic acid
occurred as contaminants
0,10, 20, 40, 80, or 120 mg/kg-d
Gavage
GDs 6-15
Doses
0
10
20
40
80
120
Mortality (incidence)
F
0/6
0/6
0/6
6/6
6/6
6/6
C-46

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-10. Evidence pertaining to mortality in animals3 (continued)
Reference and study design
Results
Angerhofer et al. (1986)
Rats, Sprague-Dawley, 39-51 mated
females/group
Purity 90%; 10% HMX and 0.3% acetic acid
occurred as contaminants
0, 2, 6, or 20 mg/kg-d
Gavage
GDs 6-15
Doses
0
2
6
20
Mortality (incidence)
F
0/39
1/40
1/40
16/51
Not reported whether
deaths at 2 and 6 mg/kg-d
related or likely related to
RDX exposure
HMX = octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine; TWA = time-weighted average.
^Statistically significant (p < 0.05) based on analysis by the study authors.
aThe 2-year rat study by Hart (1976) was not included in this evidence table because a malfunctioning heating
system incident resulted in the premature deaths of 59 animals on Study Days 75-76 across groups, thereby
confounding mortality findings.
bDoses were calculated by the study authors.
cData for male and female rats were combined for statistical analysis.
dLevine et al. (1981a) is a laboratory report of a 13-week study of RDX in F344 rats; two subsequently published
papers (Levine et al., 1990; Levine et al., 1981b) present subsets of the data provided in the full laboratory
report.
eAnimals receiving 300 mg/kg-d died by Week 3 of the study; animals receiving 600 mg/kg-d died by Week 1 of
the study.
the species of monkey used in this study was inconsistently reported in the study as either cynomolgus (in the
methods section) or rhesus (in the summary).
Following chronic dietary exposure, an increased rate of mortality in male F344 rats at
40 mg/kg-day was largely attributed to RDX-related effects on the kidneys (Levine etal.. 1983)4
(see further discussion in Section 1.2.2). Mice were less sensitive than rats with respect to
mortality following a similar 2-year exposure to RDX. After the high dose was reduced from 175 to
100 mg/kg-day at Week 11 in a 2-year dietary study in B6C3Fi mice because of high mortality, the
mortality curve at 100 mg/kg-day in surviving mice was not significantly different from the control
group for the duration of the 2-year study (Lish etal.. 1984bl. The investigators did not identify the
probable cause of death at 175 mg/kg-day.
Increased rates of mortality were also observed in experimental animals that ingested RDX
for durations up to 6 months fLish etal.. 1984b: Levine etal.. 1983: Levine etal.. 1981a: Cholakis et
al.. 1980: von Oettingen etal.. 19491. The most detailed data on RDX-related mortality come from a
90-day gavage study in F344 rats by Crouse etal. (20061. Across groups of rats exposed to
4Deaths in high-dose (40 mg/kg-day) male rats were reported beginning around month 3 to 4 [estimated from
Volume 1, Figure 1 in Levine et al. (1983)1; the cause of death in rats that died prior to 6 months on study was
generally not determined (Levine et al., 1983). Survival rates in both male and female rats at doses <8 mg/kg-day
RDX were similar to the control [see Table 4 in Levine et al. (1983)1.
C-47

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
8-15 mg/kg-day RDX, preterm deaths occurred in male rats as early as Days 26-78 and in female
rats as early as Days 8-16 flohnson. 2015: Crouse etal.. 20061. Treatment-related mortality was
also observed in the dams of rats exposed gestationally by gavage at doses ranging from 20 to
120 mg/kg-day (Angerhofer et al.. 1986: Cholakis etal.. 19801. Deaths were also reported in one of
40 pregnant dams in both 2 and 6 mg/kg-day groups in the rat developmental toxicity study by
Angerhofer etal. (19861.
In general, the evidence suggests that mortality occurs at lower doses in rats than in mice
[e.g., comparison of rates from the 2-year dietary studies in mice by Lish et al. f!984bl and in rats
by Levine etal. f!9831] and at lower doses following gavage administration than dietary
administration [e.g., comparison of rates from the 13-week rat studies using gavage fCrouse etal..
20061 and dietary (Levine etal.. 198 lal administration]. An RDX formulation with a larger particle
size (e.g., ~200 |im) (Cholakis etal.. 19801. which could potentially reduce the ability of RDX to
enter the bloodstream, appears to produce less mortality than formulations with finer particle sizes
(e.g., median particle diameter of 20 |im) (Levine etal.. 1981al. There is evidence that mortality
may be associated with nervous system effects; several investigators reported that unscheduled
deaths were frequently preceded by convulsions or seizures fCrouse etal.. 2006: Levine etal.. 1983:
Cholakis etal.. 19801. In a number of studies, treatment-related mortality was observed at doses as
low as those associated with nervous system effects (Crouse etal.. 2006: Angerhofer et al.. 1986:
Levine etal.. 1983: Levine etal.. 1981a: Cholakis etal.. 1980: von Oettingenetal.. 19491. The
evidence for an association between nervous system effects and mortality is discussed in more
detail in Section 1.2.1, Nervous System Effects.
In humans, there were no reports of mortality in case reports involving workers exposed to
RDX during manufacture or in individuals exposed acutely as a result of accidental or intentional
ingestion (see Appendix C, Section C.2).
C.3.2. Other Noncancer Effects
There are some reports of RDX inducing systemic effects in several organs/systems,
including the eyes and the cardiovascular, musculoskeletal, immune, GI, hematological, and male
reproductive systems, and body weight However, there is less evidence for these effects compared
to organ systems described in Section 1.2. Overall, at the present time, there is inadequate
information to identify these other systemic effects as human hazards of RDX exposure. Summaries
of the evidence for other systemic effects in humans and animals are shown in Tables C-ll to C-18,
respectively. Experimental animal studies are ordered in the evidence table by type of effect, and
then by duration of exposure and by species.
Ocular Effects
There are no reports of ocular effects in human case reports or epidemiological studies. In
experimental animals, evidence of ocular effects comes from cataract findings in one 2-year
bioassay (see Table C-ll). Specifically, the incidence of cataracts was 73% in female F344 rats
C-48

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
exposed to 40 mg/kg-day RDX in the diet for 2 years, compared with 32% in the control group
fLevine etal.. 19831. After 76 weeks of exposure, the incidence of cataracts in female rats at
40 mg/kg-day (23%) was also elevated compared to controls (6%). The incidence of cataracts was
not increased in RDX-exposed male rats in the same study (Levine etal.. 19831. and other studies
have not observed ocular effects associated with RDX exposure. Only two rats (dose groups not
reported) were observed to have mild cataracts in a 90-day study of male and female F344 rats
exposed to RDX at doses up to 15 mg/kg-day by gavage; however, the study authors noted that
these observations are common in F344 rats at 4 months of age and should not be attributed to
treatment fCrouse etal.. 20061. Furthermore, cataracts were not observed in male or female F344
rats exposed to 40 mg/kg-day RDX by diet for 90 days (Cholakis etal.. 19801 or in male or female
B6C3Fi mice exposed to RDX in the diet for 2 years at doses up to approximately 100 mg/kg-day
(Lish etal.. 1984b). A statistically significant increase in the incidence of cataracts in male mice was
initially noted by Lish etal. f!984bl. but was not confirmed when mice used for orbital bleedings
were excluded from the analysis, suggesting that the effect was not treatment related. No ocular
effects were observed in monkeys exposed by gavage for 90 days at doses up to 10 mg/kg-day
fMartin and Hart. 19741.
In summary, the incidence of cataracts was statistically significantly increased in high-dose
female rats in one chronic oral study; however, this finding was not reproduced in other subchronic
and chronic studies in rats or mice. There is insufficient information to assess ocular toxicity
following exposure to RDX.
C-49

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-ll. Evidence pertaining to ocular effects in animals
Reference and study design
Results
Lish et al. (1984b)
Mice, B6C3Fi, 85/sex/group; interim
sacrifices (10/sex/group) at 6 and 12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles <66 nm
0,1.5, 7.0, 35, or 175/100 mg/kg-d (high
dose reduced to 100 mg/kg-d in Week 11
due to excessive mortality)
Diet
2 yr
Doses
0
1.5
7.0
35
175/100
Cataracts; 103 wk (incidence)a
M
2/47
2/41
0/41
2/37
2/16
F
2/50
1/37
6/52
0/46
1/26
Levine et al. (1983)
Rats, F344, 75/sex/group; interim
sacrifices (10/sex/group) at 6 and 12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles <66 nm
0, 0.3,1.5, 8.0, or 40 mg/kg-d
Diet
2 yr
Doses
0
0.3
1.5
8.0
40
Cataracts; 103 wk (incidence)
M
8/40
6/39
6/31
8/35
2/6
F
14/44
4/48
11/44
8/43
22/30*
Cholakis et al. (1980)
Rats, F344,10/sex/group
88.6% pure, with 9% HMX and 2.2% water
as contaminants; ~200 nm particle size
0,10,14, 20, 28, or 40 mg/kg-d
Diet
13 wk
No ocular effects were observed (gross examination of eye was
performed in all animals, and microscopic examination was
performed in control and 40 mg/kg-d animals).
Crouse et al. (2006)
Rats, F344,10/sex/group
99.99% pure
0, 4, 8,10,12, or 15 mg/kg-d
Gavage
13 wk
No ocular effects were observed (ophthalmic examinations were
performed in all animals within 1 wk of sacrifice, and microscopic
examination of the eye was performed in control and
15 mg/kg-d animals).
Martin and Hart (1974)
Monkeys, cynomolgus or rhesus,b
3/sex/group
Purity of test material not specified
0, 0.1,1, or 10 mg/kg-d
Gavage
13 wk
No ocular effects were observed (ophthalmoscopic examination was
performed at the end of exposure).
HMX = octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine.
^Statistically significantly different compared to the control, as determined by study authors (p < 0.05).
incidence counts exclude individuals from which blood was obtained via the orbital sinus.
bThe species of monkey used in this study was inconsistently reported in the study as either cynomolgus (in the
Methods section) or rhesus (in the Summary).
C-50

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Cardiovascular Effects
Human evidence for cardiovascular effects is limited to case reports that include
observations of transient arterial hypertension in male workers following inhalation of RDX during
manufacturing (Barsotti and Crotti. 19491. sinus tachycardia, and in one instance, premature
ventricular beats in five men following accidental ingestion of RDX at 37-250 mg/kg body weight
(Kuciikardali etal.. 20031 (see Appendix C, Section C.2).
Inconsistent observations of cardiovascular effects have been reported in animal studies
(see Table C-12). An increase in the relative heart-to-body weight ratio was observed at the highest
dose tested in B6C3Fi mice (male: 13%; female: 17%) and F344 rats (male: 22%; female: 15%)
following chronic dietary administration of RDX (Lish etal.. 1984b: Levine etal.. 19831: however,
these doses also resulted in reductions in body weight in both males and females. Dose-related
decreases in absolute heart weight in rats were reported in some subchronic (dietary) studies
(Levine etal.. 1990: Levine etal.. 1981a. b; Cholakis etal.. 19801. whereas little change or modest
increases in absolute heart weight were observed in other subchronic studies in rats or mice
fCrouse etal.. 2006: Cholakis etal.. 19801. A subchronic study in male dogs reported a 31%
increase in absolute heart weight at the highest dose tested (10 mg/kg-day) fHart. 19741.
Evidence for changes in histopathology associated with RDX exposure is limited to findings
of an increased incidence of focal myocardial degeneration in female rats (6/10 versus 2/10,
respectively) and male mice (5/10 versus 0/10, respectively) compared with controls following
exposure to RDX in the diet for 90 days (Cholakis etal.. 19801. With the exception of one male
mouse, the severity of the lesion was characterized as minimal. In each study, the finding of
myocardial degeneration was limited to one sex and to the high-dose group only; the high dose in
the male mouse study caused 40% mortality. Other studies in monkeys fMartin and Hart. 19741
and rats (von Oettingen etal.. 19491 reported no observable cardiovascular effects.
In summary, evidence for cardiovascular effects associated with RDX exposure consists of
two case reports of cardiovascular effects following acute exposure, inconsistent findings of
changes in heart weight in experimental animals, and one report of minimal histopathological
changes in a 90-day rat study that was not confirmed in other toxicity studies. There is insufficient
information to assess cardiovascular effects following exposure to RDX.
C-51

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-12. Evidence pertaining to cardiovascular effects in animals
Reference and study design
Results
Lish et al. (1984b)
Mice, B6C3Fi, 85/sex/group; interim
sacrifices (10/sex/group) at 6 and
12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles
<66 urn
0,1.5, 7.0, 35, or 175/100 mg/kg-d (high
dose reduced to 100 mg/kg-d in Week
11 due to excessive mortality)
Diet
2 yr
Doses
0
1.5
7.0
35
175/100
Absolute heart weight; 104 wk (percent change compared to control)
M
0%
4%
4%
5%
7%
F
0%
1%
5%
2%
-5%
Relative heart-to-body weight; 104 wk (percent change compared to
control)
M
0%
7%
5%
5%
13%*
F
0%
0%
6%
4%
17%*
Body weight was significantly lower at termination in males and
females exposed to 175/100 mg/kg-d (-5 and -19%, respectively).
Hart (1976)
Rats, Sprague-Dawley, 100/sex/group
Purity and particle size not specified
0,1.0, 3.1, or 10 mg/kg-d
Diet
2 yr
Doses
0
1.0
3.1
10
Myocardial fibrosis (percent incidence; number not reported)
M
20%
—
—
5%
F
5%
—
—
1%
Endocardial disease (percent incidence; number not reported)
M
1%
—
—
3%
F
0%
—
—
0%
Absolute heart weight; 104 wk (percent change compared to control)
M
0%
-6%
-2%
-5%
F
0%
13%
3%
15%
Relative heart-to-body weight; 104 wk (percent change compared to
control)
M
0%
-2%
4%
1%
F
0%
23%
13%
27%
Levine et al. (1983)
Rats, F344, 75/sex/group; interim
sacrifices (10/sex/group) at 6 and
12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles
<66 nm
0, 0.3,1.5, 8.0, or 40 mg/kg-d
Diet
2 yr
Doses
0
0.3
1.5
8.0
40
Absolute heart weight; 104 wk (percent change compared to control)
M
0%
3%
-2%
-2%
1%
F
0%
-1%
0%
-4%
-3%
Relative heart-to-body weight; 104 wk (percent change compared to
control)
M
0%
2%
6%
0%
22%
F
0%
-2%
3%
-1%
15%
C-52

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-12. Evidence pertaining to cardiovascular effects in animals
(continued)
Reference and study design
Results
Cholakis et al. (1980)
Mice, B6C3Fi, 10-12/sex/group
88.6% pure, with 9% HMX and 2.2%
water as contaminants; ~200 urn
particle size
Experiment 1: 0,10,14, 20, 28, or
40 mg/kg-d
Diet
13 wk
Doses
0
10
14
20
28
40
Absolute heart weight (percent change compared to control)
M
0%
—
—
—
7%
7%
F
0%
—
—
—
0%
0%
Relative heart weight (percent change compared to control)
M
0%



6%
0%
F
0%



-4%
0%
Experiment 2: 0, 40, 60, or 80 mg/kg-d
for 2 wk followed by 0, 320,160, or
80 mg/kg-d (TWA doses of 0, 79.6,
147.8, or 256.7 mg/kg-d for males and
0, 82.4,136.3, or 276.4 mg/kg-d for
females)3
Diet
13 wk
Doses
0
80
160
320
Focal myocardial degeneration (incidence)
M#
0/10


5/10**
F+
0/11


2/11
Absolute heart weight (percent change compared to control)
M
0%
0
0
8%
F
0%
0%
0%
8%
Relative heart-to-body weight (percent change compared to control)
M
0%
0%
-2%
6%
F
0%
0%
-2%
2%
"Includes one affected and three unaffected animals that died
prematurely.
includes one unaffected animal that died prematurely.
*Minimal severity in four rats, moderate in one.
Cholakis et al. (1980)
Rats, F344,10/sex/group
88.6% pure, with 9% HMX and 2.2%
water as contaminants; ~200 nm
particle size
0,10,14, 20, 28, or 40 mg/kg-d
Diet
13 wk
Doses
0
10
14
20
28
40
Focal myocardial degeneration, minimal severity (incidence)
M
3/10
—
—
—

1/10
F
2/10
—
—
—

6/10
Absolute heart weight (percent change compared to control)
M
0%
—
—
—
0
*
0s-
00
1
F
0%
—
—
—
-6%
-n%*
Relative heart-to-body weight (percent change compared to control)
M
0%



3%
0%
F
0%



-3%
-8%
Relative heart-to-brain weight (percent change compared to control)
C-53

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-12. Evidence pertaining to cardiovascular effects in animals
(continued)
Reference and study design
Results

M
0%



-4%
-10%*
F
0%



-5%
-11%*
Cholakis et al. (1980)
Rats, CD, two-generation study; F0:
22/sex/group; Fl: 26/sex/group; F2:
10/sex/group
88.6% pure, with 9% HMX and 2.2%
water as contaminants; ~200 urn
particle size
F0 and Fl parental animals: 0, 5,16, or
50 mg/kg-d
Diet
F0 exposure: 13 wk premating, and
during mating, gestation, and lactation
of Fl; Fl exposure: 13 wk after
weaning, and during mating, gestation,
and lactation of F2; F2 exposure: until
weaning
No cardiac effects were observed (microscopic examination of heart
was performed in randomly selected F2 animals).
Doses
0
5
16
50
Absolute heart weight (percent change compared to control)
F2 M
0%
3.2%
-6.5%
—
F2 F
0%
15%
-3.7%

Crouse et al. (2006)
Rats, F344,10/sex/group
99.99% pure
0, 4, 8,10,12, or 15 mg/kg-d
Gavage
13 wk
Doses
0
4
8
10
12
15
Cardiomyopathy (incidence)
M
2/10
—
—


3/8
F
0/10
—
—


1/6
Absolute heart weight (percent change compared to control)
M
0%
-2%
-7%
1
1
11%
F
0%
-2%
0%
8%
7%
6%
Relative heart-to-body weight (percent change compared to control)
M
0%
4%
2%
1%
-1%
8%
F
0%
-2%
-7%
-6%
-9%
-16%*
Levine et al. (1990); Levine et al.
(1981a); Levine et al. (1981b)b
Rats, F344,10/sex/group; 30/sex for
control
84.7 ± 4.7% purity, ~10% HMX, median
particle diameter 20 nm, ~90% of
particles <66 nm
0,10, 30,100, 300, or 600 mg/kg-d
Diet
13 wk
All animals in the 300 and 600 mg/kg-d groups died prior to study
termination.
Doses
0
10
30
100
300
600
Chronic focal myocarditis (incidence)
M
8/30
8/10
6/10
1/10
1/10
0/10
F
8/30
3/10
1/10
1/10
1/10
1/9
Absolute heart weight (percent change compared to control)
M
0%
-2%
-10%
-15%

—
F
0%
-3%
0%
-5%

—
C-54

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-12. Evidence pertaining to cardiovascular effects in animals
(continued)
Reference and study design
Results

Relative heart-to-body weight (percent change compared to control)
M
0%
2%
-4%
3%

—
F
0%
-2%
0%
-3%

—
von Oettingen et al. (1949)
Rats (sex/strain not specified); 20/group
Purity and particle size not specified
0,15, 25, or 50 mg/kg-d
Diet
13 wk
The study authors reported that there were no cardiac effects
(microscopic examination of the heart was performed in all rats; data
were not shown).
Hart (1974)
Dogs, beagle, 3/sex/group
Premix with ground dog chow
containing 20 mg RDX/g-chow, 60 g of
dog food; purity and particle size not
specified
0, 0.1,1, or 10 mg/kg-d
Diet
13 wk
Doses
0
0.1
1
10
Focal hyalinization of the heart (incidence)
M
0/3
—
—
0/3
F
0/3
—
—
1/3
Absolute heart weight (percent change compared to control)
M
0%


31%
F
0%


5.7%
Martin and Hart (1974)
Monkeys, cynomolgus or rhesus,0
3/sex/group
Purity of test material not specified
0, 0.1,1, or 10 mg/kg-d
Gavage
13 wk
Doses
0
0.1
1
10
Myocarditis (percent change compared to control)
M
1/3
—

1/3
F
0/3
—

0/3
Absolute heart weight (percent change compared to control)
M
0%
7%
1
5%
F
0%
10%
12%
-12%
HMX = octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine; TWA = time-weighted average.
^Statistically significantly different compared to the control, as determined by study authors (p < 0.05).
aDoses were calculated by the study authors.
bLevine et al. (1981a) is a laboratory report of a 13-week study of RDX in F344 rats; two subsequently published
papers (Levine et al., 1990; Levine et al., 1981b) present subsets of the data provided in the full laboratory report.
cThe species of monkey used in this study was inconsistently reported in the study as either cynomolgus (in the
Methods section) or rhesus (in the Summary).
Note: A dash ("—") indicates that the study authors did not measure or report a value for that dose group.
Musculoskeletal Effects
Evidence of musculoskeletal effects in humans consists of case reports that include
observations of muscle twitching, myalgia/muscle soreness, and muscle injury as indicated by
C-55

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
elevated levels of aspartate aminotransferase (AST), creatine phosphokinase, and myoglobinuria
fTestud etal.. 2006: Kiiciikardali etal.. 2003: Hettand Fichtner. 2002: Hollander and Colbach. 1969:
Stone etal.. 1969: Merrill. 19681 (see Appendix C, Section C.2). Histological evaluations of
musculature or skeletal tissue did not reveal any alterations in mice (Lish etal.. 1984b 1 or rats
(Levine etal.. 1983: Hart. 19761 following chronic oral exposure to RDX, in mice and rats following
subchronic exposure (Cholakis etal.. 19801. or in dogs following a 90-day dietary exposure (Hart.
19741. In summary, based on a limited number of case reports of muscle involvement involving
acute exposure to RDX, there is insufficient information to assess musculoskeletal effects following
exposure to RDX.
Immune System Effects
RDX is structurally similar to various drugs known to induce the autoimmune disorder
systemic lupus erythematosus (SLE). Based on the initial identification of three cases of SLE at
one U.S. Army munitions plant, further investigation was conducted on a population of
69 employees at five U.S. Army munitions plants with potential exposure to RDX fHathawav and
Buck. 19771: no additional cases of SLE were identified. Increased white blood cell (WBC) counts
have been reported in some case reports of individuals (troops during the Vietnam War) who
ingested or inhaled RDX or C-4 (91% RDX) (Knepshield and Stone. 1972: Hollander and Colbach.
1969: Stone etal.. 1969: Merrill. 19681.
In animal studies (see Table C-13), increased WBC count in female rats following
subchronic dietary exposure to RDX was the only dose-related immune effect reported fLevine et
al.. 1990: Levine etal.. 1981a. b); WBC counts in male rats were unaffected. Conversely, decreased
WBC counts were reported in male and female rats in a 2-year study (Hart. 19761. Changes in
spleen weights were observed across studies, but the responses were not consistent and did not
appear to be dose related. For example, in 90-day studies, Cholakis etal. (19801 identified a
statistically significant decrease in absolute spleen weight in female F344 rats at 40 mg/kg-day,
while Crouse etal. (20061 observed a statistically significant increase in spleen weight at
>10 mg/kg-day. Across studies, there was no significant or dose-dependent pattern of response to
suggest that the WBC changes reflect RDX-induced immunotoxicity. No dose-related immune
effects from oral exposure to RDX were observed in other animal studies, including a 90-day study
in F344 rats that evaluated structural measures of immunotoxicity (including RBC and WBC
populations, proportion of cell surface markers, cellularity in proportion to organ weight, B and
T cells in the spleen, and CD4/CD8 antigens of maturing lymphocytes in the thymus) (Crouse etal..
20061. Routine clinical and histopathology evaluations of immune-related organs in a
two-generation study in rats f Cholakis etal.. 19801 and chronic studies in rats fLevine etal.. 19831
and mice fLish etal.. 1984b 1 provide no evidence of immunotoxicity associated with oral (dietary)
exposure to RDX. None of the available studies included evaluation of more sensitive measures of
functional immune system changes that would be more likely to detect immunosuppression,
unintended immune stimulation, autoimmunity, or dysregulated inflammation.
C-56

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
One special case of an immune response potentially relevant to RDX is neuroinflammation
that may result from recurrent seizures. Neuroinflammation is a key characteristic of most
neurological conditions, including seizure and epilepsy fEvo etal.. 2017: Dev etal.. 20161.
Hypothetically, RDX-induced seizures could indirectly trigger acute immune and inflammatory
responses from microglia within the brain; additionally, chronic neuroinflammation may result
from recurrent seizures. While this hypothesized relationship between an inflammatory response
and seizures may be relevant to the convulsant effects of RDX, the relationship to the less severe
manifestations of RDX neurotoxicity is unclear. Further, there is no available information on
neuroinflammatory responses after exposure to RDX.
In summary, evidence for immunotoxicity associated with RDX exposure is limited to
findings from several case reports indicating increased WBC counts in Vietnam War troops who
ingested or inhaled RDX and changes in WBC counts in rats in two studies (Levine etal.. 1981a. b;
Hart. 19761 that were not consistent in direction or necessarily across sexes, and were not
supported by dose-dependent patterns of response across the RDX database. Finally, no
dose-related immune effects were observed in a 90-day rat study that evaluated structural, but not
functional, immunotoxicity endpoints fCrouse etal.. 20061. Therefore, there is insufficient
information to assess immunotoxicity following exposure to RDX.
C-57

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-13. Evidence pertaining to immune effects in animals
Reference and study design
Results
Lish et al. (1984b)
Mice, B6C3Fi, 85/sex/group; interim
sacrifices (10/sex/group) at 6 and
12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles
<66 urn
0,1.5, 7.0, 35, or 175/100 mg/kg-d (high
dose reduced to 100 mg/kg-d in Week
11 due to excessive mortality)
Diet
2 yr
No immune effects were observed with routine hematology, clinical
chemistry, or histopathology evaluations.
Doses
0
1.5
7.0
35
175/100
WBC count; 105 wk (percent change compared to control)
M
0%
-13%
-8%
-16%
-30%
F
0%
12%
39%*
28%
0%
Absolute spleen weight; 105 wk (percent change compared to control)
M
0%
24%
31%
-10%
-28%
F
0%
4%
15%
-17%
16%
Relative spleen weight; 105 wk (percent change compared to control)
M
0%
26%
32%
-11%
-21%
F
0%
4%
15%
-17%
44%
Hart (1976)
Rats, Sprague-Dawley, 100/sex/group
Purity and particle size not specified
0, 1.0, 3.1, or 10 mg/kg-d
Diet
2 yr
Doses
0
1.0
3.1
10
WBC count; 104 wk (percent change compared to control)
M
0%
-13%
-22%*
-34%*
F
0%
5%
-32%*
-12%
Absolute spleen weight; 104 wk (percent change compared to control)
M
0%
-11%
-16%
-4%
F
0%
58%
8%
37%
Relative spleen weight; 104 wk (percent change compared to control)
M
0%
-11%
-14%
1%
F
0%
11%
19%
55%
C-58

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-13. Evidence pertaining to immune effects in animals (continued)
Reference and study design
Results
Levine et al. (1983)
Rats, F344, 75/sex/group; interim
sacrifices (10/sex/group) at 6 and
12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles
<66 urn
0, 0.3,1.5, 8.0, or 40 mg/kg-d
Diet
2 yr
No immune effects were observed with routine hematology, clinical
chemistry, or histopathology evaluations.
Doses
0
0.3
1.5
8.0
40
WBC count; 105 wk (percent change compared to control)
M
0%
-11%
103%a
184%a
15%
F
0%
7%
12%
354%a
251%a
Absolute spleen weight; 105 wk (percent change compared to control)
M
0%
5%
-10%
-32%
-49%
F
0%
-28%
-44%
-35%
17%
Relative spleen weight; 105 wk (percent change compared to control)
M
0%
9%
4%
-29%
-38%
F
0%
-34%
-45%
-36%
9%
Cholakis et al. (1980)
Mice, B6C3Fi, 10-12/sex/group
88.6% pure, with 9% HMX and 2.2%
water as contaminants; ~200 urn
particle size
Experiment 1: 0,10,14, 20, 28, or
40 mg/kg-d
Diet
13 wk
Doses
0
10
14
20
28
40
Absolute spleen weight (percent change compared to control)
M
0%
—
—
—
18%
13%
F
0%
—
—
—
-2%
-8%
Relative spleen weight (percent change compared to control)
M
0%



24%
14%
F
0%



-3%
-5%
Experiment 2: 0, 40, 60, 80 mg/kg-d for
2 wk followed by 0, 320,160, or
80 mg/kg-d (TWA doses of 0, 79.6,
147.8, or 256.7 mg/kg-d for males and
0, 82.4,136.3, or 276.4 mg/kg-d for
females)b
Diet
13 wk
Doses
0
80
160
320
WBC count (percent change compared to control)
M
0%
-27%
-12%
30%
F
0%
-17%
3%
-3%
Absolute spleen weight (percent change compared to control)
M
0%
17%
0%
-17%
F
0%
-22%
0%
0%
Relative spleen weight (percent change compared to control)
M
0%
25%
5%
0%
F
0%
-12%
0%
-3%
C-59

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-13. Evidence pertaining to immune effects in animals (continued)
Reference and study design
Results
Cholakis et al. (1980)
Rats, F344,10/sex/group
88.6% pure, with 9% HMX and 2.2%
water as contaminants; ~200 urn
particle size
0,10,14, 20, 28, or 40 mg/kg-d
Diet
13 wk
Doses
0
10
14
20
28
40
WBC count (percent change compared to control)
M
0%
—
—
—
-12%
7%
F
0%
—
—
—
17%
30%
Absolute spleen weight (percent change compared to control)
M
0%
—
—
—
2%
-4%
F
0%
—
—
—
-10%
-12%*
Relative spleen weight (percent change compared to control)
M
0%



5%
5%
F
0%



-8%
-8%
Cholakis et al. (1980)
Rats, CD, two-generation study; F0:
22/sex/group; Fl: 26/sex/group;
F2: 10/sex/group
88.6% pure, with 9% HMX and 2.2%
water as contaminants; ~200 urn
particle size
F0 and Fl parental animals: 0, 5,16, or
50 mg/kg-d
Diet
F0 exposure: 13 wk premating, and
during mating, gestation, and lactation
of Fl; Fl exposure: 13 wk after
weaning, and during mating, gestation,
and lactation of F2; F2 exposure: until
weaning
No immune effects were observed upon routine histopathology
evaluation.
C-60

