aKKQ
rlk
EPA/635/R-18/211Fa
www.epa.gov/iris
Toxicological Review of Hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
(CASRN 121-82-4]
August 2018
Integrated Risk Information System
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
-------
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
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
CONTENTS
AUTHORS | CONTRIBUTORS | REVIEWERS ix
PREFACE xiv
PREAMBLE TO IRIS TOXICOLOGICAL REVIEWS xvii
EXECUTIVE SUMMARY xxv
LITERATURE SEARCH STRATEGY | STUDY SELECTION AND EVALUATION xxx
1. HAZARD IDENTIFICATION 1-1
1.1. OVERVIEW OF CHEMICAL PROPERTIES AND TOXICOKINETICS 1-1
1.1.1. Chemical Properties 1-1
1.1.2. Toxicokinetics 1-3
1.1.3. Description of Toxicokinetic Models 1-3
1.2. PRESENTATION AND SYNTHESIS OF EVIDENCE BY ORGAN/SYSTEM 1-4
1.2.1. Nervous System Effects 1-4
1.2.2. Urinary System (Kidney and Bladder) Effects 1-24
1.2.3. Prostate Effects 1-41
1.2.4. Developmental Effects 1-47
1.2.5. Liver Effects 1-54
1.2.6. Other Noncancer Effects 1-66
1.2.7. Carcinogenicity 1-69
1.3. INTEGRATION AND EVALUATION 1-80
1.3.1. Effects Other Than Cancer 1-80
1.3.2. Carcinogenicity 1-84
1.3.3. Susceptible Populations and Life Stages for Cancer and Noncancer Outcomes 1-85
2. DOSE-RESPONSE ANALYSIS 2-1
2.1. ORAL REFERENCE DOSE FOR EFFECTS OTHER THAN CANCER 2-1
2.1.1. Identification of Studies for Dose-Response Analysis of Selected Effects 2-1
2.1.2. Methods of Analysis 2-5
2.1.3. Derivation of Candidate Values 2-12
2.1.4. Derivation of Organ/System-Specific Reference Doses 2-19
2.1.5. Selection of the Overall Reference Dose 2-21
iii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
2.1.6. Comparison with Mortality Doses Expected to be Lethal to 1% of the Animals
(LDois) 2-22
2.1.7. Uncertainties in the Derivation of the Reference Dose 2-26
2.1.8. Confidence Statement 2-27
2.1.9. Previous Integrated Risk Information System (IRIS) Assessment 2-28
2.2. INHALATION REFERENCE CONCENTRATION FOR EFFECTS OTHER THAN CANCER 2-28
2.2.1. Previous Integrated Risk Information System (IRIS) Assessment 2-29
2.3. ORAL SLOPE FACTOR FOR CANCER 2-29
2.3.1. Analysis of Carcinogenicity Data 2-29
2.3.2. Dose-Response Analysis—Adjustments and Extrapolation Methods 2-30
2.3.3. Derivation of the Oral Slope Factor 2-32
2.3.4. Uncertainties in the Derivation of the Oral Slope Factor 2-34
2.3.5. Previous Integrated Risk Information System (IRIS) Assessment 2-37
2.4. INHALATION UNIT RISK FOR CANCER 2-37
2.5. APPLICATION OF AGE-DEPENDENT ADJUSTMENT FACTORS 2-38
REFERENCES R-l
iv
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
TABLES
Table ES-1. Organ/system-specific reference doses (RfDs) and overall RfD for hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX) xxvi
Table ES-2. Summary of reference dose (RfD) derivation xxvii
Table LS-1. Inclusion-exclusion criteria for health effect studies xxxiii
Table LS-2. Studies determined not to be informative because of significant issues with
design, conduct, or reporting xxxvii
Table LS-3. Considerations and relevant experimental information for evaluation of
experimental animal studies xxxix
Table LS-4. Summary of experimental animal database xli
Table LS-5. Experimental animal studies considered less informative because of certain
study design, conduct, or reporting limitations xlvi
Table 1-1. Chemical identity and physicochemical properties of hexahydro-l,3,5-trinitro-
1,3,5-triazine (RDX) from EPA's Chemistry Dashboard 1-2
Table 1-2. Evidence pertaining to nervous system effects in humans 1-5
Table 1-3. Evidence pertaining to nervous system effects in animals 1-6
Table 1-4. Evidence pertaining to kidney effects in humans 1-25
Table 1-5. Evidence pertaining to urinary system (kidney and bladder) effects in animals 1-26
Table 1-6. Six-, 12-, and 24-month incidence of kidney endpoints in male F344 rats
reported for statistical evaluation in Levine et al. (1983) 1-34
Table 1-7. Six-, 12-, and 24-month incidence of urinary bladder endpoints in male F344 rats
reported for statistical evaluation in Levine et al. (1983) 1-36
Table 1-8. Two-year prostate inflammation incidence in male F344 rats (Levine et al., 1983) 1-42
Table 1-9. Evidence pertaining to prostate effects in animals 1-43
Table 1-10. Evidence pertaining to developmental effects in animals 1-48
Table 1-11. Evidence pertaining to liver effects in humans 1-56
Table 1-12. Evidence pertaining to liver effects in animals 1-57
Table 1-13. Liver tumors observed in chronic animal bioassays 1-71
Table 1-14. Lung tumors observed in chronic animal bioassays 1-74
Table 2-1. Information considered for evaluation of studies that examined convulsions 2-3
Table 2-2. Summary of derivation of point of departures (PODs) following oral exposure to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) 2-7
Table 2-3. Effects and corresponding derivation of candidate values 2-17
Table 2-4. Organ/system-specific reference doses (RfDs) and overall RfD for hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX) 2-19
Table 2-5. Comparison of dose levels associated with mortality and convulsions in selected
studies 2-23
Table 2-6. Summary of dose-response evaluation for mortality following oral exposure to
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) 2-25
Table 2-7. Model predictions and oral slope factors (OSFs) for hepatocellular and
alveolar/bronchiolar adenomas or carcinomas in female B6C3Fi mice
administered hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) in the diet for 2 years
(Lish et al., 1984) 2-33
Table 2-8. Summary of uncertainty in the derivation of the cancer risk value for hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX) 2-35
v
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
FIGURES
Figure LS-1. Summary of literature search and screening process for hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) xxxii
Figure 1-1. Exposure response array of nervous system effects following oral exposure 1-12
Figure 1-2. Exposure-response array of urinary system (kidney and bladder) effects 1-37
Figure 1-3. Exposure-response array of prostate effects 1-44
Figure 1-4. Exposure response array of developmental effects following oral exposure 1-52
Figure 1-5. Exposure response array of liver effects following oral exposure 1-64
Figure 2-1. Conceptual approach to dose-response modeling for oral exposure 2-6
Figure 2-2. Candidate values with corresponding point of departure (POD) and composite
uncertainty factor (UF) 2-18
vi
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
ABBREVIATIONS
AAP Army ammunition plant
ACGIH American Conference of Governmental
Industrial Hygienists
AChE acetylcholinesterase
ADAF age-dependent adjustment factor
AIC Akaike's information criterion
ALP alkaline phosphatase
ALT alanine aminotransferase
AOP adverse outcome pathway
AST aspartate aminotransferase
atm atmosphere
ATSDR Agency for Toxic Substances and
Disease Registry
AUC area under the curve
AUCTotai area under the curve for blood
concentration versus time from the
time of dosing to the time RDX is
completely eliminated
BDNF brain-derived neurotrophic factor
BMD benchmark dose
BMDL benchmark dose lower confidence limit
BMDS Benchmark Dose Software
BMDU benchmark dose upper bound
BMR benchmark response
BUN blood urea nitrogen
BW body weight
BW" :i:i body weight scaling to the 0.33 power
BW2/3 body weight scaling to the 2/3 power
BW3/4 body weight scaling to the % power
BWa animal body weight
BWh human body weight
CAAC Chemical Assessment Advisory
Committee
CASRN Chemical Abstracts Service registry
number
CI confidence interval
Cmax peak concentration
CNS central nervous system
CSF cerebrospinal fluid
CYP450 cytochrome P450
DAF dosimetric adjustment factor
d.f. degrees of freedom
DMSO dimethylsulfoxide
DNA deoxyribonucleic acid
DNX l-nitro-3,5-dinitroso-
1,3,5-triazacyclohexane
DTIC Defense Technical Information Center
EEG electroencephalogram
EPA Environmental Protection Agency
ER extra risk
FOB functional observational battery
FUDS
GABA
GD
GI
HED
HERO
HI
HMX
i.p.
i.v.
IRIS
IUR
KAD
KAS
Kel
KfC
KQB
KQC
KQF
KQL
KQR
KQS
KVB
KVF
KVL
KVR
KVS
KW
LDoi
LDH
LDLoi
LOAEL
LOD
miRNA
MNX
MOA
MRL
NA
NCE
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
micro RNA
hexahydro-l-nitroso-3,5-dinitro-
1,3,5-triazine
mode of action
minimal risk level
not available
normochromatic erythrocyte
vii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
NCEA
National Center for Environmental
TS
terminal sacrifice
Assessment
TSCATS
Toxic Substances Control Act Test
ND
not determined
Submissions
NIOSH
National Institute for Occupational
TWA
time-weighted average
Safety and Health
U.S.
United States of America
NOAEL
no-observed-adverse-effect level
UCM
Unregulated Contaminant Monitoring
NPL
National Priorities List
UF
uncertainty factor
NTP
National Toxicology Program
UFa
animal-to-human uncertainty factor
NZW
New Zealand White
UFd
database deficiencies uncertainty factor
OR
odds ratio
UFh
human variation uncertainty factor
ORD
Office of Research and Development
UFl
LOAEL-to-NOAEL uncertain factor
OSF
oral slope factor
UFS
subchronic-to-chronic uncertainty
OSHA
Occupational Safety and Health
factor
Administration
WBC
white blood cell
PB
tissue:blood partition coefficient for
brain
WHO
World Health Organization
PBPK
physiologically based pharmacokinetic
PCE
polychromatic erythrocyte
PECO
populations, exposures, comparitors,
and outcomes
PEL
permissible exposure limit
PF
tissue:blood partition coefficient for fat
PL
tissue:blood partition coefficient for
liver
PND
postnatal day
POD
point of departure
PR
tissue:blood partition coefficient for
richly perfused tissue
PS
tissue:blood partition coefficient for
slowly perfused tissue
PWG
Pathology Working Group
RBC
red blood cell
RDX
Royal Demolition eXplosive
(hexahydro-l,3,5-trinitro-
1,3,5-triazine)
RfC
inhalation reference concentration
RfD
oral reference dose
RNA
ribonucleic acid
SAB
Science Advisory Board
SD
standard deviation
SDMS
spontaneous death or moribund
sacrifice
SDWA
Safe Drinking Water Act
SE
standard error
SGOT
glutamic oxaloacetic transaminase, also
known as AST
SGPT
glutamic pyruvic transaminase, also
known as ALT
SLE
systemic lupus erythematosus
SS
scheduled sacrifice
t%
half life
TLV
threshold limit value
TNT
trinitrotoluene
TNX
hexahydro-l,3,5-trinitroso-
1,3,5-triazine
viii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
AUTHORS | CONTRIBUTORS | REVIEWERS
Assessment Team
Louis D'Amico, Ph.D. (Assessment Manager) U.S. EPA/Office of Research and Development (ORD)/
Todd Blessinger, Ph.D. National Center for Environmental Assessment (NCEA)
Ravi Subramaniam, Ph.D.
Christopher Brinkerhoff, Ph.D. Former ORISE Postdoctoral Fellow at
U.S. EPA/ORD/NCEA
Currently at U.S. EPA/Office of Chemical Safety and
Pollution Prevention
Contributors
Rob DeWoskin, Ph.D. (retired)
Belinda Hawkins, Ph.D.
Karen Hogan, M.S.
Andrew Kraft, Ph.D.
Jordan Trecki, Ph.D. (formerly with U.S. EPA)
Scott Wesselkamper, Ph.D.
Tammy Stoker, Ph.D.
Charles Wood, DVM, PhD, DACVP
Anne Loccisano, Ph.D.
Kelly Garcia
Carolyn Gigot
U.S. EPA/ORD/NCEA
U.S. EPA/ORD/National Health and Environmental
Research Laboratory (NHEERL)
Former ORISE Postdoctoral Fellow at
U.S. EPA/ORD/NCEA
U.S. EPA Environmental Health Assessment Support
Associate (ORAU Student Services Contractor)
Production Team
Maureen Johnson
Vicki Soto
Dahnish Shams
U.S. EPA/ORD/NCEA
Contractor Support
Heather Carlson-Lynch, S.M., DABT
Julie Melia, Ph.D., DABT
Megan Riccardi, M.S.
Pam Ross, M.S.
Robin Blain, Ph.D.
SRC, Inc., North Syracuse, NY
ICF International, Fairfax, VA
IX
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Executive Direction
Tina Bahadori, Sc.D. (Center Director, National Program Director, HHRA) U.S. EPA/ORD/NCEA
Kenneth Olden, Ph.D., Sc.D., L.H.D. (Center Director—Retired)
Michael Slimak, Ph.D. (Former Center Director)
John Vandenberg, Ph.D. (Former National Program Director, HHRA)
Lynn Flowers, Ph.D., DABT (Former Associate Director for Health)
Kristina Thayer, Ph.D. (IRIS Division Director)
Vincent Cogliano, Ph.D. (Former IRIS Division Director)
James Avery, Ph.D. (IRIS Division Deputy Director)
Gina Perovich, M.S. (Former IRIS Division Deputy Director)
Samantha Jones, Ph.D. (IRIS Associate Director for Science)
Susan Rieth, MPH (Quantitative Modeling Branch Chief)
Internal Review Team
General Toxicity/Immunotoxicity/Cancer Workgroup U.S. EPA/ORD/NCEA
Epidemiology Workgroup
Neurotoxicity Workgroup
Pharmacokinetics Workgroup
Reproductive and Developmental Toxicity Workgroup
Scoping and Problem Formulation Workgroup
Statistics Workgroup
Toxicity Pathways Workgroup
Reviewers
This assessment was provided for review to scientists in EPA's program and regional offices.
Comments were submitted by:
Office of the Administrator/Office of Children's Health Protection
Office of Chemical Safety and Pollution Prevention/Office of Pesticide Programs
Office of Land and Emergency Management
Office of Land and Emergency Management/Federal Facilities Forum
Office of Water
Region 2, New York, NY
Region 8, Denver, CO
This assessment was provided for review to other federal agencies and the Executive Office of the
President (EOP). A summary and EPA's disposition of major comments from the other federal
agencies and the EOP is available on the Integrated Risk Information System (IRIS) website.
Comments were submitted by:
Department of Defense
Department of Energy
Department of Health and Human Services/Agency for Toxic Substances and Disease Registry
Department of Health and Human Services/National Institute of Environmental Health
Sciences/National Toxicology Program
Department of Health and Human Services/National Institute for Occupational Safety and Health
National Aeronautics and Space Administration
Executive Office of the President/Council on Environmental Quality
Executive Office of the President/Office of Management and Budget
x
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
This assessment was released for public comment on March 10, 2016 and comments were due on
May 9, 2016. The public comments are available on Regulations.gov. A summary and EPA's
disposition of the comments from the public is included in the external review draft assessment on
the IRIS website. Comments were received from the following entities:
Nancy Beck, American Chemistry Council (currently Deputy Assistant Administrator, Office of Chemical Safety
and Pollution Prevention, U.S. EPA)
Andy Nong, Health Canada
Johns Hopkins Bloomberg School of Public Health, Special Studies in Risk Assessment Class as submitted by
Mary Fox
Ron Melnick
U.S. Army Public Health Command and Uniformed Services University of the Health Sciences
Larry Williams
Anonymous member of the public
A public science meeting was held on May 10, 2016 to obtain public input on the IRIS Toxicological
Review of RDX (Public Comment Draft). Public commenters, stakeholders, and members of the
scientific community were joined by independent experts identified by the National Academies'
National Research Council (identified by * below) in a discussion of key science topics. Discussants
and public commenters were:
Desmond Bannon
Nancy Beck
Maria Braga*
Marion Ehrich*
Kendall Frazier*
Andy Nong*
Karen Regan*
Evelyn Tiffany-Castiglioni*
Larry Williams
U.S. Army Public Health Center
American Chemistry Council (currently Deputy Assistant Administrator,
Office of Chemical Safety and Pollution Prevention, U.S. EPA)
Uniformed Services University of the Health Sciences
Virginia-Maryland College of Veterinary Medicine (Virginia Tech)
GlaxoSmithKline
Health Canada
Research Pathology Associates
Texas A&M University
This assessment was peer reviewed by independent, expert scientists external to EPA convened by
EPA's Science Advisory Board (SAB), the SAB Chemical Assessment Advisory Committee
Augmented for Review of the Draft IRIS RDX Assessment. A peer-review meeting was held on
December 12-14, 2016. The reportofthe SAB's review ofthe EPA's draft Toxicological Review of
RDX, dated September 27, 2017, is available on the IRIS website. A summary and EPA's disposition
of the comments received from the SAB is included in Appendix E.
Dr. Kenneth Ramos (Chair)
Dr. Hugh A. Barton
Associate Vice-President of Precision Health Sciences and Professor of
Medicine, University of Arizona Health Sciences, Tucson, AZ
Associate Research Fellow, Pharmacokinetics, Dynamics, and Metabolism,
Pfizer Inc., Groton, CT
XI
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Dr. Maarten C. Bosland
Dr. Mary Boudreau
Dr. James V. Bruckner
Dr. George Cobb
Dr. David Eastmond
Dr. Joanne English
Dr. Alan Hoberman
Dr. Jacqueline Hughes-Oliver
Dr. Susan Laffan
Dr. Lawrence Lash
Dr. Stephen Lasley
Dr. Melanie Marty
Dr. Marvin Meistrich
Dr. Marilyn Morris
Dr. Victoria Persky
Dr. Isaac Pessah
Dr. Kenneth M. Portier
Dr. Samba Reddy
Dr. Stephen M. Roberts
Dr. Thomas Rosol
Dr. Alan Stern
Dr. Robert Turesky
Professor of Pathology, College of Medicine, University of Illinois at Chicago,
Chicago, IL
Research Toxicologist, Division of Biochemical Toxicology, National Center
for Toxicological Research, U.S. Food and Drug Administration, Jefferson, AR
Professor, Department of Pharmacology & Toxicology, College of Pharmacy,
University of Georgia, Athens, GA
Professor, Environmental Science, College of Arts and Sciences, Baylor
University, Waco, TX
Professor and Chair, Department of Cell Biology and Neuroscience,
Toxicology Graduate Program, University of California at Riverside, Riverside,
CA
Independent Consultant, Menlo Park, CA
Toxicologist, Research, Charles River Laboratories, Inc., Horsham, PA
Professor, Statistics Department, North Carolina State University, Raleigh, NC
Safety Assessment, GlaxoSmithKline, King of Prussia, PA
Professor, Department of Pharmacology, Wayne State University School of
Medicine, Wayne State University, Detroit, MI
Professor of Pharmacology and Assistant Head, Cancer Biology &
Pharmacology, College of Medicine, University of Illinois at Chicago, Peoria, IL
Adjunct Professor, Environmental Toxicology, University of California at
Davis, Davis, CA
Professor, Experimental Radiation Oncology, M.D. Anderson Cancer Center,
University of Texas, Houston, TX
Professor of Pharmaceutical Sciences, School of Pharmacy and
Pharmaceutical Sciences, University at Buffalo, State University of New York,
Buffalo, NY
Professor, Epidemiology & Biostatistics Program, School of Public Health,
University of Illinois at Chicago, Chicago, IL
Professor, Molecular Biosciences, School of Veterinary Medicine, University of
California at Davis, Davis, CA
Independent Consultant, Athens, GA
Professor, Neuroscience and Experimental Therapeutics, College of Medicine,
Texas A&M University, Bryan, TX
Professor, Center for Environmental and Human Toxicology, University of
Florida, Gainesville, FL
Professor, Veterinary Biosciences, College of Veterinary Medicine, Ohio State
University, Columbus, OH
Chief, Bureau for Risk Analysis, Division of Science, Research and
Environmental Health, New Jersey Department of Environmental Protection,
Trenton, NJ
Professor, Masonic Cancer Center and Department of Medicinal Chemistry,
College of Pharmacy, University of Minnesota, Minneapolis, MN
Xll
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
The post-external review draft of the assessment was provided for review to scientists in EPA's
program and regional offices, and to other federal agencies and the Executive Office of the President
(EOP). A summary and EPA's disposition of major comments from the other federal agencies and
the EOP is available on the IRIS website. Comments were submitted by:
EPA, Office of Land and Emergency Management
EPA, Region 2
Department of Defense
xiii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
PREFACE
This Toxicological Review critically reviews the publicly available studies on
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX, Royal Demolition eXplosive, or cyclonite) to identify
its adverse health effects and characterize exposure-response relationships. This assessment was
prepared under the auspices of the U.S. Environmental Protection Agency (EPA) Integrated Risk
Information System (IRIS) Program. It updates a previous IRIS assessment of RDX that included an
oral reference dose (RfD) for effects other than cancer (posted in 1988), a determination on the
carcinogenicity of RDX, and derivation of an oral slope factor (OSF) to quantify the cancer risk
associated with RDX exposure (posted in 1990). New information has become available, and this
assessment reviews information on all health effects by all exposure routes.
A public meeting was held in December 2013 to obtain input on preliminary materials for
RDX, including draft literature searches and associated search strategies, evidence tables, and
exposure-response arrays prior to the development of the IRIS assessment All public comments
provided on the preliminary materials were taken into consideration in developing the draft
assessment. A second public meeting was held in May 2016 to discuss key science topics on the
public comment draft assessment. These topics included (1) suppurative prostatitis as a marker for
hazard to the urogenital system following RDX exposure, (2) evaluation and use of RDX
physiologically based pharmacokinetic (PBPK) models, (3) neurotoxicity observed with RDX and
consideration of dose and duration of exposure and the potential relationship to mortality, and
(4) other science topics in the RDX assessment. Independent experts identified by the National
Academies' National Research Council joined members of the scientific community, stakeholders,
and the general public in the discussion of these science topics. The complete set of public
comments submitted in connection with the December 2013 and May 2016 public meetings is
available on the docket at https://www.regulations.gov (Docket ID No. EPA-HQ-ORD-2013-0430).
Organ/system-specific reference values are calculated based on effects in the nervous
system, urinary system (kidney and bladder), and prostate. These reference values may be useful
for cumulative risk assessments that consider the combined effect of multiple agents acting on the
same biological system.
This assessment was conducted in accordance with EPA guidance, which is summarized in
the Preamble to IRIS Toxicological Reviews and cited at appropriate places in this assessment. The
findings of this assessment and related documents produced during its development are available
on the IRIS website (https://www, ep a. gov /iris). Appendices containing information on
assessments by other health agencies, details of the literature search strategy, toxicokinetic
information, summaries of supplementary toxicity information, and dose-response modeling are
provided as Supplemental Information to this assessment (see Appendices A to D).
xiv
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
The IRIS Program released preliminary assessment materials for RDX in December 2013
and the draft assessment for public comment in March 2016, during the period of development and
implementation of systematic review methods by the IRIS Program. The approach to
implementation is to use procedures and tools available at the time, without holding assessments
until new methods become available. Accordingly, the IRIS Program conducted literature searches
and evaluated studies using tools and documentation standards then available. Updated problem
formulation materials and systematic review protocol development began with assessments
started in 2015, after this assessment was well into assessment development. Implementation of
systematic review is a process of continual improvement and this assessment represents a step in
the evolution of the IRIS Program.
Uses and Environmental Occurrence
RDX is a military munitions explosive with limited civilian commercial uses (Gadagbui etal..
20121. In the United States, RDX is produced at Army ammunition plants and is not manufactured
commercially. RDX production peaked in the 1960s, with 180 million pounds per year produced
from 1969 to 1971. Yearly total production dropped to 16 million pounds in 1984 (ATSDR. 2012).
According to the EPA ChemView Tool (https://chemview.epa.gov/chemview). the aggregate
national production volume in 2015 was between 1 million and 10 million pounds.
RDX can be released into environmental media (air, water, soil) as a result of waste
generated during manufacture, packing, or disposal of the pure product, or use and disposal of
RDX-containing munitions (ATSDR. 2012: Gadagbui et al.. 2012: ATSDR. 1999.1993.1992). RDX is
mobile in soil, and leaching into groundwater has been reported in samples from military facilities
(Best etal.. 1999a: Godeiohann etal.. 1998: Bart etal.. 1997: Steuckartetal.. 1994: Spanggord etal..
1980). RDX transport in soil is generally through dissolution by precipitation and subsequent
downward movement, including migration to groundwater aquifers, and not much via surface
runoff (U.S. EPA. 2012b). Discussion of RDX properties and fate and transport is available in U.S.
EPA (2012b) and on the EPA's Chemistry Dashboard at https://comptox.epa.gov/dashboard/. RDX
has been detected in plants irrigated or grown with RDX-contaminated water (Best etal.. 1999b:
Simini and Checkai. 1996: Harvey et al.. 1991) and has also been detected in indoor air samples
from military facilities where RDX is produced (Bishop etal.. 1988).
Exposures to RDX among the general population are likely to be confined to individuals in
or around active or formerly used military facilities where RDX is or was produced, stored, or used.
Oral, inhalation, and dermal routes of exposure may be relevant
As of 2018, RDX was detected in surface water, groundwater, sediment, or soil at 32 active
EPA National Priorities List (NPL) sites. The NPL serves as a list of sites with known or threatened
releases of hazardous substances, pollutants, or contaminants throughout the United States and its
territories. The NPL aids the Agency in identifying the most serious sites that may warrant cleanup.
The majority of the NPL sites where RDX was listed are associated with military facilities. Based on
Department of Defense records, Gadagbui etal. (2012) reported that RDX contamination is present
xv
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
on 76 active military sites, 9 closed sites, and 15 sites under the Formerly Used Defense Sites
(FUDS) program. Not all sites under the FUDS program have been sampled, and additional sites
with RDX contamination in this program could be identified.
As of 2018, RDX was not regulated under the Safe Drinking Water Act (SDWA), although it
was included as a contaminant to be monitored under the Unregulated Contaminant Monitoring
(UCM) Rule by EPA's Office of Water from 2007 to 2011. Contaminants included in the UCM
program are suspected of being present in drinking water but do not have existing health-based
standards set under the SDWA. RDX has also been included on the Office of Water's Drinking Water
Contaminant Candidate List since the initial listing was published in 1998. The presence of a
chemical on the list suggests that it is known or anticipated to occur in public water systems.
Assessments by Other National and International Health Agencies
RDX has been evaluated by the Agency for Toxic Substances and Disease Registry, National
Institute for Occupational Safety and Health, Occupational Safety and Health Administration, and
Australian National Industrial Chemicals Notification and Assessment Scheme. The results of these
assessments (as of 2018) are presented in Appendix A of the Supplemental Information. It is
important to recognize that the assessments performed by other health agencies may have been
prepared for different purposes and may use different methods. In addition, newer studies may be
included in the IRIS assessment.
For additional information about this assessment or for general questions regarding IRIS,
please contact EPA's IRIS Hotline at 202-566-1676 (phone) or hotline.iris@epa.gov.
xvi
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Note: The Preamble summarizes the objectives and scope of the IRIS Program, general principles and
systematic review procedures used in developing IRIS assessments¦, and the overall development
process and document structure.