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-13. Evidence pertaining to immune effects in animals (continued)
Reference and study design
Results
Crouse et al. (2006)
Rats, F344,10/sex/group
99.99% pure
0, 4, 8,10,12, or 15 mg/kg-d
Gavage
13 wk
No effects were observed on thymus or spleen histology, RBC or WBC
populations, or lymphocyte populations.
Doses
0
4
8
10
12
15
WBC count (percent change compared to control)
M
0%
-5%
-12%
-7%
1%
-3%
F
0%
22%
45%
12%
52%
29%
Absolute spleen weight (percent change compared to control)
M
0%
-3%
-6%
3%
1%
5%
F
0%
1%
8%
23%*
17%*
24%*
Relative spleen weight (percent change compared to control)
M
0%
3%
4%
7%
-1%
2%
F
0%
1%
0%
6%
-1%
-2%
Absolute thymus weight (percent change compared to control)
M
0%
-1%
3%
-10%
-12%
-25%
F
0%
-7%
12%
19%
32%
19%
Relative thymus weight (percent change compared to control)
M
0%
-1%
3%
-10%
-12%
-25%
F
0%
-7%
4%
4%
12%
-6%
Levine et al. (1990); Levine et al.
(1981a); Levine et al. (1981b)c
Rats, F344,10/sex/group; 30/sex for
control
84.7 ± 4.7% purity, ~10% HMX, median
particle diameter 20 urn, ~90% of
particles <66 urn
0,10, 30,100, 300, or 600 mg/kg-d
Diet
13 wk
Data were not reported for rats in the 300 or 600 mg/kg-d dose groups
because all of the rats died before the 13-wk necropsy.
Doses
0
10
30
100
300
600
WBC count (percent change compared to control)
M
0%
4%
7%
15%
—
—
F
0%
23%*
24%*
62%*
—
—
Absolute spleen weight (percent change compared to control)
M
0%
-11%
-16%
-34%
—
—
F
0%
2%
12%
0%
—
—
Relative spleen weight (percent change compared to control)
M
0%
-9%
-12%
-21%

—
F
0%
2%
12%
3%

—
C-61

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-13. Evidence pertaining to immune effects in animals (continued)
Reference and study design
Results
von Oettingen et al. (1949)
Rats, sex/strain not specified, 20/group
90-97% pure, with 3-10% HMX;
particle size not specified
0,15, 25, or 50 mg/kg-d
Diet
13 wk
Doses
0
15
25
50
WBC count (percent change compared to control)
M
0%
-30%
7%
-6%
Hart (1974)
Dogs, beagle, 3/sex/group
Premix with ground dog chow
containing 20 mg RDX/g-chow, 60 g of
dog food; purity and particle size not
specified
0, 0.1,1, or 10 mg/kg-d
Diet
13 wk
Doses
0
0.1
1
10
WBC count (percent change compared to control)
M
0%
5%
2%
-19%
F
0%
-2%
24%
6%
Absolute spleen weight (percent change compared to control)
M
0%


123%
F
0%


-11%
Martin and Hart (1974)
Monkeys, cynomolgus or rhesus,d
3/sex/group
Purity of test material not specified
0, 0.1,1, or 10 mg/kg-d
Gavage
13 wk
Doses
0
0.1
1
10
WBC count (percent change compared to control)
M
0%
-32%
0%
-3%
F
0%
-38%
-1%
-41%
HMX = octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine; TWA = time-weighted average.
^Statistically significantly different compared to the control, as determined by study authors (p < 0.05).
Standard deviations accompanying the mean response in a given dose group were high, suggesting uncertainty in
the accuracy of the reported percent change compared to control.
bDoses were calculated by the study authors.
cLevine et al. (1981a) is a laboratory report of a 13-week study of RDX in F344 rats; two subsequently published
papers (Levine et al., 1990; Levine et al., 1981b) present subsets of the data provided in the full laboratory report.
dThe species of monkey used in this study was inconsistently reported in the study as either cynomolgus (in the
Methods section) or rhesus (in the Summary).
Note: A dash ("—") indicates that the study authors did not measure or report a value for that dose group.
Gastrointestinal Effects
Clinical signs of nausea and/or vomiting have been frequently identified in case reports of
accidental or intentional RDX poisonings, and have generally been concurrent with severe
neurotoxicity (Kasuske etal.. 2009: Davies etal.. 2007: Kiiciikardali etal.. 2003: Hettand Fichtner.
2002: Ketel and Hughes. 1972: Knepshield and Stone. 1972: Hollander and Colbach. 1969: Stone et
al.. 1969: Merrill. 1968: Kaplan etal.. 1965: Barsotti and Crotti. 19491 (see Appendix C, Section C.2).
C-62

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Additionally, Kiiciikardali etal. f20031 reported several cases of erosive gastroduodenitis in
individuals acutely exposed to RDX, suggesting severe irritation of the GI tract by direct contact
with RDX. In animal studies (see Table C-14), vomiting has also been observed following oral
exposure in swine (single-dose study) (Musick etal.. 20101. dogs fHart. 19741. and monkeys
(Martin and Hart. 19741. One subchronic oral (diet) rat study from the early literature reported
congestion of the GI tract at doses also associated with elevated mortality (von Oettingen et al..
19491: however, none of the subsequent subchronic or chronic animal studies reported histological
changes of the GI tract related to RDX administered via gavage or the diet fCrouse etal.. 2006: Lish
etal.. 1984b: Levine etal.. 1983: Hart. 1974: Martin and Hart. 19741.
In summary, evidence for GI tract effects associated with RDX exposure consists largely of
reports of nausea and vomiting in humans acutely exposed to RDX and similar reports of vomiting
in swine, dogs, and monkeys. There is some evidence that direct contact with RDX in the GI tract
may cause erosive gastroduodenitis in poisoning victims; however, histopathological changes were
not generally reported in experimental animals exposed to RDX in the diet. There is insufficient
information to assess gastrointestinal toxicity following exposure to RDX.
C-63

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-14. Evidence pertaining to gastrointestinal effects in animals
Reference and study design
Results
Lish et al. (1984b)
Mice, B6C3Fi, 85/sex/group; interim sacrifices
(10/sex/group) at 6 and 12 mos
89.2-98.7% pure, with 3-10% HMX as contaminant; 83-89%
of particles <66 urn
0,1.5, 7.0, 35, or 175/100 mg/kg-d (high dose reduced to
100 mg/kg-d in Week 11 due to excessive mortality)
Diet
2 yr
No Gl tract effects were observed as clinical signs
or on gross pathology or histopathology
examination.
Levine et al. (1983)
Rats, F344, 75/sex/group; interim sacrifices (10/sex/group) at
6 and 12 mos
89.2-98.7% pure, with 3-10% HMX as contaminant; 83-89%
of particles <66 nm
0, 0.3,1.5, 8.0, or 40 mg/kg-d
Diet
2 yr
No Gl tract effects were observed as clinical signs
or on gross pathology or histopathology
examination.
Crouse et al. (2006)
Rats, F344,10/sex/group
99.99% pure
0, 4, 8,10,12, or 15 mg/kg-d
Gavage
13 wk
No Gl tract effects were observed on gross
pathology or histopathology examination.
Increased salivation and blood stains around the
mouth were noted (affected doses and incidences
were not reported); it is not clear whether these
effects occurred in animals also experiencing
convulsions.
von Oettingen et al. (1949)
Rats (sex/strain not specified); 20/group
90-97% pure, with 3-10% HMX; particle size not specified
0,15, 25, or 50 mg/kg-d
Diet
13 wk
Congestion of the Gl tract was observed in 50 and
100 mg/kg-d rats that also exhibited mortality
(40%) and severe neurotoxicity.
Martin and Hart (1974)
Monkeys, cynomolgus or rhesus,3 3/sex/group
Purity of test material not specified
0, 0.1,1, or 10 mg/kg-d
Gavage
13 wk
Vomiting was observed more frequently in the 1
and 10 mg/kg-d groups compared to the control
or 0.1 mg/kg-d groups, although some episodes
occurred during the intubation procedure.
C-64

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-14. Evidence pertaining to gastrointestinal effects in animals
(continued)
Reference and study design
Results
Hart (1974)
Dogs, beagle, 3/sex/group
Premix with ground dog chow containing 20 mg RDX/g-chow,
60 g of dog food; purity and particle size not specified
0, 0.1,1, or 10 mg/kg-d
Diet
13 wk
Some nausea and vomiting were reported
(incidences and affected dose groups were not
reported).
HMX = octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine.
aThe species of monkey used in this study was inconsistently reported in the study as either cynomolgus (in the
Methods section) or rhesus (in the Summary).
Hematological Effects
Elevated prevalence odds ratios (ORs) for hematological abnormalities (i.e., neutropenia,
low platelet count, or macrocytosis; see Table C-15 for criteria used to define abnormal) were
observed in a case-control study of men (24 cases, 199 controls) exposed to RDX in ordnance
factories fWest and Stafford. 19971 (see Table C-15). The prevalence OR for an association between
RDX exposure and hematological abnormalities was 1.7 (95% confidence interval [CI] 0.7-4.2) for
men with >50 hours of low-intensity exposure (based on 22 cases), while the prevalence OR was
1.2 (95% CI 0.3-5.3) for men with >50 hours of high-intensity exposure (based on 2 cases). The
ORs from this study must be interpreted with caution given the small sample size, wide CIs, and
lack of identification of coexposures. No changes in hematological parameters (including
hemoglobin, hematocrit, and reticulocyte count) were observed in a cross-sectional epidemiologic
study of 69 workers exposed to RDX by inhalation (RDX exposure range: undetectable
[<0.01 mg/m3] to 1.6 mg/m3) fHathawav and Buck. 19771. Humans who ingested or inhaled
unknown amounts of RDX or C-4 (~91% RDX) for an acute duration displayed temporary
hematological alterations, including anemia, decreased hematocrit, hematuria, and
methemoglobinemia (Kasuske etal.. 2009: Kuctikardali etal.. 2003: Knepshield and Stone. 1972:
Hollander and Colbach. 1969: Stone etal.. 1969: Merrill. 19681. In other case reports, normal blood
counts were observed in accidentally exposed individuals (Testud et al.. 1996a: Goldberg etal..
1992: Woody etal.. 1986: Ketel and Hughes. 1972: Kaplan etal.. 19651 (see Appendix C,
Section C.2).
In animals, hematological alterations were observed following oral exposure in chronic and
subchronic studies in both sexes of rats (F344 or Sprague-Dawley) and B6C3Fi mice (see
Table C-16). Increases in platelet count were observed in male and female mice and rats in some
subchronic and chronic studies at doses ranging from 0.3 to 320 mg/kg-day fLish etal.. 1984b:
C-65

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Levine etal.. 1983: Cholakis etal.. 19801: however, changes were generally inconsistent across
studies and were not generally dose dependent Similarly, decreased hemoglobin levels/anemia
were observed in some chronic and subchronic studies (Levine etal.. 1983: Cholakis etal.. 1980:
von Oettingen etal.. 19491. particularly at doses >15 mg/kg-day, but trends in hemoglobin levels
across studies did not show a consistent relationship with dose. Other hematological parameters,
including WBC counts, reticulocyte counts, and hematocrit, showed conflicting results between
studies, marginal responses, or inconsistent changes with increasing dose. Other subchronic
studies in rats and dogs fCrouse etal.. 2006: Hart. 1974: von Oettingen et al.. 19491 did not identify
any changes in hematological parameters.
In summary, evidence for hematological effects associated with RDX exposure in humans
comes from several case reports that found transient fluctuations in hematological endpoints after
acute exposures. Hematological findings from a case-control study and a cross-sectional study
were inconsistent Incidences of anemia observed in human case reports may reflect coexposures
to trinitrotoluene (TNT); anemia has been reported in F344 rats exposed to TNT, but not RDX
fLevine etal.. 1990: Levine etal.. 1981a. b). The small number of cases in the case-control study
and exposed individuals in the cross-sectional study contribute to the difficulty in interpreting the
results across studies (see Table C-15). In general, animal studies of chronic and subchronic
durations showed no consistent, dose-related pattern of increase or decrease in hematological
parameters. There is insufficient information to assess hematological toxicity following exposure to
RDX.
C-66

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-15. Evidence pertaining to other noncancer effects (hematological) in
humans
Reference and study design
Results
Hematological effects
West and Stafford (1997) (United Kingdom)
Case-control study of ordnance factory
workers, 32 cases with abnormal and
322 controls with normal hematology test
drawn from 1991 study of 404 workers at
ammunitions plant; participation rate 97% of
cases, 93% of controls. Analysis limited to men
(29 cases, 282 controls). Analysis specific to
RDX: 22 low- and 2 high-intensity cases;
182 low- and 17 high-intensity controls.
Exposure measures: Exposure determination
based on employee interviews and job title
analysis; data included frequency (hrs/d, d/yr),
duration (yr), and intensity (low [1-10 ppm],
moderate [10-100 ppm], and high
[100-1,000 ppm], based on ventilation
considerations).
Effect measures: Hematology tests; blood
disorder defined as neutropenia (2.0 x 109/L),
low platelet count (<150 x 109/L), or
macrocytosis (mean corpuscular volume =
99 fL or >6% macrocytes).
Analysis: Unadjusted OR.
Hematological abnormality (neutropenia, low platelet count, or
macrocytosis) (OR; 95% CI [number of exposed cases])
Low intensity, 50-hr duration
1.7; 0.7, 4.2 [22]
Medium intensity, 50-hr duration
1.6; not reported [not
reported]
High intensity, 50-hr duration
1.2; 0.3, 5.3 [2]
C-67

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-15. Evidence pertaining to other noncancer effects (hematological)
in humans (continued)
Reference and study design
Results
Hathawav and Buck (1977) (United States)
Cross-sectional study, 2,022 workers,
1,491 participated (74% response rate).
Analysis limited to whites; 69 exposed to RDX
alone and 24 exposed to RDX and HMX;
338 not exposed to RDX, HMX, or TNT.
Exposure measures: Exposure determination
based on job title and industrial hygiene
evaluation. Exposed subjects assigned to two
groups: 0.01 mg/m3 (mean for
employees with exposures >LOD: 0.28 mg/m3).
Effect measures: Hematology tests.
Analysis: Types of statistical tests were not
reported (assumed to be t-tests for
comparison of means and x2 tests for
comparison of proportions).
Hematology tests in men (mean; standard deviation not
reported)
Test
Referent
(n = 237)
RDX exposed*
Undetected
(0.01 mg/m3
(n = 45)
Hemoglobin
15.2
14.7
15.2
Hematocrit
42
45.6
47
Reticulocyte
count
0.7
0.9
0.7
includes both workers exposed to RDX alone and RDX and
HMX.
No differences were statistically significant. Similar results in
women.
Hematology tests in men (prevalence of abnormal values)
Test
(abnormal
range)
Referent
RDX exposed*
Undetected
(0.01 mg/m3
Hemoglobin
(<14)
15/237
3/22
4/45
Hematocrit
(<40)
1/237
1/22
1/45
Reticulocyte
count (>1.5)
18/237
3/22
2/45
*lncludes both workers exposed to RDX alone and RDX and
HMX.
No differences were statistically significant. Similar results in
women.
HMX = octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine; LOD = limit of detection.
C-68

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-16. Evidence pertaining to hematological effects in animals
Reference and study design
Results
Lish et al. (1984b)
Mice, B6C3Fi, 85/sex/group; interim
sacrifices (10/sex/group) at 6 and
12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles
<66 urn
0,1.5, 7.0, 35, or 175/100 mg/kg-d (high
dose reduced to 100 mg/kg-d in
Week 11 due to excessive mortality)
Diet
2 yr
Doses
0
1.5
7.0
35
175/100
RBC count; 105 wk (percent change compared to control)
M
0%
-4%
3%
-3%
14%
F
0%
4%
-7%
5%
3%
Hemoglobin; 105 wk (percent change compared to control)
M
0%
-6%
3%
-5%
9%
F
0%
2%
-7%
3%
1%
Hematocrit; 105 wk (percent change compared to control)
M
0%
-4%
3%
-4%
9%
F
0%
3%
-6%
3%
1%
Platelets; 105 wk (percent change compared to control)
M
0%
33%
9%
21%
27%
F
0%
-14%
-7%
1%
5%
Hart (1976)
Rats, Sprague-Dawley, 100/sex/group
Purity and particle size not specified
0,1.0, 3.1, or 10 mg/kg-d
Diet
2 yr
Doses
0
1.0
3.1
10
RBC count; 104 wk (percent change compared to control)
M
0%
3%
7%
-2%
F
0%
-14%
7%
2%
Reticulocyte count; 104 wk (percent change compared to control)
M
0%
250%
500%*
850%*
F
0%
180%*
-40%
20%
Hemoglobin; 104 wk (percent change compared to control)
M
0%
3%
4%
0%
F
0%
-1%
1%
-2%
C-69

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-16. Evidence pertaining to hematological effects in animals
(continued)
Reference and study design
Results
Levine et al. (1983)
Rats, F344, 75/sex/group; interim
sacrifices (10/sex/group) at 6 and
12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles
<66 urn
0, 0.3,1.5, 8.0, or 40 mg/kg-d
Diet
2 yr
Doses
0
0.3
1.5
8.0
40
Hemoglobin levels; 105 wk (percent change compared to control)
M
0%
6%
6%
3%
-13%
F
0%
-5%
1%
-9%
-14%
RBC count; 105 wk (percent change compared to control)
M
0%
5%
2%
-1%
-9%
F
0%
-2%
2%
-9%
-13%
Platelet count; 105 wk (percent change compared to control)
M
0%
6%
-4%
-10%
-7%
F
0%
14%
-4%
5%
22%
Hematocrit; 105 wk (percent change compared to control)
M
0%
5%
5%
2%
-7%
F
0%
-5%
0%
-8%
-12%
Cholakis et al. (1980)
Mice, B6C3Fi, 10-12/sex/group
88.6% pure, with 9% HMX and 2.2%
water as contaminants; ~200 urn
particle size
0, 80, 60, or 40 mg/kg-d for 2 wk
followed by 0, 80,160, or 320 mg/kg-d
(TWA doses of 0, 79.6, 147.8, or
256.7 mg/kg-d for males and 0, 82.4,
136.3, or 276.4 mg/kg-d for females)3
Diet
13 wk
Doses
0
80
160
320
RBC count (percent change compared to control)
M
0%
-5%
-12%*
-2%
F
0%
-10%
-1%
1%
Reticulocytes (percent change compared to control)
M
0%
-36%
-13%
15%
F
0%
21%
25%
-19%
Hematocrit (percent change compared to control)
M
0%
-1%
-6%
0%
F
0%
-8%
2%
1%
Hemoglobin (percent change compared to control)
M
0%
-2%
-7%*
-3%
F
0%
-5%
4%
1%
Platelets (percent change compared to control)
M
0%
33%
28%
22%
F
0%
3%
9%
39%
C-70

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-16. Evidence pertaining to hematological effects in animals
(continued)
Reference and study design
Results
Cholakis et al. (1980)
Rats, F344,10/sex/group
88.6% pure, with 9% HMX and 2.2%
water as contaminants; ~200 urn
particle size
0,10,14, 20, 28, or 40 mg/kg-d
Diet
13 wk
Doses
0
10
14
20
28
40
RBC count (percent change compared to control)
M
0%
—
—
—
3%
-1%
F
0%
—
—
—
-1%
-7%
Hemoglobin (percent change compared to control)
M
0%
—
—
—
2%
-1%
F
0%
—
—
—
-1%
-1%
Platelet (percent change compared to control)
M
0%
—
—
—
11%
16%*
F
0%
—
—
—
-23%
-13%
Reticulocytes (percent change compared to control)
M
0%
—
—
—
26%
76%*
F
0%
—
—
—
-2%
17%
Hematocrit (percent change compared to control)
M
0%



3%
0%
F
0%



0%
-2%
Crouse et al. (2006)
Rats, F344,10/sex/group
99.99% pure
0, 4, 8,10,12, or 15 mg/kg-d
Gavage
13 wk
Doses
0
4
8
10
12
15
RBC count (percent change compared to control)
M
0%
1%
-7%
-2%
-4%
-5%
F
0%
3%
3%
-1%
2%
-2%
Hemoglobin (percent change compared to control)
M
0%
-1%
-5%
0%
-1%
-6%
F
0%
2%
4%
-1
4%
-4%
Platelet count (percent change compared to control)
M
0%
21%
11%
13%
-8%
34%
F
0%
6%
40%
47%
34%
-36%
Hematocrit (percent change compared to control)
M
0%
2%
-5%
0%
-1%
-4%
F
0%
3%
4%
0%
4%
-2%
C-71

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-16. Evidence pertaining to hematological effects in animals
(continued)
Reference and study design
Results
Levine et al. (1990); Levine et al.
(1981a); Levine et al. (1981b)b
Rats, F344,10/sex/group; 30/sex for
control
84.7 ± 4.7% purity, ~10% HMX, median
particle diameter 20 urn, ~90% of
particles <66 urn
0,10, 30,100, 300, or 600 mg/kg-d
Diet
13 wk
Data were not reported for rats in the 300 or 600 mg/kg-d dose groups
because all rats died before the 13-wk necropsy.
Doses
0
10
30
100
300
600
Hematocrit (percent change compared to control)
M
0%
-2%
-1%
-5%
—
—
F
0%
0%
-4%
-7%
—
—
Hemoglobin (percent change compared to control)
M
0%
-3%
-1%
-6%
—
—
F
0%
0%
-4%
*
0s-
00
1
—
—
RBC count (percent change compared to control)
M
0%
-2%
-2%
-5%
—
—
F
0%
-1%
-4%
-5%
—
—
Reticulocytes (percent change compared to control)
M
0%
-4%
10%
28%

—
F
0%
9%
73%
71%

—
von Oettingen et al. (1949)
Rats, sex/strain not specified, 20/group
90-97% pure, with 3-10% HMX; particle
size not specified
0,15, 25, or 50 mg/kg-d
Diet
13 wk
Doses
0
15
25
50
RBC count (percent change compared to control)
M + F
0%
-23%
-12%
-14%
Hemoglobin (percent change compared to control)
M + F
0%
-25%
-7%
-11%
Hart (1974)
Dogs, beagle, 3/sex/group
Premix with ground dog chow
containing 20 mg RDX/g-chow, 60 g of
dog food; purity and particle size not
specified
0, 0.1,1, or 10 mg/kg-d
Diet
13 wk
Doses
0
0.1
1
10
RBC count (percent change compared to control)
M
0%
-3%
3%
2%
F
0%
13%
7%
11%
Reticulocyte count (percent change compared to control)
M
0%
-66%
0%
-50%
F
0%
-17%
-50%
0%
Hematocrit (percent change compared to control)
M
0%
-4%
2%
0%
F
0%
6%
1%
7%
Hemoglobin (percent change compared to control)
C-72

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-16. Evidence pertaining to hematological effects in animals
(continued)
Reference and study design
Results

M
0%
5%
-2%
0%
F
0%
8%
-2%
8%
Martin and Hart (1974)
Monkeys, cynomolgus or rhesus,0
3/sex/group
Purity of test material not specified
0, 0.1,1, or 10 mg/kg-d
Gavage
13 wk
Histopathological examination revealed increased numbers of
degenerate or necrotic megakaryocytes in all bone marrow sections.
Doses
0
0.1
1
10
RBC count (percent change compared to control)
M
0%
-3%
2%
-3%
F
0%
0%
-1%
2%
Reticulocyte count (percent change compared to control)
M
0%
-33%
-50%
-50%
F
0%
-18%
-36%
45%
Hematocrit (percent change compared to control)
M
0%
-7%
-4%
-1%
F
0%
10%
7%
3%
Hemoglobin (percent change compared to control)
M
0%
-10%
-8%
-6%
F
0%
6%
6%
3%
HMX = octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine; TWA = time-weighted average.
^Statistically significantly different compared to the control, as determined by study authors (p < 0.05).
aDoses were calculated by the study authors.
bLevine et al. (1981a) is a laboratory report of a 13-week study of RDX in F344 rats; two subsequently published
papers (Levine et al., 1990; Levine et al., 1981b) present subsets of the data provided in the full laboratory report.
cThe species of monkey used in this study was inconsistently reported in the study as either cynomolgus (in the
Methods section) or rhesus (in the Summary).
Note: A dash ("—") indicates that the study authors did not measure or report a value for that dose group.
Reproductive Effects
Female Reproductive Effects
Reproductive function in rats was assessed in a two-generation study by Cholakis et al.
(19801. No specific effects on reproductive function were observed in FO and F1 CD rats exposed to
<16 mg/kg-day RDX. A reduction in the number of pregnancies was reported following mating of
the FO generation at the highest dose tested (50 mg/kg-day) (89% pregnancies in controls versus
69% at 50 mg/kg-day); exposure at this dose also resulted in decreased food consumption relative
C-73

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
to controls (14 and 17% less in females and males, respectively, at Week 13 of exposure), decreased
body weight relative to controls (8 and 15% less in females and males, respectively, at Week 13),
and mortality in 9% of male rats and 27% of female rats. The study authors considered the effects
on reproductive function likely due to the general toxicity of RDX rather than a direct effect of RDX
on reproduction. In a dominant lethal mutation study that used the same FO males from the two-
generation reproductive toxicity study but mated a second time to untreated female rats, pregnancy
rates were again lower in untreated females mated to high-dose (50 mg/kg-day) males (97% in
controls versus 79% in females mated with high-dose males). The study authors also attributed
this effect to adverse effects of treatment on the general health of the high-dose males fCholakis et
al„ 19801.
The limited investigation of reproductive function in RDX-exposed rats by a single
investigator (Cholakis etal.. 1980) provides insufficient information to assess female reproductive
toxicity following exposure to RDX.
Male Reproductive Effects
Evidence of male reproductive toxicity comes largely from the finding of increased
incidence of testicular degeneration (10-11%) in male B6C3Fi mice exposed to >35 mg/kg-day
RDX for 2 years in the diet compared to concurrent controls (0%) (Lish etal.. 1984b) (see
Table C-17). The biological significance of this finding is unclear. As noted by the Science Advisory
Board (SAB) in its review of external review draft of the RDX assessment (SAB. 2017). no
histopathological changes were observed in this study in animals sacrificed at 6 or 12 months,
durations longer than the 1.4-month duration of spermatogenesis in mice. Significant decreases in
testes weight would generally be expected where there is appreciable testicular degradation;
however, reductions in absolute testicular weight in male mice in Lish etal. (1984b) were small
(<6% compared to controls) and not dose related. The SAB also noted that, in general, the validity
of 2-year chronic toxicity studies for evaluating male reproductive toxicity is questionable because
of the loss of testicular function that occurs with aging rodents (SAB. 2017). In 2-year old mice,
studies have shown reductions in sperm counts and motility, hormone levels, and numbers of
spermatogonial stem cells, loss of functional ability of spermatogonial stem cells, and failure of the
somatic environment to support spermatogonial differentiation (Zhang etal.. 2006: Gosden etal..
1982: Suzuki and Withers. 1978: Bronson and Desiardins. 1977). These age-related changes in
male reproductive organs confound the interpretation of effects of a potential reproductive
toxicant
The evidence for testicular degeneration in mice suggested by Lish etal. (1984b) was
generally not supported by other studies. In particular, Cholakis etal. fl9801 found no
histopathological changes in the testes in the same strain of mice (B6C3Fi) exposed to RDX for
3 months at a dose of 320 mg/kg-day. This subchronic duration would have allowed for two
complete rounds of spermatogenic cell differentiation and therefore sufficient time for an effect of
RDX on male reproductive organs to have been detected. No dose-related histopathological
C-74

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
changes in the testes were identified in the majority of studies in rats fCrouse etal.. 2006: Levine et
al.. 1990: Levine etal.. 1981a. b; Hart. 19761 or in dogs fHart. 19741. In a 2-year bioassay in F344
rats (Levine etal.. 19831. an increased incidence of germ cell degeneration (40%) was observed in
12-month interim-sacrifice rats exposed to 40 mg/kg-day compared with controls (0%), as well as
a 14% decrease in testis weight Because RDX caused 30-40% excess mortality by 12 months at
this dose, testicular effects could have been secondary to general toxicity. Histopathological
findings obtained at 2 years were not meaningful because almost all male rats (including controls)
developed testicular masses (interstitial cell tumors)—a finding typical for rats of this strain.
Changes in testes weight were not consistent with effects of RDX on the male reproductive
system. Across studies that measured this endpoint, changes in testes weight were generally small
(<10% compared to control), not dose related, and directionally inconsistent, with testes weights
both increased and decreased relative to the control (see Table C-17).
In light of inconsistent findings of histopathology changes in the testes across studies,
questions about the validity of 2-year chronic toxicity studies for evaluating male reproductive
toxicity, and lack of supporting evidence from organ-weight measurements, there is insufficient
information to assess male reproductive toxicity following exposure to RDX.
C-75

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-17. Evidence pertaining to male reproductive effects in animals
Reference and study design
Results
Lish et al. (1984b)
Mice, B6C3Fi, 85/sex/group; interim
sacrifices (10/sex/group) at 6 and
12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles <66
Urn
0,1.5, 7.0, 35, or 175/100 mg/kg-d (high
dose reduced to 100 mg/kg-d in Week
11 due to excessive mortality)
Diet
2 yr
Doses
0
1.5
7.0
35
175/100
Testicular degeneration (incidence)

0/63
2/60
2/62
6/59
3/27a
Absolute testes weight; Week 105 (percent change compared to
control)

0%
-6%
0%
-2%
-6%
Hart (1976)
Rats, Sprague-Dawley, 100/sex/dose
Purity and particle size not specified
0,1.0, 3.1, or 10 mg/kg-d
Diet
2 yr
Doses
0
1.0
3.1
10
Absolute testes (with epididymis) weight; Week 104

0%
-2%
2%
5%
Testes were examined microscopically in control and 10 mg/kg-d
groups; no degeneration or other treatment-related effects were
observed.
Levine et al. (1983)
Rats, F344, 75/sex/group; interim
sacrifices (10/sex/group) at 6 and
12 mos
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles
<66 nm
0, 0.3,1.5, 8.0, or 40 mg/kg-d
Diet
2 yr
Doses
0
0.3
1.5
8.0
40
Testes, germ cell degeneration; 12 mos (incidence)

0/10
0/10
0/10
0/10
4/10*
Testes, germ cell degeneration; 24 mos (incidence)

0/54
0/55
0/52
0/55
0/31
Absolute gonad weight; 12 mos (percent change compared to control)

0%
0%
+1%
0%
-14%
Testes weights were not measured at termination due to testicular
masses in nearly all males.
Cholakis et al. (1980)
Mice, B6C3Fi, 10-12/sex/group
88.6% pure, with 9% HMX and 2.2%
water as contaminants; ~200 nm
particle size
Experiment 1: 0,10,14, 20, 28, or
40 mg/kg-d
Diet
13 wk
Doses
0
10
14
20
28
40
Absolute testes weight (percent change compared to control)

0%



-4%
-4%
C-76

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-17. Evidence pertaining to male reproductive effects in animals
(continued)
Reference and study design
Results
Experiment 2: 0, 40, 60, or 80 mg/kg-d
for 2 wk followed by 0, 320,160, or
80 mg/kg-d (TWA doses of 0, 79.6,
147.8, or 256.7 mg/kg-d for males and
0, 82.4,136.3, or 276.4 mg/kg-d for
females)b
Diet
13 wk
Doses
0
80
160
320
Absolute testes weight (percent change compared to control)

0%
4%
-4%
-8%
Testes were examined microscopically in control and 320 mg/kg-d
groups; no effects were observed.
Cholakis et al. (1980)
Rats, F344,10/sex/dose
88.6% pure, with 9% HMX and 2.2%
water as contaminants; ~200 nm
particle size
0,10,14, 20, 28, or 40 mg/kg-d
Diet
13 wk
Doses
0
10
14
20
28
40
Absolute testes weight (percent change compared to control)

0%



-2%
0%
Testes were examined microscopically in control and 40 mg/kg-d
groups; no effects were observed.
Crouse et al. (2006)
Rats, F344,10/sex/group
99.99% pure
0, 4, 8,10,12, or 15 mg/kg-d
Gavage
13 wk
Doses
0
4
8
10
12
15
Absolute testes weight (percent change compared to control)