PREAMBLE TO IRIS TOXICOLOGICAL REVIEWS
1. Objectives and Scope of the IRIS
Program
Soon after EPA was established in 1970, it
was at the forefront of developing risk
assessment as a science and applying it in
support of actions to protect human health
and the environment EPA's IRIS Program1
contributes to this endeavor by reviewing
epidemiologic and experimental studies of
chemicals in the environment to identify
adverse health effects and characterize
exposure-response relationships. Health
agencies worldwide use IRIS assessments,
which are also a scientific resource for
researchers and the public.
IRIS assessments cover the hazard
identification and dose-response steps of risk
assessment. Exposure assessment and risk
characterization are outside the scope of IRIS
assessments, as are political, economic, and
technical aspects of risk management. An IRIS
assessment may cover one chemical, a group
of structurally or toxicologically related
chemicals, or a chemical mixture. Exceptions
outside the scope of the IRIS Program are
radionuclides, chemicals used only as
pesticides, and the "criteria air pollutants"
(particulate matter, ground-level ozone,
carbon monoxide, sulfur oxides, nitrogen
oxides, and lead).
Enhancements to the IRIS Program are
improving its science, transparency, and
productivity. To improve the science, the IRIS
Program is adapting and implementing
principles of systematic review (i.e., using
explicit methods to identify, evaluate, and
synthesize study findings). To increase
transparency, the IRIS Program discusses key
science issues with the scientific community
and the public as it begins an assessment
External peer review, independently managed
and in public, improves both science and
transparency. Increased productivity requires
that assessments be concise, focused on EPA's
needs, and completed without undue delay.
IRIS assessments follow EPA guidance2
and standardized practices of systematic
review. This Preamble summarizes and does
not change IRIS operating procedures or EPA
guidance.
Periodically, the IRIS Program asks for
nomination of agents for future assessment or
reassessment. Selection depends on EPA's
priorities, relevance to public health, and
availability of pertinent studies. The IRIS
multiyear agenda3 lists upcoming
assessments. The IRIS Program may also
assess other agents in anticipation of public
health needs.
2. Planning an Assessment: Scoping,
Problem Formulation, and
Protocols
Early attention to planning ensures that
IRIS assessments meet their objectives and
properly frame science issues.
Scoping refers to the first step of planning,
where the IRIS Program consults with EPA's
MRIS Program website: http://www.epa.gov/iris/.
2EPA guidance documents: http://www.epa.gov/iris/hasic-inforination-ahout-integrated-risk-inforination-
system#guidance/.
3IRIS multiyear agenda: https: //www.epa.gov/iris/iris-agenda.
xvii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
program and regional offices to ascertain their
needs. Scoping specifies the agents an
assessment will address, routes and durations
of exposure, susceptible populations and life
stages, and other topics of interest
Problem formulation refers to the science
issues an assessment will address and
includes input from the scientific community
and the public. A preliminary literature
survey, beginning with secondary sources
(e.g., assessments by national and
international health agencies and
comprehensive review articles), identifies
potential health outcomes and science issues.
It also identifies related chemicals (e.g.,
toxicologically active metabolites and
compounds that metabolize to the chemical of
interest).
Each IRIS assessment comprises multiple
systematic reviews for multiple health
outcomes. It also evaluates hypothesized
mechanistic pathways and characterizes
exposure-response relationships. An
assessment may focus on important health
outcomes and analyses rather than expand
beyond what is necessary to meet its
objectives.
Protocols refer to the systematic review
procedures planned for use in an assessment.
These protocols include strategies for
literature searches, criteria for study inclusion
or exclusion, considerations for evaluating
study methods and quality, and approaches
for extracting data. Protocols may evolve as an
assessment progresses and new agent-specific
insights and issues emerge.
3. Identifying and Selecting
Pertinent Studies
IRIS assessments conduct systematic
literature searches with criteria for inclusion
and exclusion. The objective is to retrieve the
pertinent primary studies (i.e., studies with
original data on health outcomes or their
mechanisms). PECO statements (Populations,
Exposures, Comparitors, and Outcomes)
govern the literature searches and screening
criteria. "Populations" and animal species
generally have no restrictions. "Exposures"
refers to the agent and related chemicals
identified during scoping and problem
formulation and may consider route, duration,
or timing of exposure. "Comparitors" means
studies that allow comparison of effects across
different levels of exposure. "Outcomes" may
become more specific (e.g., from "toxicity" to
"developmental toxicity" to "hypospadias") as
an assessment progresses.
For studies of absorption, distribution,
metabolism, and excretion, the first objective
is to create an inventory of pertinent studies.
Subsequent sorting and analysis facilitates
characterization and quantification of these
processes.
Studies on mechanistic events can be
numerous and diverse. Here too, the objective
is to create an inventory of studies for later
sorting to support analyses of related data.
The inventory also facilitates generation and
evaluation of hypothesized mechanistic
pathways.
The IRIS Program posts initial protocols
for literature searches on its website and adds
search results to EPA's HERO database.4 The
IRIS Program then takes extra steps to ensure
identification of pertinent studies by
encouraging the scientific community and the
public to identify additional studies and
ongoing research; searching for data
submitted under the Toxic Substances Control
Act or the Federal Insecticide, Fungicide, and
Rodenticide Act; and considering
late-breaking studies that would impact the
credibility of the conclusions, even during the
review process.5
4. Evaluating Study Methods and
Quality
IRIS assessments evaluate study methods
and quality, using uniform approaches for
4Health and Environmental Research Online: https: //hero.epa.gov/hero/.
5IRIS "stopping rules": littps://www.epa.gov/sites/production/files/2014-06/documents/
iris stoppingrules.pdf.
xviii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
each group of similar studies. The objective is
that subsequent syntheses can weigh study
results on their merits. Key concerns are
potential bias (factors that affect the
magnitude or direction of an effect) and
insensitivity (factors that limit the ability of a
study to detect a true effect).
For human and animal studies, the
evaluation of study methods and quality
considers study design, exposure measures,
outcome measures, data analysis, selective
reporting, and study sensitivity. For human
studies, this evaluation also considers
selection of participant and referent groups
and potential confounding. Emphasis is on
discerning bias that could substantively
change an effect estimate, considering also the
expected direction of the bias. Low sensitivity
is a bias towards the null.
Study-evaluation considerations are
specific to each study design, health effect, and
agent Subject-matter experts evaluate each
group of studies to identify characteristics that
bear on the informativeness of the results. For
carcinogenicity, neurotoxicity, reproductive
toxicity, and developmental toxicity, there is
EPA guidance for study evaluation (U.S. EPA.
2005a. 1998. 1996. 1991). As subject-matter
experts examine a group of studies, additional
agent-specific knowledge or methodologic
concerns may emerge and a second pass
become necessary.
Assessments use evidence tables to
summarize the design and results of pertinent
studies. If tables become too numerous or
unwieldy, they may focus on effects that are
more important or studies that are more
informative.
The IRIS Program posts initial protocols
for study evaluation on its website, then
considers public input as it completes this
step.
5. Integrating the Evidence of
Causation for Each Health
Outcome
Synthesis within lines of evidence. For
each health outcome, IRIS assessments
synthesize the human evidence and the animal
evidence, augmenting each with informative
subsets of mechanistic data. Each synthesis
considers aspects of an association that may
suggest causation: consistency,
exposure-response relationship, strength of
association, temporal relationship, biological
plausibility, coherence, and "natural
experiments" in humans (U.S. EPA. 2005a.
§2.5: 1994. §2.1.31.
Each synthesis seeks to reconcile
ostensible inconsistencies between studies,
taking into account differences in study
methods and quality. This leads to a
distinction between conflicting evidence
(unexplained positive and negative results in
similarly exposed human populations or in
similar animal models) and differing results
[mixed results attributable to differences
between human populations, animal models,
or exposure conditions; (U.S. EPA. 2005a.
§2,5)].
Each synthesis of human evidence
explores alternative explanations (e.g., chance,
bias, or confounding) and determines whether
they may satisfactorily explain the results.
Each synthesis of animal evidence explores
the potential for analogous results in humans.
Coherent results across multiple species
increase confidence that the animal results are
relevant to humans.
Mechanistic data are useful to augment the
human or animal evidence with information
on precursor events, to evaluate the human
relevance of animal results, or to identify
susceptible populations and life stages. An
agent may operate through multiple
mechanistic pathways, even if one hypothesis
dominates the literature (U.S. EPA. 2005a.
§2.4.3.31.
Integration across lines of evidence. For
each health outcome, IRIS assessments
integrate the human, animal, and mechanistic
evidence to answer the question: What is the
nature of the association between exposure to
the agent and the health outcome?
For cancer, EPA includes a standardized
hazard descriptor in characterizing the
strength of the evidence of causation. The
objective is to promote clarity and consistency
xix
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
of conclusions across assessments fU.S. EPA.
2005a. §2.51.
Carcinogenic to humans: Convincing
epidemiologic evidence of a causal
association; or strong human evidence of
cancer or its key precursors, extensive
animal evidence, identification of mode of
action and its key precursors in animals,
and strong evidence that they are
anticipated in humans.
Likely to be carcinogenic to humans: Evidence
that demonstrates a potential hazard to
humans. Examples include a plausible
association in humans with supporting
experimental evidence, multiple positive
results in animals, a rare animal response,
or a positive study strengthened by other
lines of evidence.
Suggestive evidence of carcinogenic potential:
Evidence that raises a concern for humans.
Examples include a positive result in the
only study, or a single positive result in an
extensive database.
Inadequate information to assess carcinogenic
potential: No other descriptors apply.
Examples include little or no pertinent
information, conflicting evidence, or
negative results not sufficiently robust for
not likely.
Not likely to be carcinogenic to humans: Robust
evidence to conclude that there is no basis
for concern. Examples include no effects in
well-conducted studies in both sexes of
multiple animal species, extensive
evidence showing that effects in animals
arise through modes of action that do not
operate in humans, or convincing evidence
that effects are not likely by a particular
exposure route or below a defined dose.
If there is credible evidence of
carcinogenicity, mutagenicity is evaluated
because this influences the approach to
dose-response assessment and subsequent
application of adjustment factors for
exposures early in life fU.S. EPA. 2005a. §3.3.1.
§3.5: 2005b. §51.
6. Selecting Studies for Derivation of
Toxicity Values
The purpose of toxicity values (slope
factors, unit risks, reference doses, reference
concentrations; see Section 7 of the Preamble)
is to estimate exposure levels likely to be
without appreciable risk of adverse health
effects. EPA uses these values to support its
actions to protect human health.
The health outcomes considered for
deriving toxicity values may depend on the
hazard descriptors. For example, IRIS
assessments generally derive cancer values
for agents that are carcinogenic or likely to be
carcinogenic, and sometimes for agents with
suggestive evidence (U.S. EPA. 2005a. §31.
Derivation of toxicity values begins with a
new evaluation of studies, as some studies
used qualitatively for hazard identification
may not be useful quantitatively for
exposure-response assessment Quantitative
analyses require quantitative measures of
exposure and response. An assessment weighs
the merits of the human and animal studies, of
various animal models, and of different routes
and durations of exposure (U.S. EPA. 1994.
§2.11. Study selection is not reducible to a
formula, and each assessment explains its
approach.
Other biological determinants of study
quality include appropriate measures of
exposure and response, investigation of early
effects that precede overt toxicity, and
appropriate reporting of related effects (e.g.,
combining effects that comprise a syndrome,
or benign and malignant tumors in a specific
tissue).
Statistical determinants of study quality
include multiple levels of exposure (to
characterize the shape of the
exposure-response curve) and adequate
exposure range and sample sizes [to minimize
extrapolation and maximize precision; (U.S.
EPA. 2012a. §2.11],
Studies of low sensitivity may be less
useful if they fail to detect a true effect or yield
toxicity values with wide confidence limits.
xx
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
7. Deriving Toxicity Values
General approach. EPA guidance
describes a two-step approach to
dose-response assessment: analysis in the
range of observation, then extrapolation to
lower levels. Each toxicity value pertains to a
route (e.g., oral, inhalation, dermal) and
duration or timing of exposure [e.g., chronic,
subchronic, gestational; (U.S. EPA. 2002. §41].
IRIS assessments derive a candidate value
from each suitable data set Consideration of
candidate values yields a toxicity value for
each organ or system. Consideration of the
organ/system-specific values results in the
selection of an overall toxicity value to cover
all health outcomes. The
organ/system-specific values are useful for
subsequent cumulative risk assessments that
consider the combined effect of multiple
agents acting at a common anatomical site.
Analysis in the range of observation.
Within the observed range, the preferred
approach is modeling to incorporate a wide
range of data. Toxicokinetic modeling has
become increasingly common for its ability to
support target-dose estimation, cross-species
adjustment, or exposure-route conversion. If
data are too limited to support toxicokinetic
modeling, there are standardized approaches
to estimate daily exposures and scale them
from animals to humans fU.S. EPA. 2011.
2006: 2005a. §3.1: 1994. §31.
For human studies, an assessment may
develop exposure-response models that
reflect the structure of the available data (U.S.
EPA. 2005a. §3.2.1). For animal studies, EPA
has developed a set of empirical
("curve-fitting") models6 that can fit typical
data sets (U.S. EPA. 2005a. §3.2.2). Such
modeling yields a point of departure, defined
as a dose near the lower end of the observed
range, without significant extrapolation to
lower levels [e.g., the estimated dose
associated with an extra risk of 10% for animal
data or 1% for human data, or their 95% lower
confidence limits; (U.S. EPA. 2012a. §2.2.1:
2005a. §3.2.41],
When justified by the scope of the
assessment, toxicodynamic ("biologically
based") modeling is possible if data are
sufficient to ascertain the key events of a mode
of action and to estimate their parameters.
Analysis of model uncertainty can determine
the range of lower doses where data support
further use of the model (U.S. EPA. 2005a.
§3.2.2.. §3.3.21.
For a group of agents that act at a common
site or through common mechanisms, an
assessment may derive relative potency
factors based on relative toxicity, rates of
absorption or metabolism, quantitative
structure-activity relationships, or
receptor-binding characteristics (U.S. EPA.
2005a. §3.2.61.
Extrapolation: slope factors and unit
risks. An oral slope factor or an inhalation unit
risk facilitates subsequent estimation of
human cancer risks. Extrapolation proceeds
linearly (i.e., risk proportional to dose) from
the point of departure to the levels of interest
This is appropriate for agents with direct
mutagenic activity. It is also the default if there
is no established mode of action (U.S. EPA.
2005a. §3.3.1. §3.3.31.
Differences in susceptibility may warrant
derivation of multiple slope factors or unit
risks. For early-life exposure to carcinogens
with a mutagenic mode of action, EPA has
developed default age-dependent adjustment
factors for agents without chemical-specific
susceptibility data (U.S. EPA. 2005a. §3.5:
2005b. §51.
If data are sufficient to ascertain the mode
of action and to conclude that it is not linear at
low levels, extrapolation may use the
reference-value approach (U.S. EPA. 2005a.
§3.3.41.
Extrapolation: reference values. An oral
reference dose or an inhalation reference
concentration is an estimate of human
exposure (including in susceptible
populations) likely to be without appreciable
risk of adverse health effects over a lifetime
(U.S. EPA. 2002. §4.2). Reference values
generally cover effects other than cancer. They
benchmark Dose Software: http: //www.epa.gov/hinds/.
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
are also appropriate for carcinogens with a
nonlinear mode of action.
Calculation of reference values involves
dividing the point of departure by a set of
uncertainty factors (each typically 1, 3, or 10,
unless there are adequate chemical-specific
data) to account for different sources of
uncertainty and variability (U.S. EPA. 2014:
2002. §4.4.51.
Human variation: An uncertainty factor covers
susceptible populations and life stages
that may respond at lower levels, unless
the data originate from a susceptible study
population.
Animal-to-human extrapolation: For reference
values based on animal results, an
uncertainty factor reflects cross-species
differences, which may cause humans to
respond at lower levels.
Subchronic-to-chronic exposure: For chronic
reference values based on subchronic
studies, an uncertainty factor reflects the
likelihood that a lower level over a longer
duration may induce a similar response.
This factor may not be necessary for
reference values of shorter duration.
Adverse-effect level to no-observed-adverse-
effect level. For reference values based on
a lowest-observed-adverse-effect level, an
uncertainty factor reflects a level judged to
have no observable adverse effects.
Database deficiencies: If there is concern that
future studies may identify a more
sensitive effect, target organ, population,
or life stage, a database uncertainty factor
reflects the nature of the database
deficiency.
8. Process for Developing and Peer
Reviewing IRIS Assessments
The IRIS process (revised in 2009 and
enhanced in 2013) involves extensive public
engagement and multiple levels of scientific
review and comment IRIS Program scientists
consider all comments. Materials released,
comments received from outside EPA, and
disposition of major comments (steps 3, 4, and
6 below) become part of the public record.
Step 1: Draft development. As outlined in
Section 2 of this Preamble, IRIS Program
scientists specify the scope of an
assessment and formulate science issues
for discussion with the scientific
community and the public. Next, they
release initial protocols for the systematic
review procedures planned for use in the
assessment IRIS Program scientists then
develop a first draft, using structured
approaches to identify pertinent studies,
evaluate study methods and quality,
integrate the evidence of causation for
each health outcome, select studies for
derivation of toxicity values, and derive
toxicity values, as outlined in Preamble
Sections 3-7.
Step 2: Agency review. Health scientists
across EPA review the draft assessment.
Step 3: Interagency science consultation.
Other federal agencies and the Executive
Office of the President review the draft
assessment
Step 4: Public comment, followed by
external peer review. The public reviews
the draft assessment IRIS Program
scientists release a revised draft for
independent external peer review. The
peer reviewers consider whether the draft
assessment assembled and evaluated the
evidence according to EPA guidance and
whether the evidence justifies the
conclusions.
Step 5: Revise assessment. IRIS Program
scientists revise the assessment to address
the comments from the peer review.
Step 6: Final agency review and
interagency science discussion. The IRIS
Program discusses the revised assessment
with EPA's program and regional offices
and with other federal agencies and the
Executive Office of the President
xxii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Step 7: Post final assessment. The IRIS
Program posts the completed assessment
and a summary on its website.
9. General Structure of IRIS
Assessments
Main text. IRIS assessments generally
comprise two major sections: (1) Hazard
Identification and (2) Dose-Response
Assessment Section 1.1 briefly reviews
chemical properties and toxicokinetics to
describe the disposition of the agent in the
body. This section identifies related chemicals
and summarizes their health outcomes, citing
authoritative reviews. If an assessment covers
a chemical mixture, this section discusses
environmental processes that alter the
mixtures humans encounter and compares
them to mixtures studied experimentally.
Section 1.2 includes a subsection for each
major health outcome. Each subsection
discusses the respective literature searches
and study considerations, as outlined in
Preamble Sections 3 and 4, unless covered in
the front matter. Each subsection concludes
with evidence synthesis and integration, as
outlined in Preamble Section 5.
Section 1.3 links health hazard
information to dose-response analyses for
each health outcome. One subsection
identifies susceptible populations and life
stages, as observed in human or animal
studies or inferred from mechanistic data.
These may warrant further analysis to
quantify differences in susceptibility. Another
subsection identifies biological considerations
for selecting health outcomes, studies, or data
sets, as outlined in Preamble Section 6.
Section 2 includes a subsection for each
toxicity value. Each subsection discusses study
selection, methods of analysis, and derivation
of a toxicity value, as outlined in Preamble
Sections 6 and 7.
Front matter. The Executive Summary
provides information historically included in
IRIS summaries on the IRIS Program website.
Its structure reflects the needs and
expectations of EPA's program and regional
offices.
A section on systematic review methods
summarizes key elements of the protocols,
including methods to identify and evaluate
pertinent studies. The final protocols appear
as an appendix.
The Preface specifies the scope of an
assessment and its relation to prior
assessments. It discusses issues that arose
during assessment development and
emerging areas of concern.
This Preamble summarizes general
procedures for assessments begun after the
date at the end of this Preamble. The Preface
identifies assessment-specific approaches that
differ from these general procedures.
10. References
U.S. EPA. (1991). Guidelines for
developmental toxicity risk assessment
(pp. 1-83). (EPA/600/FR-91/001).
Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment
Forum.
http://cfj3ub.epa.gov/ncea/cfm/recordis
plav.cfm?deid=23162
U.S. EPA. (1994). Methods for derivation of
inhalation reference concentrations and
application of inhalation dosimetry [EPA
Report] (pp. 1-409). (EPA/600/8-
90/066F). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office
of Research and Development, Office of
Health and Environmental Assessment,
Environmental Criteria and Assessment
Office.
https://cfj3ub.epa.gov/ncea/risk/recordis
plav.cfm?deid=71993&CFTD=51174B29&
CFTQKEN=25006317
U.S. EPA. (1996). Guidelines for reproductive
toxicity risk assessment (pp. 1-143).
(EPA/630/R-96/009). Washington, DC:
U.S. Environmental Protection Agency,
Risk Assessment Forum.
U.S. EPA. (1998). Guidelines for neurotoxicity
risk assessment. Fed Reg 63:26926-
26954.
xxiii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
U.S. EPA. (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. (2005a). 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. (2005b). Supplemental guidance for
assessing susceptibility from early-life
exposure to carcinogens (pp. 1-125).
(EPA/630/R-03/003F). Washington, DC:
U.S. Environmental Protection Agency,
Risk Assessment Forum.
U.S. EPA. (2006). Approaches for the
application of physiologically based
pharmacokinetic (PBPK) models and
supporting data in risk assessment (final
report) [EPA report] (pp. 1-123).
(EPA/600/R-05/043F). Washington, DC:
U.S. Environmental Protection Agency,
Office of Research and Development,
National Center for Environmental
Assessment
http://cfpub.epa.gov/ncea/cfm/recordis
plav.cfm?deid=l 57668
U.S. EPA. (2011). Recommended use of body
weight 3/4 as the default method in
derivation of the oral reference dose (pp.
1-50). (EPA/100/R11/0001).
Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment
Forum, Office of the Science Advisor.
https://www.epa.gov/risk/recommende
d-use-bodv-weight-34-default-method-
derivation-oral-reference-dose
U.S. EPA. (2012). Benchmark dose technical
guidance (pp. 1-99). (EPA/100/R-
12/001). Washington, DC: U.S.
Environmental Protection Agency, Risk
Assessment Forum.
U.S. EPA. (2014). Guidance for applying
quantitative data to develop data-derived
extrapolation factors for interspecies and
intraspecies extrapolation. (EPA/100/R-
14/002F). Washington, DC: Risk
Assessment Forum, Office of the Science
Advisor.
https://www.epa.gOv/sites/production/f
iles/2015-01/documents/ddef-final.pdf
August 2016
xxiv
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
EXECUTIVE SUMMARY
Summary of Occurrence and Health Effects
RDX is a synthetic chemical used primarily as a military explosive. RDX releases have
been reported in air, water, and soil, and exposure is likely limited to individuals in
or around military facilities where RDX is or was produced, used, or stored. Oral
exposure may occur from drinking contaminated groundwater or ingesting crops
irrigated with contaminated water. Inhalation or dermal exposures are more likely
in occupational settings.
Epidemiological studies provide only limited information on worker populations
exposed to RDX; several case reports describe effects primarily in the nervous system
following acute exposure to RDX. Animal studies of ingested RDX demonstrate
toxicity, including effects on the nervous system, urinary system (kidney and
bladder), and prostate.
Results from animal studies provide suggestive evidence of carcinogenic potential for
RDX based on evidence of positive trends in liver and lung tumor incidence in
experimental animals. There are no data on the carcinogenicity of RDX in humans.
ES.1. EVIDENCE FOR HAZARDS OTHER THAN CANCER: ORAL EXPOSURE
Nervous system effects are a human hazard of RDX exposure. Several human case reports
and animal studies provide consistent evidence of an association between RDX exposure and effects
on the nervous system, including findings related to the induction of seizures, abnormal electrical
activity, convulsions, tremors, and a reduced threshold for seizure induction by other stimuli;
behavioral effects that may be related to seizures such as hyperirritability, hyper-reactivity, and
other behavioral changes. Mechanistic data support the hypothesis that RDX-induced seizures and
related behavioral effects likely result from inhibition of gamma-aminobutyric acid (GABA)ergic
signaling in the limbic system. Some investigators reported that unscheduled deaths in
experimental animals exposed to RDX were frequently preceded by convulsions or seizures.
Urinary system effects are a potential human hazard of RDX exposure based largely on
observations of histopathological changes in the kidney and urinary bladder of male rats exposed to
RDX at doses higher than those associated with nervous system effects. The available evidence
indicates that male rats are more sensitive than females, and rats are more sensitive than mice to
RDX-related urinary system toxicity. There is suggestive evidence of male prostate effects
associated with RDX exposure based on an increased incidence of suppurative prostatitis in male
rats exposed to RDX in the diet for 2 years, in one of the few studies that evaluated the prostate.
There is no known mode of action (MOA) for effects of RDX exposure on the urinary system or
prostate, although there are studies indicating GABA helps regulate urinary system and prostate
function. Evidence for effects on other organs/systems, or developmental effects, was more limited
than for the endpoints summarized above.
xxv
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
ES.1,1, Oral Reference Dose (RfD) for Effects Other Than Cancer
Organ-specific RfDs were derived for hazards associated with RDX exposure (see
Table ES-1). These organ- or system-specific reference values may be useful for subsequent
cumulative risk assessments that consider the combined effect of multiple agents acting at a
common site.
Table ES-1. Organ/system-specific reference doses (RfDs) and overall RfD for
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
Effect
Basis
RfD (mg/kg-day)
Study exposure
description
Confidence
Nervous system
Convulsions
4 x 10"3
Subchronic
Medium
Urinary system
Kidney medullary papillary necrosis
1 x 10"2
Chronic
Medium
Prostate
Suppurative prostatitis
8 x 10"4
Chronic
Low
Overall RfD
Nervous system effects
4 x 10"3
Subchronic
Medium
The overall RfD (see Table ES-2) is derived to be protective of all types of hazards
associated with RDX exposure. Although the RfD for prostate effects results in a smaller value, it
was not selected as the overall RfD due to uncertainties in the evaluation of this endpoint ("low
confidence"). The effect of RDX on the nervous system was chosen as the basis for the overall RfD
because nervous system effects were observed most consistently across studies, species, and
exposure durations, and because they represent a sensitive human hazard of RDX exposure.
Evidence for effects of RDX on the urinary system and prostate is more limited relative to the
effects of RDX on the nervous system. Incidence of seizures or convulsions as reported in a
subchronic gavage study (Crouse et al.. 20061 was selected for deriving the overall RfD because this
endpoint was measured in a study that was well conducted, used a test material of high purity
(99.99%), and had five closely spaced dose groups that supported characterization of the
dose-response curve. In contrast, most other studies used a technical grade with ~10% or more
impurities. Benchmark dose (BMD) modeling was used to derive the point of departure (POD) for
RfD derivation (expressed as the lower confidence limit on the benchmark dose [BMDL05]). A 5%
response level was chosen because of the severity of the endpoint.
xxvi
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table ES-2. Summary of reference dose (RfD) derivation
Critical effect
Point of departure3
UF
Chronic RfD
Confidence
Nervous system effects (convulsions)
90-d F344 rat study
Crouse et al. (2006)
BMDLqb-hed: 1.3 mg/kg-d
300
4 x 10"3 mg/kg-d
Medium
AUC = area under the curve; BMDL = benchmark dose lower confidence limit.
aA benchmark response (BMR) of 5% was used to derive the BMD and BMDL. The resulting POD was converted to a
BMDL05-hed using a PBPK model based on modeled arterial blood concentration. The concentration was derived
from the AUC of modeled RDX concentration in arterial blood, which reflects the average blood RDX concentration
for the exposure duration normalized to 24 hr.
A PBPK model was used to extrapolate the BMDL05 derived from a rat study to a human
equivalent dose (HED) based on RDX arterial blood concentration, which was then used for RfD
derivation.
The overall RfD, 4 x 10"3 mg/kg-day, was calculated by dividing the BMDL05 expressed as a
human equivalent dose (BMDLos-hed) for nervous system effects by a composite uncertainty factor
(UF) of 300 to account for extrapolation from animals to humans (3), interindividual differences in
human susceptibility (10), and uncertainty in the database (10).
Because a subchronic-to-chronic uncertainty factor (UFs) of 1 was applied to the POD based
on evidence that nervous system effects (in particular convulsions) are more strongly driven by
dose than duration of exposure, the RfD may be appropriate for assessing health risks of less-than-
lifetime as well as chronic durations of exposure.
The overall confidence in the RfD is medium based on high confidence in the principal study
(Crouse etal.. 2006) and medium to low confidence in the database. Confidence in the database is
reduced largely because of (1) differences in test material used across studies (i.e., differences in
formulation and particle size that may have affected RDX absorption and subsequent toxicity),
(2) uncertainties in the influence of oral dosing methods (in particular, based on evidence that
bolus dosing of RDX resulting from gavage administration induces neurotoxicity at doses lower
than administration in the diet), and (3) significant limitations in the available studies to fully
characterize subconvulsive neurological effects as well as developmental neurotoxicity.
ES.2. EVIDENCE FOR HAZARDS OTHER THAN CANCER: INHALATION EXPOSURE
No studies were identified that provided useful information on the effects observed
following inhalation exposure to RDX. Of the available human epidemiological studies of RDX, none
provided data that could be used for dose-response analysis of inhalation exposures. The single
experimental animal study involving inhalation exposure is not publicly available and was excluded
from consideration due to significant study limitations, including small numbers of animals tested,
lack of controls, and incomplete reporting of exposure levels. Therefore, the available health effects
xxvii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
literature does not support the identification of hazards following inhalation exposure to RDX nor
the derivation of an inhalation reference concentration (RfC).
While inhalation absorption of RDX particulates is a plausible route of exposure, there are
no toxicokinetic studies of RDX inhalation absorption to support development of an inhalation
model. Therefore, a PBPK model for inhaled RDX was not developed to support route-to-route
extrapolation of an RfC from the RfD.
ES.3. EVIDENCE FOR HUMAN CARCINOGENICITY
Under EPA's cancer guidelines (U.S. EPA. 2005a). there is suggestive evidence of carcinogenic
potential for RDX. RDX induced benign and malignant tumors in the liver and lungs of mice (Parker
etal.. 2006: Lish etal.. 1984) or rats (Levine etal.. 1983) following long-term administration in the
diet. The potential for carcinogenicity applies to all routes of human exposure.
ES.4. QUANTITATIVE ESTIMATE OF CARCINOGENIC RISK FROM ORAL EXPOSURE
A quantitative estimate of carcinogenic risk from oral exposure to RDX was based on the
increased incidence of hepatocellular adenomas or carcinomas and alveolar/bronchiolar adenomas
or carcinomas in female B6C3Fi mice observed in the carcinogenicity bioassay in mice (Lish etal..
19841. This 2-year dietary study included four dose groups and a control group, adequate numbers
of animals per dose group (85/sex/group, with interim sacrifices of 10/sex/group at 6 and
12 months), and detailed reporting of methods and results (including individual animal data). The
initial high dose (175 mg/kg-day) was reduced to 100 mg/kg-day at Week 11 due to high mortality.
When there is suggestive evidence of carcinogenicity to humans, EPA generally would not
conduct a dose-response assessment and derive a cancer value. However, when the evidence
includes a well-conducted study (as is the case with RDX), 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 (U.S. EPA. 2005a).
An OSF was derived that considered the combination of female mouse liver and lung
tumors. In modeling these data sets, the highest dose group was excluded because of the initial
high mortality (loss of almost half the mice in that dose group). BMD and benchmark dose lower
confidence limit (BMDL) estimates were calculated that correspond to a 10% extra risk (ER) of
either tumor. The BMDLio so derived was extrapolated to the HED using body-weight scaling to the
% power (BW3/4), and an OSF was derived by linear extrapolation from the BMDLio expressed as an
HED (BMDLio-hed). The OSF is 0.08 per mg/kg-day, based on the liver and lung tumor response in
female mice (Lish etal.. 1984).
ES.5. QUANTITATIVE ESTIMATE OF CARCINOGENIC RISK FROM INHALATION EXPOSURE
An inhalation unit risk (IUR) value was not calculated because inhalation carcinogenicity
data for RDX are not available. While inhalation absorption of RDX particulates is a plausible route
of exposure, there are no toxicokinetic studies of RDX inhalation absorption to support an
xxviii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
inhalation model. Therefore, a PBPK model for inhaled RDX was not developed to support
route-to-route extrapolation of an IUR from the OSF. Thus, a quantitative cancer assessment was
not conducted for inhalation exposure.
ES.6. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
Little information is available on populations that may be especially vulnerable to the toxic
effects of RDX. Life stage, particularly childhood, susceptibility has not been well-studied in human
or animal studies of RDX toxicity. In rats, transfer of RDX from the dam to the fetus during
gestation and to pups via maternal milk has been reported; however, reproductive and
developmental toxicity studies did not identify effects in offspring at doses below those that also
caused maternal toxicity. Yet, based on the primary mode of action for RDX exposure-induced
nervous system effects (GABA receptor antagonism), and the fact that GABAergic signaling plays a
prominent role in nervous system development, a significant concern is raised regarding the
potential for developmental neurotoxicity. In addition, data on the incidence of convulsions and
mortality provide some indication that pregnant animals may be a susceptible population, although
the evidence is inconclusive. Data to suggest that males may be more susceptible than females to
noncancer toxicity associated with RDX are limited. Some evidence suggests that cytochrome P450
(CYP450) enzymes may be involved in the metabolism of RDX, indicating a potential for genetic
polymorphisms in these metabolic enzymes to affect susceptibility to RDX. Similarly, individuals
with epilepsy or other seizure syndromes that have their basis in genetic mutation to GABAa
receptors (GABA receptors that are ligand-gated ion channels, also known as ionotropic receptors)
may represent another group that may be susceptible to RDX exposure; however, there is no
information to indicate how genetic polymorphisms may affect susceptibility to RDX.
xxix
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
LITERATURE SEARCH STRATEGY |
STUDY SELECTION AND EVALUATION
LS.l LITERATURE SEARCH AND SCREENING STRATEGY
A literature search and screening strategy was applied to identify literature related to
characterizing the health effects of RDX. This strategy consisted of a search of online scientific
databases and other sources, casting a wide net to identify all potentially pertinent studies. In
subsequent steps, references were screened to exclude papers not pertinent to an assessment of
the chronic health effects of RDX, and the remaining references were sorted into categories for
further evaluation.
The literature search for RDX was conducted in four online scientific databases: PubMed,
Toxline, Toxcenter, and Toxic Substances Control Act Test Submissions (TSCATS). The initial
search was performed in April 2012, and literature search updates were conducted in February
2013, January 2014, January 2015, and May 2016. Searches of TSCATS were performed in February
2013, January 2015, and May 2016 only. In addition, a post-peer-review literature search was
conducted in November 2017 (described below). The detailed pre-peer-review search approach
for these databases, including the query strings, and the numbers of citations identified per
database are provided in Appendix B, Table B-l. The Department of Defense has conducted several
unpublished toxicological studies on RDX; to ensure that all such studies were located, the Defense
Technical Information Center (DTIC) database, a central online repository of defense-related
scientific and technical information within the Department of Defense, was also searched. A
separate strategy was applied in searching DTIC because of limitations in the classification and
distribution of materials in DTIC; the detailed search strategy is described in Appendix B, Table B-2.
Searches of the five online databases identified 1,247 citations (after electronically eliminating
duplicates). The computerized database searches were supplemented by reviewing online
regulatory sources, performing "forward" and "backward" searches of Web of Science (see
Appendix B, Table B-3), and adding additional references that were identified during the
development of the Toxicological Review (including submissions from the Department of Defense);
34 citations were obtained using these additional search strategies. In total, 1,281 citations were
identified using online scientific databases and additional search strategies.
The EPA requested public submissions of additional information in 2010 (75 FR 76982;
December 10, 2010). No submissions were received in response to these calls for data. EPA also
issued a request to the public for additional information in a Federal Register Notice in 2013 (78 FR
48674; August 9, 2013) and established a docket for public comment (EPA-HQ-ORD-2013-0430;
available at www.regulations.gov) maintained through the development of the assessment.
xxx
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
The citations identified using the search strategy described above were screened based on
title and abstract, and when needed, full text for pertinence to examining the health effects of
chronic RDX exposure. The process for screening the literature is described below and is shown
graphically in Figure LS-1 and on the RDX project page on EPA's Health and Environmental
Research Online (HERO) website at:
https: //hero.epa.gov/index.cfm/proiect/page/proiect id/2216.7 The objective of this manual
screen was to identify sources of primary human health effects data (i.e., human data and pertinent
data from in vivo animal models) and other sources of primary data that inform the assessment of
RDX health effects (i.e., genotoxicity and other mechanistic studies and toxicokinetic studies).
These data sources are represented by the bottom three boxes in Figure LS-1. Inclusion and
exclusion criteria used to manually screen the references to identify health effect studies (i.e., the
green boxes with dashed boarders in Figure LS-1) are provided in Table LS-1.
All studies that provided data on adsorption, distribution, metabolism, or excretion, PBPK
models, or relevant RDX MOA were tracked and considered in the assessment.
Reviews and other sources of RDX information (e.g., studies with exposure level
information) that did not meet the inclusion criteria for primary health effect studies in Table LS-1
were tracked as "Secondary Literature and Sources of Other RDX Information," and were
considered as appropriate during development of this assessment Studies identified as
"Excluded/Not on Topic" (see exclusion criteria in Table LS-1) were not further considered in this
assessment.
7HERO 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
(NCEA). The database includes more than two million scientific references, including articles from the
peer-reviewed literature. New studies are added continually to HERO.
Studies were assigned (or "tagged") to a given category in HERO that best reflected the primary content of the
study. In general, studies were not assigned multiple tags to simplify the tracking of references.
Nevertheless, the inclusion of a citation in a category (or tag) did not preclude its use in one or more other
categories. For example, Woody et al. fl 986). a case report of accidental ingestion of RDX by a child, was
tagged to the human case reports under Supplementary Studies in Figure LS-1. This case report also provides
pharmacokinetic data and was a pertinent source of information on RDX toxicokinetics, but was not assigned
a second tag for toxicokinetics.
xxxi
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Combined Dataset
n=l,281
Additional Search Strategies
(see Table B-3 for methods and results)
n=34
Pubmed
n=698
n=l,247 (After duplicates removed electronically)
Toxline
n=433
Database Searches (through May 2016)
(see Tables B-l and B-2 for keywords and limits)
Toxcenter
n=32
DTIC
n=85
TSCATS l/2/8e
n=2
<---
Supplementary Studies
Other Sources of Supplementary
Sources of Health Effects Data
Manual Screening For Pertinence
(Title/Abstract/Full Text)
Secondary Literature and Sources of
Other RDX Information (n=190)
113 Exposure levels
7 Mixtures only studies (applies to animal
studies only)
16 Regulatory documents
51 Reviews; risk assessments; editorials
3 Other
Excluded/Not on Topic (n=907)
32 Abstract only; inadequate reporting in
abstract; no abstract
244 Treatment/remediation
154 Chemical, physical or explosive
properties
258 Laboratory methods
182 Not chemical specific
37 Other
Sources of Health Effects Data
(n=21)
4 Human health effects studies
17 Animal toxicology studies
Other Sources of Supplementary
Health Effects Data (n=116)
9 Acute/short-term animal studies
16 Human case reports
91 Studies in ecological and other
nonmammalian species
Sources of Mechanistic and
Toxicokinetic Data (n=47)
11 Genotoxicity studies
17 Other mechanistic studies
19 Toxicokinetic studies
Figure LS-1. Summary of literature search and screening process for
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX).a
aThe numbers on this figure match the HERO project page as of 8/1/2018. See text for search strategy and results
of an updated literature search conducted in November 2017 (post-peer review).
xxxii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table LS-1. Inclusion-exclusion criteria for health effect studies3
Inclusion criteria
Exclusion criteria
Population
• Humans
• Standard mammalian
animal models, including
rat, mouse, rabbit,
guinea pig, monkey, dog
• In vitro studies—tracked
as supplementary
information
• Ecological and
nonmammalian
species—tracked as
supplementary
information
Exposure
• Exposure is to RDX
• Study population is not exposed to RDX
• Exposure is measured in
• Exposure to a mixture only (applied to animal studies only)
an environmental
• Exposure via injection (e.g., intravenous)b
medium (e.g., air, water,
diet)
• Exposure via oral or
inhalation routes
Outcome
• Study includes a
measure of one or more
health effect endpoints,
including effects on the
nervous, urinary,
musculoskeletal,
cardiovascular, immune,
and gastrointestinal
systems, reproduction,
development, liver, eyes,
and cancer
• Mechanistic and
toxicokinetic
studies—tracked as
supplementary
information
xxxiii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table LS-1. Inclusion-exclusion criteria for health effect studies3 (continued)
Inclusion criteria
Exclusion criteria
Other
• Reviews, regulatory documents (i.e., not primary sources of
health effect data)b
• Exposure levelsb
• Not on topic, including:
o Abstract only, inadequately reported abstract, or no
abstract, and not considered further because study was not
potentially relevant
o Bioremediation, biodegradation, or chemical or physical
treatment of RDX and other munitions, including evaluation
of wastewater treatment technologies and methods for
remediation of contaminated water and soil
o Chemical, physical, or explosive properties, including studies
of RDX crystal quality, energetics characteristics, sublimation
kinetics, isotope ratios, and thermal decomposition and
other explosive properties
o Analytical methods for measuring/detecting/remotely
sensing RDX in environmental media, and use in sample
preparations and assays
o Not chemical specific (studies that do not involve testing of
RDX)
o Other studies not informative for evaluating RDX health
effects and not captured by other exclusion criteria,
including:
¦ Superfund site records of decision that describe
remedial action plans for waste sites
¦ Characterization of waste sites contaminated by
explosives
¦ Foreign language studies where translation was not
warranted because, based on title or abstract, the
added value to the evaluation of RDX health effects was
considered small (e.g., Chinese paper of case reports of
RDX poisonings)
¦ Duplicate studies not previously identified during
electronic screening
alnclusion/exclusion criteria were designed to identify sources of primary human health effects data (i.e., human
data and pertinent data from in vivo animal models).
bStudies that met this exclusion criterion were not considered a primary source of health effects or
supplementary health effects data; however, these studies were tracked and considered as other sources of
information potentially useful in assessing the health effects of RDX, including potential MOAs.
xxxiv
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
The results of this literature screening are described below and graphically in Figure LS-1:
• Twenty-one references (including both human and animal studies) were identified as
sources of health effects data and were considered for data extraction to evidence tables
and exposure-response arrays.
• Twenty-five references were identified as sources of supplementary health effects data,
including human case reports and experimental animal studies involving acute or
short-term exposures or dermal exposure. Studies investigating the effects of
acute/short-term and dermal exposures and case reports are generally less pertinent for
characterizing health hazards associated with chronic oral and inhalation exposure.
Therefore, information from these studies was not extracted into evidence tables.
Nevertheless, these studies were still considered as possible sources of supplementary
health effects information.
• Ninety-one references provided information on nonmammalian species (tagged as
ecosystem studies) that can inform the hazard evaluation or potential MOA, and specifically
in the case of RDX, the conservation of neurotoxic response across phylogenetically diverse
organisms. Information from these studies was not extracted into evidence tables;
however, these studies were tracked as supplementary health effects information.
• Forty-seven references were identified as sources of mechanistic and toxicokinetic data;
these included 19 studies describing PBPK models and other toxicokinetic information,
11 studies providing genotoxicity information, and 17 studies pertaining to other
mechanistic information. Information from these studies was not extracted into evidence
tables; however, these studies supplemented the assessment of RDX health effects.
Specifically, mechanistic studies were used in evaluating potential MOAs and to develop the
mechanistic evidence stream that was considered in the overall integration of evidence for
assessing hazard. Toxicokinetic data were used to inform extrapolation of experimental
animal findings to humans.
• One hundred and ninety references were identified as secondary literature (e.g., reviews
and other agency assessments), peer-review reports of primary (unpublished) health effect
studies, or contextual information (e.g., studies with RDX exposure information). These
references were kept as additional resources for development of the Toxicological Review.
• Nine hundred and seven references were identified as not being pertinent (or not on topic)
to an evaluation of the chronic health effects of RDX and were excluded from further
consideration (see Figure LS-1 and Table LS-1 for exclusion criteria).
LS.1.1. Post-Peer-Review Literature Search Update
A post-peer-review literature search update was conducted in PubMed, Toxline, TSCATS,
and DTIC for the period May 2016 to November 2017 using a search strategy consistent with
previous literature searches (see Appendix B, Tables B-l and B-2). Toxcenter, used in previous
searches, was not searched in the November update. Toxcenter is a proprietary, fee-based database
xxxv
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
produced by Chemical Abstract Service. Evaluation of the references retrieved using Toxcenter in
searches conducted through May 2016 revealed that this database did not locate pertinent
references not already identified by other online databases. Results of the November 2017
literature search update are summarized in Appendix B, Tables B-l and B-2.
Consistent with the IRIS Stopping Rules
(https://www.epa.gov/sites/production/files/2014-06/documents/iris stoppingrules.pdf).
manual screening of the literature search update focused on identifying new studies that might
change a major conclusion of the assessment No potentially pertinent references were identified in
the post-peer-review literature search.
The documentation and results for the literature search and screen, including the specific
references identified using each search strategy and tags assigned to each reference based on the
manual screen, can be found on the HERO website on the RDX project page at:
(https://hero.epa.gov/index.cfm/project/page/project id/2216).
LS.2 SELECTION OF CRITICAL STUDIES AND STUDY EVALUATION
LS.2.2. Selection of Critical Studies
In order to systematically summarize the important information from the primary health
effects studies in the RDX database, evidence tables were constructed in a standardized tabular
format as recommended by the NRC (20111. Of the studies that were retained after the literature
search and screen, 21 were categorized as "Sources of Health Effects Data" (see Figure LS-1 and
Table LS-1) and were considered for extraction into evidence tables for hazard identification in
Section 1.
A study was not subject to a more thorough review of study quality and was not presented
in evidence tables if flaws in its design, conduct, or reporting were so great that the results would
not be considered credible (e.g., studies where concurrent control information is lacking). Such
study design flaws are discussed in a number of EPA's guidelines (see
https: //www.epa.gov/iris /backgrd.html and Section 4 of the Preamble). For RDX, four studies
were considered uninformative and were removed from further consideration in the assessment
because of fundamental issues with study design, conduct, or reporting. The specific studies and
basis for considering the studies to be uninformative are summarized in Table LS-2.
The health effects literature for RDX is not extensive. Except for the studies listed in
Table LS-2 (i.e., those determined to be uninformative), all human and experimental animal studies
of RDX involving repeated exposure were considered in assessing the evidence for health effects
associated with chronic exposure to RDX.
xxxvi
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table LS-2. Studies determined not to be informative because of significant
issues with design, conduct, or reporting
Reference
Rationale for exclusion
Haskell Laboratories (1942);
repeated-dose studies in dogs and rats
Inadequate reporting of study design (e.g., limited exposure
information, breed of dog was not reported) and results; sections of
document were illegible. Deficiencies in experimental design of
dog study (e.g., investigation of only blood pressure in three dogs
exposed for 2,14, or 16 wk; no separate control). Rat study
included only 10 rats treated 41 times with RDX over an unspecified
exposure duration; only body weight and survival findings were
reported.
von Oettingen et al. (1949);
10-wk oral study in rats
No control group; strain of rat was not reported.
Note: Other studies included in the paper by von Oettingen et al.
(1949) were retained; results of these studies are included in
evidence tables.
ATSDR (1996);
disease prevalence study in residential
population
Study of a population residing in two neighborhoods where RDX
had been detected in well water. The study was conducted 7 yr
after residents were given the opportunity to connect to a
municipal water supply. Only one target-area household reported
using private well water for bathing and cooking at the time of the
health study. The study was not considered informative because
the design was not able to adequately define the exposed
population.
Unpublished report (dated 1944) from the
DTIC database;
human and animal data
One section of the report describes a human case series with no
referent group. Issues with the inhalation experimental animal
studies included lack of control groups, incomplete information on
exposure levels, and inadequate reporting of results. (Because this
report is classified as a limited distribution document in the DTIC
database, it was not added to the HERO project page for RDX.)
Studies that contain pertinent information for the toxicological review and augment hazard
identification conclusions, such as genotoxicity and other mechanistic studies, studies describing
the toxicokinetics of RDX, human case reports, and experimental animal studies involving
exposures of acute/short-term duration or routes of exposure other than oral and inhalation, were
not included in evidence tables. Nevertheless, these studies were considered, where relevant, in the
evaluation of RDX health hazards.
xxxvii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
LS.2.3. Study Evaluation
For this assessment, primary sources of health effects data consisted of three human
studies8 and 16 reports9 presenting results of experimental animal studies. These studies were
evaluated using the study quality considerations described below that addressed aspects of design,
conduct, or reporting that could affect the interpretation of results, overall contribution to the
synthesis of evidence, and determination of hazard potential as noted in various EPA guidance
documents (U.S. EPA. 2005a. 2002.19941. The objective was to identify the stronger, more
informative studies based on a uniform evaluation of quality characteristics across studies of
similar design.
Additionally, a number of general questions, presented in Table LS-3, were considered in
evaluating the animal studies. Much of the key information for conducting this evaluation can be
determined based on study methods and how the study results were reported. Importantly, the
evaluation at this stage does not consider the direction or magnitude of any reported effects.
8Two reports with human data were determined not to be informative; see Table LS-2. The study by ATS PR
(19961 was included in HERO and in Figure LS-1. The unpublished report from the DTIC database was not
included in either HERO or Figure LS-1 because this report is classified as a limited distribution document in
DTIC. This accounts for the three human studies being reviewed for study evaluation rather than the four
identified in the literature search (see Figure LS-1],
9One of 17 animal toxicity studies identified in Figure LS-1 [Haskell Laboratories (19421] was determined to
be uninformative (see Table LS-2], This study was included in HERO and Figure LS-1, but was not considered
a primary source of health effects data and was not carried forward for further review.
Also note that the number of reports of experimental animal studies does not equal the number of studies for
several reasons. The results of some studies were documented in multiple reports [e.g., a 2-year study in
F344 rats by Levine etal. (19831 was published in three volumes]. The Cholakis et al. (19801 study included,
in a single report, subchronic studies in rats and mice, a two-generation reproductive toxicity study in rats,
and developmental toxicity studies in rats and rabbits. A13-week toxicity study of RDX in rats was reported
initially as a laboratory report study (Levine et al.. 1981 al and results were subsequently included in two
published papers. A Pathology Working Group review of the female mouse liver tumor data in the Lish etal.
(19841 2-year bioassay was provided as a study report and subsequently as a published paper.
xxxviii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table LS-3. Considerations and relevant experimental information for
evaluation of experimental animal studies
Methodological
feature
Considerations
(relevant information extracted into evidence tables)
Test animal
Suitability of the species, strain, sex, and source of the test animals
Experimental design
Suitability of animal age/life stage at exposure and endpoint testing; periodicity and
duration of exposure (e.g., h/d, d/wk); timing of endpoint evaluations; and sample size
and experimental unit (e.g., animals, dams, litters)
Exposure
Characterization of test article source, composition, purity, and stability; suitability of the
control (e.g., vehicle control); documentation of exposure techniques (e.g., route,
chamber type, gavage volume); verification of exposure levels (e.g., consideration of
homogeneity, stability, analytical methods)
Endpoint evaluation
Suitability of specific methods for assessing the endpoint(s) of interest
Results presentation
Data presentation for endpoint(s) of interest (including measures of variability) and for
other relevant endpoints needed for results interpretation (e.g., maternal toxicity,
decrements in body weight in relation to organ weight)
Information on study features related to this evaluation is reported in evidence tables and
was considered in the synthesis of evidence. Discussion of study strengths and limitations (that
ultimately supported preferences for the studies and data relied upon) were included in the text
where relevant. If EPA's interpretation of a study differed from that of the study authors, the
assessment discusses the basis for the difference.
The general findings of this evaluation are presented in the remainder of this section and
discussed in the relevant health effect sections in Section 1.2.
Human Studies
The body of literature on RDX includes three studies of populations occupationally exposed
to RDX [one case-control and two cross-sectional studies; fWest and Stafford. 1997: Ma and Li.
1993: Hathaway and Buck. 1977)]. To varying degrees, these epidemiology studies are limited in
their ability to assess the relationship between RDX exposure and the incidence of human health
effects. Some studies lacked information related to study design, such as a clear definition of the
study population, while others did not include a comprehensive exposure assessment or details
regarding potential confounders. All three studies had small sample sizes (60-69 exposed workers
in the cross-sectional studies and 32 cases in the case-control study), which limits their statistical
power when comparing exposed workers or cases and unexposed or control participants.
The study by Ma and Li (1993) of Chinese industrial workers provided limited information
on participant recruitment, selection, and participation rate; the available information was not
adequate to evaluate the potential for selection bias. Also, no information on adjustment for
coexposure to trinitrotoluene (TNT) or other neurological risk factors (e.g., alcohol consumption)
was provided. The study by Hathaway and Buck (1977) included details on exposure assessment,
xxxix
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
but did not provide information on length of employment or other metrics that could be used to
ascertain duration of exposure. In the case-control study by West and Stafford (1997). RDX was
identified as one of the many chemicals that workers may have been exposed to in the ordnance
factory. Thus, there is a potential for coexposure to other chemicals that may elicit the observed
effects. The methodological limitations in these three studies were considered in the synthesis of
evidence for each of the health effects and in reaching determinations of hazard (see Section 1.2).
In addition to the three epidemiological studies, the human health effects literature includes
16 case reports that describe effects following acute exposure to RDX. Case reports can suggest
organ systems and health outcomes that might be related to RDX exposure but are often anecdotal,
and typically describe unusual or extreme exposure situations; thus, they provide little information
that would be useful for characterizing chronic health effects or deriving toxicity values. Therefore,
RDX case reports were only briefly reviewed; a critical evaluation was not undertaken. A summary
of these case reports is provided in Appendix C, Section C.2.
Experimental Animal Studies
The oral toxicity database for RDX includes three chronic studies in rats and mice, eight
subchronic studies in rats, mice, dogs, and monkeys, two shorter term studies in dogs and rats, one
two-generation reproductive toxicity study in the rat, four developmental toxicity studies in rats
and rabbits, and a single-exposure study of audiogenic seizures in rats (see Table LS-4).
With the exception of two studies (Levine etal.. 1990: von Oettingen et al.. 1949). these
toxicity studies are available only as unpublished contract laboratory reports. Peer reviews of four
unpublished studies identified as most informative to the assessment of the health effects of
RDX—the 2-year bioassays by Levine etal. (1983) and Lish etal. (1984). the subchronic toxicity
study by Crouse etal. (2006). and the collection of repeated-dose studies reported in Cholakis et al.