0%
-3%
-5%
-4%
-4%
-8%
Levine et al. (1990); Levine et al.
(1981a); Levine et al. (1981b)c
Rats, F344,10/sex/group; 30/sex for
control
84.7 ± 4.7% purity, ~10% HMX, median
particle diameter 20 nm, ~90% of
particles <66 nm
0,10, 30,100, 300, or 600 mg/kg-d
Diet
13 wk
Doses
0
10
30
100
300
600
Testes, germ cell degeneration (incidence)

0/10
0/10
0/10
0/10

—
Absolute testes weight (percent change compared to control)

0%
1%
1%
-2%

—
Note: all animals in the 300- and 600-mg/kg-d groups died within the
first 3 weeks of the study. Therefore, data for these groups were not
reported here.
C-77

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-17. Evidence pertaining to male reproductive effects in animals
(continued)
Reference and study design
Results
Hart (1974)
Dogs, beagle, 3/sex/dose
Premix with ground dog chow
containing 20 mg RDX/g-chow, 60 g of
dog food; purity and particle size not
specified
0, 0.1,1, or 10 mg/kg-d
Diet
13 wk
Doses
0
0.1
1
10
Absolute testes (with epididymis) weight (percent change compared to
control)

0%


51%
Testes were not examined microscopically.
HMX = octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine; TWA = time-weighted average.
^Statistically significantly different compared to the control, as determined by study authors (p < 0.05).
aAlthough the study authors did not observe a statistically significant increase in the incidence of testicular
degeneration, they determined that the incidences at the 35 and 175/100 mg/kg-d dose groups were "notable"
when compared to concurrent control incidence.
bDoses were calculated by the study authors.
cLevine et al. (1981a) is a laboratory report of a 13-week study of RDX in F344 rats; two subsequently published
papers (Levine et al., 1990; Levine et al., 1981b) present subsets of the data provided in the full laboratory
report.
Note: A dash ("—") indicates that the study authors did not measure or report a value for that dose group.
Body-Weight Effects
Changes in body-weight gain were reported in experimental animal studies involving
chronic and subchronic exposure to ingested RDX, generally at doses that were also associated with
other RDX-related toxicity. For example, terminal body weights were more than 10% lower than
controls in female mice exposed to 100 mg/kg-day Lish etal. (1984bl and in rats exposed to
>40 mg/kg-day f Levine etal.. 1990: Levine etal.. 1983: Cholakis etal.. 19801. In these studies, RDX
doses at which >10% decrements in body weight were observed were also associated with elevated
mortality, and in the Levine etal. (19831 study, with severe kidney and urinary bladder toxicity in
male rats. For the most part at lower doses, there were no apparent patterns of treatment-related
body-weight changes across dose groups or sexes within a study, or across studies. One exception
is the 90-day gavage study in rats by Crouse etal. f20061. where a dose-related increase in
body-weight gain was observed in female rats exposed to RDX (26% at the highest dose of
15 mg/kg-day). The study authors did not provide an explanation for the body-weight increase.
In summary, available studies provide evidence that RDX exposure affects body weight in
mice and rats, but these effects appear to be secondary to effects on other primary targets of RDX
toxicity.
C-78

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-18. Evidence pertaining to body-weight effects in animals
Reference and study design
Results
Lish et al. (1984b)
Doses
0
1.5
7.0
35
175/100
Mice, B6C3Fi, 85/sex/group; interim
sacrifices (10/sex/group) at 6 and 12 mos.
89.2-98.7% pure, with 3-10% HMX as
contaminant; 83-89% of particles <66 urn
0,1.5, 7.0, 35, or 175/100 mg/kg-d (high
dose reduced to 100 mg/kg-d in Week 11
due to excessive mortality)
Diet
2 yr
Final body weight (g) (mean ± SD)
M
38.0 ±3.8
36.5 ±2.8
37.8 ±3.8
37.1 ±2.7
36.1 ±3.3*
F
41.7 ±5.1
41.9 ±5.6
41.6 ±6.0
41.1 ±5.5
33.7 ±4.9*
Final body weight (percent change compared to control)
M
0%
-4%
-1%
-2%
-5%
F
0%
0%
0%
-1%
-19%
Hart (1976)
Rats, Sprague-Dawley, 100/sex/group
Purity and particle size not specified
0,1.0, 3.1, or 10 mg/kg-d
Diet
2 yr
Doses
0
1.0
3.1
10
Final body weight (g) (mean ± SE)
M
660.6 ±21
675.0 ± 28
643.0 ± 26
613.8 ±32
F
450.2 ± 23
399.4 ± 14
368.5 ± 17*
408.2 ± 20
Final body weight (percent change compared to control)
M
0%
2%
-3%
-7%
F
0%
-11%
-18%
-9%
Levine et al. (1983)
Doses
0
0.3
1.5
8.0
40
Rats, F344, 75/sex/group; interim
sacrifices (10/sex/group) at 6 and 12 mos.
89.2-98.7% pure, with 3-10% HMX as
contaminant: 83-89% of particles <66 nm
0, 0.3,1.5, 8.0, or 40 mg/kg-d
Diet
2 yr
Final body weight (g) (mean ± SD)
M
409 ± 43
411 ±55
377 ± 58
397 ±41
323 ± 50*
F
303 ± 23
301 ± 25
292 ± 34
291 ±38
255 ±18*
Final body weight (percent change compared to control)
M
0%
0%
-8%
-3%
-21%
F
0%
-1%
-4%
-4%
-16%
Cholakis et al. (1980)
Doses
0
80
160
320
Mice, B6C3Fi, 10-12/sex/group
88.6% pure, with 9% HMX and 2.2% water
as contaminants; ~200 nm particle size 0,
80, 60, or 40 mg/kg-d for 2 wk followed by
0, 80,160, or 320 mg/kg-d (TWA doses of
0, 79.6,147.8, or 256.7 mg/kg-d for males
and 0, 82.4,136.3, or 276.4 mg/kg-d for
females)3
Diet
13 wk
Final body weight (g) (mean ± SE)
M
26.5 ±0.4
27.1 ±0.4
27.1 ±0.4
27.3 ± 1.1
F
26.0 ±0.4
25.8 ±0.3
26.3 ±0.4
27.5 ±0.4*
Final body weight (percent change compared to control)
M
0%
2%
2%
3%
F
0%
-1%
1%
6%
C-79

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-18. Evidence pertaining to body-weight effects in animals (continued)
Reference and study design
Results
Cholakis et al. (1980)
Doses
0
10
14
20
28
40
Rats, F344,10/sex/group
88.6% pure, with 9% HMX and 2.2% water
as contaminants; ~200 urn particle size
0,10,14, 20, 28, or 40 mg/kg-d
Diet
13 wk
Final body weight (g) (mean ± SE)
M
306.2 ±
3.7
308.4 ±
5.2
308.3 ±
2.7
303.3 ±
10.0
295.7 ±
3.0
280.6 ±
3.3*
F
178.7 ±
2.6
176.3 ±
3.4
175.6 ±
1.7
175.5 ±
2.2
171.1 ±
3.1
168.8 ±
3.7
Final body weight (percent change compared to control)
M
0%
1%
1%
-1%
-3%
-8%
F
0%
-1%
-2%
-2%
-4%
-6%
Cholakis et al. (1980)
Doses
0
5
16
50
Rats, CD, two-generation study;
F0: 22/sex/group; Fl: 26/sex/group;
F2: 10/sex/group
88.6% pure, with 9% HMX and 2.2% water
as contaminants; ~200 urn particle size
F0 and Fl parental animals: 0, 5,16, or
50 mg/kg-d
Diet
F0 exposure: 13 wk premating, and during
mating, gestation, and lactation of Fl;
Fl exposure: 13 wk after weaning, and
during mating, gestation, and lactation of
F2; F2 exposure: until weaning
Final body weight (g) (mean ± SE)
M (F0)
530 ± 12
534 ± 12
528 ±9
453 ± 9*
F (F0)
285 ±5
285 ±5
280 ±3
262 ±5*
M (Fl)
435 ± 10
426 ±8
408 ±7
397*
F (Fl)
258 ±5
255 ±4
247 ±4
233**
* Mean of four males from one litter or two females from one litter.
Final body weight (percent change compared to control)
M (F0)
0%
1%
0%
-15%
F (F0)
0%
0%
-2%
-8%
M (Fl)
0%
-2%
-6%
-9%
F (Fl)
0%
-1%
-4%
-10%
Crouse et al. (2006)
Doses
0
4
8
10
12
15
Rats, F344,10/sex/group
99.99% pure
0, 4, 8,10,12, or 15 mg/kg-d
Gavage
13 wk
Final body weight (g) (mean ± SD)
M
317.5 ±
14.1
299.1 ±
26.8
289.2 ±
24.5*
308 ±
29.3
321.6 ±
19.3
325.4 ±
15.2
F
175.6 ±
7.4
174.7 ±
7.1
190.6 ±
23.5
200.9 ±
17.5*
207.8 ±
12.1*
220.8 ±
20.7*
Final body weight (percent change compared to control)
M
0%
-6%
-9%
-3%
1%
2%
F
0%
-1%
9%
14%
18%
26%
C-80

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-18. Evidence pertaining to body-weight effects in animals (continued)
Reference and study design
Results
Levine et al. (1990); Levine et al. (1981a);
Doses
0
10
30
100
300
600
Levine et al. (1981b)b
Rats, F344,10/sex/group; 30/sex for
control
84.7 ± 4.7% purity, ~10% HMX, median
particle diameter 20 nm, ~90% of particles
<66 nm
0,10, 30,100, 300, or 600 mg/kg-d
Diet
13 wk
Final body weight (g) (mean ± SD)
M
301.6 ±
23.2
292.1 ±
23.9
276.1 ±
25.0*
251.0*
—
—
F
170.7 ±
12.1
172.2 ±
6.9
170.0 ±
10.7
173.2 ±
7.9
—
—
Final body weight (percent change compared to control)
M
0%
-3%
-8%
-17%
—
—
F
0%
1%
0%
1%
—
—
Hart (1974)
Doses
0
0.1
1
10
Dogs, beagle, 3/sex/group
Premix with ground dog chow containing
20 mg RDX/g-chow, 60 g of dog food;
purity and particle size not specified
0, 0.1,1, or 10 mg/kg-d
Diet
13 wk
Terminal body weight (kg) (mean)
M
8.2
8.6
8.0
10.5
F
8.1
7.9
8.5
8.1
Final body weight (percent change compared to control)
M
0%
5%
-2%
28%
F
0%
-2%
5%
0%
Martin and Hart (1974)
Doses
0
0.1
1
10
Monkeys, cynomolgus or rhesus,0
3/sex/group
Purity and particle size not specified
0, 0.1,1, or 10 mg/kg-d
Gavage
13 wk
Final body weight (kg) (mean)
M
3.8
3.7
3.7
3.8
F
2.6
2.7
2.8
2.6
Final body weight (percent change compared to control)
M
0%
-3%
-3%
0%
F
0%
4%
8%
0%
HMX = octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine; SD = standard deviation; SE = standard error;
TWA = time-weighted average.
^Significantly different than control (p < 0.05).
aDoses were calculated by the study authors.
bLevine et al. (1981a) is a laboratory report of a 13-week study of RDX in F344 rats; two subsequently published
papers (Levine et al., 1990; Levine et al., 1981b) present subsets of the data provided in the full laboratory report.
cThe species of monkey used in this study was inconsistently reported in the study as either cynomolgus (in the
Methods section) or rhesus (in the Summary).
Note: A dash ("—") indicates that the study authors did not measure or report a value for that dose group.
C.3.3. Genotoxicity
Hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
RDX has tested negative in a variety of in vitro tests for genotoxicity, including mutation
assays in multiple strains of Salmonella typhimurium (with or without metabolic activation),
C-81

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
recombination in Saccharomyces cerevisiae strain D3, and forward mutations in both V79 Chinese
hamster lung cells and mouse lymphoma L5178Y cells. However, in genotoxicity assays designed to
be more sensitive, RDX did show some positive results. For example, when the concentration of S9
was doubled, the mutagenicity of RDX was about twice that of background. RDX also showed
positive mutagenic results with metabolic activation in a chemiluminescent assay (Mutatox assay).
In some cases, the interpretation of testing data for RDX was complicated by the tendency of the
compound to precipitate outof DMSO solution (the usual vehicle) at concentrations >250 ng/mL
fReddv etal.. 20051. As with other studies of RDX, the purity of the test compound was unknown in
several (particularly older) studies. A summary of the results of in vitro genotoxicity studies of RDX
is presented in Table C-19.
RDX has produced negative results in all reverse mutation assays in S. typhimurium that use
standard levels of the metabolic activation system (S9). No evidence of reverse mutation was
observed inS. typhimurium (strains TA98, TA100, TA1535, TA1537, andTA1538), either with or
without the addition of S9 metabolic activating mixture (Neuwoehner etal.. 2007: George etal..
2001: Lachance etal.. 1999: Tan etal.. 1992: Cholakis etal.. 1980: Whongetal.. 1980: Cotruvo etal..
1977: Simmon et al.. 19771. One exception is a finding of weak mutagenic activity of RDX to
S. typhimurium strainTA97a (mutagenicity index = 1.5-2.0) (Pan etal.. 2007a). However, this assay
used a high percentage of S9 fraction (9% instead of 4%), indicating that extensive metabolic
activation is needed to elicit a mutagenic response.
RDX did not cause gene recombination in S. cerevisiae strain D3 at concentrations up to
23 |ig/mL, with or without metabolic activation (Cotruvo etal.. 1977: Simmon etal.. 1977).
Simmon etal. f 19771 noted that the negative findings should be considered in the context of the low
concentrations tested. RDX was negative in assays with S. choleraesius and Escherichia coli with and
without metabolic activation (Neuwoehner etal.. 2007). Similarly, RDX did not induce forward
mutations (hypoxanthine-guanine phosphoribosyltransferase locus) in V79 Chinese hamster lung
cells, with or without metabolic activation, although minimal cytotoxicity was observed at 180 |iM
fLachance etal.. 19991. However, RDX produced revertants in two of three trials in the Mutatox
assay with the bacterium Vibrio fischeri when tested at doses up to 2.5 ng/tube, with and without
S9 fArfsten etal.. 19941. In the presence of S9, a dose-response was observed; in the absence of S9,
no dose-response relationship was detected fArfsten et al.. 19941. RDX did not induce forward
mutations in mouse lymphoma L5178Y cells with or without metabolic activation (Reddv etal..
2005). During an accompanying range-finding study, precipitates occurred at doses >250 |ig/mL,
suggesting that concentrations of RDX in DMSO reported beyond 250 ng/mL may not be accurate.
C-82

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-19. Summary of in vitro studies of the genotoxicity of hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
Endpoint
Test system
Dose/
concentration3
Results'5
Comments
Reference
Without
activation
With
activation
Genotoxicity studies in prokaryotic organisms
Reverse
mutation
Salmonella typhimurium
TA1535, TA1537, TA1538, TA98,
TA100
1,000 ng/plate
"
"
Metabolic activation with S9
Cholakis et al.
(1980)
Reverse
mutation
5. typhimurium TA1535,
TA1537, TA1538 TA100, TA98
14 ng/plate


Effect of disinfection treatments on
mutagenicity tested: RDX was not mutagenic
in any strain before or after disinfection
treatment with chlorine or ozone
Simmon et al.
(1977)
Reverse
mutation
S. typhimurium TA98, TA100
250 ng/plate
-
-
Study authors noted that results were
consistent with literature
George et al.
(2001)
Reverse
mutation
S. typhimurium TA98, TA100
1 mg/plate
-
-
Metabolic activation with S9
Tan et al.
(1992)
Reverse
mutation
S. typhimurium TA98, TA100
1,090 ng/plate
-
-
High S9 activation (9%) used
Pan et al.
(2007a)
Reverse
mutation
5. typhimurium TA97a
32.7 ng/plate

+
High S9 activation (9%) used; study authors
concluded that RDX "required intensive
metabolic activation" to exhibit mutagenicity
in this strain
Pan et al.
(2007a)
Reverse
mutation
5. typhimurium TA1535,
TA1537, TA1538 TA100, TA98
Up to
2.5 mg/plate


Results were reported qualitatively only;
quantitative results were not presented. Not
clear if assay was also performed without S9
Whong et al.
(1980)
Reverse
mutation
Vibrio fischeri
0.004 ng/tube
+
+
Mutatox assay with metabolic activation (S9)
Arfsten et al.
(1994)
C-83

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-19. Summary of in vitro studies of the genotoxicity of hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
(continued)
Endpoint
Test system
Dose/
concentration3
Results'5
Comments
Reference
Without
activation
With
activation
Reverse
mutation (umu
test)
Salmonella choleraesuis subsp.
choleraesuis (prior 5.
typhimurium) TA1535/pSK1002
20.6 Hg/mL


No observed effect concentration; tested at
highest concentration where the induction
rate was below 1.5 for the first time and the
growth factor was below 0.5
Neuwoehner
et al. (2007)
Reverse
mutation
(NM2009 test)
5. choleraesius subsp. chol.
NM2009,
TA1535/pS K1002/pN M12
20.6 Hg/mL


No observed effect concentration; tested at
highest concentration where the induction
rate was below 1.5 for the first time and the
growth factor was below 0.5
Neuwoehner
et al. (2007)
Induction of
the sfiA gene
(SOS
chromotest)
Escherichia coli PQ37
20.6 Hg/mL


No observed effect concentration; tested at
highest concentration where the induction
rate was below 1.5 for the first time and the
growth factor was below 0.5
Neuwoehner
et al. (2007)
Reverse
mutation
S. typhimurium, TA98, TA100
24.8 Hg/mL
-
-
No observed effect concentration; metabolic
activation with S9
Neuwoehner
et al. (2007)
Reverse
mutation
S. typhimurium TA98, TA100
2.6 ng/mL
"
"
No observed effect concentration; metabolic
activation with S9; minimal cytotoxicity was
observed at 180 nM
Lachance et al.
(1999)
Reverse
mutation
5. typhimurium TA1535,
TA1536, TA1537, TA1538
TA100, TA98
30.8 Hg/mL
"
"
Metabolic activation with S9
Cotruvo et al.
(1977)
Genotoxicity studies in nonmammalian eukaryotic organisms
Recombination
induction
Saccharomyces cerevisiae D3
23 ng/mL


Study authors concluded that this
microorganism did not appear to be useful
for detecting mutagenicity in several
compounds tested
Simmon et al.
(1977)
C-84

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-19. Summary of in vitro studies of the genotoxicity of hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
(continued)
Endpoint
Test system
Dose/
concentration3
Results'5
Comments
Reference
Without
activation
With
activation
Recombination
induction
S. cerevisiae D3
30.8 Hg/mL
-
-
Metabolic activation with S9
Cotruvo et al.
(1977)
Genotoxicity studies in mammalian cells
Forward
mutation
Chinese hamster lung
fibroblasts V79 cells
40 Hg/mL
-
-
Minimal cytotoxicity observed at 40 ng/mL
(limit of solubility)
Lachance et al.
(1999)
Mutation
L5178Y mouse lymphoma cells
500 Hg/mL
"
"
No or low cytotoxicity seen at these
concentrations; however, precipitate was
observed at >250 ng/mL
Reddv et al.
(2005)
Unscheduled
DNA synthesis;
DNA repair
WI-38 cells, human diploid
fibroblasts
4,000 Hg/mL


Precipitates were observed at concentrations
of RDX >40 ng/mL
Dillev et al.
(1979)
DNA = deoxyribonucleic acid.
aLowest effective dose for positive results; highest dose tested for negative results.
b+ = positive; ± = equivocal or weakly positive; — = negative.
C-85

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
RDX did not induce unscheduled deoxyribonucleic acid (DNA) synthesis, with or without
metabolic activation, using human diploid fibroblasts (WI-38 cells) when tested in DMSO at
concentrations up to 4,000 |ig/mg; however, precipitation of RDX at high concentrations in cell
culture media makes interpretation of these results difficult (Dilley etal.. 19791. Only two in vivo
genotoxicity studies are available; these are summarized in Table C-20. RDX did not decrease the
ratio of polychromatic erythrocytes (PCEs) to normochromatic erythrocytes (NCEs), nor did it
induce micronucleated PCEs in an in vivo mouse bone marrow micronucleus assay in young adult
male CD-I mice fReddvetal.. 20051. RDX was considered negative for the induction of dominant
lethal mutations in male CD rats fed RDX at nominal doses from 0 to 50 mg/kg-day for 15 weeks
prior to mating with untreated virgin females (Cholakis etal.. 19801. Females sacrificed at
midgestation showed no statistically significant effects on number of corpora lutea, implants, or live
or dead embryos (Cholakis etal.. 19801.
Metabolites ofhexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
Several metabolites of RDX, iV-nitroso derivatives of the parent compound (mononitroso,
dinitroso, andtrinitroso compounds, abbreviatedMNX, DNX, andTNX, respectively) (Musicketal..
2010: Major etal.. 20071 have been tested directly for genotoxicity (Pan etal.. 2007a: George etal..
2001: Snodgrass. 19841. Minipigs were used to detect these trace metabolites because the swine
model of the GI tract more closely resembles that of humans (Major etal.. 20071: an identification
and quantification of RDX metabolites in humans has not been conducted. A summary of the results
of in vitro and in vivo genotoxicity studies of metabolites of RDX is presented in Table C-21.
Pan etal. f2007al studied the mutagenicity of two metabolites, MNX and TNX. These
metabolites were not mutagenic in S. typhimurium strain TA97a at normal levels of S9, but were
clearly mutagenic at enhanced concentrations of S9 (4% versus 9% S9). The observation that these
metabolites were positive in S. typhimurium strain TA97a is likely due to this strain's higher
sensitivity for frameshift mutations that occur at a cluster of cytosine residues in the mutated gene
for histidine synthesis in this strain (Pan etal.. 2007a). These metabolites were also weakly
mutagenic in S. typhimurium strain TA102, again with high levels of S9. Strain TA102 was
developed with an A-T base pair at the site of mutation and its sensitivity was increased by the
addition of some 30 copies of a plasmid containing the mutant gene that is available for back
mutation. This strain is sensitive to many oxidative mutagenic compounds (Levin etal.. 1982).
Other metabolites with potential human relevance identified in the urine of miniature pigs have not
been assessed for their genotoxicity (Major etal.. 2007). In assays with S. typhimurium strains
TA98 and TA100, TNX was positive in strain TA100 with and without S9, but not in strain TA98;
MNX and DNX were not mutagenic in either strain fGeorge etal.. 20011.
C-86

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-20. Summary of in vivo studies of the genotoxicity of hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
Endpoint
Test system
Dose/
concentration
Results
Comments
Reference
In vivo genotoxicity studies in mammalian systems
Micronucleus
formation
CD-I mouse bone marrow
Single dose of
62.5, 125, or
250 mg/kg
No significant decrease in
PCE:NCE ratios; no induction of
micronucleated PCE at any
dose
250 mg/kg was maximum
tolerated dose determined in
dose range-finding study
Reddv et al. (2005)
Dominant
lethal
mutations
Male CD rats dosed and mated
with untreated female rats
0, 5,16, or
50 mg/kg-d for
15 wk
No statistically or biologically
significant effects on fertility;
determined to be negative for
the induction of lethal
mutations
Males in the high-dose group
experienced lower food
consumption and weight gain
compared with all other
groups
Cholakis et al.
(1980)
C-87

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table C-21. Summary of in vitro and in vivo studies of the genotoxicity of hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX) metabolites
Endpoint
Test system
Dose/
concentration3
Results'5
Comments
Reference
Without
activation
With
activation
Genotoxicity studies in prokaryotic organisms
Reverse
mutation
Salmonella typhimurium TA97a,
TA102
22 ng/plate
-
+
Mono and trinitroso metabolites (MNX
and TNX); high S9 activation (9%) used
Pan et al. (2007a)
Reverse
mutation
S. typhimurium TA98, TA100
500 ng/plate
+
+
Positive in TA100 (but not in TA98) only
for TNX; MNX, and DNX were negative
George et al.
(2001)
Reverse
mutation
5. typhimurium TA1535,
TA1537, TA1538, TA98, TA100
NR
-
-
Mononitroso metabolite, MNX;
metabolic activation with S9
Snodgrass (1984)
Genotoxicity studies in mammalian cells—in vitro
Forward
mutation
Mouse lymphoma thymidine
kinase
NR
+
+
Mononitroso metabolite, MNX;
metabolic activation with S9
Snodgrass (1984)
Chromosomal
aberrations
Chinese hamster ovary cells
NR
-
+
Mononitroso metabolite, MNX;
metabolic activation with S9
Snodgrass (1984)
Unscheduled
DNA synthesis;
DNA repair
Primary rat hepatocytes
NR
+
ND
Mononitroso metabolite, MNX;
additional metabolic activation not
required with S9
Snodgrass (1984)
In vivo genotoxicity studies in mammalian systems
Dominant
lethal
mutations
Male mice dosed and mated
with untreated female mice
NR

ND
Mononitroso metabolite, MNX;
additional metabolic activation not
required with S9
Snodgrass (1984)
aLowest effective dose for positive results; highest dose tested for negative results; NR = not reported.
b+ = positive; ± = equivocal or weakly positive; — = negative; ND = not determined.
C-88

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
The genotoxicity of MNX was positive in three out of five assays conducted for the U.S. Army
fSnodgrass. 19841. MNX was positive with or without metabolic activation in the mouse lymphoma
forward mutation assay at the thymidine kinase locus, for chromosomal aberrations in Chinese
hamster ovary cells, and in the primary rat hepatocyte unscheduled DNA synthesis assay. MNX was
not considered positive inS. typhimurium (strains TA98, TA100, TA1535, TA1537, andTA1538),
either with or without the addition of S9 metabolic activating mixture or in an in vivo dominant
lethal mutation assay in mice. However, this study is of limited use due to a significant lack of
details including information on dosing, raw data, and statistical reporting.
In summary, RDX is not mutagenic or genotoxic in vitro or in vivo in typical assays used to
detect genotoxicity. Two in vitro studies using more sensitive assays and conditions for detecting
mutagenicity found positive results for RDX. Several studies showed that the /V-nitroso metabolites
are genotoxic, but the formation and quantification of these metabolites in humans is not known.
C-89

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
APPENDIX D. DOSE-RESPONSE MODELING FOR
THE DERIVATION OF REFERENCE VALUES FOR
EFFECTS OTHER THAN CANCER AND THE
DERIVATION OF CANCER RISK ESTIMATES
This appendix provides technical details on evaluating dose-response and determining
points of departure (POD) for relevant toxicological endpoints. The endpoints were modeled using
the U.S. Environmental Protection Agency (EPA) Benchmark Dose Software (BMDS, Versions 2.4).
Sections D.l (noncancer) and D.2 (cancer) describe the common practices used in evaluating the
model fit and selecting the appropriate model for determining the POD, as outlined in the
Benchmark Dose Technical Guidance Document (U.S. EPA. 2012).
D.l. BENCHMARK DOSE MODELING SUMMARY FOR NONCANCER
ENDPOINTS
The noncancer endpoints that were selected for dose-response modeling are presented in
Table D-l. For each endpoint, the doses and response data used for the modeling are presented.
D-l

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-l. Noncancer endpoints selected for dose-response modeling for
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
Endpoint and

Dose

reference
Species/sex
(mg/kg-d)
Incidence/total (%)
Convulsions
F344 rat/female
0
0/10 (0%)
Crouse et al. (2006)a

4
0/10 (0%)


8
2/10 (20%)


10
3/10 (30%)


12
5/10 (50%)


15
5/10 (50%)

F344 rat/male
0
0/10 (0%)


4
0/10 (0%)


8
1/10 (10%)


10
3/10 (30%)


12
8/10 (80%)


15
7/10 (70%)

F344 rat/male and
0
0/20 (0%)

female, combined
4
0/20 (0%)


8
3/20 (15%)


10
6/20 (30%)


12
13/20 (65%)


15
12/20 (60%)
Convulsions
F344 rat/female
0
0/24 (0%)
Cholakis et al. (1980)
(gestational exposure)
0.2
0/24 (0%)


2
1/24 (4%)


20
18/24 (75%)
Urinary bladder:
F344 rat/male
0
0/54 (0%)
hemorrhagic/suppurative

0.3
2/55 (4%)
cystitis

1.5
1/52 (2%)
Levine et al. (1983)

8
1/51 (2%)


40
18/31(58%)
Prostate suppurative
F344 rat/male
0
2/54 (4%)
inflammation

0.3
4/55 (7%)
Levine et al. (1983)

1.5
9/52 (17%)


8
12/55 (22%)


40
19/31 (61%)
aFor convulsions in Crouse et al. (2006), the incidence rates across doses were determined to be not statistically
significantly different between the males and females using an exact Wald x2 test (p > 0.05). The data were
combined across sex for this endpoint prior to modeling.
In addition to the endpoints presented in Table D-l, the combined incidence of seizures and
mortality was modeled for Crouse etal. (20061 to determine the effect of possible underestimation
of seizures, as discussed in Section 2.1.6. Table D-2 presents the data on this combined incidence.
D-2

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-2. Convulsion or mortality endpoints from Crouse etal. (2006)
selected for dose-response modeling for hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX)
Endpoint and
reference
Species/sex
Dose
(mg/kg-d)
Incidence/total (%)a
Convulsion or
F344 rat/female
0
0/10 (0%)
mortality

4
0/10 (0%)
Johnson (2015)

8
3/10 (30%)


10
5/10 (50%)


12
9/10 (90%)


15
8/10(80%)

F344 rat/male
0
0/10 (0%)


4
0/10 (0%)


8
2/10 (20%)


10
4/10 (40%)


12
8/10(80%)


15
7/10 (70%)

F344 rat/male and
0
0/20 (0%)

female, combined
4
0/20 (0%)


8
5/20 (25%)


10
9/20 (45%)


12
17/20 (85%)


15
15/20 (75%)
incidence was defined for each animal as the presence of convulsion or mortality. The incidence rates across
doses for this endpoint were determined to be not statistically significantly different between the males and
females using an exact logistic regression-based test (p > 0.05). The data were combined across sex for this
endpoint prior to modeling.
D.l.l. Evaluation of Model Fit and Model Selection
For each dichotomous endpoint, BMDS dichotomous models5 were fitted to the data using
the maximum likelihood method. Each model was tested for goodness of fit using a x2
goodness-of-fit test (x2 p-value < 0.10 indicates lack of fit). Other factors were also used to assess
model fit, such as scaled residuals, visual fit, and adequacy of fit in the low-dose region and in the
vicinity of the benchmark response (BMR).
From among the models exhibiting adequate fit, the best-fit model was selected for
estimating the benchmark dose (BMD). This model selection was conducted in two stages: first,
from among only the multistage models to determine a representative multistage model, and
second, from among the representative multistage model and the nonmultistage models. In each
stage, the benchmark dose lower confidence limit (BMDL) estimates (95% lower confidence limit
5Unless otherwise specified, all available BMDS dichotomous models besides the alternative and nested
dichotomous models were fitted. The following parameter restrictions were applied: for the Log-Logistic
model, restrict slope >1; for the Gamma and Weibull models, restrict power >1.
D-3