(1980)—were conducted by Versar, Inc. or Eastern Research Group, Inc. for EPA. The reports of the
peer reviews (U.S. EPA. 2017. 2012c) are available at https://epa.gov/hero. The peer reviewers
generally concluded that the 2-year bioassay reports provided useful information on the toxicity of
RDX, noting that there were limitations in interpretation due to aspects of the histopathological
analysis and the statistical approaches employed. The peer reviewers similarly determined that the
report by Crouse etal. (2006) provided useful information on RDX toxicity, including an array of
endpoints for neurotoxicity and immunotoxicity, although the assessment of neurotoxicity in the
study could have been improved with more histological evaluation as well as additional behavioral
assessment (U.S. EPA. 2012c). The peer-review report of the repeated-dose studies in Cholakis et
al. (1980) found that the studies were generally appropriate and adequate for evaluating the
toxicity of RDX, with experimental design and reporting consistent with the standards in place at
the time the experiments were conducted (U.S. EPA. 2017).
xl
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table LS-4. Summary of experimental animal database
Study category
Study duration, species/strain, and oral administration method
Chronic
2-Yr studv in B6C3Fi mice (diet) (Lish et al., 1984)
2-Yr studv in Sprague-Dawley rats (diet) (Hart, 1976)
2-Yr studv in F344 rats (diet) (Levine et al., 1983)
Subchronic
13-Wk studv in B6C3Fi mice, experiment 1 (diet) (Cholakis et al., 1980)
13-Wk studv in B6C3Fi mice, experiment 2 (diet) (Cholakis et al., 1980)
13-Wk studv in F344 rats (diet) (Cholakis et al., 1980)
13-Wk studv in F344 rats (diet) (Levine et al., 1990; Levine et al., 1981a, b)
13-Wk studv in F344 rats (gavage) (Crouse et al., 2006)
13-Wk studv in rats, strain not specified (diet) (von Oettingen et al., 1949)
13-Wk studv in beagle dogs (diet) (Hart, 1974)
13-Wk studv in monkeys (gavage) (Martin and Hart, 1974)
6-Wk studv in dogs, breed not specified (diet) (von Oettingen et al., 1949)
30-D studv in Sprague-Dawley rats (gavage) (MacPhail et al., 1985)
Reproductive
2-Generation reproductive toxicity studv in CD rats (diet) (Cholakis et al., 1980)
Developmental
Developmental studv (GDs 6-19) in F344 rats (gavage) (Cholakis et al., 1980)
Developmental studv (GDs 6-15) in Sprague-Dawlev rats, range-finding (gavage) (Angerhofer
et al., 1986)
Developmental studv (GDs 6-15) in Sprague-Dawlev rats (gavage) (Angerhofer et al., 1986)
Developmental studv (GDs 7-29) in NZW rabbits (gavage) (Cholakis et al., 1980)
Nervous system
8-h studv of audiogenic seizures in Long-Evans rats (gavage) (Burdette et al., 1988)a
Acute EEG and in vitro studies of RDX evoked seizure activity in male Sprague-Dawley rats
(Williams et al., 2011)
EEG = electroencephalogram; GD = gestational day; NZW = New Zealand White.
aAs an 8-h study, Burdette et al. (1988) was tagged in Figure LS-1 and the HERO database as "Other Sources of
Supplementary Health Effects Data," but was nevertheless included in the evidence table for nervous system
effects of RDX as the only study to examine potential effects of RDX on seizure threshold.
Only one unpublished inhalation study of RDX (dated 1944) was identified. This inhalation
study was considered uninformative and was excluded from consideration in developing the
Toxicological Review because of study design issues (including lack of a control group, incomplete
information on exposure levels, and inadequate reporting, see Appendix B and Table LS-2).
Therefore, evaluation of the experimental animal database for RDX is limited to studies of oral
toxicity. An evaluation of the oral toxicity literature, organized by general methodological features,
is provided in the remainder of this section.
Test Animal
The RDX database consists of health effect studies conducted in multiple strains of rats
(F344, Sprague-Dawley, CD), mice (B6C3Fi), dogs (beagle), and monkeys. The species and strains
of animals used are consistent with those typically used in laboratory studies and thus considered
xli
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
relevant to assessing the potential human health effects of RDX. The species, strain, and sex of the
animals used are recorded in the evidence tables.
Other studies of RDX were identified that used nonstandard species, including deer mice
[Peromyscus maniculatus), western fence lizards (Sceloporus occidentalis), prairie voles (Microtus
ochrogaster), and northern bobwhite quail (Col in us virginianus). These studies provide information
relevant to RDX toxicokinetics and mechanism of action on the nervous system, but not health
effects data. Therefore, these studies (tagged under Supplementary Studies; see Figure LS-1) are
not included in evidence tables, but are discussed where relevant in the assessment
Experimental Design
General aspects of experimental design were evaluated for all studies that included health
effects data to determine whether they were appropriate for evaluating specific endpoints. Key
features of the experimental design, including the periodicity and duration of exposure, timing of
exposure (e.g., gestational days for developmental studies), experimental group sample sizes, and
interim sacrifices are summarized in the evidence tables. Note that sample size was not a basis for
excluding a study from consideration, as studies with a small number of animals can still inform the
consistency of effects observed for a specific endpoint. Nevertheless, the informativeness was
reduced for studies with small sample sizes, for example three animals/sex/group in the case of
Hart (19741 and Martin and Hart (19741. Elements of the experimental setup that could influence
interpretation of study findings are discussed in the relevant hazard identification sections of the
assessment.
Exposure
Studies were evaluated with respect to the reliability of the reported exposure to RDX,
focusing on considerations related to properties of the test material and confirmation of the
administered dose.
Two properties of the RDX test materials that varied across experimental animal studies
and that were considered in evaluating the evidence for RDX hazards are the particle size and
purity of the test material. The purity of RDX used in health effects studies varied from 84 to
99.99%. The major contaminant was octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine (HMX), which
is produced during manufacturing. The majority of studies used RDX with ~10% impurities; only
Crouse etal. (20061 used 99.99% pure RDX as a test material in their study. The toxicity of HMX
was assessed by the IRIS Program in 1988
(https://cfj3ub.epa.gov/ncea/iris2/chemicalLanding.cfm7substance nmbr=3111: histopathological
changes in the liver in male F344 rats and in the kidney in female rats were reported in a 13-week
feeding study. No chronic-duration studies were available to evaluate the carcinogenicity of HMX.
The presence of the impurities introduces some uncertainty in attributing the observed effects to
RDX. However, consistency in the doses at which some toxic effects were seen across studies
suggests that the uncertainty associated with the use of less pure test materials may be relatively
xlii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
small. Evidence of neurotoxic effects in the study with 99.99% pure RDX occurred at doses of
8-15 mg/kg-day; studies with less pure RDX reported similar symptoms at doses >20 mg/kg-day.
Note that the test materials employed in these studies (i.e., with ~10% impurities) are consistent
with the purity of RDX that would be released into the environment.
Differences in milling procedures used to generate the test material resulted in the use of
RDX of varying particle sizes across studies. Some studies used a test material with a relatively fine
particle size (majority of particles <66 |im in size), while others used a test material with
comparatively coarse particle size (~200 |im particle size). Differences in particle size across
studies could result in different rates of absorption of RDX into the blood stream, which could
account for differences in response observed across studies, including neurotoxicity.
Information on test material purity and particle size, as provided by study authors, is
reported in the evidence tables, and was considered in evaluating the toxicity of RDX. The lack of
characterization of the test material in the studies by Hart (1974). Hart (1976). and Martin and Hart
(1974) was considered a deficiency.
Only four studies assayed dose preparations to determine how close the actual RDX
concentrations were to target (nominal) concentrations (Crouse etal.. 2006: Lish etal.. 1984:
Levine etal.. 1983: Cholakis etal.. 19801. Cholakis etal. fl9801 described the largest difference
between target and actual dose concentrations; assays of the suspensions prepared for the oral
(gavage) developmental toxicity study showed RDX dosing suspensions ranging from 36 to 501% of
the target concentrations [see Appendix I of Cholakis etal. (1980)]. Assays of RDX-treated feed
used in the 90-day studies in rats and mice and the two-generation reproductive toxicity study in
rats showed RDX concentrations that were 78 to 209% of target concentrations [see Appendix I of
Cholakis etal. fl9801]. The study authors stated, "maintaining uniform suspensions was not always
easy." In the 90-day oral (gavage) toxicity study in rats (Crouse et al.. 2006). fresh dose
suspensions were prepared monthly, mixed with a magnetic stir bar until a uniform suspension
was obtained, and remixed each day during the dosing procedure; each dose suspension was
analyzed prior to use. RDX concentrations varied from 83 to 114%; the 114% suspension was
adjusted to 100% before administration (Crouse etal.. 2006). In 30 assays performed over the
course of a 24-month bioassay in mice, Lish etal. (1984) determined dietary concentrations of RDX
to be 73 to 103% of target concentrations. In 32 assays performed over a 24-month bioassay in
rats, Levine etal. (1983) reported that dietary concentrations of RDX were 67 to 122% of target
concentrations. In the remaining studies, failure to analyze or report actual concentrations of RDX
in the dosing suspension or test diet is considered a deficiency.
Endpoint Evaluation Procedures
Some methodological considerations used to evaluate studies of RDX toxicity are outcome
specific, particularly effects on the nervous system and development. Outcome-specific
methodological considerations are discussed in the relevant health effect sections in Section 1.2.
For example, many of the studies that noted neurotoxicity in the form of seizures or convulsions
xliii
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
were not designed to assess that specific endpoint and reported the number of animals with
seizures as part of clinical observations that, in general, were recorded only once daily. This
frequency of observations could have missed neurobehavioral events or failed to identify subtler
subconvulsive behaviors. While these studies can provide qualitative evidence of neurotoxicity,
they may have underestimated the true incidence of seizures or convulsive behaviors because they
were not designed to systematically evaluate neurotoxic outcomes.
Results Presentation
In evaluating studies, consideration was given to whether data were reported for all
endpoints specified in the methods section and for all study groups, and whether any data were
excluded from presentation or analysis. For example, it was noted where histopathological analysis
was limited to control and high-dose groups, a study reporting feature that limited the ability to
identify dose-related trends. In limited cases, EPA performed additional statistical analysis to
identify trends or refine analyses consistent with EPA guidance (e.g., analyzing developmental data
sets on a per litter basis rather than by individual fetus). Study results have been extracted and
presented in evidence tables.
Notable Features of the Hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) Database
Three 2-year toxicity bioassays of RDX are available as unpublished laboratory studies (Lish
etal.. 1984: Levine etal.. 1983: Hart. 19761. The bioassays by Levine etal. (19831 in the rat and by
Lish etal. (19841 in the mouse were conducted in accordance with Food and Drug Administration
Good Laboratory Practices in place at the time of the studies. Both studies included interim
sacrifices (at 6 and 12 months). Complete histopathological examinations were performed on all
animals in the control and high-dose groups; however, only a subset of tissues was examined in the
mid-dose groups (including brain, gonads, heart, liver, kidneys, spleen, and spinal cord in both
species, and lungs and tissue masses in the mouse), limiting the ability to identify dose-related
trends for tissues with incomplete histopathology. Additionally, in the mouse bioassay by Lish et al.
(19841. the initial high dose (175 mg/kg-day) was reduced to 100 mg/kg-day at Week 11 because
of high mortality, thereby reducing the number of high-dose animals on study for the full 2 years of
dosing (see Table LS-5).
An earlier unpublished 2-year study in rats by Hart (19761 used a dose range that was
lower than the Levine etal. (19831 and Lish etal. (19841 bioassays. Histopathology findings were
limited by the lack of pathology examinations in the mid-dose groups and lack of individual time of
death, which impacts the ability to interpret the histopathology data. In addition, a heating system
malfunction on Days 75-76 of the study resulted in the death of 59 rats from the control and
treatment groups, thereby reducing the number of animals in the study (see Table LS-5).
Experimental animal toxicity studies of RDX involving less-than-lifetime exposure (Crouse
etal.. 2006: Angerhofer et al.. 1986: MacPhail etal.. 1985: Levine etal.. 1981a: Cholakis etal.. 1980:
Hart. 1974: Martin and Hart. 1974: von Oettingen et al.. 19491 were published or reported between
xliv
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
the years 1949 and 2006, and differences in robustness of study design, conduct, and reporting
reflect that time span. All but two of the eight short-term and subchronic toxicity studies of RDX
are available as unpublished laboratory studies; published studies include von Oettingen et al.
(19491 and Levine etal. (1981al. a laboratory report of a 13-week study of RDX in F344 rats with
subsets of the data subsequently published as Levine etal. (1981bl and Levine etal. (19901. Most
studies conducted histopathological examinations on only some of the experimental groups (e.g.,
control and high dose).
Some of the more important limitations in study design, conduct, and reporting of
experimental animal toxicity studies of RDX are summarized in Table LS-5. Limitations of these
studies, as well as the study evaluation consideration described in this section, were considered in
evaluating and synthesizing the evidence for each of the health effects in Section 1.2.
xlv
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table LS-5. Experimental animal studies considered less informative because
of certain study design, conduct, or reporting limitations
References
Study design, conduct, and reporting limitations
Lish et al. (1984);
2-yr mouse study
The initial high dose (175 mg/kg-d) was reduced to 100 mg/kg-d at Week 11
due to high mortality. Mortality of surviving mice was similar to controls
after dose reduction.
Hart (1976);
2-yr rat study
A heating system malfunction on Days 75-76 of the study resulted in the
deaths of 59 rats from the control and treatment groups. Dead animals
were subsequently eliminated from the analysis. Interpretation of the
histopathology findings was limited by the lack of pathology examinations
in the mid-dose groups and lack of individual time of death. Test material
poorly characterized; purity was not reported.
Cholakis et al. (1980);
13-wk mouse study (Experiment 1)
The dose range was too low to produce effects in mice. Assays of
RDX-treated feed showed RDX concentrations between 123 and 209% of
target concentrations. Histopathological examinations were not
performed.
Cholakis et al. (1980);
13-wk mouse study (Experiment 2)
Nonstandard dosing regimen followed: 0, 40, 60, or 80 mg/kg-d for 2 wk.
For the next 11 wk, the dosing was inverted, so that the 40 mg/kg-d group
received 320 mg/kg-d, the 60 mg/kg-d group received 160 mg/kg-d, and the
80 mg/kg-d group continued to receive the same dose. The rationale for
this dosing regimen was not provided in the study report.
Cholakis et al. (1980);
Developmental study in rats
Large differences were reported between target and actual dose
concentrations in the suspensions prepared for oral (gavage)
administration; actual RDX concentrations in dosing suspensions ranged
from 36 to 501% of the target concentrations.
Levine et al. (1981a);
13-wk rat study
Analysis of one lot of rodent feed showed measurable levels of
contaminants, including chlorinated pesticides (dieldrin, heptachlor
epoxide, beta-hexachlorocyclohexane, and
dichlorodiphenyltrichloroethane), polychlorinated biphenyls, and
organophosphates (methyl parathion, carbophenothion, and disulfeton).
Martin and Hart (1974);
13-wk monkey study
The species of monkey is unclear (either cynomolgus or rhesus). Some test
subjects may have had variable dosing due to emesis. Small sample size per
dose group (n = 3). Test material poorly characterized; purity was not
reported.
von Oettingen et al. (1949);
12-wk rat study
The strain of rat was not reported. Only gross observations were made at
necropsy.
von Oettingen et al. (1949);
6-wk dog study
The breed of dog was not reported. Only gross observations were made at
necropsy.
xlvi
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
1.HAZARD IDENTIFICATION
1.1. OVERVIEW OF CHEMICAL PROPERTIES AND TOXICOKINETICS
1.1.1. Chemical Properties
Hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) is a member of the nitramine class of organic
nitrate explosives (Boileau etal.. 2003: Bingham etal.. 20011 and is not found naturally in the
environment. RDX is a white, crystalline solid (Bingham etal.. 2001). It has low solubility in water
(Yalkowskv and He. 2003) and slowly volatilizes from water or moist soil (ATSDR. 2012). The
normalized soil organic carbon/water partition coefficient values for RDX indicate a potential for
RDX to be mobile in soil fSpanggord et al.. 1980). The vapor pressure suggests that RDX will exist
as particulate matter in air and be removed by both wet and dry deposition fSpanggord etal..
1980). Information on physiochemical properties for RDX is available at U.S. Environmental
Protection Agency (EPA)'s Chemistry Dashboard (https://comptox.epa.gov/dashboard/) and is
summarized in Table 1-1.
RDX degrades in the environment and can be subject to both photolysis (Sikka etal.. 1980:
Spanggord et al.. 1980) and biodegradation (Funk etal.. 1993: McCormick etal.. 1981). RDX is
metabolized by microbial nitroreductases to form the /V-nitroso derivatives
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; (Taligamaetal.. 2013: Halasz et
al.. 2012: Smith etal.. 2006: Meyer etal.. 2005: Beller andTiemeier. 2002)].
4-Nitro-2,4-diazabutanal (NDAB) and methylenedinitramine (MEDINA) have also been detected as
microbial metabolites of RDX (Halasz etal.. 2012: Fuller etal.. 2010).
1-1
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-1. Chemical identity and physicochemical properties of hexahydro-
l,3,5-trinitro-l,3,5-triazine (RDX) from EPA's Chemistry Dashboard
Characteristic or property
Value
Chemical structure
0^N X"0"
^ N
t
.N„
r i
It 1
0 0"
CASRN
121-82-4
Synonyms
l,3,5-triaza-l,3,5-trinitrocyclohexane; 1,3,5-triazine, hexahydro-l,3,5-trinitro-;
l,3,5-trinitro-l,3,5-triazacyclohexane; l,3,5-trinitro-l,3,5-triazinane;
1,3,5-trinitrohexahydro-1,3,5-triazine; 1,3,5-trinitrohexahydro-s-triazine;
1,3,5-trinitroperhydro-1,3,5-triazine; cyclonite; cyclotrimethylenenitramine;
cyclotrimethylenetrinitramine; hexahydro-l,3,5-trinitro-l,3,5-s-triazine;
hexahydro-l,3,5-trinitro-1,3,5-triazine; hexahydro-l,3,5-trinitro-s-triazine;
hexogen; perhydro-l,3,5-trinitro-l,3,5-triazine; RDX; Research Development
Explosive; Royal Demolition explosive; sym-trimethylene trinitramine;
s-triazine, hexahydro-l,3,5-trinitro-; trimethylenetrinitramine;
trinitrocyclotrimethylene triamine; trinitrotrimethylenetriamine (see
https://comptox.epa.gov/dashboard for additional synonyms)
Molecular formula
CsHgNgOs
Molecular weight
222.117
Average experimental value3
Average predicated value3
Flash point (°C)
—
388
Boiling point (°C)
—
407
Melting point (°C)
205
162
Log Kow
0.87
-0.425
Water solubility (mol/L)
2.69 x 10"4
8.37 x 10"3
Density (g/cm3)
1.84
Henry's law constant
(atm-m3/mole)
2.53 x 10"6
Vapor pressure (mm Hg at 20°C)
4.10 x 10
3.76 x 10"9
atm = atmosphere; CASRN = Chemical Abstracts Service registry number.
aMedian values and ranges for physical chemical properties of RDX are also provided on the Chemistry Dashboard
at https://comptox.epa.gov/dashboard/.
1-2
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
1.1.2. Toxicokinetics
RDX is absorbed following exposure by oral and inhalation routes (see Appendix C,
Section C.l.l). Studies in experimental animals indicate that oral absorption rates can range from
approximately 50 to 90% (Krishnan etal.. 2009: Guo etal.. 1985: Schneider et al.. 1978.1977). with
the rate and extent of absorption dependent on the physical form of RDX (i.e., the increased surface
area associated with finely powdered RDX allows for increased absorption) and the dosing
preparation or matrix (Bannon etal.. 2009a: Krishnan etal.. 2009: Crouse etal.. 2008: Bannon.
2006: Guo etal.. 1985: MacPhail etal.. 1985: Schneider et al.. 1977). Dermal absorption of RDX has
been demonstrated in in vitro studies using human and pig skin (Reddv etal.. 2008: Reifenrath et
al.. 20081.
RDX is systemically distributed, including to the brain (i.e., RDX can cross the blood-brain
barrier), heart, kidney, liver, and fat (Musick etal.. 2010: Bannon etal.. 2006: MacPhail etal.. 1985:
Schneider et al.. 1977). In rats, RDX can be transferred from dam to fetus across the placental-blood
barrier, and has been identified in maternal milk fHess-Ruth et al.. 20071.
The metabolism of RDX in humans has not been investigated. Studies in experimental
animals indicate that metabolism of RDX is extensive and includes denitration, ring cleavage, and
generation of CO2 possibly through cytochrome P450 [CYP450; (Musick etal.. 2010: Major etal..
2007: Fellows etal.. 2006: Bhushan etal.. 2003: Schneider et al.. 1978.19771],
RDX and its metabolites are eliminated primarily via urinary excretion and exhalation of
CO2 (Sweeney etal.. 2012a: Musick etal.. 2010: Krishnan et al.. 2009: Major et al.. 2007: Schneider
etal.. 1977). Estimated elimination half-lives (ty2; estimated ty2 values based on RDX concentrations
in blood) indicate that RDX is more rapidly metabolized in mice than in rats and humans; estimated
ty2 values were 1.2 hours for mice, 5-10 hours for rats, and 15-29 hours for humans (Sweeney et
al.. 2012b: Krishnan etal.. 2009: Ozhan etal.. 2003: Woody et al.. 1986: Schneider et al.. 1977).
A more detailed summary of RDX toxicokinetics is provided in Appendix C, Section C.l.
1.1.3. Description of Toxicokinetic Models
A physiologically based pharmacokinetic (PBPK) model to simulate the pharmacokinetics of
RDX in rats was first developed by Krishnan et al. (2009) and revised to extend the model to
humans and mice (Sweeney etal.. 2012a: Sweeney et al.. 2012b). The Sweeney etal. (2012a) model
consists of six main compartments: blood, brain, fat, liver, and lumped compartments for rapidly
perfused tissues and slowly perfused tissues, and can simulate RDX exposures via the intravenous
or oral route. This model assumes that the distribution of RDX to tissues is flow-limited, and
represents oral absorption as first-order uptake from the gastrointestinal (GI) tract into the liver,
with 100% of the dose absorbed. RDX is assumed to be cleared by first-order metabolism in the
liver. The model does not represent the kinetics of any RDX metabolites. The Sweeney et al.
(2012a) and Sweeney et al. (2012b) PBPK models were evaluated and subsequently modified by
the EPA for use in dose-response modeling in this assessment (see Appendix C, Section C.l.5).
1-3
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
1.2. PRESENTATION AND SYNTHESIS OF EVIDENCE BY ORGAN/SYSTEM
In experimental animal studies, RDX test material administered in toxicology studies
included formulations that ranged in purity (from 84 to 99.99%) and in particle size (from <66 to
~200 |im particle size). Differences in test material purity and particle size were taken into
consideration while evaluating RDX toxicity findings, as discussed in the literature search section
and incorporated in the synthesis of evidence.
Mortality has been reported in the animal toxicology studies conducted for RDX. Due to the
serious nature associated with a frank effect such as mortality, EPA specifically evaluated the
database with respect to mortality (see Appendix C, Section C.3.1). In brief, mortality was observed
following exposure to a range of doses in chronic-duration studies, in studies up to 6 months in
duration, and during gestation (Lish etal.. 1984: Levine etal.. 1983: Levine etal.. 1981a: Cholakis et
al.. 1980: von Oettingen et al.. 1949). In further analyzing the available evidence, mortality
occurred at lower doses in rats compared with mice and following gavage administration compared
with dietary administration. Additionally, greater mortality occurred with administration of RDX in
the form of relatively finer particle sizes, likely due to faster dissolution of RDX leading to higher
blood concentrations. Some investigators attributed the mortality to RDX-related cancer or
noncancer effects (e.g., kidney or nervous system effects); others identified no cause for the animal
deaths. Typically, evidence related to various hazards is presented and synthesized in distinct
organ- or system-specific sections. However, in this case, the assessment does not present
mortality in a hazard section by itself due to the likelihood that events leading to mortality fall
under other specific hazards. Mortality evidence is considered in discussions of the evidence for
organ/system-specific hazards where applicable.
1.2.1. Nervous System Effects
In humans, nervous system effects following RDX exposure have been observed in multiple
case reports, and the association between RDX exposure and neurobehavioral effects has been
examined in a single cross-sectional occupational epidemiology study. Information relevant to an
examination of the association between RDX exposure and nervous system effects also comes from
experimental animal studies involving chronic, subchronic, and gestational exposure to ingested
RDX. No developmental neurotoxicity studies were available and minimal information was
available to evaluate potential cognitive or behavioral effects associated with RDX exposure. A
summary of nervous system effects associated with RDX exposure is presented in Tables 1-2 and
1-3 and Figure 1-1. Experimental animal studies are ordered in the evidence table and
exposure-response array by duration of exposure and then species.
1-4
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-2. Evidence pertaining to nervous system effects in humans
Reference and study design
Results
Ma and Li (1993) (China)
Cross-sectional study, 60 workers from
the same plant exposed to RDX (30 in
Group A [26 males; 4 females]; 30 in
Group B [24 males; 6 females]),
compared to 32 workers with similar
age, education level, and length of
employment from same plant with no
exposure to RDX (27 males; 5 females).
Exposure measures: Details of
exposure measurement were not
provided; two groups of workers
exposed to the following mean RDX
concentrations in air (basis for dividing
workers into two exposure groups was
not provided).
Concentration (mg/m3)
(mean ± standard deviation):
Group A: 0.407 (±0.332)
Group B: 0.672 (±0.556)
Effect measures:3 Five neurobehavioral
function tests and five additional
memory subtests.
Analysis: Variance (F-test); unadjusted
linear regression, multiple regression,
and correlation analysis.
Neurobehavioral function tests, scaled scores (mean ± standard
deviation)
Test
Control
Group A
Group B
Memory retention*
111.3 ±9.3
96.9 ±9.6
91.1 ± 10.3
Simple reaction time
(milliseconds)
493 ±199
539 ±183
578 ± 280
Choice reaction time
(milliseconds)
763 ±180
775±161
770 ±193
Block design*
(elapsed time)
18.0 ±5.4
16.0 ±4.3
13.5 ±6.7
Letter cancellation
(quality per unit time)
1,487 ± 343
1,449 ± 331
1,484 ± 443
*p < 0.01 (overall F-test); no statistically significant differences between
Group A and Group B.
Lower score indicates worse performance.
Memory retention subtests, scaled scores (mean ± standard deviation)
Subtest
Control
Group A
Group B
Directional memory*
23.5 ±3.6
17.2 ±4.9
18.1 ± 5.7
Associative learning*
24.9 ±5.1
20.0 ±4.3
18.5 ±4.6
Image free recall*
24.1 ± 3.8
20.9 ±4.1
20.4 ± 3.3
Recognition of
nonsense pictures*
26.3 ±3.6
23.2 ±4.9
21.6 ±4.3
Associative recall of
portrait characteristics*
26.3 ±3.3
20.3 ± 4.4
18.5 ±4.3
*p < 0.01 (overall F-test); no statistically significant differences between
Group A and Group B.
Lower score indicates worse performance.
Total behavioral score negatively correlated with exposure index (high
exposure correlated with poor performance).
aSymptom data were not included in evidence table because of incomplete reporting.
1-5
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-3. Evidence pertaining to nervous system effects in animals
Reference and study design
Results
Convulsions and neurobehavioral effects
Lish et al. (1984)
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 wk 11
due to excessive mortality)
Diet
2 yr
One male in the 35 mg/kg-day dose group and one female in the
175/100 mg/kg-day group convulsed near the end of the study.
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
No nervous system effects, as evidenced by clinical signs or changes
in appearance or behavior, were reported.
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
Tremors, convulsions, and hyper-responsiveness to stimuli were
noted in males and females at 40 mg/kg-day; no incidence data were
reported.
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
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
Hyperactivity and/or nervousness observed in 50% of the high-dose
males; no signs observed in females;b no incidence data were
reported.
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 nervous system effects, as evidenced by clinical signs or changes
in appearance or behavior, were reported.
1-6
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-3. Evidence pertaining to nervous system effects in animals
(continued)
Reference and study design
Results
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 nervous system effects were reported.
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
8b
10
12
15
Convulsions (incidence)
M
0/10
0/10
1/10
3/10
8/10
7/10
F
0/10
0/10
2/10
3/10
5/10
5/10
Tremors (incidence)
M
0/10
0/10
0/10
0/10
2/10
3/10
F
0/10
0/10
0/10
0/10
0/10
1/10
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
Hyper-reactivity to approach was observed in rats (sex not specified)
receiving >100 mg/kg-d; no incidence data were reported.
Tremors and convulsions were observed prior to death in one female
and two male rats receiving 600 mg/kg-dd (600 mg/kg-d was lethal to
all rats).
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
Hyperirritability and convulsions were observed in the 25 and
50 mg/kg-d groups;b no incidence data were reported.
1-7
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-3. Evidence pertaining to nervous system 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 dog food; purity
and particle size not specified
0, 0.1,1, or 10 mg/kg-d
Diet
13 wk
No nervous system effects, as evidenced by clinical signs or changes
in appearance or behavior, were reported.
Martin and Hart (1974)
Monkeys, cynomolgus or rhesus,e
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"
CNS effects characterized as depression, trembling, shaking, jerking,
or convulsions (incidence)
M
0/3
0/3
0/3
2/3
F
0/3
0/3
0/3
3/3
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
size not specified
0 or 50 mg/kg-d
Diet
6 d/wk for 6 wk
Treated dogs exhibited convulsions, excitability, ataxia, and
hyperactive reflexes;b no incidence data were reported.
MacPhail etal. (1985)
Rats, Sprague-Dawley derived CD,
8-10 males or females/group
Purity 84 ± 4.7%; <66 nm particle size
0,1, 3, or 10 mg/kg-d
Gavage
30 d
No changes in motor activity, flavor aversion, schedule-controlled
response, or acoustic startle response were reported.
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
Convulsions (incidence)
F
0/24
0/24
1/24
18/25
1-8
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-3. Evidence pertaining to nervous system effects in animals
(continued)
Reference and study design
Results
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
Convulsions preceding death were observed at >40 mg/kg-d;
no incidence data were reported.
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
Convulsions and hyperactivityb were observed at 20 mg/kg-d;
no incidence data were reported.
Burdette et al. (1988)
Rats, Long-Evans, 10-21 males/group
Exp 1: 0,10, 20, or 60 mg/kg-d
Exp 2: 0,12.5, 25, or 50 mg/kg-d
Experiments 1 and 2 conducted using the
same study design, each with a control
group
Gavage (single exposure)
8-h after exposure, rats placed in
observation chamber; 0-64 kHz, 95 dB
ultrasonic cleaner turned on for 1 min or
until seizure initiated with uncontrolled
running (whichever occurred first)
Doses
0
10
12.5
20
25
50
60
Number of spontaneous seizures during 8-h interval between dosing
and audiogenic seizure testing (mean)
M
0
0.17 ±0.2
1.4 ±0.2*
4.5 ±0.6*
—
Note: first seizures in all 3 treatment groups observed within first 2 h
after RDX exposure.
Prevalence of audiogenic seizures (incidence)+
M
0/31 1/10 0/10 3/10
4/10
10/12* 13/16*
fValues estimated from graph using Grab It! Software and numbers of
animals from Figure 2 of the paper. Statistical significance indicated
by study authors; spontaneous seizures—p < 0.012; audiogenic
seizures—p < 0.017.
Brain weight
Lish et al. (1984)
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 wk 11
due to excessive mortality)
Diet
2 yr
Doses
0
1.5
7
35
175/100
Absolute brain weight (percent change compared to control)
M
0%
-0.2%
0.61%
0.81%
-1%
F
0%
-2%
-2%
-4%*
-3%*
Relative brain weight (percent change compared to control)
M
0%
4%
2%
2%
5%
F
0%
-4%
-1%
-3%
18%*
1-9
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-3. Evidence pertaining to nervous system 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
40
Absolute brain weight (percent change compared to control)
M
0%
2%
-1%
2%
2%
F
0%
-0.3%
-0.4%
1%
2%*
Relative brain weight (percent change compared to control)
M
0%
0%
8%
2%
22%*
F
0%
-1%
3%
4%
20%*
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 brain weight (percent change compared to control)
M
0%
-
-
-
2%
2%
F
0%
-
-
-
4%
2%
Relative brain weight (percent change compared to control)
M
0%
-
-
-
6%
2%
F
0%
-
-
-
0%
3%
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
Absolute brain weight (percent change compared to control)
M
0%
0%
2%
10%
F
0%
0%
4%
2%
Relative brain weight (percent change compared to control)
M
0%
-3%
1%
8%
F
0%
0%
3%
-4%
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
Absolute brain weight (percent change compared to control)
M
0%
-
-
-
3%
0%
F
0%
-
-
-
0%
0%
Relative brain weight (percent change compared to control)
M
0%
-
-
-
7%*
10%*
F
0%
-
-
-
5%
6%
1-10
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-3. Evidence pertaining to nervous system 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
Doses
0
4
8
10
12
15
Absolute brain weight (percent change compared to control)
M
0%
-1%
-0.3%
2%
5%*
7%*
F
0%
-2%
6%
1%
4%
6%
Relative brain weight (percent change compared to control)
M
0%
6%
10%
5%
3%
4%
F
0%
-2%
-2%
-12%*
-12%*
-15%*
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
Absolute brain weight (percent change compared to control)
M
0%
1%
0.53%
-6%
-
-
F
0%
-1%
1%
2%
-
-
Relative brain weight (percent change compared to control)
M
0%
4%
7%
14%
-
-
F
0%
0.3%
2%
5%
-
-
CNS = central nervous system; F = female; GD = gestational day;
HMX = octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine; M = male; TWA = time-weighted average.
Note: A dash indicates that the study authors did not measure or report a value for that dose group.
^Statistically significant (p < 0.05) based on analysis by study authors.
aDoses were calculated by the study authors.
bMortality was reported in some RDX-treated groups in this study.
cLevine et al. (1981a) is a laboratory report of a 13-wk 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.
dDiscrepancies in the doses at which convulsions occurred were identified in the technical report. The nervous
system effects reported in this table and in the corresponding exposure-response array are those provided in
the results section of the technical report (Levine et al., 1981a) and in the published paper (Levine et al., 1990).
In other sections of the technical report, the study authors reported that hyperactivity to approach and
convulsions were observed in rats receiving >30 mg/kg-day (abstract and executive summary), or that mortality
was observed in rats receiving 100 mg/kg-day and that hyperactivity to approach, tremors, and convulsions
were observed in animals exposed to lethal doses (discussion).
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).
1-11
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
1000 q
100 -
• signifies ntly changed
O not significantly changed
-3 10
tu)
E
o
Q
1 :
6 6 o o
0.1
JU U
ra —
Chronic
15 «
"a —
(Ll —
Subchronic
Convulsions and/or Seizures
Gestational (dams)
— ID
"qj Si
Absolute Brain Weight
Figure 1-1. Exposure response array of nervous system effects following oral exposure.3
aBecause convulsions and seizures are rare in experimental animals, any occurrence in an RDX-exposed group was considered treatment-related. Given the
severity of this endpoint, a response in treated groups was determined to be significant (filled circles) in the array where there was any occurrence of
convulsions and/or seizures reported in the study, whether or not the incidence was statistically significantly elevated over the control.
1-12
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Observational Studies in Humans
In a cross-sectional study by Ma and Li (19931. neurobehavioral effects were evaluated in
Chinese workers occupationally exposed to RDX. Memory retention and block design scores10 were
significantly lower among exposed workers (mean RDX air concentrations in two exposed groups:
0.407 and 0.672 mg/m3) compared to unexposed workers from the same plant. However, no
significant differences were observed between the groups on other neurobehavioral tests (e.g.,
simple and choice reaction times, and letter cancellation test; see Table 1-2). This study did not
consider potential confounders such as alcohol consumption or coexposure to trinitrotoluene
(TNT), and there was limited information characterizing exposure to RDX.
Case reports suggest an association between RDX exposure (via ingestion, inhalation, and
possibly dermal exposure) and neurological effects (see Appendix C, Section C.2). Severe
neurological disturbances include tonic-clonic seizures (formerly known as grand mal seizures) in
factory workers (Testud et al.. 1996a: Testud et al.. 1996b: Kaplan etal.. 1965: Barsotti and Crotti.
19491: seizures and convulsions in exposed soldiers serving in Vietnam fKetel and Hughes. 1972:
Knepshield and Stone. 1972: Hollander and Colbach. 1969: Stone etal.. 1969: Merrill. 1968):
seizures, dizziness, headache, and nausea following nonwartime/nonoccupational exposures
(Kasuske etal.. 2009: Davies etal.. 2007: Kuctikardali etal.. 2003: HettandFichtner. 2002: Harrell-
Bruder and Hutchins. 1995: Goldberg et al.. 19921: and seizures in a child following ingestion of
plasticized RDX from the mother's clothing (Woody et al.. 19861.
Studies in Experimental Animals
Nervous system effects in experimental animals include an array of behavioral changes
consistent with the induction of seizures by RDX exposure, and have been observed in the majority
of chronic, subchronic, and developmental studies examining oral exposure to RDX (see Table 1-3
and Figure 1-1). Although study authors interchangeably used the terms seizures and convulsions,
seizures, which result from abnormal electrical activity in the brain, can outwardly manifest in a
variety of ways, including as convulsions. However, seizures can also manifest as facial twitches or
tremors, or more subtly as increased irritability or aggression, absence of response to external
stimuli, or they may go unnoticed. While behavioral methods exist to capture a spectrum of
responses known to occur as a result of this aberrant neuronal activity, the most reliable detection
methods are electrophysiological (Racine. 19721. Only one acute exposure study, testing a single,
high dose of RDX, included electrophysiologic recordings (Williams etal.. 20111.
10The memory quotient index measured short-term hearing memory, visual memory, combined hearing and
visual memory, and learning ability. The block design index measured visual perception and design
replication, and the ability to analyze spatial relationships.
1-13
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Convulsions (a sudden and irregular movement of a limb or of the body) have been
reported in studies with different animal species and experimental designs. In every study that
reported convulsions, the incidence of convulsions increased with dose. In 2-year dietary studies in
rats (F344 and Sprague-Dawley) and mice (B6C3Fi), convulsions were observed beginning at doses
of 35-40 mg/kg-day, but not at lower doses (Lish etal.. 1984: Levine etal.. 1983: Hart. 19761.11
Subchronic dietary exposure to RDX was also associated with convulsions in the rat, although doses
reported to increase convulsive activity were inconsistent across studies. Convulsions were
reported in RDX-exposed rats at subchronic doses as low as 8 and 25 mg/kg-day (Crouse etal..
2006: von Oettingen et al.. 19491. In three other studies of nonpregnant, adult rats involving
exposure durations of 30-90 days, no evidence of seizures, convulsions, or tremors was reported at
doses ranging from 1 to 50 mg/kg-day [(MacPhail etal.. 1985: Cholakis etal.. 19801: both
unpublished technical reports].12 Levine etal. f 19901 reported convulsions in rats following
subchronic exposure only at a dose of 600 mg/kg-day (a dose associated with 100% mortality);
however, the unpublished technical report of this study (Levine etal.. 1981a) reported convulsions
at 600 and >30 mg/kg-day, thereby reducing confidence in the identification of the dose level at
which nervous system effects were observed in this study. RDX exposure (by gavage) during
gestation in the rat was associated with induction of seizures or convulsions in the dams at doses
ranging from 2 to 40 mg/kg-day [(Angerhofer et al.. 1986: Cholakis etal.. 1980): unpublished
technical reports], demonstrating that effects on the nervous system can be observed following
exposure durations as short as 10-14 days. Convulsions were also reported in dogs exposed to
50 mg/kg-day RDX for 6 weeks (von Oettingen et al.. 1949). but not 10 mg/kg-day for 13 weeks
[(Hart. 1974): unpublished technical report]; however, five of six monkeys exhibited convulsions
following a gavage dose of 10 mg/kg-day for 13 weeks [fMartin and Hart. 19741: unpublished
technical report]. Linkage of these convulsions to seizure activity was most directly demonstrated
by Williams etal. (2011). who observed abnormal electroencephalogram (EEG) activity consistent
with seizure activity that coincided with physical manifestations ranging from subtle convulsive
behaviors (e.g., twitches) to tonic-clonic seizures in rats acutely exposed to 75 mg/kg-day RDX via
gavage.
"The 2-year dietary studies in F344 rats by Levine etal. (19831 and B6C3Fi mice by Lish et al. (19841 were
available only as laboratory reports. An external peer review was sought by EPA in July 2012 to evaluate the
accuracy of experimental procedures, results, and interpretation and discussion of the findings presented. A
report of this peer review organized by Versar, Inc. is available on the Health and Environmental Research
Online (HERO) database (U.S. EPA. 2012c). The 2-year dietary study in Sprague-Dawley rats by Hart (19761
is available as an unpublished technical report.
12The series of nine toxicology studies reported in Cholakis et al. (19801 were available only as laboratory
reports. An external peer review was sought by EPA in 2017 to evaluate the accuracy of experimental
procedures, results, and interpretation and discussion of the findings presented in six of the nine studies
(90-day toxicity study in rats, initial 90-day toxicity study in mice, supplemental 90-day toxicity study in mice,
teratology study in rats, teratology study in rabbits, and two-generation reproductive toxicity study in rats).
A report of this peer review organized by Eastern Research Group, Inc. is available on the HERO database
(U.S. EPA. 20171.
1-14
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
In the only study addressing susceptibility to seizures (chemicals that may alter seizure
frequency, severity, duration, or threshold), Burdette etal. (1988) found that seizure occurrence
was more frequent in Long-Evans rats exposed to a single dose of 50 or 60 mg/kg RDX by gavage
when challenged with an audiogenic stimulus 8 and 16 hours after treatment [Note: some
uncertainty exists regarding the administered dose as neither the purity nor the specific particle
size of the RDX used in the experiments by Burdette etal. (1988) was reported]. No audiogenic
seizures were observed at the earlier 2- and 4-hour postdosing test periods even though RDX
plasma concentrations were elevated throughout the testing period, which could suggest that the
blockade of gamma-aminobutyric acid (GABA)ergic signaling by RDX (see Mechanistic Evidence
section) needs to be sustained for some minimal duration to induce these types of effects. In a
complementary experiment, Long-Evans rats treated daily with 6 mg/kg-day RDX for up to 18 days
required fewer stimulation trials compared to animals not treated with RDX to exhibit amygdaloid
kindled seizures compared to controls. These findings provide evidence that RDX exposure can
reduce the seizure threshold for other proconvulsant stimuli, an adverse effect (U.S. EPA. 1998).
Most animal studies reported convulsions and/or seizures as clinical observations; thus,
interpretation of these observations is limited because the nature and severity of convulsions and
seizures were not more fully characterized. The 90-day study by Crouse etal. f2006113 was one of
the few studies that collected and reported incidence data for convulsions and tremors, and
demonstrated a clear dose-related increase in convulsions and tremors in male and female F344
rats associated with RDX exposure via gavage (see Table 1-3). Tremors were reported following
administration of >12 mg/kg-day, persisting throughout the 90-day study. Convulsions were
observed at >8 mg/kg-day in male and female rats; information on convulsion duration and onset
after the start of dosing was not reported fCrouse et al.. 20061.
In general, gavage dosing induced convulsions at lower doses than did dietary
administration. For example, in the subchronic gavage study by Crouse etal. (2006) and the
developmental gavage study by Cholakis etal. (1980). convulsions were observed in 1-3 F344
rats/group at doses of 2-8 mg/kg-day; at doses of 15-20 mg/kg-day, convulsions were observed in
approximately 60-70% of the animals. Consistent with this pattern, even an acute (single dosing)
gavage study reported seizures in 2/10 rats shortly after exposure to 12.5 mg/kg-day, and
approximately 80% of rats developed spontaneous seizures shortly after exposure to
25-50 mg/kg-day (Burdette etal.. 1988): the longevity of the seizure behaviors was also highly
dose-dependent In contrast, in a 2-year dietary study by Levine etal. (1983). convulsions were
reported only at a dose of 40 mg/kg-day; no convulsions were observed at lower doses
(<8 mg/kg-day). The difference in response between gavage and dietary administration may be
13The 13-week gavage study in F344 rats by Crouse etal. (2006) was available only as a laboratory report. An
external peer review was organized by Versar, Inc. in July 2012 to evaluate the accuracy of experimental
procedures, results, and interpretation and discussion of the findings presented. The U.S. EPA (2012c) report
of this peer review is available on the HERO database.
1-15
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
due to the bolus dosing resulting from gavage administration and the comparatively faster
absorption and higher peak blood concentrations of RDX.
Several experimental animal studies documented that unscheduled deaths were frequently
preceded by convulsions or seizures. In a 2-year study in rats, Levine etal. (19831 noted that
tremors and/or convulsions were often seen in high-dose animals prior to their death. In a rat
developmental toxicity study (Cholakis etal.. 19801. investigators concluded that early deaths in
dams were preceded by convulsions based on the observation of convulsions in one rat prior to
death, and a similar appearance (e.g., dried blood around the mouth and nose) in other dams that
died during the study. Convulsions preceding death were also observed in pregnant
Sprague-Dawley rats exposed to RDX during gestation (Angerhofer etal.. 19861. Burdette et al.
(19881 reported that 9/28 rats died during spontaneous seizure within 8 hours of administration
(by gavage) of a single dose of 50 or 60 mg/kg RDX.
The 90-day Crouse etal. (2006) study provides the most detailed information on the
relationship between convulsions and mortality (see Appendix C, Table C-10 for additional
information on evidence of mortality associated with RDX exposure). Convulsions (3/20) and
preterm deaths (2/2 0)14 were observed in male and female rats exposed to 8 mg/kg-day RDX; the
incidences of both convulsions and preterm deaths were higher in dose groups with greater
exposures. Investigators stated that nearly all observed preterm deaths in rats exposed to the three
higher doses (10,12, and 15 mg/kg-day RDX) for 90 days were preceded by neurotoxic signs such
as rearing behavior, tremors, and convulsions; however, preterm death did not occur in all animals
that convulsed. Convulsions were not typically observed during a functional observational battery
(FOB) test conducted after exposure, possibly due to the time needed to complete exposures before
beginning behavioral testing (convulsions typically occurred shortly after dosing). Of the
100 RDX-treated rats in the Crouse etal. (2006) study, convulsions were documented in 34 male
and female rats across the five dose groups (with convulsions initially observed anywhere from day
7 to 87); based on additional information provided as a memorandum by study investigators
(lohnson. 2015a). 26 of these 34 rats (76%) survived to the end of the 90-day study. In general,
higher doses of RDX were associated with fewer days of exposure before the first convulsion was
observed. Of the eight rats that exhibited convulsions before preterm death, convulsions were
documented anywhere from the same day that the animal died to 8 weeks before death. Of the
26 rats that seized and survived to Day 90, the first seizures were observed as early as Day 10 and
as late as Day 87. Thus, while an increase in mortality was observed in the Crouse etal. (2006)
study at the same dose as convulsions, the additional information provided by lohnson (2015a)
does not show as clear a correspondence between convulsions (and other neurotoxic signs) and
mortality. Analysis of these data is limited to the extent that convulsions may have occurred at
times when animals were not observed and therefore may be undercounted in the individual
14At the 8 mg/kg-day dose level, the three rats that convulsed survived to the end of the study; no convulsions
were observed in the two rats that died before study termination.
1-16
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
animal data; however, Tohnson f2015al noted that it is unlikely that seizure observations were
missed because seizures generally occurred soon after dosing.
A few studies reported mortality that was not specifically associated with neurological
effects [see Appendix C, Table C-10; (Angerhofer et al.. 1986: Levine etal.. 1981a: von Oettingen et
al.. 19491]: however, in these studies, animals may not have been monitored for clinical
observations or monitored with sufficient frequency to have observed convulsive activity before
death. There were no reports of mortality subsequent to convulsions in case reports of nervous
system effects in workers exposed to RDX during manufacture and in individuals exposed acutely
as a result of accidental or intentional ingestion (see Appendix C, Section C.2).
Additional neurobehavioral effects associated with RDX exposure in rats included increased
hyperactivity, hyper-reactivity to approach, fighting, and irritability at doses similar to those that
induced tremors, convulsions, and seizures [20-100 mg/kg-day; fLevine etal.. 1990: Angerhofer et
al.. 1986: Levine etal.. 1983: Levine etal.. 1981a. b; von Oettingen et al.. 1949)]. Hyperactivity and
nervousness were also reported in male mice that received a subchronic exposure to
320 mg/kg-day RDX (Cholakis etal.. 19801. No changes in motor activity, flavor aversion,
scheduled-controlled behavior, or acoustic startle response were observed in a 30-day gavage
study in rats at relatively low doses (<10 mg/kg-day), although changes in acoustic startle response
in acute exposures at higher doses (12.5-50 mg/kg) were noted (MacPhail etal.. 1985). No
significant changes in behavioral or neuromuscular activity were observed in rats following
exposure to <15 mg/kg-day for 90 days (Crouse et al.. 2006). Crouse etal. (2006) observed that
stained haircoats and increased barbering in female F344 rats receiving 15 mg/kg-day may have
been caused by the gavage dosing procedure alone.
Changes in absolute and relative brain weight were mixed across studies, and no studies
included histopathologic evaluation of neuronal damage. Elevated absolute brain weights were
reported in subchronic assays in B6C3Fi mice and F344 rats (Crouse etal.. 2006: Levine etal..
1990: Levine etal.. 1981a. b; Cholakis etal.. 1980): however, the changes were not consistently
observed across studies. Relative brain weights in some studies showed correspondingly greater
increases compared to absolute brain weight (Crouse etal.. 2006: Levine etal.. 1983: Cholakis etal..
1980). but these changes were likely due to changes in body weight in the study, and were not as
useful a measure of effects of RDX on brain weights as absolute brain weight. In 2-year oral studies,
a decrease in absolute brain weight of female B6C3Fi mice (3-4% relative to control) was reported
at doses >35 mg/kg-day (Lish etal.. 1984). whereas an increase in absolute brain weight (2%
relative to control) was observed in F344 rats at a dose of 40 mg/kg-day (Levine etal.. 1983). Less
emphasis is placed on evidence of organ-weight changes from chronic (2-year) studies because
normal physiological changes associated with aging and intercurrent disease may contribute to
interanimal variability that could confound organ-weight interpretation (Sellers etal.. 2007).
In some studies, seizures appeared soon after dosing, suggesting that seizure induction was
more strongly correlated with dose level than with duration of exposure. This observation is
1-17
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
consistent with the findings of Williams etal. f20111. who demonstrated that RDX is rapidly
absorbed and crosses the blood-brain barrier following oral administration in rats, and that
distribution of RDX (8 ng/g wet weight) to the brain correlated with seizure onset However, the
incomplete or slow reversibility of the blockade of GABA receptor signaling after removal of RDX in
the in vitro study by Williams etal. (20111 suggests that some effects might persist without the
continued presence of RDX in the brain, which could permit cumulative effects.
While a dose-response relationship was observed consistently within studies, a dose that
induced convulsions in animals in one study did not necessarily induce convulsions at the same
dose in another study. This lack of consistency may be attributed, at least in part, to differences in
the purity or particle size of the test material across studies. Assuming that increased particle size
(and the corresponding reduction in available surface area compared with smaller particle sizes)
results in slowed absorption and distribution to the brain, studies that used a larger particle size
may be expected to produce less neurotoxicity in test animals. The mouse study by Cholakis et al.
(1980) used a relatively large RDX particle size (200 |im) compared to the rat study by Levine et al.
(1983) that used a smaller (<66 |im) particle size. This may explain why the Cholakis etal. (1980)
subchronic dietary study in the mouse (doses up to 320 mg/kg-day RDX) and rat (doses up to
40 mg/kg-day) failed to report seizures or convulsions. Finally, differences in study design may
have contributed to differences in reported neurological responses in subchronic- and
chronic-duration studies. In particular, the observation protocols for clinical signs [e.g.,
observations performed once daily in the morning in Levine etal. (1983)] may not have been
sufficiently frequent to accurately measure the incidence of seizures or other nervous system
effects.
The lack of developmental neurotoxicity studies was identified as a data gap within the
available studies on RDX. A pilot study in rats did not directly investigate potential RDX nervous
system effects but did find RDX in the brains of offspring rats as well as milk from dams treated
with RDX during gestation (Hess-Ruth etal.. 2007). Studies on chemicals with similar modes of
action to RDX (e.g., bicuculline), combined with demonstrated transfer of RDX to perinatal rodent
brains, suggest a potential for RDX to be harmful during brain development, possibly at a lower
dose than required for neurotoxicity in adults. Further discussion on the potential developmental
neurotoxicity of RDX can be found in Susceptible Populations and Life Stages for Cancer and
Noncancer Outcomes (see Section 1.3.3).
Mechanistic Evidence
Studies that have explored the mode of action (MOA) of RDX on the central nervous system
(CNS) have focused on the potential impacts on neurotransmission. These studies implicate an
MOA for RDX-induced seizures involving distribution of RDX from the blood to the brain (across the
blood-brain barrier) and subsequent effects on neurotransmission, specifically GABA-mediated
signaling in limbic regions of the brain. There is significant evidence from the scientific literature to
suggest that RDX neurotoxicity results from interactions of RDX with the GABAa receptor. GABA is
1-18
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
a major inhibitory neurotransmitter in the brain, and the GABAa receptor has been implicated in
susceptibility to seizures (Galanopoulou. 2008). A large literature base exists to support the
relationship between blockade of GABAAergic neurotransmission and seizure induction, and
GABAAergic pharmaceuticals are routinely used to suppress seizures in the treatment of epilepsy
and other disorders (perhaps most recognizably, drugs in the benzodiazepine family).
In research conducted by the U.S. Army Center for Health Promotion and Preventative
Medicine, Williams etal. (2011) and Bannon etal. (2009a) showed a correlation between blood and
brain concentrations of RDX in rats that received a single oral dose of RDX (>98-99.5% purity) by
gavage, which closely correlated with the time of seizure onset. RDX (75 mg/kg) was distributed to
the brain in direct proportion to levels found in the blood, while time to seizure onset was reduced
as RDX brain levels increased (Williams etal.. 2011). Similarly, oral exposure to RDX (via a gel
capsule: 3 or 18 mg/kg) resulted in quick absorption followed by transport to the brain and
subsequent alterations in neurotransmission (Bannon et al.. 2009a).
In receptor binding studies, RDX showed significant affinity for GABAa receptors (Williams
etal.. 2011: Williams and Bannon. 2009). Specifically, RDX showed an affinity for the picrotoxin
convulsant site of the GABA channel, with nearly 100-fold less potency than picrotoxin itself.