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
on the BMD, as estimated by the profile likelihood method) and Akaike's information criterion (AIC)
values of the models considered in that stage were used to make the selection, as follows. If the
BMDL estimates were "sufficiently close," that is, differed by threefold or less, the model selected
was the one that yielded the lowest AIC value. If the BMDL estimates were not sufficiently close, the
model with the lowest BMDL was selected. The model selected in the second stage was considered
the best-fit model.
D.1.2. Modeling Results
The tables that follow summarize the modeling results for the noncancer endpoints
modeled.
Nervous System Effects
Table D-3 (and Figure D-l) presents the BMD modeling results for incidence of convulsions
for male and female F344 rats combined based on data from Crouse etal. (20061. using BMRs of 10,
5, and 1% extra risk (ER). Table D-4 (and Figure D-2) presents the BMD modeling results for
incidence of convulsions for female F344 rats based on data from Cholakis etal. f 19801. using BMRs
of 10, 5, and 1% ER. Table D-5 (and Figure D-3) presents the BMD modeling results for combined
incidence of convulsions and mortality for male and female rats combined based on data from
Crouse etal. (2006).
D-4

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-3. Model predictions for convulsions in male and female F344 rats
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by gavage for 90
days (Crouse et al.. 2006): BMR = 5% ER
Model3
Goodness of fit
BMDspct
(mg/kg-d)
BMDLsPct
(mg/kg-d)
Basis for model selection
p-value
AIC
Gamma
0.484
101.79
5.78
3.80
The Quantal-Linear model did not
fit the data adequately
(goodness-of-fit p-value <0.10),
so it was excluded from
consideration. Of the higher
degree multistage models, the
Multistage 3° model was selected
based on lowest AIC. From
among the Multistage 3° and
nonmultistage models, the
Multistage 3° model was selected
based on lowest BMDL (BMDLs
differed by more than threefold).
Logistic
0.231
104.55
5.13
3.49
LogLogistic
0.512
101.66
5.74
3.85
Probit
0.291
103.61
5.29
3.43
LogProbit
0.557
101.25
6.01
4.20
Weibull
0.369
102.91
5.11
3.18
Multistage 4°b
0.502
100.91
5.19
2.65
Multistage 3°
0.502
100.91
5.19
2.66
Multistage 2°
0.364
103.03
3.47
2.20
Quantal-Linear
0.0369
111.56
1.13
0.860
aSelected model in bold; scaled residuals for selected model for doses 0, 4, 8,10,12, and 15 mg/kg-d were 0,
-0.69, -0.25, -0.06,1.62, -1.08, respectively. The BMDio and BMDLio values for the selected model were
6.60 and 4.59 mg/kg-d, respectively; the BMDoi and BMDLoi values for the selected model were 3.02 and
0.569 mg/kg-d, respectively.
bFor the Multistage 5° model, the beta coefficient estimates were 0 (boundary of parameters space). The models
in this row reduced to the Multistage 4° model.
D-5

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Multistage Model, with BMR of 5% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
0.8
0
Multistage
16:55 02/07 2018
Figure D-l. Plot of incidence rate by dose, with the fitted curve of the
Multistage 3° model, for convulsions in male and female F344 rats exposed to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by gavage for 90 days (Crouseet
al.. 20061. BMR = 5% ER; dose shown in mg/kg-day.
Multistage Model (Version: 3.4; Date: 05/02/2014)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
Benchmark Dose Computation
BMR = 5% Extra risk
BMD = 5.19399
BMDL at the 95% confidence level = 2.65815
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0
0
Beta(l)
0
0.00163806
Beta(2)
0
0.00485933
Beta(3)
0.000366065
0
D-6

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-47.08
6



Fitted model
-49.46
1
4.75213
5
0.45
Reduced
model
-71.53
1
48.8965
5
<.0001
d.f. = degrees of freedom.
AIC = 100.913
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0
0
0
20
0
4
0.0232
0.463
0
20
-0.69
8
0.1709
3.418
3
20
-0.25
10
0.3065
6.131
6
20
-0.06
12
0.4688
9.375
13
20
1.62
15
0.7093
14.186
12
20
-1.08
X2 = 4.34 degrees of freedom (d.f.) = 5 p-value = 0.5021
D-7

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-4. Model predictions for convulsions in female F344 rats exposed to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by gavage on gestational days
(GDs) 6-19 fCholakis et al.. 19801: BMR = 5% ER
Model3
Goodness of fit
BMDspct
(mg/kg-d)
BMDLsPct
(mg/kg-d)
Basis for model selection
p-value
AIC
Gamma
0.989
42.003
2.31
0.759
Of the multistage models, the
Quantal-Linear model was selected
based on lowest AIC. From among
the Quantal-Linear and
nonmultistage models, the
Quantal-Linear model was selected
based on lowest BMDL (BMDLs
differed by more than threefold).
Logistic
0.526
43.556
6.53
3.90
LogLogistic
0.991
41.996
2.27
0.823
Probit
0.577
43.348
5.41
3.34
LogProbit
1.000
41.963
2.18
0.902
Weibull
0.983
42.026
2.36
0.756
Multistage 3°b
0.960
42.113
2.51
0.747
Multistage 20b
0.960
42.113
2.51
0.747
Quanta l-Linear
0.669
42.077
0.915
0.628
aSelected model in bold; scaled residuals for selected model for doses 0, 0.2, 2, and 20 mg/kg-d were 0.00, -0.52,
-1.03, and 0.49, respectively. The BMDio and BMDLio values for the selected model were 1.88 and 1.29 mg/kg-d,
respectively; the BMDoi and BMDLoi values for the selected model were 0.179 and 0.123 mg/kg-d, respectively.
bThe Multistage 2° and Multistage 3° models may appear to be equivalent; however, differences exist in digits not
displayed in the table.
D-8

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Quantal Linear Model, with BMR of 5% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
0.8
0.2
0
15:29 02/27 2018
Figure D-2. Plot of incidence rate by dose, with fitted curve for Quantal-Linear
model, for convulsions in female F344 rats exposed to hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) by gavage on gestational days (CDs) 6-19
fCholakiset al.. 19801: BMR = 5% ER; dose shown in mg/kg-day.
Quantal-Linear Model Using Weibull Model (Version: 2.16; Date: 2/28/2013)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-slope*dose)]
Benchmark Dose Computation
BMR = 5% Extra risk
BMD = 0.914694
BMDL at the 95% confidence level = 0.627577
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0
0.0384615
Slope
0.056077
0.0588587
Power
n/a
1
D-9

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-18.98
4



Fitted model
-20.04
1
2.11537
3
0.55
Reduced
model
-47.98
1
57.9972
3
<.0001
AIC = 42.0769
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0
0
0
24
0
0.2
0.0112
0.268
0
24
-0.52
2
0.1061
2.546
1
24
-1.02
20
0.6742
16.856
18
25
0.49
X2 = 1.56 d.f. = 3 p-value = 0.6686
D-10

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-5. Model predictions for combined incidence of convulsion and
mortality in male and female F344 rats exposed to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) by gavage for 90 days (Crouse et al.. 2006): BMR = 1% ER
Model3
Goodness of fit
BMDiPct
(mg/kg-d)
BMDLiPct
(mg/kg-d)
Basis for model selection
p-value
AIC
Gamma
0.245
99.260
3.73
2.10
The Log-Logistic and Quantal-
Linear models did not achieve an
adequate fit (goodness-of-fit
p-value <0.10). The Multistage 2°
model was excluded from model
selection because the residual in
the lowest dose group, near the
BMD, was above 1.5 in absolute
value. Of the remaining
multistage models, the
Multistage 3° model was selected
based on lowest AIC. From
among the Multistage 3° and
nonmultistage models, the
Multistage 3° model was selected
based on lowest BMDL (BMDLs
differed by more than threefold).
Dichotomous-Hill
0.436
98.317
5.22
3.04
Logistic
0.0859
102.17
1.81
0.846
LogLogistic
0.305
98.593
3.70
2.20
Probit
0.101
101.85
2.16
0.853
LogProbit
0.316
98.465
4.22
2.75
Weibull
0.152
101.16
2.45
1.24
Multistage 4°b
0.229
99.182
2.56
0.486
Multistage 3°b
0.229
99.182
2.56
0.486
Multistage 2°
0.165
102.01
1.22
0.470
Quantal-Linear
0.0052
113.90
0.144
0.113
aSelected model in bold; scaled residuals for selected model for doses 0, 4, 8,10,12, and 15 mg/kg-d were 0,
-0.88, -0.14, -0.01,1.92, and -1.55, respectively. The BMDio and BMDLio values for the selected model were
5.60 and 3.85 mg/kg-d, respectively; the BMDos and BMDLos values for the selected model were 4.41 and
2.25 mg/kg-d, respectively.
bThe Multistage 3° and Multistage 4° models may appear to be equivalent; however, differences exist in digits not
displayed in the table.
D-ll

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Multistage Model, with BMR of 1% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
Multistage
1
0.8
0.6
0.4
0.2
0
2
4
6
8
10
12
14
Figure D-3. Plot of incidence rate by dose, with fitted curve for Multistage 3°
model, for combined incidence of convulsion and mortality in male and female
F344 rats exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by gavage
for 90 days fCrouse et al.. 20061: BMR = 1% ER; dose shown in mg/kg-day.
Multistage Model (Version: 3.4; Date: 05/02/2014)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
Benchmark Dose Computation
BMR = 1% Extra risk
BMD = 2.56012
BMDL at the 95% confidence level = 0.486284
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0
0
Beta(l)
0
0.0272036
Beta(2)
0
0.00626035
Beta(3)
0.000598962
0
D-12

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of
parameters
Deviance
Test d.f.
p-value
Full model
-44.71
6



Fitted model
-48.59
1
7.76102
5
0.17
Reduced model
-9.88
1
70.3406
5
<0.0001
AIC = 99.1817
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0
0
0
20
0
4
0.0376
0.752
0
20
-0.88
8
0.2641
5.282
5
20
-0.14
10
0.4506
9.012
9
20
-0.01
12
0.6448
12.896
17
20
1.92
15
0.8675
17.351
15
20
-1.55
X2 = 6.88 d.f. = 5 p-value = 0.2294
D-13

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Urinary System (Bladder) Effects
Table D-6. Model predictions for hemorrhagic/suppurative cystitis of the
urinary bladder in male F344 rats exposed to hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) by diet for 24 months fLevine et al.. 19831: BMR = 10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLiopct
(mg/kg-d)
Basis for model selection
p-value
AIC
Gamma
0.373
87.850
21.3
11.3
The Quantal-Linear model did not
achieve an adequate fit
(goodness-of-fit p-value <0.10).
Of the remaining multistage
models, the Multistage 3° model
was selected based on lowest
AIC. From among the Multistage
3° and nonmultistage models, the
Multistage 3° model was selected
based on lowest AIC.
Logistic
0.394
86.310
19.2
15.6
LogLogistic
0.373
87.850
22.9
11.1
Probit
0.321
86.683
16.8
13.6
LogProbit
0.373
87.850
19.5
10.7
Weibull
0.373
87.850
24.3
11.5
Multistage 3°
0.543
85.909
20.0
11.6
Multistage 2°
0.343
87.038
14.5
10.1
Quantal-Linear
0.0181
95.014
7.00
4.86
aSelected model in bold; scaled residuals for selected model for doses 0, 0.3,1.5, 8, and 40 mg/kg-d were -0.98,
1.06, 0.09, -0.21, 0.02, respectively.
Multistage Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
0.8
Multistage
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
O
5
10
15
20
25
30
35
40
dose
15:28 02/09 2018
Figure D-4. Plot of incidence rate by dose, with fitted curve for Multistage 3°
model, for hemorrhagic/suppurative cystitis of the urinary bladder in male
F344 rats exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for
24 months (Levine etal.. 1983): BMR = 10% ER; dose shown in mg/kg-day.
D-14

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Multistage Model (Version: 3.4; Date: 05/02/2014)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseA1-beta2*doseA2...)]
Benchmark Dose Computation
BMR = 10% Extra risk
BMD = 19.9607
BMDL atthe 95% confidence level = 11.5693
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.0175684
0.0172098
Beta(l)
0
0
Beta(2)
0
0
Beta(3)
0.0000132481
0.0000133069
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-39.54
5



Fitted model
-40.95
2
2.83327
3
0.42
Reduced
model
-73.82
1
68.5588
4
<.0001
AIC = 85.9086
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0176
0.949
0
54
-0.98
0.3
0.0176
0.966
2
55
1.06
1.5
0.0176
0.916
1
52
0.09
8
0.0242
1.235
1
51
-0.21
40
0.5792
17.955
18
31
0.02
X2 = 2.15 d.f. = 3 p-value = 0.5428
D-15

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Prostate Effects
Table D-7 (and Figure D-5) presents the BMD model results for incidence of suppurative
inflammation of the prostate in male F344 rats based on data from Levine etal. (19831. using a BMR
of 10% ER.
Table D-7. Model predictions for prostate suppurative inflammation in male
F344 rats exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for
24 months fLevine et al.. 19831: BMR = 10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLioPct
(mg/kg-d)
Basis for model selection
p-value
AIC
Gammab
Multistage 2°
Quantal-Linear
Multistage 3°
Multistage 4°
0.288
200.37
4.61
3.24
The Log-Probit model is selected
based on lowest BMDL (BMDLs
differ by more than threefold.
The Multistage models had the
same AIC values and BMDLs, so
selection from among the
Multistage models was
unnecessary.)
Logistic
0.102
203.50
10.8
8.58
LogLogistic
0.328
200.05
3.33
2.09
Probit
0.116
203.10
9.91
7.96
LogProbit
0.204
202.03
1.67
0.469
Weibull0
0.288
200.37
4.61
3.24
aSelected model in bold; scaled residuals for selected model for doses 0, 0.3,1.5, 8, and 40 mg/kg-d were -0.289,
0.172, 0.846, -1.298, and 0.819, respectively. The BMDos and BMDLos values for the selected model were
0.702 and 0.122 mg/kg-d, respectively; the BMDoi and BMDLoi values for the selected model were 0.137 and
0.00906 mg/kg-d, respectively.
bThe Gamma model had a power parameter estimate of 1 (boundary of parameter space). The Multistage 2°, 3°,
and 4° models had b2, b3, and b4 coefficients of 0 (boundary of parameter space). The models in this row are
equivalent to the Quantal-Linear model.
The Weibull model may appear equivalent to the Quantal-Linear model; however, differences exist in digits not
displayed in the table.
D-16

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
LogProbit Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BIN
0.8
0.7
0.6
1 05
o
.CD
^	0.4
c
o
o
£	0.3
LL.
0.2
0.1
13:39 02/14
Figure D-5. Plot of inddence rate by dose, with fitted curve for the Log-Probit
model, for prostate suppurative inflammation in male F344 rats exposed to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24 months (Levine et
al.. 19831: BMR = 10% ER; dose shown in mg/kg-day.
Probit Model (Version: 3.3; Date: 2/28/2013)
The form of the probability function is:
P[response] = Background + (1-Background) * CumNorm(Intercept+Slope*Log(Dose)), where
CumNorm(.) is the cumulative normal distribution function.
Slope parameter is not restricted.
Benchmark Dose Computation
BMR = 10% Extra risk
BMD = 1.67454
BMDL at the 95% confidence level = 0.468688
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.0452202
0.037037
Intercept
-1.4970E+00
-1.3564E+00
Slope
0.417872
0.36341
LogProbit
BMD
D-17

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of
parameters
Deviance
Test d.f.
p-value
Full model
-96.3905
5



Fitted model
-98.0147
3
3.24837
2
0.1971
Reduced model
-118.737
1
44.6933
4
<0.0001
AIC = 202.029
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0452
2.442
2
54
-0.289
0.3
0.0669
3.682
4
55
0.172
1.5
0.1332
6.927
9
52
0.846
8
0.2982
16.402
12
55
-1.298
40
0.5396
16.726
19
31
0.819
X2 = 3.18 d.f. = 2 p-value = 0.2035
Mortality: Dose-Response Analysis and Benchmark Dose (BMD) Modeling Documentation
This appendix also presents a quantitative dose-response analysis of mortality incidence
from studies identified in Section 2.1.6 (see Table D-8).
D-18

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-8. Mortality data selected for dose-response modeling for hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX)



Incidence/total (%) or
Reference
Species/sex
Dose
mean ± SD (number of animals)
Lish et al. (1984b)
B6C3Fi
0 mg/kg-d
1/85 (1%)
(mortality at 11 wk)
mouse/male
1.5
0/85 (0%)


7
0/85 (0%)


35
0/85 (0%)


175/100
30/85 (35%)
Lish et al. (1984a)
B6C3Fi
0 mg/kg-d
0/85 (1%)
(mortality at 11 wk)
mouse/female
1.5
0/85 (0%)


7
0/85 (0%)


35
0/85 (0%)


175/100
36/85 (42%)
Levine et al. (1981b)a
F344 rat/female
0 mg/kg-d
0/30 (0%)


10
1/10 (10%)


30
0/10 (0%)


100
5/10 (50%)


300
10/10 (100%)


600
10/10 (100%)

F344 rat/male
0 mg/kg-d
0/30 (0%)


10
0/10 (0%)


30
0/10 (0%)


100
8/10(80%)


300
10/10 (100%)


600
10/10 (100%)

F344 rat/male
0 mg/kg-d
0/60 (0%)

and female,
10
1/20 (5%)

combined
30
0/20 (0%)


100
13/20 (65%)


300
20/20(100%)


600
20/20(100%)
von Oettingen et al.
Rats, sex/strain
0 mg/kg-d
0/20 (0%)
(1949)
not specified
15
0/19 (0%)b


25
8/20 (40%)


50
8/20 (40%)
Cholakis et al. (1980)
CD rat/female
0 mg/kg-d
0/22 (0%)
(2-generation study)

5
0/22 (0%)


16
0/22 (0%)


50
6/22 (27%)
D-19

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-8. Mortality data selected for dose-response modeling for
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) (continued)



Incidence/total (%) or
Reference
Species/sex
Dose
mean ± SD (number of animals)
Levine et al. (1983)
F344 rat/male
0 mg/kg-d
0/75 (0%)
(mortality at 13 wk)

0.3
0/75 (0%)


1.5
0/75 (0%)


8
0/75 (0%)


40
0/75 (0%)

F344 rat/female
0 mg/kg-d
0/75 (0%)


0.3
0/75 (0%)


1.5
0/75 (0%)


8
0/75 (0%)


40
0/75 (0%)
Cholakis et al. (1980)
F344 rat/male
0 mg/kg-d
0/10 (0%)
(13-wk study)

10
0/10 (0%)


14
0/10 (0%)


20
0/10 (0%)


28
0/10 (0%)


40
0/10 (0%)

F344 rat/female
0 mg/kg-d
0/9 (0%)c


10
0/10 (0%)


14
0/10 (0%)


20
0/10 (0%)


28
0/10 (0%)


40
0/10 (0%)
Crouse et al. (2006)
F344 rat/female
0 mg/kg-d
0/10 (0%)


4
0/10 (0%)


8
1/10 (20%)


10
2/10 (20%)


12
5/10 (50%)


15
4/10 (40%)

F344 rat/male
0 mg/kg-d
0/10 (0%)


4
0/10 (0%)


8
1/10 (10%)


10
3/10 (30%)


12
2/10 (20%)


15
3/10 (30%)

F344 rat/male
0 mg/kg-d
0/20 (0%)

and female,
4
0/20 (0%)

combined
8
2/20 (10%)


10
5/20 (25%)


12
7/20 (35%)


15
7/20 (35%)
D-20

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-8. Mortality data selected for dose-response modeling for
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) (continued)



Incidence/total (%) or
Reference
Species/sex
Dose
mean ± SD (number of animals)
Cholakis et al. (1980)
F344
0 mg/kg-d
0/24 (0%)
(gestational exposure)
rats/female
0.2
0/24 (0%)

(gestational
2
0/24 (0%)

exposure)
20
5/ 24 (21%)
Angerhofer et al. (1986)
SD rat/female
0 mg/kg-d
0/39 (0%)

(gestational
2
1/40 (3%)

exposure)
6
1/40 (3%)


20
16/51 (31%)
Cholakis et al. (1980)
New Zealand
0 mg/kg-d
0/11 (0%)

White
0.2
0/11 (0%)

rabbit/female
2
0/11 (0%)

(gestational
20
0/12 (0%)

exposure)


SD = standard deviation.
aFor Levine et al. (1981a) and Crouse et al. (2006), the incidence rates across doses were determined to be not
statistically significantly different between the males and females using an exact Cochran-Mantel-Haenszel test
(p > 0.10). The data were combined across sex for each of these endpoints prior to modeling.
bFor von Oettingen et al. (1949), one death was reported in the 15 mg/kg-d dose group. However, this death was
most likely not related to RDX, so the animal that died was excluded.
Tor Cholakis et al. (1980) (13-week study in female rats), one accidental death was reported in the 0 mg/kg-d dose
group. The animal that died was excluded.
Tables D-9 to D-12 present the BMD modeling results for incidence of mortality from
Crouse etal. (2006). von Oettingen etal. (1949). Levine etal. (1983). and Angerhofer et al. (1986).
The following data sets were not modeled because each had either no response or a positive
response only in the highest dose group: 11-week mortality data from Lish etal. (1984b). both male
(one death in control group) and female; 13-week mortality data from Levine etal. fl9831. both
male and female; mortality in female CD rats and male and female F344 rats from Cholakis et al.
(1980): and mortality in female F344 rats during gestational exposure from Cholakis etal. (1980).
D-21

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-9. Model predictions for combined mortality in male and female F344
rats exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 13
weeks (Levine et al.. 1981b): BMR = 1% ER
Model3
Goodness of fit
BMDiPct
(mg/kg-d)
BMDLiPct
(mg/kg-d)
Basis for model selection
p-value
AIC
Gamma
0.401
41.100
49.8
12.7
The Quantal-Linear model was
excluded because the fit has a
residual below -2 at 4 mg/kg-d.
The Multistage 4° model was
selected as the representative
multistage model based on
lowest AIC. From among the
Multistage 4° and nonmultistage
models, the Multistage 4° model
was selected based on lowest
BMDL (BMDLs differed by more
than threefold).
Logistic
0.346
41.429
18.3
8.25
LogLogistic
0.257
43.098
73.2
16.9
Probit
0.328
41.727
15.0
6.82
LogProbit
0.257
43.098
58.6
19.5
Weibull
0.257
43.101
56.6b
6.51b
Multistage 2°
0.424
42.942
7.72
2.01
Quantal-Linear
0.139
50.257
1.12
0.818
Multistage 3°
0.503
41.520
7.71
2.08
Multistage 4°
0.535
40.935
7.85
2.15
Multistage 5°
0.371
42.928
7.86
2.15
aSelected model in bold; scaled residuals for selected model for doses 0,10, 30,100, 300, and 600 mg/kg-d were
0.00,1.48, -0.97, 0.05, 0.00, and 0.00, respectively. The BMDio and BMDLio estimates for the selected model
were 47.2 and 22.2 mg/kg-d, respectively; the BMDos and BMDLos estimates for the selected model were 32.4 and
11.0 mg/kg-d, respectively.
bThe convergence parameter was increased to 2 x 10"8 to obtain convergence.
Multistage Model, with BMR of 1% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BIV
Bli/IDL BMD
1 3:54 02/27 201 4
300
dose
Figure D-6. Plot of incidence rate by dose, with the fitted curve of the
Multistage 2° model, for combined mortality in male and female F344 rats
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 13 weeks
(Levine et al.. 1981b): BMR = 1% ER; dose shown in mg/kg-day.
D-22

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Multistage Model (Version: 3.3; Date: 02/28/2013)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
Benchmark Dose Computation
BMR =1% extra risk
BMD = 7.85287
BMDL at the 95% confidence level = 2.15059
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0
0
Beta(l)
0.00127544
1.9710E+17
Beta(2)
0
0
Beta(3)
0
0
Beta(4)
9.0721E-09
0
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of
parameters
Deviance
Test d.f.
p-value
Full model
-16.9192
6



Fitted model
-18.4677
2
3.09685
4
0.5418
Reduced model
-102.298
1
170.758
5
<0.0001
AIC = 40.9353
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0
0
0
60
0
10
0.0128
0.255
l
20
1.484
30
0.0446
0.892
0
20
-0.966
100
0.6447
12.894
13
20
0.05
300
1
20
20
20
0
600
1
20
20
20
0
X2 = 3.14 d.f. = 4 p-value = 0.5352
D-23

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-10. Model predictions for mortality (number found dead) in rats
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) in the diet for
13 weeks (von Oettingen et al.. 1949): BMR = 1% ER
Model3
Goodness of fit
BMDiPct
(mg/kg-d)
BMDLiPct
(mg/kg-d)
Basis for model selection
p-value
AIC
Gamma
0.0341
66.088
3.14
0.648
All of the models besides
Dichotomous-Hill had goodness-
of-fit p-values <0.10 and thus did
not provide an adequate fit to the
data. For the Dichotomous-Hill
model, the slope parameter
achieved the BMDS internal
upper bound (18), so the results
from this model were not
reliable. No model was selected.
Dichotomous-Hill
0.984
57.888
17.2
10.9
Logistic
0.0044
70.074
3.40
2.08
LogLogistic
0.0397
65.853
3.39
0.529
Probit
0.0056
69.283
3.28
1.94
LogProbit
0.0426
65.464
5.67
0.409
Weibull
0.0349
66.233
2.30
0.641
Multistage 3°a
0.0351
66.517
1.22
0.628
Multistage 2°a
0.0351
66.517
1.22
0.628
Quantal-Linear
0.0995
64.639
0.919
0.623
aThe Multistage 2° and Multistage 3° models may appear to be equivalent; however, differences exist in digits not
displayed in the table.
Dichotomous-Hill Model, with BMR of 1 % Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
0.7
0.6
0.5
0.4
0.3
0.2
0.1
O
O
20
30
40
50
14:56 02/04 2016
Figure D-7. Plot of incidence rate by dose, with fitted curve for
Dichotomous-Hill model, for mortality (number found dead) in rats exposed
to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) in the diet for 13 weeks (von
Oettingen et al.. 1949): BMR = 1% ER; dose shown in mg/kg-day.
D-24

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-ll. Model predictions for combined mortality (number found dead)
in male and female F344 rats exposed to hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) by gavage for 90 days (Crouse et al.. 2006): BMR = 1% ER
Model3
Goodness of fit
BMDiPct
(mg/kg-d)
BMDLiPct
(mg/kg-d)
Basis for model selection
p-value
AIC
Gamma
0.794
93.263
3.46
0.840
The Multistage 2° model was
selected as the representative
multistage model based on
lowest AIC. From among the
Multistage 2° and nonmultistage
models, the Multistage 2° model
was selected based on lowest
BMDL (BMDLs differed by more
than threefold).
Logistic
0.474
95.709
2.11
1.11
LogLogistic
0.794
93.332
3.17
0.872
Probit
0.574
94.797
2.40
1.07
LogProbit
0.854
92.832
3.96
1.48
Weibull
0.743
93.698
2.76
0.641
Multistage 2°
0.858
91.926
2.11
0.463
Quantal-Linear
0.535
95.345
0.405
0.288
Multistage 5°b
0.731
93.851
2.42
0.433
Multistage 4°b
0.731
93.851
2.42
0.433
Multistage 3°
0.731
93.851
2.42
0.439
Dichotomous-Hill
0.998
93.343
5.96
1.95
aSelected model in bold; scaled residuals for selected model for doses 0, 4, 8,10,12, and 15 mg/kg-d were 0.00,
-0.86, -0.46, 0.53, 0.72, and 0.45, respectively. The BMDio and BMDLio values for the selected model were
6.82 and 4.41 mg/kg-d, respectively.
bThe Multistage 4° and Multistage 5° models may appear to be equivalent; however, differences exist in digits not
displayed in the table.
D-25

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Multistage Model, with BMR of 1 % Extra Risk for the BMD arid 0.95 Lower Confidence Limit for the Bh
Multistage
0.6
0.5
0.4
0.3
0.2
0.1
O
BMDI
BMD
O
2
4
6
8
10
12
14
Figure D-8. Plot of inddence rate by dose, with the fitted curve of the
Multistage 2° model, for mortality in male and female F344 rats exposed to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by gavage for 90 days (Crouseet
al.. 2006): BMR = 1% ER; dose shown in mg/kg-day.
Multistage Model (Version: 3.3; Date: 02/28/2013)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
Benchmark Dose Computation
BMR = 1% Extra risk
BMD = 2.10625
BMDL at the 95% confidence level = 0.462994
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0
0
Beta(l)
0
0.0134587
Beta(2)
0.00226548
0.00141278
D-26

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of
parameters
Deviance
Test d.f.
p-value
Full model
-43.6462
6



Fitted model
-44.963
1
2.63354
5
0.7563
Reduced model
-55.6472
1
24.0019
5
0.0002169
AIC = 91.926
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0
0
0
20
0
4
0.0356
0.712
0
20
-0.859
8
0.135
2.699
2
20
-0.458
10
0.2027
4.054
5
20
0.526
12
0.2784
5.567
7
20
0.715
15
0.3993
7.987
7
20
-0.451
X2 = 1.94 d.f. = 5 p-value = 0.8576
D-27

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-12. Model predictions for mortality in female Sprague-Dawley rats
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by gavage on
Gestation Days 6-15 (Angerhofer et al.. 1986): BMR = 1% ER
Model3
Goodness of fit
BMDiPct
(mg/kg-d)
BMDLiPct
(mg/kg-d)
Basis for model selection
p-value
AIC
Gamma
0.314
89.496
5.15
0.538
The Multistage 3° model was
selected as the representative
multistage model based on
lowest AIC. From among the
Multistage 3° and nonmultistage
models, the Multistage 3° model
was selected based on lowest
BMDL (BMDLs differed by more
than threefold).
Logistic
0.667
87.213
3.88
2.16
LogLogistic
0.312
89.473
4.88
0.560
Probit
0.643
87.196
3.37
1.87
LogProbit
0.319
89.522
5.58
0.885
Weibull
0.309
89.458
4.62
0.541
Quantal-Linear
0.450
87.502
0.652
0.452
Multistage 3°
0.655
86.906
1.68
0.588
Multistage 2°
0.554
87.291
1.78
0.555
aSelected model in bold; scaled residuals for selected model for doses 0, 2, 6, and 20 mg/kg-d were 0.00, 0.76,
-0.52, and 0.04, respectively. The BMDio and BMDLio values for the selected model were 10.9 and 6.09 mg/kg-d,
respectively; the BMDos and BMDLos estimates for the selected model were 5.23 and 7.29 mg/kg-d, respectively.
Multistage Model, with BMR of 1 % Extra Risk for the BMD arid 0.95 Lower Confidence Limit for the BIY
0.5
Multistage
1 0:1 8 05/22 201 4
10
dose
Figure D-9. Plot of incidence rate by dose, with the fitted curve of the
Multistage 3° model, for mortality in female Sprague-Dawley rats exposed to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by gavage on Gestation Days
6-15 (Angerhofer et al.. 19861: BMR = 1% ER; dose shown in mg/kg-day.
D-28

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Multistage Model (Version: 3.3; Date: 02/28/2013)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
Benchmark Dose Computation
BMR = 1% Extra risk
BMD = 1.68097
BMDL atthe 95% confidence level = 0.587568
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0
0.00807857
Beta(l)
0.00588873
0.00216407
Beta(2)
0
0
Beta(3)
0.0000319123
0.0000406218
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of
parameters
Deviance
Test d.f.
p-value
Full model
-41.0771
4