Consistent with the observations of abnormal electrical activity after in vivo RDX exposure (see
discussion in previous section), in vitro RDX treatment of brain slices from the basolateral
amygdala inhibited GABAa-mediated inhibitory postsynaptic currents and initiated seizure-like
electrical activity. Thus, RDX exposure appears to reduce the inhibitory effects of GABAergic
neurons, resulting in a loss of inhibitory tone and enhanced excitability that can eventually lead to
seizures (Williams etal.. 2011: Williams and Bannon. 2009). The limbic system, and the amygdala
and hippocampus in particular, are known to be critical to the development of seizures in various
human conditions (e.g., epilepsy) and animal models (Teffervs et al.. 2012: Gilbert. 1994).
Consistent with the in vitro observations by Williams etal. (2011). Burdette etal. (1988) also
implicated the limbic system in seizures caused by RDX exposure. Burdette etal. (1988) reported
that the pattern of the seizure behaviors manifest in response to RDX exposure mimicked the
sequence of behavioral stages observed following repeated electrical stimulation of temporal lobe
structures by Racine (1972). In addition, amygdaloid kindled rats (rats subjected to patterns of
electrical stimulation to this limbic region, which promotes the development of seizures) exhibited
proconvulsant activity at a dose that was approximately half of the dose necessary for RDX to
induce spontaneous seizures [rats treated with RDX also required fewer electrical stimulations to
trigger kindled seizures (Burdette etal.. 1988)]. These latter findings occurred at lower doses than
RDX-induced increases in audiogenic seizures (Burdette etal.. 1988). further suggesting a primaiy
role for the limbic regions (brain structures involved in sound-induced seizures may be indirectly
affected). Potential limbic system involvement is also suggested given its role in integrating
emotional and behavioral responses (including aggression) and the anecdotal observations of
hyperactivity, hyper-responsiveness to approach, and irritability noted across several studies of
1-19
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
RDX toxicity fLevine etal.. 1990: Levine etal.. 1983: Levine etal.. 1981a. b; Cholakis etal.. 1980: von
Oettingen et al.. 1949).
It is possible to construct a hypothetical MOA for RDX-induced seizure activity based on the
evidence summarized above. These steps are consistent with ongoing efforts to identify an adverse
outcome pathway (AOP) for ionotropic GABA-receptor antagonism, reviewed in Gong etal. (20151
and Collier etal. (20161 and described in greater detail in the draft AOP available at
https: //aopwiki.org/. Following distribution of RDX to the brain:
1) Parent RDX acts as a GABAa-receptor antagonist [supported by Schneider et al. (19771 and
Williams etal. (20111]. binding noncompetitively to the picrotoxin convulsant site of the
GABAa receptor [supported by Williams and Bannon (20091 and Williams etal. (20111],
2) RDX binding to the GABAa receptor at the picrotoxin site blocks the conduction of chloride
through the ion channel.
3) Reduced chloride conduction results in reduced GABA-mediated inhibition of neuronal
signaling, often manifesting as a reduction in spontaneous inhibitory postsynaptic currents
(sIPSCs). Williams etal. (20111 observed a reduction in the amplitude and frequency of
sIPSCs in whole-cell in vitro recordings of neurons in brain slices from the rat basolateral
amygdala after exposure to RDX. In addition, RDX treatment of slices inhibited
GABA-induced currents.
4) Reduced inhibitory tone (e.g., reduced sIPSCs) increases the likelihood of action potentials
by decreasing the resting potential of neuronal membranes (depolarization).
5) As a group of neurons begins firing abnormally and excessively (e.g., due to the reduced
inhibitory tone, which typically would hyperpolarize, or reset, the membrane after firing),
they can begin firing in a synchronized manner and initiate a wave of depolarization; these
events can be detected electrophysiologically. Williams etal. (2011) observed a pattern of
seizure-like neuronal discharges after in vivo RDX exposure and in vitro from slices of the
basolateral amygdala in rats after adding RDX (the in vitro effects were not reversible after
40 minutes of washout).
The steps above provide a biologically plausible sequence of mechanistic events that result
in the generation of seizure-like neuronal activity. Reduction of the inhibitory GABAergic signaling
is common to many convulsants, as summarized in Kalueff (2007). Some organochlorine
insecticides, including alpha-endosulfan, dieldrin, and lindane, also exert neurotoxic effects through
interaction with the GABAa receptor, and can produce a range of hyperexcitability effects (including
convulsions) in mammals (Vale etal.. 2003: Bloomquist. 1992: Sunol etal.. 1989). The interaction
of RDX with the GABAa receptor is directly supported by receptor-binding assays (Williams etal..
2011). Although these binding assays were performed on rat receptors, it is plausible that the
results are relevant to human neurotoxicity. Seizures have been observed in many species,
including humans, rats, mice, dogs, lizards, and birds at varying dosages and durations of exposure
1-20
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
fOuinn etal.. 2013: McFarland etal.. 2009: Tohnson et al.. 2007: Bruchim etal.. 2005: Kiiciikardali et
al.. 2003: Woody etal.. 1986: Lish etal.. 1984: Berry et al.. 1983: Levine etal.. 1983). A more recent
meta-analysis of toxicogenomic data across a phylogenetically diverse set of organisms (rat, quail,
fathead minnow, earthworm, and coral) demonstrated that neurotoxic responses are conserved in
more highly related species and that binding to the GABAa receptor is a common molecular
initiating event (Garcia-Revero etal.. 2011). While these lines of evidence do not preclude a role of
other receptors unscreened for RDX binding affinity, they support a primary role for the GABAergic
pathway described above in the development of RDX neurotoxicity.
As mentioned previously, the GABAa receptor is also a target of many anticonvulsant
therapies [e.g., benzodiazepines, propofol, barbiturates; (Meldrum and Rogawski. 2007: Mohler.
2006)]. Additional support for the involvement of GABAergic signaling in the neurotoxicity of RDX
comes from human case reports. In multiple case reports, medical intervention included treatment
with benzodiazepines (commonly diazepam or lorazepam) to treat seizing patients (Kasuske etal..
2009: Davies etal.. 2007: Kiiciikardali et al.. 2003: Hett and Fichtner. 2002: Woody etal.. 1986).
Benzodiazepines act in large part by enhancing the effects of GABA at the GABAa receptor by
increasing chloride conductance, resulting in anticonvulsant and relaxant effects (Goodman etal..
1996).
Some other proconvulsant agents with minimal direct toxicity to nerve cells, such as sarin
and some organophosphate pesticides, are known to act through inhibition of acetylcholinesterase
(AChE) activity (McDonough and Shih. 1997). Some of the clinical signs observed following RDX
exposure are similar to the clinical signs associated with organophosphate pesticides and nerve
agents (Crouse etal.. 2006: Burdette etal.. 1988: Barsotti and Crotti. 1949). However, the limited
data available for RDX do not support AChE inhibition as a contributing mechanism because
(1) blood and brain levels of AChE are unaffected by RDX (Williams etal.. 2011: Williams and
Bannon. 2009) and (2) in vitro neurotransmitter receptor binding studies do not reveal any affinity
of RDX for acetylcholine receptors (Williams etal.. 2011: Williams and Bannon. 2009). Additionally,
common AChE-induced symptoms (salivation and lacrimation) have not routinely been observed
(Williams etal.. 2011). RDX showed no affinity for other receptors that are known targets of
convulsants, including the glutamate family of receptors, nicotinic receptors, glycine receptors, and
several monoamine receptors (Williams etal.. 2011: Williams and Bannon. 2009).
In a microarray experiment, Bannon et al. (2009a) found that RDX caused a down
regulation of an abundance of genes in the cerebral cortex related to neurotransmission, including
those encoding proteins involved in synaptic transmission and vesicle transport. Genes encoding
proteins involved in the glutamate pathway were also underexpressed, indicating a possible
mechanism of action for RDX via excessive glutamate stimulation. The study authors speculated
that this depression of the major excitatory neurotransmitter system could be a negative response
to the increase in seizure likelihood from RDX influx into the brain. Molecular changes in response
to RDX have been described by Zhang and Pan (2009b). who observed significant changes in
1-21
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
microRNA (miRNA) expression in the brains of B6C3Fi mice fed 5 mg RDX/kg in the diet [estimated
dose: 0.75-1.5 mg/kg-day; (Bannon etal.. 2009a)] for 28 days. One miRNA, miR-206, was
upregulated 26-fold in RDX-exposed brains; brain-derived neurotrophic factor (BDNF) was
identified as a downstream gene target of this miRNA, along with two other miRNAs that were
upregulated in RDX-exposed brains [miR-30a and miR-195; (Zhang and Pan. 2009a. b)]. BDNF is a
member of the neurotrophin family of growth factors, promoting the survival and differentiation of
existing and new neurons. Deng etal. (2014) conducted miRNA and mRNA profiling in rats to
identify targets up or downregulated after 48-hour exposure to RDX, finding that many of the gene
targets of these miRNAs were associated with nervous system function, and may contribute to the
neurotoxicity of RDX. However, while effects of RDX on BDNF expression or other downstream
targets may play a role in RDX neurotoxicity, the use of miRNAs as predictors of toxicity has not
been demonstrated, and downstream targets of miRNA require verification fBannon etal.. 2009bl.
Despite this uncertainty, the potential for RDX exposure to modulate the expression or function of
BDNF and other factors crucial to normal brain development raises concern for the possibility of
neurotoxic effects with developmental exposure. Overall, the contribution, if any, of aberrant
expression of a suite of miRNAs to the MOA for RDX neurotoxicity is unknown.
Some uncertainty remains regarding how the mechanistic understanding of RDX
neurotoxicity may inform longer-term or cumulative exposures. To some extent, RDX binding at
the picrotoxin convulsant site of the GABA channel may inform the relationship between exposure
to the chemical and the time when a seizure is observed. Many of the available studies reported
that seizures or convulsions were typically observed shortly after exposure, and several studies
associated seizures with blood (and, correspondingly, brain) levels of RDX, indicating that a major
contributing factor to the seizurogenic effects of RDX exposure appears to be the transient presence
of RDX at target sites in the brain. Observations by Crouse etal. (2006). clarified in Tohnson
(2015a). showed that the median time to seizure after dosing in F344 rats is 55 minutes (range of
20-85 minutes); peak brain concentrations of RDX in F344 rats after single oral doses occurred
within the first 3-4 hours after dosing (Bannon etal.. 2009a). In addition, seizure intensity and the
longevity of seizure-related behaviors was directly related to RDX dose, even with acute exposure
(Burdette et al.. 1988). These observations are all consistent with the presumed primary MOA. In
general, across the RDX database, neurotoxicity, including induction of convulsions and seizures,
appears to be more strongly correlated with dose than duration of exposure. Crouse etal. (2006)
reported that 80-90% of rats exposed to 12 and 15 mg/kg-day exhibited signs of neurotoxicity
beginning on Day 0 of the study and continuing for the study duration. However, some uncertainty
regarding the influence of exposure duration remains. Gerkin etal. (2010) demonstrated that
young C57/B16 mice injected intraperitoneally with picrotoxin to induce seizures had a significantly
increased frequency of elevated neuronal activity ("up state"), and firing rates were significantly
increased in neocortical neurons up to 24 hours after exposure, despite the rapid clearance (within
a few hours) of picrotoxin (Soto-Otero etal.. 1989). This extended period of elevated neuronal
1-22
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
activity might increase the likelihood that a subsequent stimulus could trigger a seizure. While the
study authors did not look at longer durations post exposure, these observations with picrotoxin
may be consistent with the lack of complete reversibility of GABAergic signaling inhibition after
removal of RDX in Williams etal. (20111. Thus, there remains the possibility that, in a chronic
exposure scenario with repeated exposure to RDX and binding at the same site as picrotoxin, a
generalized increase in elevated neuronal activity could increase the likelihood of seizures
developing over time, or have other longer-term effects on normal brain function. While duration
of exposure alone generally did not appear to be predictive of seizures [e.g., in Crouse etal. (2006).
of the rats that survived the 90-day study, the range of time to onset of first observed convulsion
after gavage exposure to 10 mg/kg-day RDX was as early as Day 7 and as late as Day 87], exposure
to higher doses of RDX was associated with fewer days of exposure before the first convulsion was
observed. The variation in time between the start of the experiment and the onset to first seizure
with increasing dose could either reflect the increased probability of action potentials with greater
decreases in inhibitory tone at higher doses or indicate a cumulative component of RDX
neurotoxicity not accounted for by currently available mechanistic understanding.
Recent research in experimental animals has provided greater insight to inform a
mechanistic basis of RDX neurotoxicity. While other possible MOA(s) may contribute to the overall
neurotoxicity of RDX, the demonstrated affinity of RDX for the GABAa receptor, the evidence of
supportive electrophysiological changes in vivo or with direct application of RDX, and the
toxicokinetic evidence of distribution of RDX to the brain provide a mechanistic basis for the
association of seizures with exposure to RDX. This MOA is like other well-studied convulsants and
relevant to humans. The available information supports that RDX-induced seizures and related
behavioral effects likely result from inhibition of GABAergic signaling within limbic regions of the
brain.
Integration of Nervous System Effects
Evidence for nervous system effects associated with exposure to RDX comes from studies in
both humans and animals. One occupational study reported memory impairment and decrements
in certain neurobehavioral tests in workers exposed to RDX compared to controls (Ma and Li.
19931. and human case reports provide other evidence of an association between acute RDX
exposure and neurological effects. There was consistent evidence of neurotoxicity associated with
exposure to RDX; 11 of 16 repeated-dose animal studies (of varying design) reported neurological
effects (some severe), including seizures, convulsions, tremors, hyperirritability, hyper-reactivity,
and behavioral changes associated with RDX exposure (Crouse etal.. 2006: Angerhofer et al.. 1986:
Levine etal.. 1983: Levine etal.. 1981b: Cholakis etal.. 1980: von Oettingen et al.. 1949). In most of
these studies, the occurrence of neurological effects was dose related. In those studies that found
no evidence of RDX-associated neurotoxicity fMacPhail etal.. 1985: Cholakis etal.. 1980: Hart.
1976. 1974). differences in dosing, particle size, and purity of the RDX administered potentially
account for the lack of effect. Seizures resulting from RDX exposure likely result from inhibition of
1-23
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
GABAergic signaling due to the interaction of RDX with the GABAa receptor. The proconvulsant
effects of RDX exposure are specific to CNS toxicity, as supported by observations of aberrant brain
electrical activity corresponding with physical seizure behaviors f Williams et al.. 20111. as well as
evidence of decreases in the seizure threshold for other centrally acting convulsants, including
amygdaloid kindling and audiogenic stimuli (Burdette etal.. 19881.
Together, toxicological information in animals and humans, supported by toxicokinetic and
mechanistic information, provides a coherent identification of nervous system effects as a human
hazard of RDX exposure.
1.2.2. Urinary System (Kidney and Bladder) Effects
The association between RDX exposure and effects on clinical measures of kidney function
was examined in one occupational epidemiology study. Case reports, involving accidental exposure
to ingested or inhaled RDX, offer some information on the potential for acute exposure to RDX to
affect the kidney in humans. Organ weight and histopathology findings from experimental animal
studies involving subchronic and chronic exposure to ingested RDX also provide data relevant to an
examination of the association between RDX exposure and urinary system (kidney and bladder)
effects. A summaiy of these effects associated with RDX exposure is presented in Tables 1-4 to 1-7
and Figure 1-2. Experimental animal studies are ordered in the evidence table and
exposure-response array by duration of exposure and then by species.
Human case reports of individuals accidently exposed to unknown amounts of RDX by
ingestion or inhalation provide some evidence that RDX affects the kidney. Reported symptoms
included decreased urine output fKetel and Hughes. 1972: Knepshield and Stone. 1972: Hollander
and Colbach. 1969: Merrill. 19681. blood in urine fKasuske etal.. 2009: Knepshield and Stone. 1972:
Hollander and Colbach. 1969: Merrill. 19681. proteinuria fKasuske etal.. 2009: Kuctikardali et al..
2003: Ketel and Hughes. 1972: Hollander and Colbach. 1969: Merrill. 19681. glucosuria
(Kuctikardali et al.. 2003). elevated blood urea nitrogen (BUN) levels (Hollander and Colbach. 1969:
Merrill. 1968). and one case of acute renal failure requiring hemodialysis following accidental
inhalation of RDX fKetel and Hughes. 1972). In many of these case reports, renal parameters
returned to normal within a few days following exposure. No changes in renal parameters were
reported in other individuals exposed to unknown amounts of RDX (Stone etal.. 1969: Kaplan etal..
19651. In a cross-sectional epidemiologic study of workers from five U.S. Army munitions plants
(69 exposed to RDX alone and 24 exposed to RDX and octahydro-l,3,5,7-tetranitro-
1,3,5,7-tetrazocine [HMX]; RDX exposure range: undetectable [<0.01 mg/m3] to 1.6 mg/m3), no
statistically significant differences in BUN or total serum protein between nonexposed and
RDX-exposed groups were observed [(Hathaway and Buck. 1977): see Table 1-4], As it is a
cross-sectional study, no information was provided on the length of employment or other proxies
that could be used to indicate exposure duration or cumulative exposure.
1-24
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-4. Evidence pertaining to kidney effects in humans
Reference and study design
Results
Hathawav and Buck (1977)
Cross-sectional study, 2,022 workers,
1,491 participated (74% response rate).
Analysis group: limited to whites;
69 workers exposed to RDX alone and
24 workers exposed to RDX and HMX,
compared to 338 workers 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:
undetected (0.01 mg/m3 (mean
for employees with exposures >LOD:
0.28 mg/m3).
Effect measures: Renal function tests
(blood)
Analysis: Types of statistical tests were
not reported (assumed to be t-tests for
comparison of means and x2 tests for
comparison of proportions).
Renal function tests: mean (standard deviation not reported)
Test
Referent
(n = 237)
RDX exposed males*
Undetected (0.01 mg/m3
[n = 45)
BUN
15.5
15.6
16.4
Total protein
7.2
7.2
7.3
Referent
(n = 101)
RDX exposed females*
Undetected (0.01 mg/m3
[n = 25)
BUN
13.2
8
12.6
Total protein
7.3
7.6
7.2
includes both workers exposed to RDX alone and RDX and HMX.
No differences were statistically significant in men or women.
BUN = blood urea nitrogen; LOD = limit of detection.
1-25
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-5. Evidence pertaining to urinary system (kidney and bladder) effects
in animals
Reference and study design
Results
Histopathological lesions
Lish et al. (1984)
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 wk 11 due
to excessive mortality)
Diet
2 yr
The incidence of cytoplasmic vacuolization of renal tubules was
greater for RDX-treated males than the control group males after
6 mos of treatment. However, at 12 and 24 mos of treatment, this
lesion was observed as frequently in controls as males treated with
RDX. There was no increase in incidence of this lesion in females at
any time point.
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
Histopathological examination of kidney did not reveal any significant
differences compared to controls; lesions observed were not
attributed to RDX treatment; incidence data were reported only for
control and 10 mg/kg-d groups.
1-26
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-5. Evidence pertaining to urinary system (kidney and bladder)
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
Note: More detailed histopathological
results, including interim sacrifice data at
6 and 12 mos, are provided in Tables 1-6
to 1-8.
Data for male rats sacrificed on schedule (SS) and those that died
spontaneously or were sacrificed moribund (SDMS) (summarized
below) were analyzed separately. There were no treatment-related
changes in incidence of kidney or urinary bladder lesions in females.
Doses
0
0.3
1.5
8.0
40
Kidney, medullary papillary necrosis; 24 mos (incidence)
SS
0/38
0/36
0/25
0/29
0/4
SDMS
0/17
1/19
0/27
0/26
18/27*
Sum
0/55
1/55
0/52
0/55
18/31*
Kidney, suppurative pyelitis; 24 mos (incidence)
SS
0/38
0/36
0/25
0/29
0/4
SDMS
0/17
1/19
0/27
1/26
5/27*
Sum
0/55
1/55
0/52
1/55
5/31*
Kidney, uremic mineralization; 24 mos (incidence)
SS
1/38
0/36
0/25
0/29
0/4
SDMS
0/17
1/19
2/27
0/26
13/27
Sum
1/55
1/55
2/52
0/55
13/31
Urinary bladder, luminal distention; 24 mos (incidence)
SS
0/38
0/36
0/25
0/29
1/4*
SDMS
0/16
2/19
1/27
3/22
24/28*
Sum
0/54
2/55
1/52
3/51
25/32*
Urinary bladder, cystitis hemorrhagic/suppurative; 24 mos
(incidence)
SS
0/38
0/36
0/25
1/29
0/4
SDMS
0/16
2/19
1/27
0/22
18/27*
Sum
0/54
2/55
1/52
1/51
18/31*
1-27
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-5. Evidence pertaining to urinary system (kidney and bladder)
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
0, 80, 60, or 40 mg/kg-d for 2 wks
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
Tubular nephrosis (incidence)
M
0/10
-
-
4/9*
F
0/11
1/11
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
Histopathological examination of kidney did not reveal any significant
differences compared to controls; incidence data were reported only
for control and 40 mg/kg-day groups.
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
Data were reported only for F2 generation controls and 5 and
16 mg/kg-day groups.
Doses
0
5
16
50
Renal tubule cysts, cortex (incidence)
M
4/10
4/10
8/10
-
F
3/10
4/10
8/10
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
Histopathological examination of kidney did not reveal any significant
differences compared to controls; incidence data were reported only
for control and 15 mg/kg-day groups.
1-28
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-5. Evidence pertaining to urinary system (kidney and bladder)
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
Doses
0
10
30
100
Nephropathy, chronic, unilateral (incidence)
M
7/30
0/10
2/10
1/10
F
4/30
0/10
0/10
1/10
Nephropathy, chronic, bilateral (incidence)
M
22/30
8/10
7/10
1/10
F
13/30
2/10
5/10
1/10
Microcretions, focal, unilateral (incidence)
M
0/30
0/10
0/10
0/10
F
4/30
5/10
0/10
1/10
Microcretions, focal, bilateral (incidence)
M
0/30
0/10
0/10
0/10
F
21/30
4/10
8/10
6/10
Note: Incidence data not presented for 300 and 600 mg/kg-day dose
groups because all rats died by Week 3 at these doses.
Hart (1974)
Dogs, beagle, 3/sex/group
Premix with ground dog chow containing
20 mg RDX/g-chow, 60 g dog food; purity
and particle size not specified
0, 0.1,1, or 10 mg/kg-d
Diet
13 wk
Histopathological examination of kidney did not reveal any significant
differences compared to controls; incidences were reported only for
control and 10 mg/kg-d groups.
Martin and Hart (1974)
Monkeys, cynomolgus or rhesus/
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
Medulla; mineralization, minimal to mild (incidence)
M + F
0/6
1/6
0/6
4/6
Dilated tubules, mild to moderate (incidence)
M + F
4/6
3/6
6/6
3/6
Multinucleated cells, tubules, minimal to moderate (incidence)
M + F
5/6
0/6
3/6
6/6
Eosinophilic inclusions, minimal to moderate (incidence)
M + F
2/6
0/6
0/6
3/6
1-29
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-5. Evidence pertaining to urinary system (kidney and bladder)
effects in animals (continued)
Reference and study design
Results
Kidney weightd
Lish et al. (1984)
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 wk 11 due
to excessive mortality)
Diet
2 yr
Doses
0
1.5
7.0
35
175/100
Absolute kidney weight at 104 wk (percent change compared to
control)
M
0%
-1%
4%
9%*
19%*
F
0%
3%
1%
1%
-2%
Relative kidney weight at 104 wk (percent change compared to
control)
M
0%
3%
6%
11%*
27%*
F
0%
1%
1%
2%
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
Absolute kidney weight (percent change compared to control)
M
0%
-3%
-7%
2%
F
0%
14%
-4%
8%
Relative kidney weight (percent change compared to control)
M
0%
-1%
-4%
4%
F
0%
22%
3%
18%
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 kidney weight at 105 wk (percent change compared to
control)
M
0%
2%
-7%
1%
0%
F
0%
3%
3%
2%
2%
Relative kidney weight at 105 wk (percent change compared to
control)
M
0%
1%
0%
2%
20%*
F
0%
3%
6%
5%
21%*
1-30
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-5. Evidence pertaining to urinary system (kidney and bladder)
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 kidney weight (percent change compared to control)
M
0%
-
-
-
18%
2%
F
0%
-
-
-
-8%
-5%
Relative kidney weight (percent change compared to control)
M
0%
-
-
-
29%
0%
F
0%
-
-
-
-8%
-3%
Experiment 2: 0, 40, 60, or 80 mg/kg-d for
2 wks 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
Absolute kidney weight (percent change compared to control)
M
0%
8%
11%
13%
F
0%
-5%
-3%
0%
Relative kidney weight (percent change compared to control)
M
0%
5%
9%
10%
F
0%
-5%
-4%
-5%
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
Absolute kidney weight (percent change compared to control)
M
0%
-
-
-
-2%
-5%
F
0%
-
-
-
1%
0%
Relative kidney weight (percent change compared to control)
M
0%
-
-
-
1%
5%
F
0%
-
-
-
6%
6%
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
Absolute kidney weight (percent change compared to control)
M
0%
6%
-12%
-
F
0%
-4%
-21%*
1-31
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-5. Evidence pertaining to urinary system (kidney and bladder)
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
Doses
0
4
8
10
12
15
Absolute kidney weight (percent change compared to control)
M
0%
-3%
-4%
-1%
3%
5%
F
0%
2%
5%
13%*
10%
15%*
Relative kidney weight (percent change compared to control)
M
0%
3%
6%
2%
1%
3%
F
0%
1%
-3%
-1%
-6%
-7%*
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
Doses
0
10
30
100
300
600
Absolute kidney weight (percent change compared to control)
M
0%
1%
1%
-9%
-
-
F
0%
1%
3%
-1%
-
-
Relative kidney weight (percent change compared to control)
M
0%
5%
7%
10%
-
-
F
0%
3%
5%
2%
-
-
Hart (1974)e
Dogs, beagle, 3/sex/group
Premix with ground dog chow containing
20 mg RDX/g-chow, 60 g 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 kidney weight (percent change compared to control)
M
0%
-
-
38%
F
0%
-18%
1-32
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-5. Evidence pertaining to urinary system (kidney and bladder)
effects in animals (continued)
Reference and study design
Results
Martin and Hart (1974)e
Monkeys, cynomolgus or rhesus,e
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
Absolute kidney weight (percent change compared to control)
M + F
0%
-2%
-3%
4%
F = female; M = male; SDMS = spontaneous death or moribund sacrifice; SS = scheduled sacrifice; TWA = time-
weighted average.
Note: A dash indicates that the study authors did not measure or report a value for that dose group.
^Statistically significant (p < 0.05) based on analysis by study authors.
aDoses were calculated by the study authors.
bLevine et al. (1981a) is a laboratory report of a 13-wk 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.
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).
dAn analysis by Craig et al. (2014) found a statistically significant correlation between absolute, but not relative,
kidney weights and renal histopathology. Therefore, only absolute kidney-weight data from RDX studies are
presented in Figure 1-2.
eKidney-weight data from the Hart (1974) and Martin and Hart (1974) studies were considered less informative
than other studies. Hart (1974) reported organ-weight data for high-dose dogs (3/sex/group) only, and the
kidney weights from Martin and Hart (1974) were highly variable across monkeys (e.g., kidney weights for the
control animals ranged from 4.9 to 13.1 g). Therefore, kidney-weight data from these two studies were not
presented in the exposure-response array for urinary system effects (see Figure 1-2).