Fitted model
-41.4531
2
0.752152
2
0.6866
Reduced model
-57.4292
1
32.7043
3
<0.0001
AIC = 86.9063
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0
0
0
39
0
2
0.012
0.478
l
40
0.759
6
0.0413
1.654
l
40
-0.519
20
0.3114
15.881
16
51
0.036
X2 = 0.85 d.f. = 2 p-value = 0.6549
D-29

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
D.2. BENCHMARK DOSE MODELING SUMMARY FOR CANCER ENDPOINTS
The cancer endpoints in the mouse that were selected for dose-response modeling are
presented in Table D-13. For each endpoint, the doses and tumor incidence data used for the
modeling are presented.
Table D-13. Cancer endpoints selected for dose-response modeling for
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
Endpoint and reference
Species/sex
Dose (mg/kg-d)a
Incidence/total
Hepatocellular tumors: adenomas or
B6C3Fi/female
0
1/67 (l%)b
carcinomas
mouse
1.5
4/62 (6%)
Lish et al. (1984b); Parker et al. (2006)

7
5/63 (8%)


35
10/64 (16%)
Alveolar/bronchiolar tumors:
B6C3Fi/female
0
7/65(11%)
adenomas or carcinomas
mouse
1.5
3/62 (5%)
Lish et al. (1984b)

7
8/64 (13%)


35
12/64 (19%)
PWG = Pathology Working Group.
aThe highest dose group (175/100 mg/kg-d) was excluded from the analysis because approximately half the
animals in the group died from overdosing, which possibly introduces bias in the estimate of response rate in that
group. See Section 2.3.1 for more details.
bFor female mouse hepatocellular tumors from Lish et al. (1984b), tumor incidence and totals are those reported in
the PWG re-evaluation (Parker et al., 2006).
D.2.1. Evaluation of Model Fit and Model Selection for Mouse Tumor Data
First, the survival curves were compared across dose groups for female mice in Lish et al.
f!984bl in the study to determine whether time of death should be incorporated in the
dose-response analysis of tumors. A log-rank test on the Kaplan-Meier survival curves per dose
was used to do the comparison, excluding the high-dose group. The test yielded a nonsignificant
result (p-value = 0.64), so a time-to-tumor analysis was not necessary for this study.
Therefore, non-time-dependent dose-response analyses were conducted using standard
BMDS models. For each tumor type, BMDS multistage-cancer models6 were fitted to the data using
the maximum likelihood method. Each model was tested for goodness of fit using a x2
goodness-of-fit test (x2 p-value <0.057 indicates lack of fit). Other factors were used to assess model
fit, including scaled residuals, visual fit, and adequacy of fit in the low-dose region and in the
vicinity of the BMR. The BMDL estimate and AIC value were used to select a best-fit model from
6The coefficients of the multistage-cancer models were restricted to be nonnegative (beta values >0).
7 A significance level of 0.05 from U.S. EPA (2012) is used for selecting cancer models because the model
family (multistage) is selected a priori.
D-30

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
among the models exhibiting adequate fit If the BMDL estimates were "sufficiently close"
(i.e., differed by threefold or less), then the model selected was the one that yielded the lowest AIC
value. If the BMDL estimates were not sufficiently close, then the lowest BMDL was selected as the
POD.
After selecting models for the two endpoints, the results were combined using the MS-
COMBO procedure in BMDS, which calculates the BMD and BMDL for combinations of tumors under
the assumption that the tumor incidences are mutually independent. Bogen and Seilkop (19931
analyzed the correlation among tumor rates in treated and untreated animals from 24 previously
published studies and observed that all correlations among the rates of different tumor types were
low. Based on these findings, NRC (19941 noted that the assumption of independence in incidence
between tumor types is reasonable when no evidence exists to the contrary. Because there is no
evidence thattumor types related to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) exposure are
correlated, it is reasonable to assume that liver and lung tumor rates subsequent to RDX exposure
are independent
In its analysis of combined tumor data, the MS-COMBO procedure does not rely on the
modeling of pooled incidence of "tumor-bearing" animals (i.e., animals with one or more tumors).
As noted by NRC (1994) and Bogen (1990). an approach based on such a pooled incidence does not
reflect possible differences in dose-response relationships across sites and would tend to
underestimate the overall risk of tumor incidence when tumors occur independently across sites.
Instead, the procedure does a "composite potency analysis," as described in Bogen f!9901.
Specifically, the combined incidence model is based on the probability of combined incidence
across tumor types as derived using the multistage model for the individual incidence of each
tumor type and the assumption of independence. This model has the multistage form:
P(x) = 1 - exp[—(/?o + fax + fi2x2 + •••)]	(D-l)
where the coefficients [}j are specified by
Po = Id=iPoi,Pl = Ei=i/?1 i>p2 = Zi=i/?2i> etc->
t is the number of tumors under consideration, and /3xi is the xth parameter (x = 0,1, 2,...) for tumor
/. In the case of Lish etal. (1984b). there are two tumor types, so t = 2. Note that the degree of
multistage model used in the combined analysis can vary across sites.
D.2.2. Modeling Results for Female Mouse Tumor Data
The BMD modeling results for mouse tumor data sets are provided in Tables D-14 to D-16
(and Figures D-10 to D-12).
D-31

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Mouse Tumor Data— Benchmark Dose (BMD) Modeling Results
Table D-14. Model predictions for combined alveolar/bronchiolar adenoma
and carcinoma in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) by diet for 24 months fLish etal.. 1984bl. with highest
dose dropped; BMR = 10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLiopct
(mg/kg-d)
Basis for model selection
p-value
AIC
Multistage lob
Multistage 2°
Multistage 3°
0.375
184.58
29.9
14.9
All of the models reduced to the
Multistage 1° model, so this model was
selected.
aSelected model in bold; scaled residuals for selected model for doses 0,1.5, 7, and 35 mg/kg-d were 0.68,
-1.12, 0.48, -0.06, respectively.
bFor the Multistage 2° and 3° models, the b2 and b3 coefficient estimates were 0 (boundary of parameter space).
The models in this row reduced to the Multistage 1° model.
Multistage Cancer Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
0.3
0.25
0.1
0.05
O
Multistage Cancer
Linear extrapolation
14:45 01/13 2016
Figure D-10. Plot of incidence rate by dose, with fitted curve for Multistage-
Cancer 1° model, for combined alveolar/bronchiolar adenoma and carcinoma
in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX) by diet for 24 months (Lish etal.. 1984b). with highest dose dropped;
BMR = 10% ER; dose shown in mg/kg-day.
Multistage Model (Version: 3.4; Date: 05/02/2014)
The form of the probability function is:
D-32

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
The parameter betas are restricted to be positive.
Benchmark Dose Computation
BMR = 10% Extra risk
BMD = 29.8728
BMDL at the 95% confidence level = 14.8898
Benchmark dose upper bound (BMDU) at the 95% confidence level = 169.654
Taken together, (14.8898,169.654) is a 90% two-sided confidence interval for the BMD
Multistage Cancer Slope Factor = 0.00671603
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.0841575
0.084635
Beta(l)
0.00352698
0.00347086
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-89.22
4



Fitted model
-90.29
2
2.14531
2
0.34
Reduced model
-92.36
1
6.29109
3
0.1
AIC = 184.582
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0842
5.47
7
65
0.68
1.5
0.089
5.517
3
62
-1.12
7
0.1065
6.815
8
64
0.48
35
0.1905
12.193
12
64
-0.06
X2 = 1.96 d.f. = 2 p-value = 0.3749
D-33

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-15. Model predictions for combined hepatocellular adenoma and
carcinoma in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) by diet for 24 months (Parker et al.. 2006: Lish et al.. 1984b).
with highest dose dropped; BMR = 10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLiopct
(mg/kg-d)
Basis for model selection
p-value
AIC
Multistage l°b
Multistage 2°
Multistage 3°
0.390
136.52
25.5
14.2
All of the models reduced to the
Multistage 1° model, so this model was
selected.
aSelected model in bold; scaled residuals for selected model for doses 0,1.5, 7, and 35 mg/kg-d were -0.97, 0.82,
0.47, -0.23, respectively.
bFor the Multistage 2° and 3° models, the b2 and b3 coefficient estimates were 0 (boundary of parameter space).
The models in this row reduced to the Multistage 1° model.
0.3
0.25
0.2
"G
0.1
0.05
O	5	10	15	20	25	30	35
dose
14:48 01/13 2016
Figure D-ll. Plot of incidence rate by dose, with fitted curve for Multistage-
Cancer 1° model, for combined hepatocellular adenoma and carcinoma in
female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
by diet for 24 months (Parker etal.. 2006: Lish etal.. 19R4bl. with highest
dose dropped; BMR = 10% ER; dose shown in mg/kg-day.
Multistage Model (Version: 3.4; Date: 05/02/2014)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
The parameter betas are restricted to be positive.
Multistage Cancer Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
Multistage Cancer
Linear extrapolation
D-34

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Benchmark Dose Computation
BMR = 10% Extra risk
BMD = 25.5021
BMDL at the 95% confidence level = 14.1795
BMDU at the 95% confidence level = 68.9086
Taken together, (14.1795, 68.9086) is a 90% two-sided confidence interval for the BMD
Multistage Cancer Slope Factor = 0.00705244
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.0374026
0.0421855
Beta(l)
0.00413144
0.00372219
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-65.23
4



Fitted model
-66.26
2
2.05455
2
0.36
Reduced model
-70.19
1
9.91138
3
0.02
AIC = 136.516
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0374
2.506
l
67
-0.97
1.5
0.0433
2.688
4
62
0.82
7
0.0648
4.085
5
63
0.47
35
0.167
10.688
10
64
-0.23
X2 = 1.88 d.f. = 2 P-va
ue = 0.3902
Combined results for presence of hepatocellular or alveolar/bronchiolar adenoma or
carcinoma in B6C3Fi female mice exposed to RDX by diet for 24 months, with highest dose
dropped; BMR = 10% ER
BMD = 13.8 mg/kg-day; BMDL = 8.53 mg/kg-day
D-35

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
MS-COM BO Results
BMR of 10% Extra Risk
**** Start of combined BMD and BMDL	Calculations.****
Combined Log-Likelihood	-15 6.54886486624446
Combined Log-likelihood Constant	140.87841237048966
Benchmark Dose Computation
Specified effect =	0.1
Risk Type	=	Extra risk
Confidence level =	0.95
BMD =	13.7575
BMDL =	8.53099
Multistage Cancer Slope Factor =	0.011722
Analysis to Address Discrepancy in Sample Size
As reported in Parker etal. (20061. the Pathology Working Group (PWG)'s re-evaluation of
female mouse liver tumors from Lish etal. f!984bl yielded different sample sizes in two dose
groups than did the original study, a discrepancy that could not be resolved (see Table 1-13 of the
Toxicological Review). To determine how this discrepancy affected the modeling results, the liver
tumor data were remodeled using the incidence frequencies from Parker etal. (20061 and the
sample sizes from Lish etal. (1984b). The BMD model results from this reanalysis are provided in
Table D-16, along with the plot and output for the selected model. The BMD and BMDL from this
reanalysis were 25.7 and 14.3 mg/kg-day, respectively, which are close to the BMD and BMDL (25.5
and 14.2 mg/kg-day, respectively) from the original analysis using incidence data from Parker et al.
f20061. In addition, MS-COMBO was also run with the results from the liver tumor reanalysis
combined with the BMD model results for lung tumors. The BMD and BMDL from this combined
analysis were 13.8 and 8.56 mg/kg-day, which were close to the BMD and BMDL (13.8 and
8.53 mg/kg-day, respectively) from the original MS-COMBO analysis. Therefore, it was determined
that the sample size discrepancy had little effect on the BMDL and consequently the oral slope
factor (OSF).
D-36

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-16. Model predictions for combined hepatocellular adenoma and
carcinoma in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) by diet for 24 months (Parker et al.. 2006: Lish et al.. 1984b).
with highest dose dropped and sample sizes from Lish etal. f!984bl: BMR =
10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLiopct
(mg/kg-d)
Basis for model selection
p-value
AIC
Multistage lob
Multistage 2°
Multistage 3°
0.413
136.50
25.7
14.3
All of the models reduced to the
Multistage 1° model, so this model was
selected.
aSelected model in bold; scaled residuals for selected model for doses 0,1.5, 7, and 35 mg/kg-d were -0.95, 0.8,
0.43, -0.22, respectively.
bFor the Multistage 2° and 3° models, the b2 and b3 coefficient estimates were 0 (boundary of parameter space).
The models in this row reduced to the Multistage 1° model.
0.3
0.25
0.2
"G
0.1
0.05
O	5	10	15	20	25	30	35
dose
08:16 03/14 2018
Figure D-12. Plot of incidence rate by dose, with fitted curve for Multistage-
Cancer 1° model, for combined hepatocellular adenoma and carcinoma in
female B6C3F1 mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
by diet for 24 months (Parker et al.. 2006: Lish et al.. 1984b). with highest
dose dropped and sample sizes from Lish et al. (1984b): BMR = 10% ER; dose
shown in mg/kg-day.
Multistage Model (Version: 3.4; Date: 05/02/2014)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
Multistage Cancer Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
Multistage Cancer
Linear extrapolation
D-37

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
The parameter betas are restricted to be positive.
Benchmark Dose Computation
BMR = 10% extra risk
BMD = 25.7422
BMDL at the 95% confidence level = 14.2717
BMDU atthe 95% confidence level = 70.4126
Taken together, (14.2717, 70.4126) is a 90% two-sided confidence interval for the BMD
Multistage Cancer Slope Factor = 0.00700686
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.0378291
0.041973
Beta(l)
0.00409291
0.00372237
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-65.28
4



Fitted model
-66.25
2
1.93249
2
0.38
Reduced model
-70.1
1
9.6454
3
0.02
AIC = 136.497
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0378
2.459
l
65
-0.95
1.5
0.0437
2.711
4
62
0.8
7
0.065
4.16
5
64
0.43
35
0.1662
10.64
10
64
-0.22
X2 = 1.77 d.f. = 2 P-value = 0.413
D-38

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
MS-COM BO Results
BMR of 10% Extra Risk
**** Start of combined BMD and BMDL Calculations.****
Combined Log-Likelihood	-15 6.53930900047271
Combined Log-likelihood Constant	14 0.92 9452 6604 4825
Benchmark Dose Computation
Specified effect =	0.1
Risk Type	=	Extra risk
Confidence level =	0.95
BMD =	13.827
BMDL =	8.5 6197
Multistage Cancer Slope Factor =	0.0116796
D.2.3. Additional Dose-Response Analysis: Male Mice and Rats
This appendix also presents a quantitative dose-response analysis of incidence of lung
carcinomas in male B3C6Fi mice and incidence of liver carcinomas in male F344 rats fLevine etal..
19831 (see Table D-17). The resulting candidate OSFs are presented for comparison with OSF
estimates provided in Section 2.3.3 of the Toxicological Review.
Table D-17. Carcinoma data from Lish et al. (1984b) and Levine et al. (1983)
Endpoint and reference
Species/sex
Dose (mg/kg-d)
Incidence/total
Alveolar/bronchiolar carcinomas
B6C3Fi
0
3/63 (5%)
Lish et al. (1984b)
mouse/male
1.5
6/60(10%)


7
3/62 (5%)


35a
7/59 (12%)
Hepatocellular carcinomas
F344 rat/male
0
1/55 (2%)
Levine et al. (1983)

0.3
0/55 (0%)


1.5
0/52 (0%)


8
2/55 (4%)


40
2/31b (6%)
aThe highest dose (175/100 mg/kg-d) was excluded from the analysis because approximately half the animals in
the group died from overdosing, which possibly introduces bias in the estimate of response rate in that group.
See Section 2.3.1 for more details.
bThe denominators listed in the table for carcinomas in rats represent the number of animals that were alive 1 year
after dosing began.
D-39

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Male Mouse Lung Tumor Analysis
For male mice in Lish etal. f!984b). a log-rank test on the Kaplan-Meier survival curves
with the highest dose dropped, stratified by dose, yielded a nonsignificant result (p-value = 0.50),
indicating that the survival curves were similar across dose groups. Therefore, a time-to-tumor
analysis was not necessary for hepatocellular carcinomas in Lish etal. fl984b). A
non-time-dependent dose-response analysis was conducted using BMDS multistage-cancer models
with the highest dose dropped, and the model selection procedures described in Section D.2.1 were
used to select the appropriate models. Subsequently, the administered dose was converted to a
human equivalent dose (HED) on the basis of body weight scaling to the % power (BW3/4) (U.S. EPA.
19921. as described in Section 2.3.2. The POD estimate for male mouse carcinomas and OSF
calculated from this POD are provided in Table D-18; detailed BMD modeling results are provided
in Table D-19 (and Figure D-13).
Table D-18. Model predictions and oral slope factor for alveolar/bronchiolar
carcinomas in male B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) by diet for 2 years (Lish etal.. 1984b)
Tumor type
Selected
model
BMR
BMD,
mg/kg-d
BMDL,
mg/kg-d
POD =
BMDLio-hed3
mg/kg-d
Candidate OSFb
(mg/kg-d)1
Alveolar/bronchiolar
carcinomas
Multistage 1°
10% ER
42.5
24.7
3.69
0.027
BWa = animal body weight; BWh = human body weight
aBased on allometric scaling of administered RDX dose; BMDLio-hed = BMDLio x (BWa1/4/BWh1/4), BWa = 0.035 kg,
and BWh = 70 kg.
bSlope factor = BMR/BMDLio-hed, where BMR = 0.10 (10% ER).
Table D-19. Model predictions for alveolar/bronchiolar carcinoma in male
B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet
for 2 years (Lish et al.. 1984b). with highest dose dropped; BMR = 10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLioPct
(mg/kg-d)
Basis for model selection
p-value
AIC
Multistage 3°
0.402
135.85
42.5
24.7
The Multistage 3° model was selected
based on lowest AIC.
Multistage 2°
0.391
135.90
47.1
24.5
Multistage 1°
0.360
136.10
65.9
23.9
aSelected model in bold; scaled residuals for selected model for doses 0,1.5, 7, and 35 mg/kg-d were -0.55,1.11,
-0.54, 0.01, respectively.
D-40

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Multistage Cancer Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
Multistage Cancer
Linear extrapolation
0.25
0.2
0.15
0.1
0.05
0
0
5
10
15
20
25
30
35
40
Figure D-13. Plot of incidence rate by dose, with fitted curve for Multistage-
Cancer 3° model, for alveolar/bronchiolar carcinoma in male B6C3Fi mice
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for
24 months fLish et al.. 1984bl. with highest dose dropped; BMR = 10% ER;
dose shown in mg/kg-day.
Multistage Model (Version: 3.4; Date: 05/02/2014)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
The parameter betas are restricted to be positive.
Benchmark Dose Computation
BMR = 10% extra risk
BMD = 42.459
BMDL at the 95% confidence level = 24.7112
BMDU at the 95% confidence level = 81368.9
Taken together, (24.7112, 81368.9) is a 90% two-sided confidence interval for the BMD
Multistage Cancer Slope Factor = 0.00404674
D-41

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.0647969
0.0655597
Beta(l)
0
0
Beta(2)
0
0
Beta(3)
1.3765E-06
1.3607E-06
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-65.07
4



Fitted model
-65.92
2
1.71424
2
0.42
Reduced model
-66.74
1
3.35142
3
0.34
AIC = 135.847
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0648
4.082
3
63
-0.55
1.5
0.0648
3.888
6
60
1.11
7
0.0652
4.045
3
62
-0.54
35
0.1184
6.985
7
59
0.01
X2 = 1.82 d.f. = 2 p-value = 0.4021
Male Rat Liver Tumor Analysis
For male rats in the Levine etal. T19831 study, the high-dose group had a markedly lower
survival curve than the other dose groups, indicating a substantial number of early deaths in the
high-dose group. A log-rank test on the Kaplan-Meier survival curves, stratified by dose, yielded a
significant result (p-value <0.001), in which case a time-to-tumor analysis is generally preferred.
However, such an analysis was not possible because the data were insufficient to allow it. Although
tumor incidence was listed for each animal in the source article, the pathology report used a
different animal numbering system than the experimental report where the times of death were
listed, and the relationship between the two systems was not documented. This precluded the
matching of the times of death and the tumor incidence of the animals, thus precluded the use of a
time-to-tumor analysis.
D-42

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Therefore, a non-time-dependent dose-response analysis was conducted using BMDS
multistage cancer models. The model selection procedures described in Section D.2.1 were used to
select the appropriate models. To account for the difference in the survival curves across the
groups for rats, the number of animals alive at 12 months was used as the denominator in the
analysis (denominators listed in Table D-17). Because the maximum liver tumor response in the
male rat was 6.4%, a BMR of 5% was used to model male rat liver tumor data to obtain a BMD and
BMDL in the range of the experimental data, as recommended in Section 3.2 of Guidelines for
Carcinogen Risk Assessment fU.S. EPA. 20051.
To estimate the HED at the BMDL, HEDs based on both administered dose scaled by BW3/4
and physiologically based pharmacokinetic (PBPK) modeling were considered. Confidence in the
revised rat PBPK model is relatively high (see Appendix C, Section C.1.5); however, the choice of an
internal dose is not straightforward. First, evidence regarding the involvement of metabolites has
been discussed in the literature only in the context of the mouse, and the rate of metabolism
(allometrically adjusted) appears to be qualitatively slower for the rat Second, metabolism in the
model represents the total of all pathways, whereas it is only the minor iV-nitroso metabolic route,
and not the oxidative route that has been proposed as a factor in RDX-induced mouse
carcinogenicity. Third, while blood concentration of RDX as an internal dose would be more
proximally relevant to the tissue than administered dose, there are no data to indicate that the
parent RDX is directly related to its carcinogenicity. Therefore, given the uncertainties, HEDs based
on both administered dose scaled by BW3/4and area under the curve (AUC) of RDX arterial blood
concentration (calculated using the PBPK model) are presented. Extrapolation based on the
internal dose of the parent compound is accomplished by assuming toxicological equivalence when
dose is expressed in terms of the AUC of the RDX blood concentration.
The POD estimates for rat liver carcinomas and the OSFs calculated from these PODs are
provided in Table D-20; detailed BMD modeling results are provided in Table D-21 (and
Figure D-14). Results based on two dose metrics are presented: administered dose of RDX scaled
by BW3/4 (when dose is expressed in terms of mg/kg-day, this entails scaling by BW-1/4) and AUC of
RDX arterial blood concentration (using PBPK modeling).
D-43

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-20. Model predictions and oral slope factor for hepatocellular
carcinomas in male F344 rats administered hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) in the diet for 2 years (Levine et al.. 1983)
Tumor type
Selected
model
BMR
BMD,
mg/kg-d
BMDL,
mg/kg-d
POD =
BMDLos-hed,
mg/kg-d
Candidate OSFa
(mg/kg-d)"1
Hepatocellular carcinomas
Multistage 1°
5% ER
28.5
11.8
2.88,b 5.75°
0.017,b 0.009°
BWa = animal body weight; BWh = human body weight
aSlope factor = BMR/BMDLos-hed, where BMR = 0.05 (5% ER).
bBased on allometric scaling of administered RDX dose; BMDLos-hed = BMDLos x (BWa1/4/BWh1/4), BWa = 0.25 kg, and
BWh = 70 kg.
cBased on toxicological equivalence of PBPK model derived AUC of RDX blood concentration.
Table D-21. Model predictions for combined hepatocellular adenoma and
carcinoma in F344 rats exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX) by diet for 24 months fLevine et al.. 19831: BMR = 5% ER
Model3
Goodness of fit
BMDspct
(mg/kg-d)
BMDLsPct
(mg/kg-d)
Basis for model selection
p-value
AIC
Multistage l°b
Multistage 2°
Multistage 3°
Multistage 4°
0.493
49.095
28.5
11.8
All of the models reduced to the
Multistage 1° model, so this model
was selected.
aSelected model in bold. Scaled residuals for the selected model for doses 0, 0.3,1.5, 8, and 40 mg/kg-d were 0.89,
-0.67, -0.74, 0.74, and -0.26, respectively.
bFor the Multistage 2°, 3°, and 4° models, the b2, b3, and b4 coefficient estimates were 0 (boundary of parameter
space). The models in this row reduced to the Multistage 1° model.
D-44

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Multistage Cancer Model, with BMR of 5% Extra Risk for the BMD arid 0.95 Lower Confidence Limit for tl*
0.25
Multistage Cancer
Linear extrapolation
0.2
0.15
0.1
0.05
o
BMD
BMDI
0
5
10
15
20
25
30
35
40
Figure D-14. Plot of incidence rate by dose, with fitted curve for Multistage 1°
model, for combined hepatocellular adenoma and carcinoma in F344 rats
exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for 24
months fLevine et al.. 19831: BMR = 5% ER; dose shown in mg/kg-day.
Multistage Model (Version: 3.4; Date: 05/02/2014)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
The parameter betas are restricted to be positive.
Benchmark Dose Computation
BMR = 5% extra risk
BMD = 28.4525
BMDL at the 95% confidence level = 11.8487
BMDU at the 95% confidence level = 235.886
Taken together, (11.8487, 235.886) is a 90% two-sided confidence interval for the BMD
Multistage Cancer Slope Factor = 0.00421987
D-45

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.00766363
0.00949438
Beta(l)
0.00180277
0.00149364
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of
parameters
Deviance
Test d.f.
p-value
Full model
-21.0055
5



Fitted model
-22.5473
2
3.08372
3
0.3789
Reduced model
-24.4692
1
6.92747
4
0.1398
AIC = 49.0947
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0077
0.421
l
55
0.894
0.3
0.0082
0.451
0
55
-0.674
1.5
0.0103
0.538
0
52
-0.737
8
0.0219
1.203
2
55
0.735
40
0.0767
2.378
2
31
-0.255
X2 = 2.4	d.f. = 3 p-value = 0.493
D.2.4. Sensitivity Analysis on Dose-Response Modeling of Female Mouse Tumor Data
Comparison of Multistage and Nonmultistage Model Fits with and without High Dose
In their evaluation of the external review draft of the RDX assessment (SAB. 20171. the
Science Advisory Board (SAB) expressed concern about the fit of the multistage cancer models to
the mouse liver and lung tumor data from Lish etal. f!984bl and Parker et al. f2006I In the
external review draft assessment, all the dose groups were included in the dose-response modeling.
On p. 72, the SAB report stated,
The SAB expresses concern that the near linearity of the fitted multistage
dose-response models...results in a relatively poor fit (model estimates) at the
highest doses. The two fitted models used...have BMDLio estimates that are larger
D-46

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
than the two lowest nonzero doses used. The Cancer Guidelines (USEPA, 2005)
states (page 3-16): 'If the POD is above some data points, it can fail to reflect the shape
of the dose-response curve at the lowest doses and can introduce bias into subsequent
extrapolations.' This seems to be what is happening with the data on RDX-induced
adenomas and carcinomas, and the issues with the BMDLio seem to arise primarily
because the fitted multistage models (with parameter constraints invoked) lack
sufficient curvature. Larger than expected BMDLio values (the PODs) result in lower
estimated OSFs. The SAB conjectures that using a model form that allows more
curvature could provide a better fit at the mid-range and higher doses, and improve
the quality of fit As mentioned in Section 3.3.5.3, the SAB acknowledges that EPA's
standard practice is to use the multistage model for benchmark dose modeling of
cancer dose-response when there is no biological basis for choosing another model.
In this case however, the relatively poor fit of the multistage model to the
hepatocellular and alveolar/bronchiolar adenomas and carcinomas data produces
an estimate of the POD with poor properties. The SAB recommends that at a
minimum, the assessment discuss the adequacy of the fit of the multistage model to
available data. This discussion could be further supported by exploring and
reporting fits to other standard BMD model forms—engaging in a curve-fitting
exercise starting for example with the list in Table D-13 in the draft supplemental
document Although the multistage model does ensure positive slopes throughout,
the BMDS software facilitates fitting other models that also adhere to this constraint
In its key recommendations (p. 65), the SAB also stated that "the effect of
including/excluding the highest dose on model choice and the POD estimate" should be explored.
In response to these recommendations, the liver and lung tumor data were modeled using
multistage and nonmultistage models, both with all doses included and with highest dose group
dropped.
Analysis of Female Mouse Liver Tumor Data
Table D-22 summarizes the BMD model results for liver tumors for the case where all doses
are included, along with the plot and output for the Multistage 1° model,8 which was selected for
calculating the OSF in the external review draft of the assessment. The Multistage 1° model
exhibited an adequate fit to the data. Its goodness-of-fit p-value was above 0.10, and its residuals
were all below 2 in absolute value. However, as demonstrated by the plot (see Figure D-15), it did
not account for the supralinearity (steep rise) in the data near the origin, and the residual at the
control group (-1.37) was moderately high. The Log-Probit model had a substantially lower AIC
value than the Multistage 1° model and the other models; the plot for this model is provided in
Figure D-16, followed by its output As demonstrated by the plot, the Log-Probit model accounted
for supralinearity near the origin, and the residual for the control group was very low. Therefore,
this model exhibited a better fit to the data than the multistage models. However, the BMDL for this
throughout this sensitivity analysis, the term "Multistage 1°" is used to describe the Quantal-Linear model in
order to conform to the terminology used in cancer modeling; the Multistage 1° and Quantal-Linear models
are mathematically equivalent. (In many cases, the term "quantal linear" appears in BMDS output.)
D-47

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
model was exceedingly low, and the BMD:BMDL ratio was approximately 67, indicating substantial
uncertainty in the BMD estimate. Table D-23 presents the BMD model results for multistage and
nonmultistage models for liver tumors, with the highest dose dropped. Also presented are the plot
(see Figure D-17) and output for the Multistage 1° model that was selected for calculating the OSF
in the external review draft of this assessment. With the highest dose dropped, the Multistage
1° model had an AIC value that was close to the lowest among the models, and the residuals for all
the dose groups, including control, were below 1 in absolute value. Therefore, the quality of fit of
the Multistage 1° model was comparable to the best-fitting model among the nonmultistage models.
Table D-22. Model predictions for combined hepatocellular adenoma and
carcinoma in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) by diet for 24 months (Parker et al.. 2006: Lish et al.. 1984b).
with all doses included; BMR = 10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLioPct
(mg/kg-d)
p-value
AIC
Gamma
0.160
164.06
64.2
32.6
Logistic
0.0849
165.59
95.7
59.2
LogLogistic
0.178
163.81
59.1
28.5
Probit
0.0917
165.41
91.7
55.2
LogProbit
0.678
161.08
19.1
0.286
Weibull"
Multistage l°c
0.160
164.06
64.2
32.6
Multistage 4°d
Multistage 3°
Multistage 2°e
0.160
164.06
64.2
32.6
aMultistage model selected in external review draft assessment in bold.
bFor the Weibull model, the power parameter estimate was 1. The models in this row reduced to the Multistage 1°
model.
cThe Multistage 1° model may appear equivalent to the Gamma model, however differences exist in digits not
displayed in the table.
dFor the Multistage 3° and 4° models, the b3 and b4 coefficient estimates were 0 (boundary of parameters space).
The models in this row reduced to the Multistage 2° model.
eThe Multistage 2° model may appear equivalent to the Gamma and Weibull models, however differences exist in
digits not displayed in the table.
D-48

-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Quantal Linear Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
0.3
0.25
0.2
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of
parameters
Deviance
Test d.f.
p-value
Full model
-77.15
5



Fitted model
-80.03
2
5.75967
3
0.12
Reduced
model
-82.52
1
10.74
4
0.03
AIC = 164.063
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0521
3.489
l
67
-1.37
1.5
0.0544
3.373
4
62
0.35
7
0.0629
3.963
5
63
0.54
35
0.105
6.719
10
64
1.34
107
0.2047
6.347
4
31
-1.04
X2 = 5.17 d.f. = 3 p-value = 0.16
LogProbit Model, with BMR of 1 0% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
0.3
0.25
0.2
0.15
0.1
0.05
O
3MDL,
O
20
40
60
80
100
16:31 12/19 2016
Figure D-16. Plot of incidence rate by dose, with fitted curve for Log-Probit
model, for combined hepatocellular adenoma and carcinoma in female B6C3Fi
mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for
24 months (Parker et al.. 2006: Lish etal.. 1984b). with all doses included;
BMR = 10% ER; dose shown in mg/kg-day.
D-50
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Probit Model (Version: 3.3; Date: 2/28/2013)
The form of the probability function is:
P[response] = Background + (1-Background) * CumNorm(Intercept+Slope*Log(Dose)), where
CumNorm(.) is the cumulative normal distribution function.
Slope parameter is not restricted.
Benchmark Dose Computation
BMR = 10% extra risk
BMD = 19.0651
BMDL at the 95% confidence level = 0.285889
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.0147757
0.0149254
Intercept
-1.6996E+00
-1.7126E+00
Slope
0.141823
0.143382
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-77.15
5



Fitted model
-77.54
3
0.775228
2
0.68
Reduced model
-82.52
1
10.74
4
0.03
AIC = 161.078
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0148
0.99
l
67
0.01
1.5
0.0643
3.988
4
62
0.01
7
0.0909
5.727
5
63
-0.32
35
0.129
8.258
10
64
0.65
107
0.1624
5.036
4
31
-0.5
X2 = 0.78 d.f. = 2 p-value = 0.6777
D-51
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-23. Model predictions for combined hepatocellular adenoma and
carcinoma in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) by diet for 24 months (Parker et al.. 2006: Lish et al.. 1984b).
with highest dose dropped and multistage and nonmultistage models fitted;
BMR = 10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLioPct
(mg/kg-d)
p-value
AIC
Gamma
Multistage 3°b
Multistage 2°
0.390
136.52
25.5
14.2
Logistic
0.319
137.15
31.0
22.6
LogLogistic
0.400
136.44
24.6
13.0
Probit
0.327
137.08
30.4
21.4
LogProbit
0.645
136.68
13.7
1.66
Weibull0
Multistage lod
0.390
136.52
25.5
14.2
aMultistage models used in cancer assessment in bold.
bFor the Multistage 3° model, the beta coefficient estimates were 0 (boundary of parameters space). The models
in this row reduced to the Multistage 2° model.
Tor the Weibull model, the power parameter estimate was 1. The models in this row reduced to the Multistage 1°
model.
dThe Weibull and Multistage 1° models may appear equivalent to the Gamma, Multistage 2°, Multistage 3° models,
however differences exist in digits not displayed in the table.
D-52
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Quantal Linear Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
0.3
0.25
0.2
£ 0.15
•<
H
0.1
0.05
0
0	5	10	15	20	25	30	35
dose
08:37 04/13 2017
Figure D-17. Plot of incidence rate by dose, with fitted curve for Multistage 1°
model, for combined hepatocellular adenoma and carcinoma in female B6C3Fi
mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet for
24 months fParker et al.. 2006: Lish etal.. 1984bl. with highest dose dropped;
BMR = 10% ER; dose shown in mg/kg-day.
Quantal-Linear Model Using Weibull Model (Version: 2.16; Date: 2/28/2013)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-slope*dose)]
Benchmark Dose Computation
BMR = 10% extra risk
BMD = 25.5022
BMDL atthe 95% confidence level = 14.1795
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.0374028
0.0289855
Slope
0.00413143
0.00436879
Power
n/a
1
D-53
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-65.23
4



Fitted model
-66.26
2
2.05455
2
0.36
Reduced model
-70.19
1
9.91138
3
0.02
AIC = 136.516
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0374
2.506
l
67
-0.97
1.5
0.0433
2.688
4
62
0.82
7
0.0648
4.085
5
63
0.47
35
0.167
10.688
10
64
-0.23
X2 = 1.88 d.f. = 2 p-value = 0.3902
Analysis of Female Mouse Lung Tumor Data
Tables D-24 and D-25 (and Figures D-18 and D-19) present the BMD model results for
multistage and nonmultistage models for lung tumors, with all doses included and with the highest
dose dropped, respectively. Also presented for each case is the plot and output for the Multistage 1°
model. For the case with all doses included, the AIC value for the Multistage 1° model was near the
lowest from among the models, so its fit was comparable to the best-fitting model. For the case
with the highest dose dropped, the Multistage 1° model had the lowest AIC value of all the models,
so it was the best-fitting model. Therefore, in both cases the multistage model yielded a quality of
fit that was comparable to the best-fitting model.
D-54
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-24. Model predictions for combined alveolar/bronchiolar adenoma
and carcinoma in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) by diet for 24 months (Lish etal.. 1984b). with all doses
included and multistage and nonmultistage models fitted; BMR = 10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLioPct
(mg/kg-d)
p-value
AIC
Gammab
Multistage 4°c
Multistage 3°c
Multistage 2°d
0.417
218.68
52.8
27.7
Logistic
0.315
219.45
70.4
45.6
LogLogistic
0.439
218.53
49.2
24.2
Probit
0.328
219.34
67.8
42.9
LogProbit
0.339
219.93
37.0
10.2
Weibull®
Multistage 1°
0.417
218.68
52.8
27.7
aMultistage model selected in external review draft assessment draft in bold.
bThe Gamma model may appear equivalent to the Weibull model, however differences exist in digits not displayed
in the table.
Tor the Multistage 3° and 4° models, the b3 and b4 coefficient estimates were 0 (boundary of parameters space).
The models in this row reduced to the Multistage 2° model.
dThe Multistage 2° model may appear equivalent to the Weibull model, however differences exist in digits not
displayed in the table. This also applies to the Multistage 1° model.
eFor the Weibull model, the power parameter estimate was 1. The models in this row reduced to the Multistage 1°
model.
D-55
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Quantal Linear Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
0
20
40
60
80
100
Figure D-18. Plot of incidence rate by dose, with fitted curve for Multistage 1°
model, for combined alveolar/bronchiolar adenoma and carcinoma in female
B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by diet
for 24 months fLish etal.. 1984bl. with all doses included; BMR = 10% ER;
dose shown in mg/kg-day.
Quantal-Linear Model Using Weibull Model (Version: 2.16; Date: 2/28/2013)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-slope*dose)]
Benchmark Dose Computation
BMR = 10% extra risk
BMD = 52.8079
BMDL at the 95% confidence level = 27.748
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.0931678
0.119403
Slope
0.00199517
0.00140632
Power
n/a
1
D-56
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-105.78
5



Fitted model
-107.34
2
3.12764
3
0.37
Reduced model
-110.16
1
8.77367
4
0.07
AIC = 218.682
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0932
6.056
7
65
0.4
1.5
0.0959
5.944
3
62
-1.27
7
0.1057
6.768
8
64
0.5
35
0.1543
9.877
12
64
0.73
107
0.2675
8.292
7
31
-0.52
X2 = 2.84 d.f. = 3 p-value = 0.4168
D-57
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-25. Model predictions for combined alveolar/bronchiolar adenoma
and carcinoma in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) by diet for 24 months (Lish etal.. 1984b). with highest
dose dropped and multistage and nonmultistage models fitted; BMR = 10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLioPct
(mg/kg-d)
p-value
AIC
Gammab
0.161
186.58
30.0
14.9
Logistic
0.364
184.66
31.9
20.5
LogLogistic
0.162
186.57
29.7
13.9
Probit
0.366
184.65
31.7
19.7
LogProbit
0.171
186.47
28.4
9.50
Weibull0
0.161
186.58
30.0
14.9
Multistage 30d
Multistage 2°
Multistage 1°
0.375
184.58
29.9
14.9
aMultistage model used in cancer assessment in bold.
bThe Gamma model may appear equivalent to the Weibull model, however differences exist in digits not displayed
in the table.
cThe Weibull model may appear equivalent to the Gamma model, however differences exist in digits not displayed
in the table.
dFor the Multistage 2° and 3° models, the b2 and b3 coefficient estimates were 0 (boundary of parameters space).
The models in this row reduced to the Multistage 1° model.
D-58
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
0.3
0.25
-o °-2
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-89.22
4



Fitted model
-90.29
2
2.14531
2
0.34
Reduced model
-92.36
1
6.29109
3
fO.l
AIC = 184.582
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0842
5.47
7
65
0.68
1.5
0.089
5.517
3
62
-1.12
7
0.1065
6.815
8
64
0.48
35
0.1905
12.193
12
64
-0.06
X2 = 1.96 d.f. = 2 p-value = 0.3749
Discussion of Multistage and Nonmultistage Model Results for Female Mouse Tumor Data
Regarding the BMD model results on liver tumors with all doses included, the multistage
models exhibit an inferior fit to the data compared to a subset of nonmultistage models (in this
case, the Log-Probit), which would raise concerns about the appropriateness of their use. However,
with the highest dose dropped, the multistage models yielded a fit comparable to the best-fitting
model. For lung tumors, the multistage models exhibited fits comparable to the best-fitting model
among multistage and nonmultistage models, both with all doses included and with the highest
dose dropped.
As noted in the SAB report (SAB. 20171. EPA uses a linear low-dose extrapolation approach
to estimate human cancer risk when the mode of carcinogenic action for tumor incidence is
unknown, as is the case for RDX (see Section 2.3.2). Thus, the multistage models are used for
modeling cancer data because they allow for the statistical plausibility of low-dose linearity, that is,
a positive slope at the origin. When the estimate of the first-degree coefficient of the multistage
model is positive, the slope of the estimated model is positive. Furthermore, even when the
estimate of the first-degree coefficient is zero, the upper confidence bound on the slope is positive,
thus allowing for the plausibility of low-dose linearity. This property does not hold for most of the
other models in BMDS, where the slope at the origin is either zero or infinite for all possible
parameter estimates (except in some cases when models reduce to the Multistage 1° model). With
D-60
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
the highest dose dropped, the multistage models were the best-fitting or near the best-fitting for
both liver and lung tumors, and were determined to be sufficient for use for these data.
Analysis Using Female Mouse Historical Control Tumor Rates
In their evaluation of the external review draft of the RDX assessment ("SAB, 20171, the SAB
expressed concern that the liver tumor rate in the concurrent control was low compared to
historical control. As indicated in Haseman etal. f 19851. the liver tumor rate of untreated female
B6C3Fi mice in National Toxicology Program (NTP) studies had a mean of 8.3% across studies
collected, while the control group in Lish etal. f!984bl had a tumor rate of 1.5%. While the rates
between the two control groups are different, the standard deviation of the historical control
rates as indicated in Haseman et al. f 19851 is 4.8%, and the range of these rates is 0-10%.
Because the concurrent control rate fell within the historical range, this rate was determined to
be reasonable, and using concurrent control is generally preferred to using historical control.
Therefore, the concurrent control rate was determined to be appropriate for the tumor modeling
for RDXFor presentation purposes, the liver and lung tumor data in Lish etal. f1984b 1 were
remodeled using historical control. The data used for these analyses are presented in Table D-26.
Note that these data are the same as in Table D-13, except that the percentages were entered into
BMDS as the response and the control response for each tumor was set equal to the historical
control in Haseman etal. (19851.
Table D-26. Tumor data used for dose-response modeling with historical
control




Percentage with
Endpoint and reference
Species/sex
Dose (mg/kg-d)a
Total
tumors
Hepatocellular adenomas or
B6C3Fi
0
67b
8.3%c
carcinomas
mouse/female
1.5
62
6.5%
Parker et al. (2006)

7
63
7.9%


35
64
15.6%
Alveolar/bronchiolar adenomas or
B6C3Fi
0
65
6.9%
carcinomas
mouse/female
1.5
62
4.8%
Lish et al. (1984b)

7
64
12.5%


35
64
18.8%
PWG = Pathology Working Group.
aThe highest dose group was excluded from the analysis because approximately half the animals in the group died
from overdosing, which possibly introduces bias in the estimate of response rate in that group. See Section 2.3.1
for more details.
bFor female mouse hepatocellular tumors from Lish et al. (1984b), tumor incidence and totals are those reported in
the PWG re-evaluation (Parker et al.. 2006).
cThe percentages for the control group for both liver and lung tumors were obtained from Haseman et al. (1985);
the percentage for each positive dose group for each tumor was calculated by dividing the incidence in that group
by the total number of animals in the group.
D-61
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
The BMD model results are provided in Tables D-27 and D-28 (and Figures D-20 and D-21),
along with the corresponding plots and model outputs. A comparison of the BMD results is
provided in Table D-29. Use of historical control data from NTP (Haseman etal.. 19851 had a
relatively small impact on the POD used for deriving the OSF. BMDio values for liver and lung
tumors using concurrent control data and excluding the high dose from the analysis (i.e., the basis
for the OSF in the current assessment) were 25.5 and 29.9 mg/kg-day, respectively. Replacing the
incidence of tumors in concurrent controls with NTP historical control data yielded BMDio values
for liver and lung tumors of 37.5 and 24.1 mg/kg-day, respectively.
Table D-27. Model predictions for combined alveolar/bronchiolar adenoma
and carcinoma in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) by diet for 24 months fLish etal.. 1984bl. using historical
control and with highest dose dropped; BMR = 10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLioPct
(mg/kg-d)
Basis for model selection
p-value
AIC
Multistage lob
Multistage 2°
Multistage 3°
0.520
171.92
24.1
13.2
All of the models reduced to the
Multistage 1° model, so this model was
selected.
aSelected model in bold; scaled residuals for selected model for doses 0,1.5, 7, and 35 mg/kg-d were 0.10,
-0.73, 0.85, -0.21, respectively.
bFor the Multistage 2° and 3° models, the b2 and b3 coefficient estimates were 0 (boundary of parameter space).
The models in this row reduced to the Multistage 1° model.
D-62
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Multistage Cancer Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
Multistage Cancer
Linear extrapolation
0.3
0.25
0.2
0.15
0.1
0.05
0
BMDL
BMD
0
5
10
15
20
25
30
35
09:45 03/13 2018
Figure D-20. Plot of incidence rate by dose, with fitted curve for Multistage-
Cancer 1° model, for combined alveolar/bronchiolar adenoma and carcinoma
in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX) by diet for 24 months fLish etal.. 1984bl. using historical control and
with highest dose dropped; BMR = 10% ER; dose shown in mg/kg-day.
Multistage Model (Version: 3.4; Date: 05/02/2014)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
The parameter betas are restricted to be positive.
Benchmark Dose Computation
BMR = 10% extra risk
BMD = 24.1402
BMDL at the 95% confidence level = 13.1829
BMDU at the 95% confidence level = 75.0865
Taken together, (13.1829, 75.0865) is a 90% two-sided confidence interval for the BMD
Multistage Cancer Slope Factor = 0.00758556
D-63
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.0659699
0.0688823
Beta(l)
0.00436453
0.00406884
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-83.3032
4



Fitted model
-83.9602
2
1.31395
2
0.5184
Reduced model
-87.1906
1
7.77481
3
0.0509
AIC = 171.92
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0660
4.288
4.485
65
0.098
1.5
0.0721
4.468
2.976
62
-0.733
7
0.0941
6.021
8.000
64
0.847
35
0.1983
12.690
12.032
64
-0.206
X2 = 1.31 d.f. = 2 p-value = 0.5201
D-64
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-28. Model predictions for combined hepatocellular adenoma and
carcinoma in female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-
triazine (RDX) by diet for 24 months (Parker et al.. 2006: Lish et al.. 1984b).
using historical control and with highest dose dropped; BMR = 10% ER
Model3
Goodness of fit
BMDioPct
(mg/kg-d)
BMDLiopct
(mg/kg-d)
Basis for model selection
p-value
AIC
Multistage 1°
0.926
162.69
39.9
18.4
The Multistage 2° model had the lowest
AIC, so this model was selected.
Multistage 2°b
Multistage 3°
0.926
162.55
37.5
18.7
aSelected model in bold; scaled residuals for selected model for doses 0,1.5, 7, and 35 mg/kg-d were 0.26,
-0.30, 0.03, -0.00, respectively.
bFor the Multistage 3° model, the b3 coefficient estimate was 0 (boundary of parameter space). This model
reduced to the Multistage 2° model.
Multistage Cancer Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
Multistage Cancer
Linear extrapolation
0.25
0.2
~o
<
£=
o
~o
(C
0.15
LL
0.05
BMDL
3MCL
0
5
10
15
20
25
30
35
dose
09:41 03/13 2018
Figure D-21. Plot of incidence rate by dose, with fitted curve for Multistage-
Cancer 1° model, for combined hepatocellular adenoma and carcinoma in
female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
by diet for 24 months fParker et al.. 2006: Lish et al.. 1984bl. using historical
control and with highest dose dropped; BMR = 10% ER; dose shown in
mg/kg-day.
D-65
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Multistage Model (Version: 3.4; Date: 05/02/2014)
The form of the probability function is:
P[response] = background + (l-background)*[l-EXP(-betal*doseAl-beta2*doseA2...)]
The parameter betas are restricted to be positive.
Benchmark Dose Computation
BMR = 10% extra risk
BMD = 37.4534
BMDL at the 95% confidence level = 18.6687
BMDU atthe 95% confidence level = 480758
Taken together, (18.6687, 480758) is a 90% two-sided confidence interval for the BMD
Multistage Cancer Slope Factor = 0.00535657
Parameter Estimates
Variable
Estimate
Default initial parameter values
Background
0.0746974
0.0745007
Beta(l)
0
0
Beta(2)
7.51096E-005
7.52728e-005
Analysis-of-Deviance Table
Model
Log(likelihood)
Number of parameters
Deviance
Test d.f.
p-value
Full model
-79.1947
4