1-33
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-6. Six-, 12-, and 24-month incidence of kidney endpoints in male F344
rats reported for statistical evaluation in Levine et al. (1983)
Doses (mg/kg-d)
0
0.3
1.5
8.0
40
Medullary papillary necrosis (incidence)
6 mos
SS
0/10
0/10
0/10
0/10
0/10
SDMS
-
-
-
-
0/5
Sum
0/10
0/10
0/10
0/10
0/15
12 mos
SS
0/10
0/10
0/10
0/10
0/10
SDMS
-
-
0/3
-
15/19*
Sum
0/10
0/10
0/13
0/10
15/29*
24 mos
SS
0/38
0/36
0/25
0/29
0/4
SDMS
0/17
1/19
0/27
0/26
18/27*
Sum
0/55
1/55
0/52
0/55
18/31*
Pyelitis (incidence)
6 mos
SS
0/10
0/10
0/10
0/10
0/10
SDMS
-
-
-
-
0/5
Sum
0/10
0/10
0/10
0/10
0/15
12 mos
SS
0/10
0/10
0/10
0/10
0/10
SDMS
-
-
0/3
-
1/19
Sum
0/10
0/10
0/13
0/10
1/29
24 mos
SS
0/38
0/36
0/25
0/29
0/4
SDMS
0/17
1/19
0/27
1/26
5/27*
Sum
0/55
1/55
0/52
1/55
5/31*
Pyelonephritis (incidence)
6 mos
SS
0/10
0/10
0/10
0/10
0/10
SDMS
-
-
-
-
0/5
1-34
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-6. Six-, 12-, and 24-month incidence of kidney endpoints in male
F344 rats reported for statistical evaluation in Levine etal. (1983)
(continued)
Doses (mg/kg-d)
0
0.3
1.5
8.0
40
Sum
0/10
0/10
0/10
0/10
0/15
12 mos
SS
0/10
0/10
0/10
0/10
0/10
SDMS
-
-
0/3
-
1/19
Sum
0/10
0/10
0/13
0/10
1/29
24 mos
SS
0/38
0/36
0/25
1/29
0/4
SDMS
0/17
0/19
2/27
1/26
1/27
Sum
0/55
0/55
2/52
2/55
1/31
SDMS = spontaneous death or moribund sacrifice; SS = scheduled sacrifice.
Note: A dash indicates that the study authors did not measure or report a value for that dose group.
^Statistically significant (p < 0.05) based on analysis by study authors.
Source: Levine et al. (1983).
1-35
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-7. Six-, 12-, and 24-month incidence of urinary bladder endpoints in
male F344 rats reported for statistical evaluation in Levine etal. (1983)
Doses (mg/kg-d)
0
0.3
1.5
8.0
40
Luminal distention (incidence)
6 mos
SS
0/10
0/10
0/10
0/10
0/10
SDMS
-
-
-
-
0/5
Sum
0/10
0/10
0/10
0/10
0/15
12 mos
SS
0/10
0/10
0/10
0/10
0/10
SDMS
-
-
0/3
-
18/19*
Sum
0/10
0/10
0/13
0/10
18/29
24 mos
SS
0/38
0/36
0/25
0/29
1/4*
SDMS
0/16
2/19
1/27
3/22
24/28*
Sum
0/54
2/55
1/52
3/51
25/32*
Cystitis, hemorrhagic/suppurative (incidence)
6 mos
SS
0/10
0/10
0/10
0/10
0/10
SDMS
-
-
-
-
0/5
Sum
0/10
0/10
0/10
0/10
0/15
12 mos
SS
0/10
0/10
0/10
0/10
0/10
SDMS
-
-
0/3
-
17/19*
Sum
0/10
0/10
0/13
0/10
17/29
24 mos
SS
0/38
0/36
0/25
1/29
0/4
SDMS
0/16
2/19
1/27
0/22
18/27*
Sum
0/54
2/55
1/52
1/51
18/31*
SDMS = spontaneous death or moribund sacrifice; SS = scheduled sacrifice.
Note: A dash indicates that the study authors did not measure or report a value for that dose group.
^Statistically significant (p < 0.05) based on analysis by study authors.
Source: Levine et al. (1983).
1-36
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
1000
100 4
¦>
fU
T3
jp
10
E
o
Q
1 i
0.1
o
o
o
o
• significantly changed
o not significantly changed
o o
O 0 o
^ ^ ro
03 ra i-
£ E ^
a - £
00
cn
t
03
x
Chronic
03
O)
<1>
to
in
3
3
m
00
o
E
O
E
a\
t-H
o
o
00
00
cr>
_OJ
f3
03
O
E
00
03
t-H
O
LL.
03
o
fj
03
E
-Q
•4—
05*
T—1
03
00
cn
t-H
<£>
O
o
o
cr>
OJ
O)
on
0J
3
_c
O
>
u
OJ
d>
OJ
c
c
>
>
-
LT
c
¦o
-a
05
-Q
Chronic
o
oo
o
oc
(D
O
O
m m ^
a)
o
(J)
cn
CjO
O
H
t
03
X
o
u
s
ai
I
c3
>-
Subchronic
Histopathological lesions
Figure 1-2. Exposure-response array of urinary system (kidney and bladder) effects.
1-37
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Studies in experimental animals provide some evidence that RDX exposure is associated
with urinary system effects (see Table 1-5 and Figure 1-2). The strongest evidence of effects on this
organ system is the collection of histopathological changes, including increased incidences of
kidney medullary papillary necrosis and pyelitis, uremic mineralization, and bladder distention
and/or cystitis, observed in male F344 rats exposed to 40 mg/kg-day RDX in the diet for 12 months
or longer (Levine etal.. 19831. The incidences of urinary system changes were higher at 2 years
than at 12 months, but the response at both time points was robust (e.g., incidence of medullary
papillary necrosis in male 40-mg/kg-day rats: 15/29 at 12 months, 18/31 at 2 years).15 Renal
effects were considered the principal cause of treatment-related morbidity and mortality in these
high-dose males. Similar kidney lesions were not observed in male rats in any dose group at the
6-month interim sacrifice (see Tables 1-6 and 1-7). Histopathological changes reported in some
male rats in the lower dose groups (0.3,1.5, and 8 mg/kg-day) after 2 years on study were not
dose-related, few in number, and consistent with background changes seen in aged rats.
Results from Levine etal. (1983) demonstrate a marked sex difference in response to RDX
urinary system toxicity; no kidney or urinaiy bladder changes were associated with RDX exposure
in female rats. In addition, mice appear to be less sensitive to the urinary system effects of RDX
than rats; the incidences of kidney histopathological changes in male and female B6C3Fi mice
exposed to RDX in the diet for 2 years at concentrations as high as 100 mg/kg-day were similar to
controls (Lishetal.. 1984).
Histopathological findings in the urinary system from other experimental animal studies
are largely consistent with the 2-year findings from Levine etal. (1983) and Lish etal. (1984) [i.e.,
that kidney and urinary bladder system effects are generally observed after RDX exposures longer
than 6 months in duration and at high doses (e.g., >40 mg/kg-day)]. Specifically, no pattern of
histopathological changes in the kidney were reported in rats exposed to RDX for 13 weeks (Crouse
etal.. 2006: Levine etal.. 1990: Levine etal.. 1981a. b; Cholakis etal.. 1980). in a 2-year study in
Sprague-Dawley rats or 13-week study in beagle dogs that used a maximum dose of 10 mg/kg-day
(Hart. 1976.1974). or in rabbits exposed dermally to a cumulative dose of 165 mg/kg RDX in
dimethylsulfoxide received over a 4-week period [5 days/week; (McNamara et al.. 1974)]. In fact,
in the 13-week study in F344 rats (Levine etal.. 1981a) conducted by the same investigators that
conducted the 2-year study in the same strain (Levine etal.. 1983). chronic nephropathy was
observed in both control and treated animals with no evidence of a dose-related increase in
incidence.
Evidence of kidney histopathological changes in RDX-exposed animals following an
exposure duration of less than 6 months is limited to an increased incidence of tubular nephrosis
observed in B6C3Fi mice exposed for 13 weeks to 320 mg/kg-day RDX (Cholakis etal.. 1980). a
dose eightfold higher than the dose that produced kidney and urinary bladder pathology in rats
15Denominator represents scheduled sacrifice animals plus spontaneous deaths and moribund sacrifice
animals.
1-38
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
after 2 years of exposure. Increased incidence of minimal to mild mineralization of the medulla was
observed in male and female monkeys exposed to 10 mg/kg-day RDX for 90 days by gavage (Martin
and Hart. 19741. but the study authors did not identify this as treatment related. Finally, in a
2-generation study, Cholakis etal. (19801 reported an increased incidence of renal tubular
epithelial-lined cysts in the kidney cortex in F2-generation rats exposed to RDX at doses up to
16 mg/kg-day. However, the kidney findings from this two-generation study are difficult to
interpret because F2 animals were exposed for a relatively short duration (during gestation and
through weaning only) and because no histopathology was performed for the parental and F1
generations.
Other kidney endpoints—serum chemistry parameters that may indicate changes in renal
function and kidney weights—did not provide consistent evidence of treatment-related changes.
Measurement of serum chemistry parameters (including BUN and uric acid) in studies of RDX in
mice, rats, dogs, and monkeys (Crouse etal.. 2008: Levine etal.. 1990: Lish etal.. 1984: Levine etal..
1981a. b; Cholakis etal.. 1980: Hart. 1976.1974: Martin and Hart. 19741 revealed variations
(increases or decreases) from the respective control groups that were not dose related. Kidney
weights in subchronic oral toxicity studies in rats, dogs, and monkeys did not show a clear pattern
of change associated with RDX exposure. Kidney-weight changes were either not dose related or
were inconsistently increased or decreased across studies (see Table 1-5). Less emphasis is placed
on evidence of organ-weight changes from chronic (2-year) studies (Lish etal.. 1984: Hart. 1976)
because normal physiological changes associated with aging and intercurrent disease may
contribute to interanimal variability that could confound organ-weight interpretation (Sellers etal..
20071.
Exposure to HMX, the major contaminant in many of the available RDX studies, was
associated with histopathological changes in the kidney and alterations in renal function in female,
but not male, rats fed doses >450 mg/kg-day HMX for 13 weeks (see the Integrated Risk
Information System [IRIS] assessment of HMX at https: //www.epa.gov/irisl. No effects were
observed at doses <115 mg/kg-day. Given the dose levels where HMX appears to exhibit toxicity
and the percentage of HMX (up to 10%) present as an impurity in technical grade RDX that would
result in HMX exposures <60 mg/kg-day in the studies of RDX toxicity, the contribution of HMX to
the observed kidney toxicity in studies of RDX is expected to be negligible. Further, differences in
the pattern of toxicity (i.e., kidney effects observed only in RDX-exposed males and HMX-exposed
females) also suggest that HMX contaminants were not responsible for kidney effects in rats
exposed to RDX.
Mechanistic Evidence
No MOA information is available for RDX-induced urinary system effects. Mechanistic
information underlying the neurotoxicity observed with RDX exposure, and the specific affinity of
RDX to the GABAa receptor-convulsant site (Williams et al.. 2011: Williams and Bannon. 2009).
1-39
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
suggests a biologically plausible role for the GABAa receptor in RDX-related effects on the urinary
system.
GABA and GABA receptors have been identified in a number of peripheral tissues (Erdoet
al.. 1991: Ong and Kerr. 1990: Erdo. 19851. Brar etal. (20141 demonstrated that pretreatment with
picrotoxin reduced the renoprotective effects of sodium valproate (which acts on both GABAa and
GABAb receptors) in a rat model of ischemia-induced acute kidney injury, suggesting that GABAa
receptors may be important in renal function. GABA is believed to play a role in the regulation of
urination and bladder capacity [reviewed in Fowler et al. (2008) and Yoshimura and de Groat
(1997)]. In rats, injection of a GABAa receptor agonist inhibits the urination reflex (Igawa etal..
1993: Kontani et al.. 1987). GABAa agonists injected into the periaqueductal gray area in rats
inhibited reflex bladder activity, while injection of an antagonist reduced bladder capacity and
increased the frequency of bladder reflex activity fStone etal.. 20111. RDX would be expected to act
like an antagonist and increase bladder activity, although the impact of chronic exposure to RDX
acting as a GABAa receptor antagonist is not known. Evidence of GABAergic signaling regulating
bladder function, and the hypothesized disruption of that regulation by RDX via interaction with
GABAa receptors, suggests a possible MOA for the kidney and urinary bladder lesions observed in
particular by Levine etal. f 19831: however, there does not appear to be any direct evidence (basic
science or RDX-specific) to help discern the role of GABAa receptor in mediating these lesion types.
In summary, no studies are available to inform mechanistically how RDX might lead to
urinary system effects. There is evidence that RDX binds to GABAa receptors in neuronal tissues
(Williams etal.. 2011: Williams and Bannon. 2009). and it is biologically plausible that binding to
the GABA receptor could occur in other tissues as well, contributing to the observed kidney and
urinary bladder effects. However, the way(s) by which GABAa receptors may work in nonneuronal
tissues and organs is not well understood, and the MOA by which RDX induces urinary system
effects is not established.
Integration of Urinary System (Kidney and Bladder) Effects
Evidence for kidney effects resulting from RDX exposure consists of human case reports and
findings of histopathological changes in rodents. In humans, evidence for kidney effects (including
decreased urine output, blood in urine, and proteinuria) is limited to individuals with acute
accidental exposure (ingestion and inhalation) to unknown amounts of RDX. No RDX-related
changes in kidney parameters were found in a small cross-sectional study of RDX-exposed workers
(Hathaway and Buck. 1977).
The 2-year Levine etal. (1983) study in F344 rats reported histopathological changes
(papillary necrosis, pyelitis, luminal distension, and cystitis) in the kidney and urinary bladder in
approximately 50% of male rats exposed to 40 mg/kg-day (the highest dose tested in this study),
but only following exposure to RDX for longer than 6 months. Histopathological findings from
other studies in rats, mice, and dogs (Crouse etal.. 2006: Levine etal.. 1990: Levine etal.. 1981a. b;
Cholakis etal.. 1980: Hart. 1976.1974) are largely consistent with the 2-year findings from Levine
1-40
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
etal. f 19831 (i.e., that kidney and urinary bladder effects are generally observed after RDX
exposures longer than 6 months in duration and at high doses [e.g., >40 mg/kg-day]). Other
measures of kidney effects (kidney weights and serum chemistry parameters) did not provide
consistent evidence of dose-related changes associated with RDX exposure.
Histopathologic findings from 2-year studies in F344 rats (Levine etal.. 19831 and B6C3Fi
mice fLish etal.. 19841 provide evidence of sex and species differences in response to RDX. In
contrast to the substantial urinary system toxicity observed in high-dose F344 male rats that was
considered the primary cause of RDX-related morbidity and mortality (Levine etal.. 1983). no
kidney toxicity was associated with RDX in similarly exposed female rats. Additionally, mice appear
to be less sensitive than rats, based on an absence of RDX-related kidney histopathological changes
in male and female B6C3Fi mice exposed to RDX in the diet for 2 years at doses more than twofold
greater than doses that produced substantial urinary system toxicity in male rats fLish etal.. 19841.
In light of the dose-related increase in histopathological changes in the kidney and urinary
bladder in male rats in the Levine etal. (1983) study, and in particular the robust response in the
high-dose animals, urinary system effects are a potential human hazard of RDX exposure.
1.2.3. Prostate Effects
No human studies were identified that evaluate the potential of RDX to cause effects on the
prostate. There was limited information to evaluate prostate effects in animal studies, including
two 2-year dietary studies in rats and mice (Lish etal.. 1984: Levine etal.. 1983). and one 90-day
gavage study (Crouse et al.. 2006). A summary of the prostate effects associated with RDX exposure
in animals is presented in Tables 1-8 and 1-9 and Figure 1-3. Studies are ordered in the evidence
tables and exposure-response arrays by duration of exposure and then by species.
Most animal studies available did not specifically evaluate whether there were prostate
effects associated with RDX exposure. A significant, dose-related increase in the total incidence of
suppurative prostatitis was reported in male F344 rats exposed to >1.5 mg/kg-day RDX in the diet
for 2 years (Levine etal.. 1983). Neither suppurative prostatitis nor any other treatment-related
prostate effects were observed in a 2-year dietary study in mice (Lish etal.. 1984). Suppurative
prostatitis was not observed in 90-day studies in the rat involving oral (dietary or gavage) exposure
to RDX (Crouse etal.. 2006: Levine etal.. 1990: Levine etal.. 1981a. b). In the 90-day gavage study
fCrouse etal.. 20061. mild subacute inflammation was observed in the prostate of one of the rats at
terminal sacrifice (TS) in the high-dose group (15 mg/kg-day).16 The study authors considered the
single observation to be consistent with expected background incidence and not treatment related.
Excluding inflammation, there was little additional information to identify prostate effects
associated with RDX exposure. Histopathological analysis identified a statistically significant
15A reporting discrepancy exists in Crouse etal. (20061 between the results section and the summary of
histopathological findings in males in the appendix. The results section reports that mild subacute
inflammation of the prostate was presentin 1/7 males in the 15-mg/kg-day dose group atterminal sacrifice.
The summary of histopathological findings (Appendix U) reports an incidence of 1/8 at 15 mg/kg-day.
1-41
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
increase in the incidence of spermatic granuloma of the prostate in rats fed 40 mg/kg-day RDX for
up to 6 months. No gross abnormalities of the prostate were observed to accompany this finding,
nor was this endpoint observed in 12- or 24-month dietary exposures to RDX (Levine etal.. 19831.
Suppurative prostatitis is part of a continuum of inflammation. Further, suppurative
prostatitis and nonsuppurative prostatitis are not mutually exclusive; one form can evolve into
another. Levine etal. (19831 also reported the incidence of nonsuppurative (chronic-active)
inflammation as well as subacute inflammation in male rats (see Table 1-8).
Table 1-8. Two-year prostate inflammation incidence in male F344 rats
(Levine et al.. 1983)
Doses
(mg/kg-day)
0
0.3
1.5
8.0
40
Inflammation, subacute
TS
1/38
0/36
0/25
0/29
0/4
SDMS
0/16
0/19
0/27
0/26
0/27
Sum
1/54
0/55
0/52
0/55
0/31
Inflammation, chronic-active
TS
15/38
13/36
6/25
6/29
1/4
SDMS
5/16
5/19
7/27
5/26
1/27
Sum
20/54
18/55
13/52
11/55
2/31
Inflammation, suppurative
TS
0/38
1/36
2/25*
4/29*
0/4
SDMS
2/16
3/19
7/27*
8/26
19/27*
Sum
2/54
4/55
9/52*
12/55*
19/31*
All inflammation
Sum
23/54
22/55
21/52
23/55
21/31
SDMS = spontaneous death or moribund sacrifice; TS = terminal sacrifice.
^Statistically significant (p < 0.05) based on analysis by study authors.
Source: Levine et al. (1983)
1-42
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-9. Evidence pertaining to prostate effects in animals
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 nm
0, 0.3,1.5, 8.0, or 40 mg/kg-d
Diet
2 yr
Data for male rats sacrificed on schedule (SS) and those that died
spontaneously or were sacrified moribund (SDMS) (summarized
below) were analyzed separately.
0
0.3
1.5
8.0
40
Prostate, suppurative inflammation (prostatitis); 24 mos
(incidence)
SS
0/38
1/36
2/25*
4/29*
0/4
SDMS
2/16
3/19
7/27*
8/26
19/27*
Sum
2/54
4/55
9/52*
12/55*
19/31*
Spermatic granuloma of the prostate; 6 mos (incidence)
SS
0/10
2/10
2/10
1/10
6/10*
SDMS
—
—
2/5
Sum
0/10
2/10
2/10
1/10
8/15*
Lish et al. (1984)
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 wk 11 due
to excessive mortality)
Diet
2 yr
0
1.5
7.0
35
175/100
Prostate, chronic inflammation; 24 mos (incidence)3
M
1/62
1/3
0/1
1/1
0/27
Crouse et al. (2006)
Rats, F344,10/sex/group
99.99% pure
0, 4, 8,10, 12, or 15 mg/kg-da
Gavage
13 wk
Doses
0
4
8
10
12
15
Prostate, mild subacute inflammation (incidence)
M
0/10
1/8
M = male; SDMS = spontaneous death or moribund sacrifice; SS = scheduled sacrifice.
Note: A dash indicates that the study authors did not measure or report a value for that dose group.
^Statistically significant (p < 0.05) based on analysis by study authors.
Examination only required by protocol in the control and high-dose groups.
1-43
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
100
10 -
T3
¦
00
J*
OJD
E
o
a
1 -
0.1
• significantly changed
O not significantly changed
+-> m
£ 00
O G)
F ^
£ 00
CO
C
O
<72
E
4->
2
o
E
E
ID
O
VD
o
CD
4—
CNl
E
c
'
:
_o
ro
3
¦*->
u
TO
E
E
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
As noted by the Science Advisory Board (SAB) in their review of the external review draft of
the RDX assessment (SAB. 2017). the incidences of all observations of inflammation at 24 months in
the Levine etal. (1983) study were similar in all dose group (approximately 40%) except for the
high-dose group (68%). The incidence rate in the control and three lowest dose groups is lower
than the background incidence of inflammation (70.4%) in a retrospective analysis of background
lesions in male accessory sex organs of F344 rats reported by Suwaetal. (2001). The lower
incidence in F344 rats reported in Levine etal. (1983) suggests there may have been differences in
histopathological practices between those employed by Levine etal. (1983) and more recent
diagnostic criteria. For example, inflammation incidence varies across lobes of the prostate, and the
methods section in the Levine etal. (1983) report does not provide sufficient information to
determine how the prostates were evaluated for inflammation. Finally, male rats in the high-dose
group (40 mg/kg-day) were moved from group to individual housing between Weeks 30 and 40
during the study, due to a high incidence of fighting. The increased incidence of fighting may have
contributed to conditions that lead to urogenital infections in male rats (Creasy etal.. 2012).
The severity of inflammation differs in Levine etal. (1983) compared to that reported in
Suwa etal. (2001). In reviewing the background incidence of all inflammation in the prostate in
1,768 F344 rats, Suwa etal. f20011 identified an average severity grade of "mild." In Levine et al.
(1983). there is an increased incidence of suppurative prostatitis, which is more severe
(characterized by the formation of pus and a high concentration of neutrophils). There was also a
shift from chronic inflammation to suppurative inflammation with increasing dose of RDX starting
at 1.5 mg/kg-day (see Table 1-8). At the highest dose, the shift from chronic to suppurative
inflammation is clear, with only two animals exhibiting chronic inflammation and 19 identified as
having suppurative inflammation.
Some reports have hypothesized that the observation of prostate inflammation in Levine et
al. (1983) is secondary to a bacterial infection unrelated to RDX toxicity (ATSDR. 2012: Sweeney et
al.. 2012a: Crouse etal.. 2006). For example, in describing the results from the 2-year dietary study
in rats, Crouse etal. (2006) observed that the inflammation reflects a common condition in rodents,
noting that because 85% of the incidence occurred in rats found at spontaneous death or moribund
sacrifice (SDMS), it was most likely that the condition was a result of an incidental bacterial
infection. Although the proportion of suppurative prostatitis was higher in SDMS rats, there was an
increasing trend with dose in both the scheduled sacrifice (SS) and SDMS groups; the incidence of
suppurative prostatitis in the control group was 4% when the SS and SDMS groups were combined.
Additionally, the dose-related nature of the increased incidence suggests that the primary cause
(potentially leading to bacterial infection) was treatment related, as a more uniform distribution of
rats with suppurative prostatitis would be expected with a spontaneous or age-related lesion. The
dose-responsiveness could be explained if the infections were secondary to treatment-related
immunotoxicity, but there is no evidence from Levine etal. (1983) to support this possibility. A
more thorough analysis of immune endpoints in a 90-day gavage exposure of F344 rats did not
1-45
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
identify any immunotoxic effects associated with RDX fCrouse et al.. 20061. In general, causes of
prostatitis other than infection exist, including stress, endocrine effects (i.e., changing prolactin
levels), and autoimmune dysfunction (see, for example. Bosland. 1992: Gatebeck etal.. 1987: Parker
and Grabau. 19871.
Mechanistic Evidence
No MOA information is available for RDX-induced prostate effects. However, mechanistic
information underlying the neurotoxicity observed with RDX exposure, and the specific affinity of
RDX to the GABAa receptor-convulsant site (Williams et al.. 2011: Williams and Bannon. 20091.
suggests a biologically plausible role for the GABAa receptor in RDX-related effects, and provides
some potential MOA hypotheses for the effects reported in Levine etal. (1983) that do not require
bacterial infection.
One possibility is that effects would result from direct interactions with GABAa receptors
located on the prostate. GABAa receptors have been identified on the prostate (Napoleone et al..
19901. providing a potential mechanism by which RDX could interact directly with the prostate.
However, this would require that the prostate is actively maintained in a noninflamed state,
mediated by GABA; RDX binding to GABAa receptor-convulsant sites on the prostate would reduce
the inhibitory effects of the GABA receptor, leading to increased inflammation (Tohnson. 2015b).
No evidence was found to support this potential pathway leading to prostate inflammation.
Another possibility is that alterations in hormonal signaling or circulating levels of estrogen
or prolactin may lead to prostatitis. Prostate inflammation has been associated with endocrine
disruptors in the environment (Cowin etal.. 2010). and increased prolactin has been shown to
cause lateral lobe prostatitis (Stoker et al.. 1999b: Stoker etal.. 1999a: Tangbanluekal and
Robinette. 1993: Robinette. 1988). Typically, the inflammation seen is chronic and does not reverse
over time (Robinette. 1988). Functional GABAa receptors have been identified in the anterior
pituitary (Zemkova etal.. 2008: Maverhofer etal.. 2001). which also serves as the primary source of
prolactin. Thus, the prostate inflammation observed in the rat in the 2-year study by Levine et al.
(1983) could have been produced by disruption of pituitary prolactin or another hormonal signal
by interfering with normal regulatory GABA-related hormonal control. However, no direct
evidence for this hypothesized MOA is available. Levine etal. (1983) did not evaluate serum
endocrine measures or pituitary weights, nor did they observe pituitary adenomas that could
account for higher prolactin levels.
Another hypothesis is that the prostate effects could be mediated through an autoimmune
inflammatory response. GABAA-receptor transcripts have been identified in immune cells of mouse
models (Reyes-Garcia et al.. 2007: Tian etal.. 2004). and GABAA-receptor agonists have decreased
cytotoxic immune responses and hypersensitivity reactions (Tian etal.. 1999: Bergeretetal.. 1998).
In a mouse autoimmune model of multiple sclerosis, Bhatetal. f20101 found that treatment of
macrophages challenged with lipopolysaccharide with various GABA agonists decreased cytokine
production; addition of picrotoxin (which may have effects similar to those of RDX, as it binds to the
1-46
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
same site) was able to reduce this effect However, picrotoxin on its own did not significantly alter
cytokine production, suggesting that the effects are limited to reversal of agonist-induced
GABAergic activity flohnson. 2015b). If an autoimmune mechanism was contributing to the effects
observed with RDX exposure, it is unclear why inflammation would be limited to the prostate. RDX
has also tested negative in the only battery of immunotoxicity tests to which it was subjected
fCrouse etal.. 20061.