Fitted model
-79.2727
2
0.156066
2
0.9249
Reduced model
-80.8944
1
3.39949
3
0.334
AIC = 162.545
D-66
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Goodness-of-Fit Table
Dose
Est. prob.
Expected
Observed
Size
Scaled residuals
0
0.0747
5.005
5.561
67
0.258
1.5
0.0749
4.641
4.030
62
-0.295
7
0.0781
4.920
4.977
63
0.027
35
0.1560
9.986
9.984
64
-0.001
X2 = 0.15 d.f. = 2 p-value = 0.9257
Combined results for presence of hepatocellular or alveolar/bronchiolar tumors in B6C3Fi
female mice exposed to RDX by diet for 24 months, using historical control and with highest
dose dropped; BMR = 10% ER
BMD = 18.3 mg/kg-day; BMDL = 4.86 mg/kg-day
MS-COMBO Results
BMR of 10% Extra Risk
**** Start of combined BMD and BMDL Calculations.****
Combined Log-Likelihood	-163.23289572215137
Combined Log-likelihood Constant	148.36798304322497
Benchmark Dose Computation
Specified effect =	0.1
Risk Type	=	Extra risk
Confidence level =	0.95
BMD =	18.3472
BMDL =	4.85 83
Multistage Cancer Slope Factor =	0.0205833
D-67
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table D-29. Comparison of model predictions, for lung and liver tumors in
female B6C3Fi mice exposed to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
by diet for 24 months (Parker et al.. 2006: Lish et al.. 1984b). using concurrent
and historical control incidence (with highest dose dropped); BMR = 10% ER
Control
Tumor type
BMDioPct
(mg/kg-d)
BMDLioPct
(mg/kg-d)
Concurrent
Hepatocellular adenomas or carcinomas
25.5
14.2
Alveolar/bronchiolar adenomas or
carcinomas
29.9
14.9
Liver + lung
13.8
8.53
Historical3
Hepatocellular adenomas or carcinomas
37.5
18.7
Alveolar/bronchiolar adenomas or
carcinomas
24.1
13.2
Liver + lung
18.3
4.86
historical control data from NTP (Haseman et al., 1985).
D-68
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
APPENDIX E. SUMMARY OF SCIENCE ADVISORY
BOARD (SAB) PEER REVIEW COMMENTS AND
EPA'S DISPOSITION
The Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine (RDX), dated September
2016, underwent a formal external peer review in accordance with U.S. Environmental Protection
Agency (EPA) guidance on peer review (U.S. EPA. 20151. This peer review was conducted by the
Chemical Assessment Advisory Committee (CAAC) Augmented for Review of the Draft IRIS RDX
Assessment (SAB-CAAC RDX panel) of EPA's Science Advisory Board (SAB). An external peer
review workshop was held on December 12-14, 2016. Public teleconferences of the SAB-CAAC
RDX panel were held on November 17, 2016 and April 13 and 17, 2017. The SAB held a public
meeting on August 29, 2017 to conduct a quality review of the draft peer review report The final
report of the SAB was released on September 27, 2017.
The SAB was tasked with providing feedback in response to charge questions that
addressed scientific issues related to the hazard identification and dose-response assessment of
RDX. Key recommendations of the SAB and EPA's responses to these recommendations, organized
by charge question, follow.
1. Literature Search/Study Selection and Evaluation
Charge Question 1. The section on Literature Search Strategy / Study Selection and Evaluation
describes the process for identifying and selecting pertinent studies. Please comment on
whether the literature search strategy, study selection considerations, including exclusion
criteria, and study evaluation considerations, are appropriate and clearly described. Please
identify additional peer-reviewed studies that the assessment should consider.
Key Recommendation: EPA should include a literature search on the role of GABAergic systems in
brain development, and how this knowledge can inform a better understanding of the potential
developmental neurotoxicity of RDX.
Response: Along with review of references provided by the SAB in the peer review report, a
targeted literature search was performed in PubMed to identify literature to supplement discussion
in the assessment on the generalized role of GABAergic systems in brain development As a result
of this review, additional language on the role of GABAergic signaling has been added as part of a
significantly expanded discussion on potential developmental neurotoxicity of RDX in Section 1.3.3
E-l
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
of the assessment (Susceptible Populations and Life Stages for Cancer and Noncancer Outcomes).
Additional language on developmental neurotoxicity has also been included in the discussions of
nervous system hazard (1.2.1), derivation of candidate values (2.1.3), and uncertainties in the
derivation of the reference dose (2.1.7).
Key Recommendation: EPA should not exclude nonmammalian species as they may bring important
mechanistic insight into the draft assessment.
Response: The Literature Search and Screening Strategy section was revised to clarify that studies
in nonmammalian and other ecological species were not excluded from the RDX assessment and to
recognize the potential use of such studies. Along with revisions to text in the assessment,
Figure LS-1 and Table LS-1 were revised to indicate more clearly that nonmammalian/ecological
studies were tracked as supplementary studies and used as a source of information to inform the
assessment of RDX health effects and potential modes of action.
Key Recommendation: EPA should clarify its reasoning and approach for including or excluding
nonmammalian species studies and secondary references.
Response: The Literature Search and Screening Strategy section, including text (see Figure LS-1 and
Table LS-1), was revised to clarify that studies in nonmammalian/ecological species were tracked
as supplementary information. Secondary references (e.g., regulatory documents, reviews, risk
assessments) were excluded as primary sources of health effects data, but were tracked and
considered as sources of information in assessing the health effects of RDX.
2. Toxicokinetic Modeling
Charge Question 2a. Are the conclusions reached based on EPA's evaluation of the models
scientifically supported? Do the revised PBPK models adequately represent RDX
toxicokinetics? Are the model assumptions and parameters clearly presented and
scientifically supported? Are the uncertainties in the model appropriately considered and
discussed?
Key Recommendation: None provided.
Response: No response required.
Charge Question 2b. The average concentration of RDX in arterial blood (expressed as area
under the curve) was selected over peak concentration as the dose metric for interspecies
extrapolation for oral points of departure (PODs) derived from rat data. Is the choice of dose
metric for each hazard sufficiently explained and appropriate? The mouse PBPK model was
E-2
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
not used to derive PODs for noncancer or cancer endpoints because of uncertainties in the
model and because of uncertainties associated with selection of a dose metric for cancer
endpoints. Is this decision scientifically supported?
Key Recommendation: While current approaches for dose metrics are generally appropriate, the
basis for the choice of dose metric for the prostatitis endpoint should be described.
Response: The basis for selecting area under the curve (AUC) for RDX concentration in arterial
blood as the internal dose metric for analyzing effects on the prostate, as well as for analyzing
effects on the kidney and bladder, was added to Section 2.1.2, Methods of Analysis. Tissue-specific
dose metrics for these organs were not available in the PBPK model. AUC was selected as the dose
metric because effects on the prostate, kidney, and bladder were observed only after subchronic or
chronic exposure to RDX (i.e., there is no evidence that effects were associated with peak exposure)
and because greater confidence was placed in model estimates of blood AUC.
Charge Question 2c. In Section 2.1.3 of the draft assessment, an uncertainty factor of 10 for
human variation is applied in the derivation of the RfD. Does the toxicokinetic modeling
support the use of a different factor instead?
Key Recommendation: None provided.
Response: In its review of the draft RDX assessment (SAB. 20171. the SAB expressed agreement
with the uncertainty factor (UF) of 10 to account for human variation uncertainty factor (UFh).
Accordingly, no recommendations for use of an alternative UF were offered. Consistent with the
position of the SAB, no changes to the UFh were made.
3. Hazard Identification and Dose-Response Assessment
3.1. Nervous System Effects
Charge Question 3a(i). The draft assessment concludes that nervous system toxicity is a
human hazard of RDX exposure. Please comment on whether the available human, animal,
and mechanistic studies support this conclusion. Are all hazards to the nervous system
adequately assessed? Is there an appropriate endpoint to address the spectrum of effects?
Key Recommendation: Lack of studies on neurodevelopmental toxicity, as well as cognitive and
behavioral effects of RDX, should be recognized in the assessment (see discussion in Section 3.3.1.4,
Database Uncertainty Factor [UFd]).
Response: Consistent with the response to Charge Question 1, the revised assessment has an
expanded treatment on the lack of developmental neurotoxicity, and notes the general paucity of
E-3
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
information on subconvulsive and/or cognitive and behavioral effects that could be associated with
exposure to RDX. These observations have been integrated into multiple parts of the assessment,
as relevant, including: Nervous System Effects (1.2.1), Integration and Evaluation of Effects Other
than Cancer (1.3.1), Susceptible Populations and Life Stages for Cancer and Noncancer Outcomes
(1.3.3), Derivation of Candidate Values (2.1.3), and Uncertainties in the Derivation of the Reference
Dose (2.1.7).
Charge Question 3a(ii). Please comment on whether the selection of studies reporting
nervous system effects is scientifically supported and clearly described. Considering the
difference in toxicokinetics between gavage and dietary administration (described in
Appendix C, Section C.l, and in the context of specific hazards in the toxicological review), is
it appropriate to consider the Cholakis etal. (1980) study, which used gavage administration?
Is the characterization of convulsions as a severe endpoint, and the potential relationship to
mortality, appropriately described?
Key Recommendation: The problem maintaining uniform dose suspensions should be identified in
the EPA assessment as a critical study limitation that increases uncertainty in deriving the RfD
based on the Cholakis etal. f 19801 study.
Response: A thorough evaluation of methods used to prepare test diets and suspensions was
performed for all repeated-dose studies of RDX, focusing on the variability in actual concentrations
of the test material and the extent to which target (nominal) concentrations were achieved. The
results of this evaluation were documented in Experimental Animal Studies/Exposure under the
section Literature Search Strategy | Study Selection and Evaluation. The new text included
discussion of issues with maintaining uniform dose suspensions in the Cholakis etal. f!9801 study.
In addition, Section 2.1.4, Derivation of Organ/System-Specific Reference Doses was revised to
include discussion of the highly variable concentrations of RDX in the dose suspensions in Cholakis
etal. (1980) as one of the factors considered in selecting the Crouse etal. (2006) study as the basis
for the nervous system reference value.
Charge Question 3a(iii). Is the selection of convulsions as the endpoint to represent this
hazard scientifically supported and clearly described? Are the calculations of PODs for these
studies scientifically supported and clearly described? Is the calculation of the HEDs for
these studies scientifically supported and clearly described? Does the severity of
convulsions warrant the use of a benchmark response level of 1% extra risk? Is calculation
of the lower bound on the benchmark dose (BMDL) for convulsions appropriate and
consistent with the EPA's Benchmark Dose Guidance?
E-4
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Key Recommendation: EPA should consider using a benchmark response (BMR) of 5% for their
dose-response modeling of the Crouse etal. f20061 data, while addressing the uncertainty of using
data on a frank effect (convulsions in this case) as the basis of an RfD with a larger database
uncertainty factor.
Response: Consistent with the recommendations of the SAB, the BMR used for dose-response
modeling of convulsion incidence data from both the Crouse etal. (20061 and Cholakis etal. (19801
studies was changed from 1 to 5% ER. The uncertainty associated with using data on a frank effect
(convulsions) in the absence of adequate investigation of the potential for RDX to induce subclinical
cognitive and behavioral neurological effects was addressed with the application of a larger
database uncertainty factor. Section 1.2.1, Methods of Analysis/Nervous System Effects, was
revised by adding a discussion of the considerations involved in selecting both a 1 and 5% ER BMR,
and the justification for the final selection of a 5% BMR. Appendix D, Tables D-3 and D-4, were
revised to present results for the benchmark dose (BMD) analysis based on a 5% ER.
Key Recommendation: If EPA decides to use a BMR of 1% for the dose-response assessment using
Crouse etal. (20061. EPA should justify why the greater conservatism in risk assessment required
for a frank effect (due to the lack of incidence data for less severe endpoints) is better dealt with
through a lower BMR than through application of UFd-
Response: As noted above, the BMR was revised to 5% ER consistent with the recommendations of
the SAB.
Key Recommendation: If EPA decides to use a BMR of 1% for Crouse etal. (20061. EPA should
provide clear justification for why a 1% BMR is more appropriate than a 5% BMR for RDX, given the
greater uncertainty introduced into the dose-response assessment for RDX using a BMR of 1%.
Response: As noted above, the BMR was revised to 5% ER consistent with the recommendations of
the SAB.
Charge Question 3a(iv). Is the application of uncertainty factors to these PODs scientifically
supported and clearly described? The subchronic and database uncertainty factors
incorporate multiple considerations; please comment specifically on the scientific rationale
for the application of a subchronic uncertainty factor of 1 and a database uncertainty factor
of 3.
Key Recommendation: Consistent with EPA guidance for UFs, the SAB strongly suggests applying
the full default UFd of 10 to account for data gaps for developmental neurotoxicity, lack of incidence
E-5
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
data for less severe neurological effects resulting in use of a severe effect (convulsions) as a basis
for the RfD, and the proximity of lethal doses to convulsive doses.
Response: Consistent with the recommendation by the SAB, the revised assessment of RDX applies
a database deficiencies uncertainty factor (UFd) of 10, recognizing significant gaps in the available
information to evaluate RDX neurotoxicity. Clarifying text emphasizing the lack of information on
subconvulsive effects of RDX exposure has been added throughout relevant sections of the
assessment, most notably in the discussion on susceptible populations and life stages (1.3.3), and
the rationale for selecting the UFd. Additionally, clarifying language was added to the examples of
information that would reduce database uncertainty to emphasize the need for additional data that
informs subconvulsive and other cognitive and behavioral endpoints.
Key Recommendation: EPA should discuss whether potential neurodevelopmental effects of RDX
would be sufficiently addressed by the default UFd of 10, given that the mechanism of RDX argues
there would likely be developmental neurotoxic effects and the other database uncertainties
(lethality at convulsive doses, other less severe neurotoxic effects that may have a lower LOAEL)
that also need to be addressed by the UFd.
Response: EPA raised the UFd from 3 to 10, recognizing that the database for RDX contains
important gaps in the evaluation of neurotoxic effects, including a lack of developmental
neurotoxicity studies as well as studies on subconvulsive and/or cognitive or behavioral effects.
Additionally, text was added to the assessment (see Section 1.3.3) describing how bicuculline, a
GABAa receptor antagonist with a similar mode of action to RDX, caused developmental and
behavioral impairment in neonatal mice at a dose 1/10 the dose that provoked seizures in adult
mice; in the absence of any RDX-specific information, this observation is consistent with application
of a UFd of 10. The UFh accounts for variation in susceptibility in the human population (including
life-stage variability). A UFh of 10 was also applied in this assessment to account for the possibility
that the RDX database does not allow for a full characterization of the exposure-response
relationship in the most sensitive portions of the human population [for further discussion on the
application of UFh,please see U.S. EPA (2002)]. EPA concluded thatthe application of a UFd of 10
along with a UFh of 10 sufficiently addressed the uncertainty.
Key Recommendation: SAB recommends that EPA reconsider the UF for subchronic-to-chronic
extrapolation (UFs), and at a minimum, provide stronger justification for a UFs of 1.
Response: The discussion of the UFs in Section 2.1.3 was revised to consider support for both a UFs
of 1 and 3 when applied to data from gavage studies of less-than-chronic duration. A UFs of 1 was
retained based on observations that convulsions occurred shortly after dosing (minutes to hours)
across most studies and generally did not appear to be influenced by duration of exposure.
E-6
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Charge Question 3a(v). Is the organ/system-specific reference dose derived for nervous
system effects scientifically supported and clearly characterized?
Key Recommendation: EPA should justify the rationale for utilizing the dose-response data of
Crouse etal. (20061 in preference to Cholakis etal. (19801 as the primary basis for the RfD.
Response: As discussed in response to Charge Question 3a(ii), a discussion of deficiencies in the
Cholakis etal. (19801 developmental toxicity study introduced by the highly variable
concentrations of RDX in dose suspensions was added to the Literature Search Strategy | Study
Selection and Evaluation section (under Experimental Animal Studies/Exposure) and to
Section 2.1.4, Derivation of Organ/System-Specific Reference Doses. The uncertainties associated
with the administered gavage doses in the Cholakis etal. (19801 study were added as further
justification for preferring dose-response data from Crouse etal. (20061 over data from Cholakis et
al. (19801 as the basis for the nervous system reference value.
Key Recommendation: The SAB recommends increasing the UFd from 3 to 10.
Response: As discussed in response to Charge Question 3a(iv), the UFd in the revised assessment
was increased from 3 to 10.
Key Recommendation: The SAB recommends revisiting the UFs and providing a better justification,
at a minimum, for the use of a UFs of 1.
Response: As discussed in response to Charge Question 3a(iv), justification of the UFs was revised
to better support the value of 1 for this uncertainty factor.
3.2. Kidney and Other Urogenital System Effects
Charge Question 3b(i). The draft assessment concludes that kidney and other urogenital
system toxicity is a potential human hazard of RDX exposure. Please comment on whether
the available human, animal, and mechanistic studies support this conclusion. Are all
hazards to kidney and urogenital system adequately assessed? Is the selection of
suppurative prostatitis as the endpoint to represent this hazard scientifically supported and
clearly described?
Key Recommendation: Suppurative prostatitis should not be used as a surrogate marker of renal
and urogenital effects, and instead, be considered a separate hazard of RDX exposure (see also
Section 3.3.2.5.) for quantitative risk assessment.
E-7
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Response: As recommended by the SAB, the incidence of suppurative prostatitis was removed from
discussion on urinary system effects, and instead moved to a new section of the Hazard
identification, Prostate Effects (1.2.3). The revised discussion on prostate effects was expanded to
discuss prostate-specific findings and the broader continuum of inflammation. Consistent with SAB
recommendations, the incidence of suppurative prostatitis was selected as the endpoint most
representative of prostate effects. A prostate-specific reference dose was derived based on this
endpoint as well.
Key Recommendation: The description and analysis of prostatitis should be expanded to include
discussion of both chronic and suppurative inflammation.
Response: Text was added to the new prostate effects section (see Section 1.2.3) to discuss both
chronic and suppurative inflammation. A new table was added to the assessment identifying the
incidence of all types of prostate inflammation in Levine etal. (1983).
Key Recommendation: The description of the various uncertainties regarding the Levine et al.
(1983) rat study should be expanded to include commentary on lack of detail on methods used in
histopathological evaluations, lack of peer review, and the impact of the high prevalence of fighting
in highest dose rats.
Response: Additional information on potential differences in histopathology methods and the
potential impact of fighting was included in the Prostate Effects section (see Section 1.2.3), as well
as discussion on the selection of the overall reference dose (see Section 2.1.5). It should be noted
that as part of the assessment development process, EPA had the 2-year bioassays by Levine et al.
(1983) and Lish etal. (1984b), the subchronic toxicity study by Crouse etal. (2006), and the
collection of repeated-dose studies reported in (Cholakis etal., 1980) peer reviewed. The results of
these peer reviews are discussed in the Study Evaluation section of the assessment and are
available at https://hero.epa.gov/hero/.
Charge Question 3b(ii). Is the selection of the Levine etal. f 19831 study that describes kidney
and other urogenital system effects scientifically supported and clearly described?
Key Recommendation: Improve the discussion and analysis of renal effects observed in studies
other than those reported by Levine etal. (1983).
Response: Synthesis of the evidence for kidney and urinary bladder effects associated with RDX
exposure in experimental animals was substantially revised. The section title was changed to
"Urinary System (Kidney and Urinary Bladder) Effects" to reflect the revised focus of the section,
and more detailed findings from the subchronic studies were added. The evidence table pertaining
E-8
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
to urinary system effects (see Table 1-5) was revised by including additional histopathological
findings from the 13-week studies by Levine etal. f!981al and Martin and Hart f19741. The
exposure-response array (see Figure 1-2) was revised to be consistent with changes made to the
evidence table.
Key Recommendation: Include a brief discussion of the marked sex difference in the renal toxicity
in rats due to RDX exposure.
Response: Discussion of the marked sex difference in kidney response to RDX based on the Levine
etal. (1983) study was added to the synthesis text in Section 1.2.2 and to the integration of
evidence for urinary system effects.
Charge Question 3b(iii). Is the calculation of a POD for this study scientifically supported
and clearly described? Is the calculation of the HED for this study scientifically supported
and clearly described?
Key Recommendation: The SAB strongly recommends that suppurative prostatitis should be used
as a stand-alone endpoint, separate from kidney and other urogenital system endpoints for
calculation of the POD and HED.
Response: Consistent with the SAB's recommendation, EPA revised the assessment to develop a
separate evaluation of prostate effects (see Section 1.2.3). Suppurative prostatitis was identified as
the endpoint most representative of prostate effects and for calculating the related POD and human
equivalent dose (HED).
Charge Question 3b(iv). Is the application of uncertainty factors to the POD scientifically
supported and clearly described?
Key Recommendation: Develop or cite documentation for the use of organ-specific reference values
for individual chemicals, including how these would be used in assessing the combined noncancer
health impacts of multiple agents acting at a common site in cumulative risk assessments.
Response: Section 2.1.4 was expanded to provide examples of the use of organ/system-specific
reference values in cumulative risk assessments, including how these values can be used to refine
the hazard index approach described in EPA's Risk Assessment Guidance for Superfund (U.S. EPA.
19891.
Key Recommendation: A separate RfD should be derived for renal papillary necrosis and the
associated renal inflammation of the kidney and urogenital system.
E-9
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Response: An organ/system-specific oral reference dose (RfD) for effects on the urinary system
(kidney and urinary bladder) was derived separate from one for prostate effects. Section 1.3.1 and
Chapter 2 were revised accordingly to characterize effects of RDX on the prostate as separate from
the effects on the kidney and urinary bladder. As described in Sections 2.1.1 to 2.1.4, dose-response
analyses were conducted for kidney (medullary papillary necrosis) and urinary bladder
(hemorrhagic/suppurative cystitis) data sets as reported by Levine etal. (1983). A no-observed-
adverse-effect level (NOAEL) was used as the POD for incidence of medullary papillary necrosis in
the kidney because this data set was considered unsuitable for BMD modeling; incidence data for
hemorrhagic/suppurative cystitis in the urinary bladder were amenable to BMD modeling. Because
the PODs derived from these two data sets were similar, a single organ/system-specific RfD for the
urinary system was developed.
Key Recommendation: The male accessory sex glands should be designated as a separate organ
system with a separate RfD derived for suppurative prostatitis.
Response: A new section was developed in the hazard identification (see Section 1.2.3) that
specifically evaluated the available data to characterize effects of RDX on the prostate. As described
in revised text in Section 1.3.1, EPA determined that the incidence of suppurative prostatitis was
the endpoint most appropriate for representing prostate effects. The Levine etal. (1983) study was
selected for dose-response analysis, as the only study to identify an increased incidence of
suppurative prostatitis (see Sections 2.1.1 to 2.1.4). Incidence data were amenable to BMD
modeling with a BMR of 10% extra risk, and an organ/system-specific RfD for prostate effects was
derived.
Charge Question 3b(v). Is the organ/system-specific reference dose derived for kidney and
other urogenital system effects scientifically supported and clearly characterized?
Key Recommendation: Separate RfDs should be derived for renal papillary necrosis and the
associated renal inflammation and for suppurative prostatitis.
Response: As discussed in response to Charge Question 3b(iv), separate organ/system-specific RfDs
were derived for the urinary system (based on medullary papillary necrosis of the kidney and
hemorrhagic/suppurative cystitis of the urinary bladder) and prostate (based on suppurative
prostatitis). Section 1.3.1 and Chapter 2 were revised accordingly.
Key Recommendation: Available data are not consistent enough across species, doses, sex, or time
points to recommend separate candidate RfDs for other, milder renal effects (tubular nephrosis,
E-10
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
epithelial vacuolization, and mineralization) found in subchronic studies in mice, rats, and
monkeys.
Response: EPA agrees that evidence for the milder renal effects found in subchronic studies of RDX
in mice, rats, and monkeys was inconsistent Further, findings from the subchronic studies that
provided some evidence of renal effects were difficult to interpret. For example, positive findings
were observed only at a very high dose (320 mg/kg-day) in mice, minimal to mild mineralization of
the medulla in monkeys was not recognized as treatment related by the study authors, and renal
tubular epithelial-lined cysts in the kidney cortex were reported in a study that performed
histopathology only for F2-generation rats, with exposure to RDX limited to gestation and weaning.
Discussion of the findings from the subchronic studies in Section 1.2.2, Urinary System (Kidney and
Urinary Bladder) Effects, was revised to better characterize the nature of these findings. In
addition, the discussion of kidney and urinary bladder effects in Section 1.3.1 (Effects Other Than
Cancer) was revised to include characterization of subchronic findings as limited and inconsistent.
EPA agrees that the renal findings from subchronic studies do not support derivation of a candidate
reference value. Only kidney and urinary bladder data from the 2-year bioassay by Levine et al.
(1983) were used for dose-response analysis.
3.3. Developmental and Reproductive System Effects
Charge Question 3c(i). The draft assessment concludes that there is suggestive evidence of
male reproductive effects associated with RDX exposure, based on evidence of testicular
degeneration in male mice. The draft assessment did not draw any conclusions as to
whether developmental effects are a human hazard of RDX exposure. Please comment on
whether the available human, animal, and mechanistic studies support these decisions. Are
other hazards to human reproductive and developmental outcome adequately addressed?
Key Recommendation: Due to significant weaknesses of the findings in the rat and mouse studies,
RDX should not be considered as having suggestive evidence of male reproductive effects.
Response: The section on male reproductive effects was revised to be consistent with
recommendations from the SAB. In re-evaluating the available evidence, EPA concurred with the
SAB's determination that the available evidence did not support an association between RDX
exposure and male reproductive effects. The hazard determination for this hazard was revised to
"there is insufficient information to assess male reproductive toxicity following exposure to RDX."
In light of the revised hazard determination, discussion of male reproductive effects was moved to
Section C.3.2, Other Noncancer Effects, in the Supplemental Information. A summary of the
evidence for male reproductive effects was included in Section 1.2.6 of the Toxicological Review
E-ll
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
with other noncancer effects for which there is inadequate information to assess an association
with RDX exposure.
Charge Question 3c(ii). Is the selection of the Lish etal. f!984bl study that describes male
reproductive system effects scientifically supported and clearly described?
Key Recommendation: SAB finds that derivation of a reproductive-system specific toxicity value is
not justified, as there have been no convincing studies showing significant male reproductive
toxicity.
Response: The Toxicological Review was revised to remove derivation of an organ/system-specific
RfD for male reproductive effects from Chapter 2.
Charge Question 3c(iii). Is the calculation of a POD for this study scientifically supported and
clearly described? Is the calculation of the HED for this study scientifically supported and
clearly described?
Key Recommendation: No POD for reproductive endpoints should be calculated from the existing
data and therefore there is no need to calculate the HED.
Response: As noted in response to Charge Question 3c(ii), the Toxicological Review was revised to
remove derivation of an organ/system-specific RfD for male reproductive effects from Chapter 2,
including calculation of a POD and HED.
Charge Question 3c(iv). Is the application of uncertainty factors to the POD scientifically
supported and clearly described?
Key Recommendation: Since no valid POD should be calculated for reproductive endpoints, there is
no need to discuss UFs for reproductive endpoints.
Response: No response required.
Charge Question 3c(v). Is the organ/system-specific reference dose derived for reproductive
system effects scientifically supported and clearly characterized?
Key Recommendation: No RfD based on male reproductive toxicity should be calculated since no
valid POD can be estimated.
Response: As noted in response to Charge Question 3c(ii), the Toxicological Review was revised to
remove derivation of an organ/system-specific RfD for male reproductive effects from Chapter 2.
E-12
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
3.4. Oth er Non can cer Hazards
Charge Question 3d. The draft assessment did not draw any conclusions as to whether liver,
ocular, musculoskeletal, cardiovascular, immune, or gastrointestinal effects are human
hazards of RDX exposure. Please comment on whether the available human, animal, and
mechanistic studies support this decision. Are other noncancer hazards adequately
described?
Key Recommendation: For each of the other noncancer hazards discussed in the draft assessment,
add a summary statement regarding whether the available studies do, or do not, support a
conclusion that the identified toxicity is a potential human hazard. Include an explanation of the
rationale for reaching the conclusion, taking into consideration the additional information
pertaining to liver effects, the muscle injury, immune system, neuroinflammation and
gastrointestinal effects.
Response: In Section C.3.2 in the Supplemental Information, a statement was added for each
organ/system, based on the synthesis and integration of evidence from human and animal studies,
as to whether the available health effects information was adequate to support a hazard
determination. Additionally, other organ/systems were included during revision, and information
was added where warranted (e.g., neuroinflammation). Consistent across all endpoints included in
Other Noncancer Effects, there is insufficient information to assess toxicity following exposure to
RDX. The rationale for these conclusions, taking into consideration consistency of the findings
within and across studies, biological significance of the findings, and study confidence, was
provided with each organ/system-specific discussion.
Key Recommendation: Include as a potential noncancer hazard the available subchronic and
chronic data on body weight/body-weight gain, and whether the studies do, or do not, support a
conclusion that body-weight effects represent a potential systemic human hazard. Discuss the
rationale for the conclusion and explain why body-weight effects are, or are not, carried forward to
the dose-response analysis.
Response: Terminal body-weight data from subchronic and chronic studies of RDX were
summarized in an evidence table (see Table C-18), and a synthesis of the evidence for body-weight
changes associated with RDX exposure was added to Section C.3.2, Other Noncancer Effects, in the
Supplemental Information. In general, biologically significant decreases in body-weight gain
(considered to be a decrease of ~10% relative to controls) were observed at >40 mg/kg-day—RDX
doses that also produced severe toxicity in animals (e.g., serious kidney toxicity in male rats and
lethality in other studies). At lower doses, no apparent pattern of treatment-related body-weight
E-13
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
change within or across studies was observed. Because decreased body-weight gain appears to be
secondary to other primary targets of toxicity that were brought forward for dose-response
analysis, change in body weight was not considered a systemic hazard in and of itself and therefore
was not carried forward for dose-response analysis.
3.5. Cancer
Charge Question 3e(i). There are plausible scientific arguments for more than one hazard
descriptor as discussed in Section 1.3.2. The draft assessment concludes that there is
suggestive evidence of carcinogenic potential for RDX, and that this descriptor applies to all
routes of human exposure. Please comment on whether the available human, animal, and
mechanistic studies support these conclusions.
Key Recommendation: Strengthen and make more specific the justification for selecting the
"suggestive evidence of carcinogenic potential" descriptor rather than the "likely to be carcinogenic to
humans" descriptor.
Response: The justification for selecting the cancer descriptor of "suggestive evidence of
carcinogenic potential" was strengthened by adding more detailed information on the individual
cancer bioassays to both Section 1.2.7 (synthesis of evidence for carcinogenicity) and Section 1.3.2
(cancer descriptor). Specifically, the revised text included additional discussion of the mortality
time course observed in the bioassays in mice (Lish etal.. 1984b 1 and rats fLevine etal.. 19831 (see
Section 1.2.7); clarification of the evidence for rat liver tumors (see Section 1.2.7); summary of
concerns raised by the Pathology Working Group (PWG) regarding female mouse liver tumors and
observations regarding the low liver tumor incidence in female controls (see Sections 1.2.7 and
1.3.2); discussion of the evidence for cancer mode of action, including lack of evidence for precursor
events leading to liver and lung tumor response (see Section 1.3.2); and lack of a full pathology peer
review in all three cancer bioassays of RDX (see Section 1.2.7).
Charge Question 3e(ii). As noted in EPA's 2005 Guidelines for Carcinogen Risk Assessment,
"When there is suggestive evidence, the Agency generally would not attempt a
dose-response assessment, as the nature of the data generally would not support one;
however, when the evidence includes a well-conducted study, quantitative analyses may be
useful for some purposes, for example, providing a sense of the magnitude and uncertainty
of potential risks, ranking potential hazards, or setting research priorities." Does the draft
assessment adequately explain the rationale for quantitative analysis, considering the
uncertainty in the data and the suggestive nature of the weight of evidence, and is the
selection of the Lish etal. (1984b) study for this purpose scientifically supported and clearly
described?
E-14
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Key Recommendation: None provided.
Response: In its review of the draft RDX assessment fSAB. 20171. the SAB stated that EPA's
rationale for quantitative cancer dose-response analysis using the Lish etal. f!984bl study was
supported scientifically and clearly described. Accordingly, the SAB offered no recommendations
for revising the rationale for undertaking a quantitative analysis. Consistent with the position of
the SAB, no changes to this section of the assessment were made.
Charge Question 3e(iii). Are the calculations of PODs and oral slope factors scientifically
supported and clearly described?
Key Recommendation: For liver cancer, perform and discuss results from additional BMD modeling
(i.e., a sensitivity analysis documented in the Supplemental Materials) that examines and discusses:
•	The impact of low concurrent controls on model choice and the POD estimate.
•	The effect of including/excluding the highest dose has on model choice and the POD
estimate.
Response: A sensitivity analysis was conducted and added to Appendix D (see Section D.2.4) to
investigate (1) the effect of dropping the highest dose group on model choice and the POD estimate,
(2) the impact of low concurrent control liver tumor incidence on model selection and the POD
estimate, and (3) the fit of the multistage models in the low-dose region. EPA's more detailed
response to item (3) is provided under the next key recommendation.
To examine the effect of including/excluding the highest dose on model choice and the POD
estimate, EPA remodeled the female mouse liver and lung tumor data from Lish etal. f!984bl after
excluding the highest dose group from the analysis because of the high mortality in that group (i.e.,
loss of almost half the animals in that group before the dose was dropped from 175 to
100 mg/kg-day at Week 11 of the 2-year study). Excluding the high-dose group from the analysis
resulted in a change in the oral slope factor (OSF) from 0.04 to 0.08 (mg/kg-day)1. In light of the
high mortality, EPA determined that it was appropriate to exclude the high-dose group from the
analyses because mice that survived the 11-week exposure to 175 mg/kg-day RDX maynothave
constituted an unbiased representation of the population of animals exposed for the full 2-year
study (i.e., those more or less sensitive to RDX than the animals in the general population). The
rationale for dropping the high-dose group from the data set used for dose-response modeling was
provided in Section 2.3.1, and the documentation of the modeling in Appendix D was revised to
present the reanalysis with the high-dose group dropped.
E-15
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
To address the impact of low concurrent control liver tumor incidence on model selection and the
POD, EPA conducted a dose-response analysis of female mouse liver and lung tumor data using
historical control data from National Toxicology Program [NTP; (Haseman etal.. 19851]. This
analysis revealed that historical control incidence had a relatively small impact on the POD (i.e.,
BMDLio) used for deriving the OSF. BMDLio values for liver and lung tumors using concurrent
control data and excluding the high dose from the analysis were 14.2 and 14.9 mg/kg-day,
respectively. Replacing the incidence of tumors in concurrent controls with NTP historical control
data yielded BMDLio values for liver and lung tumors of 18.7 and 13.2 mg/kg-day, respectively.
EPA notes a preference for using concurrent control data, especially as in the current case where
historical control data originated from a different laboratory where there was potential for cross-
study differences in diet, laboratory, pathological evaluation, and animal provider. The analysis
using historical control data from NTP is presented as part of the sensitivity analysis added to
Appendix D, Section D.2.4; however, EPA retained the analysis using concurrent control tumor data
as the basis for the OSF.
Key Recommendation: Provide details and discuss the adequacy of the fit of the multistage model to
available data.
Response: The adequacy of the fit of the multistage model to the female mouse liver and lung tumor
data was addressed in the context of the sensitivity analysis added to Appendix D, Section D.2.4.
As discussed in Section D.2.4, EPA agreed with the SAB that the multistage models exhibited an
inferior fit to the female mouse liver tumor data with all doses included compared to a subset of
nonmultistage models in benchmark dose software (BMDS); however, with the high-dose group
dropped, the multistage models yielded a fit comparable to the best-fitting model. For female
mouse lung tumors, the multistage models exhibited fits comparable to the best-fitting model
among multistage and nonmultistage models both with all doses included and with the high-dose
group dropped.
Key Recommendation: Provide a better and more detailed description of the MS-COMBO
methodology (in the Supplemental Information document) and ensure that this description
discusses the issues below:
•	importance of the assumption of independence,
•	why this assumption is needed,
•	how this assumption might be examined statistically given adequate study
data/documentation,
E-16
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
•	whether the independence of tumor incidence assumption further constrains the tumor-
specific dose-response model form to be the same across included tumor types, and,
•	the extent to which violations of the independence of tumor incidence assumption
negatively affect the estimated POD.
Response: A more detailed and comprehensive description of the MS-COMBO methodology was
added to Appendix D, Section D.2.1; reference to this description of MS-COMBO was added to
Section 2.3.2 of the Toxicological Review. In particular, the importance and testability of the
assumption of independence was discussed in greater detail.
4. Dose-Response Analysis
Charge Question 4a. The draft assessment presents an overall oral reference dose of
3 x 10"3 mg/kg-day, based on nervous system effects as described in the Crouse etal. f20061.
Is this selection scientifically supported and clearly described, including consideration of
mortality as described in Section 2.1.6, and consideration of the organ/system-specific
reference dose derived from the toxicity study by Cholakis etal. fl9801 that is lower (by
approximately fivefold) as described in Section 2.1.4?
Key Recommendation: The SAB agrees that the overall RfD should be based on neurotoxicity as
measured by convulsions in Crouse etal. (20061. but the SAB concludes that the scientific support
for the proposed oral RfD is somewhat lacking primarily due to concerns with the choice of BMR
and the value of the database uncertainty factor and the uncertainty factor for
subchronic-to-chronic extrapolation. This deficiency needs to be rectified.
Response: As noted in responses to Charge Questions 3a(iii) and 3a(iv), EPA revised the choice of
the BMR from 1% to 5% extra risk and the UFd from 3 to 10, consistent with the recommendations
of the SAB. The rationales supporting the revised BMR and UFd are provided in the toxicological
review (see Sections 2.1.2 and 2.1.3). In addition, EPA provided stronger justification for the
uncertainty factor for UFs of 1 (see Section 2.1.3).
Key Recommendation: Since the Cholakis etal. (1980) study suffers from several quality issues, it is
appropriate to give more weight to the Crouse etal. f20061 study with respect to the quantitative
dose-response analysis. The rationale for the selection of Crouse et al. f20061 and setting aside the
Cholakis etal. (1980) study even though it reported a lower NOAEL/LOAEL, should be
strengthened and clarified.
Response: As noted in responses to Charge Questions 3a(ii) and 3a(v), a discussion of deficiencies
in the Cholakis etal. (1980) developmental toxicity study introduced by the highly variable
E-17
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
concentrations of RDX in dose suspensions was added to the Literature Search Strategy | Study
Selection and Evaluation section (under Experimental Animal Studies/Exposure). Section 2.1.4,
Derivation of Organ/System-Specific Reference Doses, was revised to provide further discussion of
the deficiencies in the Cholakis et al. (19801 study and a stronger justification for using
dose-response data from Crouse etal. (20061 [over data from Cholakis etal. (19801] as the basis for
the nervous system reference value.
Key Recommendation: The discussion and key recommendations from Section 3.3.1.3 and
Section 3.3.1.4 [of the SAB report) are all pertinent to the SAB finding that the scientific support for,
and discussion of, the proposed oral RfD for the convulsions endpoint is lacking. These
recommendations are repeated here:
•	EPA should consider using a BMR of 5% for their dose-response modeling of the Crouse et
al. (20061 data while addressing the uncertainty of using data on a frank effect (convulsions
in this case) as the basis of an RfD with a larger database uncertainty factor.
•	If EPA decides to use a BMR of 1% for the dose-response assessment using Crouse et al.
f20061. EPA should justify why the greater conservatism in risk assessment required for a
frank effect (due to the lack of incidence data for less severe endpoints) is better dealt with
through a lower BMR than through application of UFd.
•	If EPA decides to use a BMR of 1% for Crouse etal. f20061. EPA should provide in its
discussion clear justification for why a 1% BMR is more appropriate than a 5% BMR for
RDX, given the greater uncertainty introduced into the dose-response assessment for RDX
using a BMR of 1%.
•	Consistent with EPA guidance for uncertainty factors, the SAB strongly recommends that
EPA apply the full default UFd of 10 to account for data gaps for developmental
neurotoxicity, lack of incidence data for less severe neurological effects resulting in use of a
severe effect (convulsions) as a basis for the RfD, and the proximity of lethal doses to
convulsive doses.
•	EPA should discuss whether potential neurodevelopmental effects of RDX would be
sufficiently addressed by a UFd of 10, given that the mechanism of RDX argues there would
likely be developmental neurotoxic effects and the other database uncertainties (lethality at
convulsive doses, other less severe neurotoxic effects that may have a lower LOAEL) that
also need to be addressed by the UFd-
•	EPA should reconsider the value of the UFs and at a minimum provide stronger justification
for a UFs of 1.
Response: As discussed in response to Charge Questions 3a(iii) and 3a(iv), consistent with the
recommendations of the SAB, the BMR used for dose-response modeling of convulsion incidence
data was changed from 1 to 5% ER, and the UFd was increased from 3 to 10 to account for
E-18
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
significant gaps in neurotoxicity testing for RDX. As discussed in response to Charge Question
3a(iv), the UFs of 1 was retained, and the justification was revised to better support the value of 1
for this uncertainty factor.
Charge Question 4b. The draft assessment does not derive an inhalation reference
concentration as the available studies were insufficient to characterize inhalation hazard
and conduct dose-response analysis, and no toxicokinetic studies of RDX were available to
support development of a PBPK inhalation model. If you believe that the available data
might support an inhalation reference concentration, please describe how one might be
derived.
Key Recommendation: EPA should not attempt to derive an inhalation reference concentration
since neither toxicokinetic data nor exposure levels information from animal or human RDX
inhalation studies are available to make estimation possible.
Response: No response required.
Charge Question 4c. The draft assessment presents an overall oral slope factor of 0.038 per
mg/kg-day based the combination of liver and lung tumors in female mice. Is this derivation
scientifically supported and clearly described?
Key Recommendation: Acknowledge that the issues with the estimation of the POD for estimation
of the cancer slope factor as discussed in Section 3.3.5.3 [of the SAB report) and note that the
associated key recommendations for improving the presentation on the POD also apply to the
estimation of the cancer slope factor.
Response: As discussed in response to Charge Question 3e(iii), EPA addressed the SAB's concerns
regarding the derivation and presentation of the OSF as follows:
•	The female mouse liver and lung tumor data from Lish etal. (1984b) were remodeled after
excluding the highest dose group from the analysis in light of the high mortality in that
group, resulting in a change in the OSF from 0.04 to 0.08 (mg/kg-day)1.
•	A sensitivity analysis was conducted to investigate (1) the fit of the multistage models in the
low-dose region, (2) the effect of dropping the highest dose group, and (3) the impact of low
concurrent controls on model selection and the POD estimate. This sensitivity analysis was
added to Appendix D (see Section D.2.4).
Charge Question 4d. The draft assessment does not derive an inhalation unit risk because
inhalation carcinogenicity data were not available, nor were toxicokinetic studies of
E-19
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
inhalation of RDX available to support development of an inhalation PBPK model. If you
believe that the available data might support an inhalation unit risk, please describe how
one might be derived.
Key Recommendation: EPA should not attempt to derive an inhalation unit risk since there are no
study data available to make estimation possible.
Response: No response required.
5. Executive Summary
Charge Question 5. Does the executive summary clearly and adequately present the major
conclusions of the assessment?
Key Recommendation: None provided.
Response: No response required.
E-20
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
REFERENCES FOR APPENDICES
Angerhofer. R: Davis. G: Balezewski. L. (1986). Teratological assessment of Trinitro - RDX in rats
(pp. 24). (75-51-0573-86). Aberdeen Proving Ground, MD: U.S. Army Environmental
Hygiene Agency, http: //www.dtic.mil/dtic/tr/fulltext/u2/a166249.pdf.
Arfsten. DP: Davenport. R: Schaeffer. DT. (1994). Reversion of bioluminescent bacteria (Mutatox
TM) to their luminescent state upon exposure to organic compounds, munitions, and metal
salts. Biomed Environ Sci 7: 144-149.
ATSDR (Agency for Toxic Substances and Disease Registry). (2012). Toxicological profile for RDX
(Update) [ATSDR Tox Profile], (NTIS/10590213). Atlanta, GA: U.S. Department of Health
and Human Services, Public Health Service.
Bannon. PI. (2006). Toxicology study no. 87-XE-059F-06: Biomarkers of RDX exposure in breath of
swine, August - September 2006. Aberdeen Proving Ground, MD: U.S. Army Center for
Health Promotion and Preventive Medicine.
Bannon. PI: Dillman. IF: Hable. MA: Phillips. CS: Perkins. El. (2009). Global gene expression in rat
brain and liver after oral exposure to the explosive hexahydro-l,3,5-trinitro-l,3,5-triazine
(RPX). Chem Res Toxicol 22: 620-625. http://dx.doi.org/10.1021/tx800444k.
Bannon. PI: Pillman. IF: Perkins. El: Bao. W: Wolfinger. RP: Chu. T: Phillips. CS. (2006). Acute RPX
exposure and gene expression in the rat brain [Abstract], Oxford, United Kingdom: Oxford
University Press. http: //www.toxic0l0gy.0rg/pubs/d0cs/Tox/2006Tox.pdf.
Barsotti. M: Crotti. G. (1949). [Attacchi epileptici come manifestazione di intossicazione
professionale da trimetilen-trinitroamina (T4)]. Med Lav 40: 107-112.
Beller. HR: Tiemeier. K. (2002). Use of liquid chromatography/tandem mass spectrometry to detect
distinctive indicators of in situ RPX transformation in contaminated groundwater. Environ
Sci Techno1 36: 2060-2066. http://dx.d0i.0rg/l 0.1021 /esOl 57696.
Bhushan. B: Trott. S: Spain. TC: Halasz. A: Paquet. L: Hawari. 1. (2003). Biotransformation of
hexahydro-l,3,5-trinitro-l,3,5-triazine (RPX) by a rabbit liver cytochrome P450: Insight
into the mechanism of RPX biodegradation by Rhodococcus sp. strain PN22. Appl Environ
Microbiol 69: 1347-1351. http://dx.doi.Org/10.1128/AEM.69.3.1347-1351.2003.
Bogen. KT. (1990). [Excerpt], In Uncertainty in environmental health risk assessment New York,
NY: Garland Publishing.
Bogen. KT: Seilkop. S. (1993). Investigation of independence in inter-animal tumor type occurrence
within the NTP rodent-bioassay database. (UCRL-IP-115092 ON: PE94006440).
Washington, PC: U.S. Pepartment of Energy, Office of Scientific and Technical Informationg.
https://www.osti.gov/biblio / 10121101.
Bongiovanni. R: Podolak. GE: Clark. LP: Scarborough. PT. (1984). Analysis of Trace Amounts of Six
Selected Poly-Nitro Compounds in Soils. Am Ind Hyg Assoc J 45: 222-226.
http://dx.doi.org/10.1080/15298668491399686.
Bronson. FH: Pesiardins. C. (1977). Reproductive failure in aged CBF1 male mice: Interrelationships
between pituitary gonadotropic hormones, testicular function, and mating success.
Endocrinology 101: 939-945. http://dx.doi.org/10.121Q/endo-101-3-939.
Brown. RP: Pelp. MP: Lindstedt. SL: Rhomberg. LR: Beliles. RP. (1997). Physiological parameter
values for physiologically based pharmacokinetic models. Toxicol Ind Health 13: 407-484.
http://dx.doi.Org/10.1177/074823379701300401.
R-l
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Cao. CT: Reddv. G: Bannon. PI: Tohnson. MS. (2008). In vitro study of hexahydro-l,3,5-trinitro-l,3,5-
triazine (RPX) metabolism in human liver. Aberdeen Proving Ground, MD: U.S. Army Center
for Health Promotion and Preventive Medicine.
Cholakis. 1M: Wong. LCK: Van Goethem. PL: Minor. 1: Short. R: Sprinz. H: Ellis. HV. III. (1980).
Mammalian toxicological evaluation of RDX. (DAMD17-78-C-8027). Kansas City, MO:
Midwest Research Institute, http: //www.dtic.mil/dtic/tr/fulltext/u2 /a092531.pdf.
Cotruvo. TA: Simmon. VF: Spanggord. RI. (1977). Investigation of mutagenic effects of products of
ozonation reactions in water. Ann N Y Acad Sci 298:124-140.
http://dx.doi.Org/10.llll/i.1749-6632.1977.tbl9259.x.
Crouse. LCB: Michie. MW: Maior. M: Tohnson. MS: Lee. RB: Paulus. HI. (2006). Subchronic oral
toxicity of RDX in rats. (Toxicology Study No. 85-XC-5131-03). Aberdeen Proving Ground,
MD: U.S. Army Center for Health Promotion and Preventive Medicine.
http: / /www. dtic.mil /dtic/tr /fulltext/u2 /I0 5 0 9 0 3 .pdf.
Crouse. LCB: Michie. MW: Major. MA: Leach. GT: Reddv. G. (2008). Oral bioavailability of
cyclotrimethylenetrinitramine (RDX) from contaminated site soils in rats. Int J Toxicol 27:
317-322. http://dx.doi.Org/10.1080/10915810802366885.
Davies. TOT: Roberts. DM: Hittarage. A: Buckley. NA. (2007). Oral C-4 plastic explosive in humans - A
case series. Clin Toxicol 45: 454-457. http://dx.doi.org/10.1080/15563650601118044.
Dev. A: Kang. X: Oiu. 1: Du. Y: Tiang. 1. (2016). Anti-inflammatory small molecules to treat seizures
and epilepsy: From bench to bedside [Review], Trends Pharmacol Sci 37: 463-484.
http://dx.doi.Org/10.1016/i.tips.2016.03.001.
Dillev. TV: Tyson. CA: Newell. GW. (1979). Mammalian toxicological evaluation of TNT wastewaters.
Volume II. Acute and subacute mammalian toxicity of TNT and the LAP mixture.
(ADA080957. DAMD17-76-C-6050.). Menlo Park, CA: SRI International.
http: / /www.dtic. mil /docs /citations /AD AO 80 9 5 7.
Evo. UB: Murugan. M: Wu. LI. (2017). Microglia-neuron communication in epilepsy [Review], Glia
65: 5-18. http://dx.doi.org/10.1002/glia.23006.
Fellows. RI: Driver. CR: Cataldo. DA: Harvey. SD. (2006). Bioavailability of hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) to the prairie vole (Microtus ochrogaster). Environ Toxicol Chem 25:
1881-1886. http: //dx.doi.org/10.1897/05-446R1.1.
Fuller. ME: Perreault. N: Hawari. 1. (2010). Microaerophilic degradation of hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) by three Rhodococcus strains. Lett Appl Microbiol 51: 313-318.
http: //dx.doi.org/10.1111 /i.l 472-765X.2010.02897.x.
George. SE: Huggins-Clark. G: Brooks. LR. (2001). Use of a Salmonella microsuspension bioassay to
detect the mutagenicity of munitions compounds at low concentrations. MutatRes 490: 45-
56. http://dx.doi.org/10.1016/S1383-5718r00100150-9.
Goldberg. PI: Green. ST: Nathwani. D: McMenamin. 1: Hamlet. N: Kennedy. PH. (1992). RDX
intoxication causing seizures and a widespread petechial rash mimicking
meningococcaemia. J R Soc Med 85: 181.
Gosden. RG: Richardson. PW: Brown. N: Pavidson. PW. (1982). Structure and gametogenic
potential of seminiferous tubules in ageing mice. J Reprod Fertil 64: 127-133.
Guo. L: Xu. H: Chen. Y: Chang. Y. (1985). Pistribution and metabolism of tritium-labeled hexogen in
white mice. Chinese Journal of Industrial Hygiene and Occupational Piseases 3: 335-339.
Halasz. A: Manno. P: Perreault. NN: Sabbadin. F: Bruce. NC: Hawari. 1. (2012). Biodegradation of
RPX nitroso products MNX and TNX by cytochrome P450 XplA. Environ Sci Technol 46:
7245-7251. http://dx.doi.org/10.1021/es3011964.
Harrell-Bruder. B: Hutchins. KL. (1995). Seizures caused by ingestion of composition C-4. Ann
Emerg Med 26: 746-748. http://dx.doi.org/10.1016/S0196-0644r95170049-8.
Hart. ER. (1974). Subacute toxicity of RPX and TNT in dogs. Final report. (A717530). Kensington,
MP: Litton Bionetics, Inc. http://www.dtic.mil/dtic/tr/fulltext/u2/a035717.pdf.
R-2
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Hart. ER. (1976). Two-year chronic toxicity study in rats. (N00014-73-C-0162). Kensington, MD:
Litton Bionetics, Inc. http://www.dtic.mil/dtic/tr/fulltext/u2/a04016l.pdf.
Haseman. IK: Huff. IE: Rao. GN: Arnold. IE: Boorman. GA: McConnell. EE. (1985). Neoplasms
observed in untreated and corn oil gavage control groups of F344/N rats and (C57BL/6N X
C3H/HeN)Fl (B6C3F1) mice. J Natl Cancer Inst 75: 975-984.
http://dx.doi.Org/10.1093/inci/75.5.975.
Hathaway. TA: Buck. CR. (1977). Absence of health hazards associated with RDX manufacture and
use. J Occup Med 19: 269-272.
Hess-Ruth. A: Crouse. L: Roszell. L. (2007). RDX pilot development neurotoxicity test in rats.
(Toxicology Study No. 85-XC-064Y-07). Aberdeen Proving Ground: U.S. Army Center for
Health Promotion and Preventive Medicine.
Hett. DA: Fichtner. K. (2002). Aplastic explosive by mouth. J R Soc Med 95: 251-252.
http://dx.doi.Org/10.1258/irsm.95.5.251.
Hollander. AI: Colbach. EM. (1969). Composition C-4 induced seizures: A report of five cases. Mil
Med 134: 1529-1530.
Taligama. S: Kale. VM: Wilbanks. MS: Perkins. El: Meyer. SA. (2013). Delayed myelosuppression with
acute exposure to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) and environmental
degradation product hexahydro-l-nitroso-3,5-dinitro-1,3,5-triazine (MNX) in rats. Toxicol
Appl Pharmacol 266: 443-451. http://dx.doi.Org/10.1016/i.taap.2012.ll.022.
Tohnson. MS. (2015). Memorandum for National Center for Environmental Assessment (8101R/Dr.
Louis D'Amico), U.S. EPA- Office ofResearch and Development, 1200 Pennsylvania Ave.,
NW, Washington, DC 20460. Subject: Additional data from the oral subchronic toxicity of
RDX in rats, 2006. (Toxicology study no. 85-XC-5131-03), U.S. Army Center for Health
Promotion and Preventative Medicine, Aberdeen Proving Ground, Maryland; Incidence of
seizure relative to mortality events. (MCHB-IP-T). Aberdeen Proving Ground, MD:
Department of the Army, US Army Institute of Public Health.
Kaplan. AS: Berghout. CF: Peczenik. A. (1965). Human intoxication from RDX. Arch Environ Health
10: 877-883.
Kasuske. L: Schofer. TM: Hasegawa. K. (2009). Two marines with generalized seizure activity. J
Emerg Nurs 35: 542-543. http://dx.doi.Org/10.1016/i.ien.2008.05.001.
Ketel. WB: Hughes. TR. (1972). Toxic encephalopathy with seizures secondary to ingestion of
composition C-4. A clinical and electroencephalographic study. Neurology 22: 871-876.
Knepshield. TH: Stone. WT. (1972). Toxic effects following ingestion of C-4 plastic explosive. In W
Keup (Ed.), Drug Abuse: Current Concepts and Research (pp. 296-300). Springfield, IL:
Charles C. Thomas.
Krishnan. K: Crouse. LCB: Bazar. MA: Major. MA: Reddv. G. (2009). Physiologically based
pharmacokinetic modeling of cyclotrimethylenetrinitramine in male rats. J Appl Toxicol 29:
629-637. http://dx.doi.org/10.1002/\at. 1455.
Kiiciikardali. Y: Acar. HV: Ozkan. S: Nalbant. S: Yazgan. Y: Atasovu. EM: Keskin. 0: Naz. A: Akvatan. N:
Gokben. M: Danaci. M. (2003). Accidental oral poisoning caused by RDX (cyclonite): A report
of 5 cases. J Intensive Care Med 18: 42-46. http://dx.doi.org/10.1177/0885066602239123.
Lachance. B: Robidoux. PY: Hawari. I: Ampleman. G: Thiboutot. S: Sunahara. GI. (1999). Cytotoxic
and genotoxic effects of energetic compounds on bacterial and mammalian cells in vitro.
MutatRes 444: 25-39. http://dx.doi.org/10.1016/Sl 383-5718f99100073-X.
Levin. DE: Hollstein. M: Christman. MF: Schwiers. EA: Ames. BN. (1982). A new Salmonella tester
strain (TA102) with AT base pairs at the site of mutation detects oxidative mutagens. Proc
Natl Acad Sci USA 79: 7445-7449. http://dx.doi.org/10.1073/pnas.79.23.7445.
Levine. BS: Furedi. EM: Gordon. DE: Barklev. IT: Lish. PM. (1990). Toxic interactions of the munitions
compounds TNT and RDX in F344 rats. Fundam Appl Toxicol 15: 373-380.
http://dx.doi.org/10.1016/0272-0590r90190062-0.
R-3
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Levine. BS: Furedi. EM: Gordon. DE: Burns. TM: Lish. PM. (1981a). Thirteen week oral (diet) toxicity
study of trinitrotoluene (TNT), hexahydro-1, 3, 5-trinitro-l, 3, 5-triazine (RDX) and
TNT/RDX mixtures in the Fischer 344 rat. Final report (ADA108447. DAMD17-79-C-9120.
DAMD17-79-C-9161). Chicago, IL: IIT Research Institute.
http://www.dtic.mil/dtic/tr/fulltext/u2/al08447.pdf.
Levine. BS: Furedi. EM: Gordon. DE: Burns. TM: Lish. PM. (1981b). Thirteen week toxicity study of
hexahydro-1,3,5-trinitro-l,3,5-triazine (RDX) in Fischer 344 rats. Toxicol Lett 8: 241-245.
http://dx.doi.org/10.1016/0378-4274r81190108-9.
Levine. BS: Lish. PM: Furedi. EM: Rac. VS: Sagartz. TM. (1983). Determination of the chronic
mammalian toxicological effects of RDX (twenty-four month chronic
toxicity/carcinogenicity study of hexahydro-1,3,5-trinitro-l,3,5-triazine (RDX) in the
Fischer 344 rat): Final report-phase V. Chicago, IL: IIT Research Institute.
Lipscomb. TC: Poet. TS. (2008). In vitro measurements of metabolism for application in
pharmacokinetic modeling [Review], Pharmacol Ther 118: 82-103.
http://dx.doi.0rg/lO.lOl6/i.pharmthera.2OO8.Ol.OO6.
Lish. PM: Levine. BS: Furedi. EM: Sagartz. TM: Rac. VS. (1984a). Determination of the chronic
mammalian toxicological effects of RDX: Twenty-four month chronic
toxicity/carcinogenicity study of hexahydro-1,3,5-trinitro-l,3,5-triazine (RDX) in the
B6C3F1 hybrid mouse (Volumelof3) (pp. 367). (ADA181766. DAMD17-79-C-9161). Fort
Detrick, Frederick, MD: U.S. Army Medical Research and Development Command.
http: / /www. dtic.mil/dtic/tr/fulltext/u2 /al 8176 6.pdf.
Lish. PM: Levine. BS: Furedi. EM: Sagartz. TM: Rac. VS. (1984b). Determination of the chronic
mammalian toxicological effects of RDX: Twenty-four month chronic
toxicity/carcinogenicity study of hexahydro-1,3,5-trinitro-l,3,5-triazine (RDX) in the
B6C3F1 hybrid mouse (Volumesl-3). (ADA181766. DAMD17-79-C-9161). Fort Detrick,
Frederick, MD: U.S. Army Medical Research and Development Command.
Ma. B: Li. H. (1993). Neurobehavioral effects of hexogen on exposed workers. Gongye Weishengyu
Zhiyebin 19: 20-23.
MacPhail. RC: Walker. OP: Cook. L. (1985). Neurotoxicology of cyclotrimethylenetrinitramine
(RDX). Final report (ADA168266). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Health Effects Research Laboratory, Neurotoxicology Division.
http: / /www. dtic.mil /dtic/tr /fulltext/u2 /al 68 2 6 6.pdf.
Maior. MA: Reddv. G: Berge. MA: Patzer. SS: Li. AC: Gohdes. M. (2007). Metabolite profiling of
[14C]hexahydro-1,3,5-trinitro-l,3,5-triazine (RDX) in Yucatan miniature pigs. J Toxicol
Environ Health A 70: 1191-1202. http://dx.doi.org/10.1080/15287390701252717.
Martin. D: Hart. E. (1974). Subacute toxicity of RDX and TNT in monkeys (pp. 1-216). (ADA044650).
Kensington, MD: Litton Bionetics, Inc.
http://www.dtic.mil/dtic/tr/fulltext/u2/a044650.pdf.
Merrill. SL. (1968). Ingestion of an explosive material, composition C-4: A report of two cases.
USARV Med Bull 40: 5-11.
Meyer. SA: Marchand. AT: Hight. IL: Roberts. GH: Escalon. LB: Inouve. LS: MacMillan. DK. (2005). Up-
and-down procedure (UDP) determinations of acute oral toxicity of nitroso degradation
products of hexahydro-1,3,5-trinitro-l,3,5-triazine (RDX) [Abstract], J Appl Toxicol 25: 427-
434. http://dx.doi.org/10.1002 /iatl090.
Musick. TT: Berge. MA: Patzer. SS: Tilch. KR. (2010). Absorption, distribution, metabolism, and
excretion of 14C-RDX following oral administration to minipigs (DAAD05-02-P-2319).
(ADA526472). Madison, WI: Covance Laboratories Inc. http://www.dtic.mil/cgi-
bin/GetTRD oc? AD =AD A5 26472.
Neuwoehner. I: Schofer. A: Erlenkaemper. B: Steinbach. K: Hund-Rinke. K: Eisentraeger. A. (2007).
Toxicological characterization of 2,4,6-trinitrotoluene, its transformation products, and two
R-4
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
nitramine explosives. Environ Toxicol Chem 26: 1090-1099. http://dx.doi.org/10.1897/06-
471R.1.
NIOSH (National Institute for Occupational Safety and Health). (2012). NIOSH pocket guide to
chemical hazards: Cyclonite, CAS 121-82-4. Cincinnati, OH: U.S. Department of Health and
Human Services, Centers for Disease Control and Prevention, National Institute for
Occupational Safety and Health, http: / /www,cdc. gov/niosh /npg/npgdO 169.html.
NRC (National Research Council). (1994). Science and judgement in risk assessment (pp. 908-909).
Washington, DC: National Academy Press.
OSHA (Occupational Safety & Health Administration). (2012). Air contaminants. Table Z-Shipyards.
CFR 29 CFR 1915, Subpt P, App A.
Ozhan. G: Topuz. S: Alpertunga. B. (2003). Determination of cyclonite (RDX) in human plasma by
high-performance liquid chromatography. Farmaco 58: 445-448.
http://dx.doi.org/10.1016/S0014-827Xr03100069-7.
Pan. X: Ochoa. KM: San Francisco. Ml: Cox. SB: Dixon. K: Anderson. TA: Cobb. GP. (2013).
Absorption, distribution, and biotransformation of hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX) in B6C3F1 mice (Mus musculus). Environ Toxicol Chem 32: 1295-1303.
http://dx.doi.org/10.1002/etc.2188.
Pan. X: San Francisco. Ml: Lee. C: Ochoa. KM: Xu. X: Liu. I: Zhang. B: Cox. SB: Cobb. GP. (2007a).
Examination of the mutagenicity of RDX and its N-nitroso metabolites using the Salmonella
reverse mutation assay. MutatRes Genet Toxicol Environ Mutagen 629: 64-69.
http://dx.doi.Org/10.1016/i.mrgentox.2007.01.006.
Pan. X: Zhang. B: Smith. IN: San Francisco. M: Anderson. TA: Cobb. GP. (2007b). N-Nitroso
compounds produced in deer mouse (Peromyscus maniculatus) GI tracts following
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) exposure. Chemosphere 67: 1164-1170.
http://dx.doi.Org/10.1016/i.chemosphere.2006.10.077.
Paquet. L: Monteil-Rivera. F: Hatzinger. PB: Fuller. ME: Hawari. I. (2011). Analysis of the key
intermediates of RDX (hexahydro-l,3,5-trinitro-l,3,5-triazine) in groundwater: Occurrence,
stability and preservation. J Environ Monit 13: 2304-2311.
http://dx.doi.org/10.1039/cleml0329f.
Parker. GA: Reddv. G: Maior. MA. (2006). Reevaluation of a twenty-four-month chronic
toxicity/carcinogenicity study of hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) in the
B6C3F1 hybrid mouse. Int J Toxicol 25: 373-378.
http: //dx.doi.org/10.1080/10915810600846245.
Poulin. P: Krishnan. K. (1995). An algorithm for predicting tissue : blood partition coefficients of
organic chemicals from n-octanol: water partition coefficient data. J Toxicol Environ Health
46: 117-129. http://dx.doi.org/10.1080/15287399509532021.
Poulin. P: Theil. F. -P. (2000). A priori prediction of tissue : plasma partition coefficients of drugs to
facilitate the use of physiologically-based pharmacokinetic models in drug discovery. J
Pharm Sci 89: 16-35. http://dx.doi.org/10.1002/rSICni520-6017r200001189:l<:16::AID-
TPS3>:3.0.CQ:2-E.
Reddv. G: Allen. NA: Maior. MA. (2008). Absorption of 14C-cyclotrimethylenetrinitramine (RDX)
from soils through excised human skin. Toxicol Mech Meth 18: 575-579.
http://dx.doi.Org/10.1080/15376510701703466.
Reddv. G: Eisenhut. K: Morrison. SA: Kelly. TA. (1989). Toxicokinetics of 14C RDX
(cyclotrimethylenetrinitramine) in rats after intratracheal administration. Toxicologist 9:
83.
Reddv. G: Erexson. GL: Cifone. MA: Maior. MA: Leach. GT. (2005). Genotoxicity assessment of
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX). Int J Toxicol 24: 427-434.
http://dx.doi.Org/10.1080/10915810500366922.
R-5
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Reifenrath. WG: Kammen. HO: Reddv. G: Major. MA: Leach. GT. (2008). Interaction of hydration,
aging, and carbon content of soil on the evaporation and skin bioavailability of munition
contaminants. J Toxicol Environ Health A 71: 486-494.
http://dx.doi.Org/10.1080/15287390801906956.
Rispin. A: Farrar. D: Margosches. E: Gupta. K: Stitzel. K: Carr. G: Greene. M: Meyer. W: McCall. D.
(2002). Alternative methods for the median lethal dose (LD(50)) test: The up-and-down
procedure for acute oral toxicity. ILAR J 43: 233-243.
http://dx.doi.Org/10.1093/ilar.43.4.233.
SAB (Science Advisory Board). (2017). Letter to E. Scott Pruitt re: Review of EPAs draft assessment
entitled toxicological reviewofhexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) (September
2016). Available online at (accessed
Safe Work Australia. (2018). Hazardous Chemical Information System (HCIS). Exposure standard
documentation: Cyclonite [Database], Retrieved from http://hcis.safeworkaustralia.gov.au/
Schneider. NR: Bradley. SL: Andersen. ME. (1977). Toxicology of cyclotrimethylenetrinitramine:
Distribution and metabolism in the rat and the miniature swine. Toxicol Appl Pharmacol 39:
531-541. http://dx.doi.org/10.1016/0041-008Xr77190144-2.
Schneider. NR: Bradley. SL: Andersen. ME. (1978). The distribution and metabolism of
cyclotrimethylenetrinitramine (RDX) in the rat after subchronic administration. Toxicol
Appl Pharmacol 46: 163-171. http://dx.doi.org/10.1016/0041-008Xr78190147-3.
Simmon. VF: Spanggord. RT: Eckford. SL: McClurg. V. (1977). Mutagenicity of some munition
wastewater chemicals and chlorine test kit reagents (pp. 1-205). (ADA057680. DAMD17-
76-C-6013). Menlo Park, CA: SRI International.
http: / / www, dtic. mil /docs /citations /AD AO 57680.
Smith. IN: Espino. MA: Liu. I: Romero. NA: Cox. SB: Cobb. GP. (2009). Multigenerational effects in
deer mice (Peromyscus maniculatus) exposed to hexahydro-l,3,5-trinitroso-l,3,5-triazine
(TNX). Chemosphere 75: 910-914. http://dx.doi.Org/10.1016/i.chemosphere.2009.01.010.
Smith. IN: Pan. X: Gentles. A: Smith. EE: Cox. SB: Cobb. GP. (2006). Reproductive effects of
hexahydro-l,3,5-trinitroso-l,3,5-triazine in deer mice (Peromyscus maniculatus) during a
controlled exposure study. Environ Toxicol Chem 25: 446-451.
http: //dx.doi.org/10.1897 /05-277R.1.
Snodgrass. HL. Tr. (1984). Preliminary assessment of relative toxicity and mutagenicity potential of
l-nitroso-3,5-dinitro-l,3,5-triazacylcohexane (mononitroso-RDX) (Final Reported.).
(ADA149351. USAEHA-75-51-0345-85). Aberdeen Proving Ground, MD: U.S. Army
Environmental Hygiene Agency.
http: / /www, dtic. mil /docs /citations /AD A149 3 51.
Snoeck. V: Huvghebaert. N: Cox. E: Vermeire. A: Saunders. I: Remon. TP: Verschooten. F: Goddeeris.
BM. (2004). Gastrointestinal transit time of nondisintegrating radio-opaque pellets in
suckling and recently weaned piglets. J Control Release 94: 143-153.
http://dx.doi.Org/10.1016/i.iconrel.2003.09.015.
Spiteri. NT. (1982). Circadian patterning of feeding, drinking and activity during diurnal food access
in rats. Physiol Behav 28: 139-147. http: //dx.doi.org/10.1016/0031 -9384r82190115-9.
Stone. WT: Paletta. TL: Heiman. EM: Bruce. II: Knepshield. TH. (1969). Toxic effects following
ingestion of C-4 plastic explosive. Arch Intern Med 124: 726-730.
http://dx.doi.Org/10.1001/archinte.1969.00300220078015.
Suzuki. N: Withers. HR. (1978). Exponential decrease during aging and random lifetime of mouse
spermatogonial stem cells. Science 202: 1214-1215.
http://dx.doi.org/10.1126/science.725597.
Sweeney. LM: Gut. CP. Tr: Gargas. ML: Reddv. G: Williams. LR: Tohnson. MS. (2012a). Assessing the
non-cancer risk for RDX (hexahydro-l,3,5-trinitro-l,3,5-triazine) using physiologically
R-6
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
based pharmacokinetic (PBPK) modeling [Review], Regul Toxicol Pharmacol 62: 107-114.
http://dx.doi.0rg/lO.lOl6/i.yrtph.2Oll.i2.OO7.
Sweeney. LM: Okolica. MR: Gut. CP. Tr: Gargas. ML. (2012b). Cancer mode of action, weight of
evidence, and proposed cancer reference value for hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX). Regul Toxicol Pharmacol 64: 205-224.
http://dx.doi.0rg/lO.lOl6/i.yrtph.2Oi2.O7.OO5.
Tan. EL: Ho. CH: Griest. WH: Tvndall. RL. (1992). Mutagenicity of trinitrotoluene and its metabolites
formed during composting. J Toxicol Environ Health A 36: 165-175.
http://dx.doi.org/10.1080/15287399209531632.
Testud. F: Descotes. 1: Le Meur. B. (2006). Acute occupational poisoning by octogen: First case
report. Clin Toxicol 44: 189-190.
Testud. F: Glanclaude. 1. -M: Descotes. 1. (1996a). Acute hexogen poisoning after occupational
exposure. J Toxicol Clin Toxicol 34: 109-111.
http://dx.doi.org/10.3109/15563659609020244.
Testud. F: Glanclaude. 1: Imperatori. 1: Le Meur. B: Descotes. 1. (1996b). [Acute hexogen poisoning
after occupational exposure, report of 2 cases]. Arch Mai Prof Med Trav Secur Soc 57: 342-
346.
Timchalk. C: Nolan. RT: Mendrala. AL: Dittenber. DA: Brzak. KA: Mattsson. TL. (2002). A
physiologically based pharmacokinetic and pharmacodynamic (PBPK/PD) model for the
organophosphate insecticide chlorpyrifos in rats and humans. Toxicol Sci 66: 34-53.
http://dx.doi.0rg/lO.lO93/toxsci/66.l.34.
U.S. EPA (U.S. Environmental Protection Agency). (1989). Risk assessment guidance for superfund
[EPA Report], (EPA/540/1-89/002). Washington, DC.
U.S. EPA (U.S. Environmental Protection Agency). (1992). A cross-species scaling factor for
carcinogen risk assessment based on equivalence of mg/kg3/4/day [EPA Report] (pp.
24152-24161). Washington, DC.
U.S. EPA (U.S. Environmental Protection Agency). (2002). A review of the reference dose and
reference concentration processes (pp. 1-192). (EPA/630/P-02/002F). Washington, DC:
U.S. Environmental Protection Agency, Risk Assessment Forum.
http://www.epa.gov/osa/review-reference-dose-and-reference-concentration-processes.
U.S. EPA (U.S. Environmental Protection Agency). (2005). Guidelines for carcinogen risk assessment
[EPA Report] (pp. 1-166). (EPA/630/P-03/001F). Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum, http: //www2.epa.gov/osa/guidelines-
carcinogen-risk-assessment
U.S. EPA (U.S. Environmental Protection Agency). (2012). Benchmark dose technical guidance.
(EPA/100/R-12/001). Washington, DC: U.S. Environmental Protection Agency, Risk
Assessment Forum, https: //www.epa.gov/risk/benchmark-dose-technical-guidance.
U.S. EPA (U.S. Environmental Protection Agency). (2014). RDX PBPK files in acsIX format [EPA
Report],
U.S. EPA (U.S. Environmental Protection Agency). (2015). Science policy council peer review
handbook, 4th edition 2015 [EPA Report] (4th ed.). (EPA/100/B-15/001). Washington, DC:
U.S. Environmental Protection Agency, Science Policy Council.
https://www.epa.gov/osa/peer-review-handbook-4th-edition-2015.
von Oettingen. WF: Donahue. DP: Yagoda. H: Monaco. AR: Harris. MR. (1949). Toxicity and potential
dangers of cyclotrimethylenetrinitramine (RDX). J Ind Hyg Toxicol 31: 21-31.
West. RR: Stafford. DA. (1997). Occupational exposures and haematological abnormalities among
ordnance factory workers: A case control study. Leuk Res 21: 675-680.
http://dx.doi.org/10.1016/S0145-2126r97100042-8.
R-7
-------
Supplemental Information—Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Whong. W. -Z: Speciner. ND: Edwards. GS. (1980). Mutagenic activity of tetryl, a nitroaromatic
explosive, in three microbial test systems. Toxicol Lett 5: 11-17.
http://dx.doi.org/10.1016/0378-4274r80190142-3.
Woody. RC: Kearns. GL: Brewster. MA: Turlev. CP: Sharp. GB: Lake. RS. (1986). The neurotoxicity of
cyclotrimethylenetrinitramine (RDX) in a child: A clinical and pharmacokinetic evaluation.
Clin Toxicol 24: 305-319. http://dx.doi.org/10.3109/lSS636S8608992S9S.
Zhang. X: Ebata. KT: Robaire. B: Nagano. MC. (2006). Aging of male germ line stem cells in mice. Biol
Reprod 74: 119-124. http://dx.doi.org/10.1095/biolreprod.105.045591.
R-8
-------