In summary, no studies were available that inform mechanistically how RDX exposure
might lead to prostate effects. There is evidence that RDX binds to GABAa receptors in neuronal
tissues (Williams etal.. 2011: Williams and Bannon. 20091. and it is biologically plausible that
binding to the GABA receptor could occur in other tissues as well. Among the mechanistic
information presented above, MOAs that require direct action on the prostate appear less likely;
however, the ways that GABAa receptors work in nonneuronal tissues and organs is still not well
understood, and the MOA by which RDX may induce prostate effects is unknown.
Integration of Prostate Effects
Suppurative prostatitis was reported in male F344 rats chronically exposed to RDX in the
diet for 24 months fLevine etal.. 19831. No other studies of equivalent duration were performed in
rats to determine the consistency of this effect Spermatic granuloma of the prostate was identified
in F344 rats exposed to RDX for up to 6 months, but not at 12 or 24 months in the study; therefore,
the biological significance of the 6-month finding is uncertain. A 24-month study in mice (Lishet
al.. 1984) did not report prostate effects associated with RDX exposure. No other animal studies of
shorter duration identified prostate effects associated with RDX exposure. In light of the
dose-related, statistically significant increase in suppurative prostatitis, there is suggestive
evidence that prostate effects are a potential human hazard of RDX exposure.
1.2.4. Developmental Effects
No human studies were identified that evaluate the potential of RDX to cause
developmental effects. Information relevant to an examination of the association between RDX
exposure and developmental effects comes from a two-generation reproductive toxicity study in
rats and developmental studies in rats and rabbits involving oral administration of RDX during
gestation. A summary of the developmental effects associated with RDX exposure is presented in
Table 1-10 and Figure 1-4. Studies are ordered in the evidence tables and exposure-response
arrays by duration of exposure and then by species.
1-47
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-10. Evidence pertaining to developmental effects in animals
Reference and study design
Results
Prenatal mortality/offspring survival
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
Doses
0
5
16
50
Stillborn pups (incidence)
Fl
8/207
6/296
4/259
16/92*
F2
6/288
6/290
2/250
24/46*
Offspring survival at birth (percent of fetuses)
Fl
96%
98%
98%
83%*
F2
98%
98%
99%
48%*
Survival at weaning (percent ofliveborn pups)
Fl
87%
96%
90%
8%
F2
79%
86%
79%
0%
F0 maternal deaths occurred at 50 mg/kg-d. Only six Fl females in
this group survived to serve as parental animals; none of the
surviving six died during subsequent treatment.
Note: Results on a per litter basis were not provided.
Cholakis et al. (1980)
Rabbits, NZW, 11-12/group
88.6% pure, with 9% HMX and 2.2% water
as contaminants; ~200 nm particle size
0, 0.2, 2.0, or 20 mg/kg-d
Gavage
GDs 7-29
Doses
0
0.2
2
20
Early resorptions (mean percent per dam)
6%
5%
4%
1%
Late resorptions (mean percent per dam)
8%
5%
3%
3%
Viable fetuses (mean percent per dam)
85%
82%
11%
94%
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
Early resorptions (mean percent per dam)
6.0%
2.5%
4.8%
15.3%
Late resorptions (mean percent per dam)
0.5%
0.5%
0.3%
1.6%
Complete litter resorptions (number of litters)
0
0
0
2
Viable fetuses (mean percent per dam)
93.2%
97.6%
94.9%
81.4%
Significant maternal mortality (7/24 dams) occurred at 20 mg/kg-d.
1-48
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-10. Evidence pertaining to developmental effects in animals
(continued)
Reference and study design
Results
Angerhofer et al. (1986)
Rats, Sprague-Dawley, 39-51 mated
females/group (25-29 pregnant
dams/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
Resorptions (percent of total implantations)
4.8%
6.1%
5.9%
6.4%
Early resorptions (percent of total implantations)
4.8%
6.1%
5.9%
6.2%
Late resorptions (percent of total implantations)
0%
0%
0%
0.27%
Live fetuses (mean percent per litter)
100%
100%
100%
100%
Significant maternal mortality (16/51) occurred at 20 mg/kg-d.
Percent resorptions and live fetuses based on number of surviving
females at time of necropsy.
Offspring growth
Cholakis et al. (1980)
Rabbits, NZW, 11-12/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
Gavage
GDs 7-29
Doses
0
0.2
2.0
20
Fetal body weight (percent change compared to control)
0%
-6.7%
-2.3%
-9.3%
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
Fetal body weight (percent change compared to control)
0%
2%
3%
-7%
Significant maternal mortality (7/24 dams) occurred at 20 mg/kg-d.
Angerhofer et al. (1986)
Rats, Sprague-Dawley, 39-51 mated
females/group (25-29 pregnant
dams/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
Fetal body weight (percent change compared to control)
0%
-4%
-2%
-9%a
Fetal body length (percent change compared to control)
0%
-1%
-1%
-5%a
Significant maternal mortality (16/51) occurred at 20 mg/kg-d.
1-49
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-10. Evidence pertaining to developmental effects in animals
(continued)
Reference and study design
Results
Morphological development
Cholakis et al. (1980)
Rabbits, NZW, 11-12/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
Gavage
GDs 7-29
Doses
0
0.2
2.0
20
Spina bifida (incidence)
Fetuses
0/88
0/99
0/94
3/110
Litters
0/11
0/11
0/11
2/12
Misshapen eye bulges (incidence)
Fetuses
0/88
0/99
0/94
3/110
Litters
0/11
0/11
0/11
1/12
Cleft palate (incidence)
Fetuses
0/39
1/46
2/44
2/52
Litters
0/11
1/11
1/11
1/12
Enlarged front fontanel (incidence)
Fetuses
0/49
5/53
2/50
8/58
Litters
0/11
2/11
2/11
2/12
Unossified sternebrae (incidence)
Fetuses
4/49
12/53
8/50
12/58
Litters
4/11
7/11
4/11
6/12
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
No gross or soft-tissue anomalies were seen in any exposure group.
No treatment-related increase in the incidence of litters with skeletal
anomalies was observed. Significant maternal mortality (7/24 dams)
occurred at 20 mg/kg-d.
1-50
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-10. Evidence pertaining to developmental effects in animals
(continued)
Reference and study design
Results
Angerhofer et al. (1986)
Rats, Sprague-Dawley, 39-51 mated
females/group (25-29 pregnant
dams/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
No treatment-related increase in the incidence of anomalies was
observed.
Doses
0
2
6
20
Total malformations (percent of fetuses with malformations)
1%
1%
0%
2%
Significant maternal mortality (16/51) occurred at 20 mg/kg-d.
GD = gestational day; NZW = New Zealand White.
^Statistically significant (p < 0.05) based on analysis by study authors.
statistically significant dose-related trend (p < 0.05) by linear trend test, performed for this assessment. Average
fetal weights or lengths for each litter comprised the sample data for this test.
1-51
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
• signficantly changed
O not signifcantly changed
LL
_Q
CD
T3
_Q
e
CD
o
ro
00
uT
o
00
cn
i
•4—1
en
ro
i
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Animal studies have reported decreases in offspring survival following administration of
RDX. Pup survival rates in the FO and F1 generations (including both stillborn pups and postnatal
deaths through the age of weaning) were statistically significantly decreased in RDX-exposed CD
rats compared to controls in the only available two-generation reproductive toxicity study of RDX
fCholakis etal.. 19801. This observation was noted only at the highest dose tested (50 mg/kg-day)
that also produced toxicity in adults (mortality [18%], reduced body weights [8-14%], and reduced
food consumption [10-17%]). Decreased fetal viability was observed at the highest dose tested,
20 mg/kg-day, in a developmental toxicity study in F344 rats (Cholakis etal.. 1980). although no
effect on live fetuses was observed in a developmental toxicity study in Sprague-Dawley rats at the
same dose (Angerhofer etal.. 1986): both of these studies reported significant mortality (29-31%)
in dams at 20 mg/kg-day. Increased resorptions were similarly limited to the highest dose tested
[20 mg/kg-day; fCholakis etal.. 19801], Both studies started treatment with RDX on Gestational
Day (GD) 6, which may contribute to the incidence of resorptions observed in the control and
treated groups. As noted in EPA's Guidelines for Developmental Toxicity Risk Assessment (U.S. EPA.
1991). treatment beginning around the time of implantation may result in an increase in
implantation loss that reflects variability that is not treatment related. There was no evidence of
maternal toxicity, embryotoxicity, or decreased fetal viability in a teratology study of pregnant New
Zealand White (NZW) rabbits administered RDX by gavage from GD 7 to 29 at doses up to
20 mg/kg-day (Cholakis etal.. 1980). suggesting that rabbits may be less sensitive to RDX toxicity
than rats.
Statistically significant, dose-related reductions in fetal body weight and length were
reported in Sprague-Dawley rats administered RDX by gavage from GD 6 to 15 (Angerhofer etal..
19861.17 Decreased fetal body weight (9%) and body length (5%), with statistically significant
trends, were observed at 20 mg/kg-day, a dose that produced significant (31%) mortality in the
dams. A similar reduction in fetal body weight of 7% (not statistically significant) was observed in
F344 rats exposed to RDX at 20 mg/kg-day, a dose associated with 29% maternal mortality
fCholakis etal.. 1980). Dose-related reductions in fetal body weight were not observed in NZW
rabbits at doses up to 20 mg/kg-day fCholakis etal.. 1980).
No treatment-related effects on morphological development have been reported in rats
exposed to a dose as high as 20 mg/kg-day RDX, a dose that resulted in 29-31% maternal mortality
(Angerhofer et al.. 1986: Cholakis etal.. 1980). Examination of rabbits administered RDX at doses
up to 20 mg/kg-day from GD 7 to 29 also provided no evidence of treatment-related developmental
anomalies fCholakis etal.. 1980). Although increased incidences of enlarged frontal fontanel and
unossified sternebrae were observed in fetuses of all groups of NZW rabbits administered RDX
17The statistical analyses presented by the study authors were performed on a per fetus basis; EPA's
Guidelines for Developmental Toxicity Risk Assessment (U.S. EPA. 19911 recommend that fetal data be analyzed
on a per litter (rather than per fetus) basis. In a reanalysis of the Angerhofer et al. (19861 data by EPA on a
per litter basis, fetal body weight and length showed statistically significant decreasing trends.
1-53
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
fCholakis etal.. 19801. these developmental anomalies did not exhibit a dose-related increase in the
number of either fetuses or litters affected, and were thus interpreted as not being
treatment-related by the study authors fCholakis etal.. 19801. Neither individual litter data nor
historical control data from the performing laboratory were available to assist in the interpretation
of these findings. The study author's interpretation is supported by the following additional
considerations. A report of historical control incidences of fetal skeletal observations in NZW
rabbits for 224 prenatal developmental toxicology studies conducted in 8 contract research
laboratories during the period of 1988-1992 (MTA. 1992) included findings from 26,166 fetuses of
3,635 litters. Background control incidences of enlarged anterior fontanel were observed in
8 fetuses (0.031%) of 7 litters (0.193%), while sternebrae agenesis [which may not be entirely
comparable to the finding of unossified sternebrae in Cholakis etal. (1980)] was found in 10 fetuses
(0.038%) of 5 litters (0.138%). Although the use of concurrent control data is preferable for the
interpretation of developmental toxicity data, this historical information supports the low control
incidences of these findings in the Cholakis etal. (1980) study as being within typical historical
parameters. It is also noted that the nondose-related pattern of increased enlarged fontanel and
unossified sternebrae across treated groups in Cholakis etal. (1980) was similar to the pattern of
decreases in fetal body weight in the same study, suggesting a possible link between these
particular sternebral and fontanel anomalies with fetal growth status. Given the lack of
dose-related increases in the incidences of these anomalies, and patterns that mirrored fetal
body-weight decreases (which were also not dose-related), the findings of enlarged frontal fontanel
and unossified sternebrae were not considered treatment-related. Gestational administration of
RDX to NZW rabbits did not result in any other dose- or treatment-related skeletal abnormalities.
Integration of Developmental Effects
Developmental studies in rats (Angerhofer etal.. 1986: Cholakis etal.. 1980) demonstrated
effects on offspring survival, growth, and morphological development only at doses associated with
severe maternal toxicity and mortality. No dose-related developmental effects were observed in
rabbits fCholakis etal.. 1980). As noted in EPA's Guidelines for Developmental Toxicity Risk
Assessment (U.S. EPA. 1991). where adverse developmental effects are produced only at doses that
cause minimal maternal toxicity, developmental effects should not be discounted as being
secondary to maternal toxicity; however, at doses causing excessive toxicity, as is the case with
RDX, information on developmental effects may be difficult to interpret and of limited value. At this
time, the information available to assess the association between RDX exposure and developmental
effects is considered inadequate.
1.2.5. Liver Effects
One occupational epidemiology study examined the association between RDX exposure and
changes in serum liver enzymes. Case reports involving accidental exposure to RDX provide
information on the potential for acute exposure to RDX to affect the liver in humans. In addition,
1-54
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
organ weight, histopathology, and serum chemistry findings from experimental animal studies
involving subchronic and chronic exposure to ingested RDX provide data relevant to an
examination of the association between RDX exposure and liver effects. A summary of the liver
effects associated with RDX exposure is presented in Tables 1-11 and 1-12 and Figure 1-5.
Experimental animal studies are ordered in the evidence table and exposure-response array by
duration of exposure and then by species.
Reports in humans provide inconsistent evidence of liver toxicity associated with acute
exposure to RDX. Elevated serum levels of aspartate aminotransferase (AST) and/or alanine
aminotransferase (ALT) were reported in several case reports of individuals who ingested
unknown amounts of RDX [(Kucukardali et al.. 2003: Woody etal.. 1986: Knepshield and Stone.
1972: Hollander and Colbach. 1969: Stone etal.. 1969: Merrill. 1968): see Appendix C, Section C.2],
Liver biopsies did not reveal any abnormal observations f Stone etal.. 19691. In other case reports,
no significant changes in serum levels of liver enzymes were observed (Testud etal.. 1996a: Ketel
and Hughes. 1972). In a cross-sectional epidemiologic study of workers from five U.S. Army
munitions plants [69 exposed to RDX alone and 24 to RDX and HMX; RDX exposure range:
undetectable (<0.01 mg/m3) to 1.6 mg/m3; (Hathaway and Buck. 1977)]. serum chemistry analysis
(including the serum liver enzymes AST, ALT, and alkaline phosphatase [ALP]) revealed no
statistically significant differences between exposed and unexposed workers (see Table 1-11).
In experimental animals, some, but not all, subchronic studies reported increased liver
weight associated with RDX exposure (see Table 1-12 and Figure 1-5). Dose-related increases in
relative liver weight18 (11-25% in high-dose groups) were observed in male and female B6C3Fi
mice given RDX in the diet for 90 days (Cholakis etal.. 1980) and in female F344 rats in two
separate 90-day dietary studies of RDX fLevine etal.. 1990: Levine etal.. 1981a. b; Cholakis etal..
1980): however, relative liver weights were not increased in female F344 rats in another 90-day
gavage study (Crouse et al.. 2006). Male F344 rats exhibited an increase in relative liver weight
only in one of these subchronic studies (Levine etal.. 1990: Levine etal.. 1981a. b). In subchronic
studies in other species, absolute liver weights were increased in male and female monkeys
[6-16% relative to control at 1 and 10 mg/kg-day; (Martin and Hart. 1974)] and in male, but not
female, beagle dogs [53% relative to control in male dogs at 10 mg/kg-day; fHart. 19741],
18Based on an evaluation of the relationship between organ weight and body/brain weight to determine
which endpoint (organ weight, organ-to-body weight ratio, or organ-to-brain weight ratio) is likely to detect
target organ toxicity more accurately, Bailey etal. (20041 concluded that evaluation of the effects of a test
chemical on liver weight are optimally analyzed using organ-to-body weight ratios. Therefore, the analysis of
liver weight here focuses on relative weight data where study authors reported both relative and absolute
weights, although both relative and absolute data are summarized in the evidence table (see Table 1-12).
1-55
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-11. Evidence pertaining to liver effects in humans
Reference and study design
Results
Hathawav and Buck (1977) (United States)
Cross-sectional study, 2,022 workers,
1,491 participated (74% response rate).
Analysis group: 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: Liver function 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).
Mean laboratory values of liver enzymes in men (mean; standard
deviation not reported)
Test
Referent
(n = 237)
RDX exposed*
Undetected (0.01 mg/m3
[n = 45)
LDH
173
191
174
ALP
82
78
80
AST (SGOT)
22
25
21
ALT (SGPT)
21
26
18
Bilirubin
0.5
0.4
0.4
includes both workers exposed to RDX alone and RDX and HMX.
No differences were statistically significant as reported by study
authors. Similar results in women.
Liver function tests in men (prevalence of abnormally elevated
values)
Test
(abnormal
range)
Referent
RDX exposed*
Undetected (0.01 mg/m3
LDH (>250)
2/237
1/22
0/45
ALP (>1.5)
34/237
1/22
6/45
AST (SGOT)
(>35)
20/237
4/22
2/45
ALT (SGPT)
(>35)
15/237
2/22
0/45
Bilirubin
(>1.0)
5/237
1/22
1/45
includes both workers exposed to RDX alone and RDX and HMX.
No differences were statistically significant as reported by study
authors. Similar results in women.
ALP = alkaline phosphatase; ALT = alanine aminotransferase; AST = aspartate aminotransferase; LDH = lactate
dehydrogenase; LOD = limit of detection; SGOT = glutamic oxaloacetic transaminase; SGPT = glutamic pyruvic
transaminase.
1-56
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-12. Evidence pertaining to liver effects in animals
Reference and study design
Results
Liver weight
Lish et al. (1984)
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
wk 11 due to excessive mortality)
Diet
2 yr
Doses
0
1.5
7.0
35
175/100
Absolute liver weight at 104 wk (percent change compared to control)
M
0%
28%*
11%
12%
35%*
F
0%
7%
7%
15%
18%*
Relative liver weight at 104 wk (percent change compared to control)
M
0%
32%*
12%
14%
46%*
F
0%
6%
8%
18%
45%*
Note: Percent change in liver weights of male and female mice was reduced
in all dose groups when mice with liver tumors were removed from the
analysis, suggesting no real effect on liver weight.
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
Absolute liver weight (percent change compared to control)
M
0%
-6%
-6%
-6%
F
0%
7%
-11%
1%
Relative liver weight (percent change compared to control)
M
0%
-5%
-2%
-3%
F
0%
17%
-2%
13%
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 liver weight at 105 wk (percent change compared to control)
M
0%
3%
-7%
1%
-8%
F
0%
1%
-4%
3%
0%
Relative liver weight at 105 wk (percent change compared to control)
M
0%
1%
0%
2%
11%
F
0%
1%
-2%
6%
18%*
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 liver weight (percent change compared to control)
M
0%
-
-
-
-6%
-5%
F
0%
-
-
-
-4%
-1%
Relative liver weight (percent change compared to control)
M
0%
-
-
-
-4%
-4%
F
0%
-
-
-
-6%
1%
1-57
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-12. Evidence pertaining to liver effects in animals (continued)
Reference and study design
Results
Experiment 2: 0, 40, 60, or 80 mg/kg-d
for 2 wks 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
Absolute liver weight (percent change compared to control)
M
0%
2%
12%
26%*
F
0%
4%
9%
29%*
Relative liver weight (percent change compared to control)
M
0%
0%
9%
25%*
F
0%
4%
4%
22%*
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
Absolute liver weight (percent change compared to control)
M
0%
-
-
-
-2%
-5%
F
0%
-
-
-
6%
4%
Relative liver weight (percent change compared to control)
M
0%
-
-
-
2%
3%
F
0%
-
-
-
10%
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 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
Absolute liver weight (percent change compared to control)
M
0%
7%
-16%
-
F
0%
0%
-14%
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 liver weight (percent change compared to control)
M
0%
-6%
-9%
0%
7%
5%
F
0%
1%
7%
18%*
15%
28%*
Relative liver weight (percent change compared to control)
M
0%
0%
-1%
2%
5%
2%
F
0%
1%
-2%
2%
-3%
2%
1-58
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-12. Evidence pertaining to liver effects in animals (continued)
Reference and study design
Results
Levine et al. (1990); Levine et al.
(1981a); Levine etal. (1981b)b
Rats, F344, 3-4 wk old; 10/sex/group;
30/sex/group for controls
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
Absolute liver weight (percent change compared to control)
M
0%
5%
-1%
-2%
-
-
F
0%
2%
4%
16%*
-
-
Relative liver weight (percent change compared to control)
M
0%
9%
6%
20%
-
-
F
0%
3%
5%
19%*
-
-
Hart (1974)c
Dogs, beagle, 3/sex/group
Premix with ground dog chow
containing 20 mg RDX/g-chow, 60 g
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 liver weight (percent change compared to control)
M
0%
-
-
53%
F
0%
3%
Martin and Hart (1974)c
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
Absolute liver weight (percent change compared to control)
M + F
0%
2%
6%
16%
Histopathological lesions
Lish et al. (1984)
Mice, B6C3Fi, 85/sex/group; interim
sacrifices (10/sex/group) at 6 and
12 mo
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
wk 11 due to excessive mortality)
Diet
2 yrs
Histopathological lesions in liver other than adenomas and carcinomas
were not significantly different compared to controls, as reported by study
authors.
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 yrs
Histopathological examination performed only for controls and 10 mg/kg-d
rats; no significant differences compared to controls were reported by
study authors.
1-59
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-12. Evidence pertaining to liver effects in animals (continued)
Reference and study design
Results
Levine et al. (1983)
Rats, F344, 3-4 wk old; 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
Microgranulomas (incidence)
M
0/38
0/36
0/25
0/29
0/4
F
10/43
19/45
12/42
17/41
4/28
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 wks
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
Liver microgranulomas; mild (incidence)
M
2/10
-
-
1/9
F
2/11
-
-
7/11*
Increased karyomegaly of hepatocytes (incidence)
M
0/10
-
-
5/9*
F
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
Liver granulomas; mild (incidence)
M
0/10
-
-
-
-
1/10
F
-
-
-
-
-
-
Liver portal inflammation (incidence)
M
2/10
-
-
-
-
3/10
F
1/10
-
-
-
-
7/10
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
Histopathology examination of the 15-mg/kg-d group showed one male
with mild liver congestion and one female with a moderate-sized focus of
basophilic cytoplasmic alteration; neither finding was attributed by study
authors to RDX treatment.
1-60
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-12. Evidence pertaining to liver effects in animals (continued)
Reference and study design
Results
Levine et al. (1990); Levine et al.
(1981a); Levine etal. (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
Histopathological examination of liver did not reveal any significant
differences compared to controls, as reported by study authors. No
histopathology findings available for the 300 or 600 mg/kg-day dose groups
because all rats in these groups died before the 13-wk necropsy.
Hart (1974)
Dogs, beagle, 3/sex/group
Premix with ground dog chow
containing 20 mg RDX/g-chow, 60 g
dog food; purity and particle size not
specified
0, 0.1,1, or 10 mg/kg-d
Diet
13 wk
Histopathological examination performed only for controls and 10 mg/kg-d
dogs; no significant differences compared to controls were reported.
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
An increase in the amount of iron-positive material in liver cord cytoplasm
was reported in monkeys treated with 10 mg/kg-d RDX, which the study
authors considered to be of uncertain toxicological significance. Because
iron-positive stain was present in controls and no further characterization
of the staining was provided in the study report, the toxicological
significance of this finding could not be determined.
Serum chemistry
Lish et al. (1984)
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
wk 11 due to excessive mortality)
Diet
2 yr
Doses
0
1.5
7.0
35
175/100
Serum cholesterol at 105 wk (percent change compared to control)
M
0%
11%
-11%
5%
39%
F
0%
5%
15%
25%
38%
Serum triglycerides at 105 wk (percent change compared to control)
M
0%
21%
-20%
10%
-25%
F
0%
34%
28%
41%
28%
1-61
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-12. Evidence pertaining to liver 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
Serum cholesterol at 104 wk (percent change compared to control)
M
0%
15%
38%
19%
-6%
F
0%
6%
3%
-7%
-9%
Serum triglycerides at 104 wk (percent change compared to control)
M
0%
14%
-15%
-12%
-52%
F
0%
18%
5%
-42%
-51%*
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
Serum cholesterol (percent change compared to control)
M
0%
-3%
-10%*
-16%*
-18%*
-11%*
F
0%
-1%
-8%
-4%
-4%
-1%
Serum triglycerides (percent change compared to control)
M
0%
1%
1%
-7%
-2%
-19%
F
0%
-16%
-21%
7%
-37%
18%
Levine et al. (1990); Levine et al.
(1981a); Levine etal. (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
Doses
0
10
30
100
300
600
Serum triglyceride levels (percent change compared to control)
M
0%
-14%
-34%
-62%*
-
-
F
0%
-12%
-29%
-50%*
1-62
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-12. Evidence pertaining to liver effects in animals (continued)
Reference and study design
Results
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
Serum biochemistry analysis revealed scattered deviations, but study
authors indicated they appear to have no toxicological significance.
Doses
0
0.1
1
10
Serum cholesterol (percent change compared to control)
M
0%
-17%
-2%
-7%
F
0%
7%
7%
7%
F = female; M = male; TWA = time-weighted average.
Note: A dash indicates that the study authors did not measure or report a value for that dose group.
^Statistically significant (p < 0.05) based on analysis by study authors.
aDoses were calculated by the study authors.
bLevine et al. (1981a) is a laboratory report of a 13-wk 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.
cLiver-weight data from the Hart (1974) and Martin and Hart (1974) studies were considered less informative than
other studies. Hart (1974) reported organ-weight data for high-dose dogs (3/sex/group) only, and the liver weights
from Martin and Hart (1974) were highly variable across monkeys (e.g., liver weights for the control animals ranged
from 46 to 141 g). Therefore, liver-weight data from these two studies were not presented in the
exposure-response array for liver effects (see Figure 1-5).
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).
1-63
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
1000
100 ,
lie
~W3
£
O 1
a 1
o.i
1 significantly changed o not significantly changed x not determined
4 © a
6 (i
CD —
fD
X
on
oo
en
Chronic
o
oo
cr,
o
JZ
o
3
c
E
o
00
cn
o
.c
o
o
00
m
c
_c
u
o
ffr
0/
c
o
cn
cn
Subchronic
1s Relative liver weight
Chronic
3
O
E
o
00
CTi
O
u
3
o
E,
o
00
cr.
g tf
o
in
cri
x
T3
Subchronic
Histo pathology
&
o
£
Chronic
VO
8
(N
I
¦o
Subchronic
4, Cholesterol
Chronic
o
L-.
Subchr
•I Trigylcerides
Serum biochemistry changes
Note: Filled circle indicates that response was statistically significantly different from the control.
X - Not considered due to confounding caused by presence of tumors.
Figure 1-5. Exposure response array of liver effects following oral exposure.
1-64
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Chronic RDX exposures in B6C3Fi mice and F344 or Sprague-Dawley rats showed a less
consistent pattern of liver-weight increases. Interpretation of liver-weight increases in the 2-year
mouse study is complicated by the incidence of adenomas and carcinomas in each dose group; the
apparent increase in liver weights in male and female mice exposed to RDX in diet fLish etal.. 19841
was reduced when mice with liver adenomas or carcinomas were removed from the analysis. In a
2-year rat study (Levine etal.. 19831. relative liver weights were increased in high-dose
(40 mg/kg-day) males and females (by 11 and 18% compared to controls, respectively), likely
reflecting the depressed weight gain in the high-dose rats (2-30% in males and 10-15% in females).
In evaluating organ-weight data across studies of all durations, less emphasis is placed on evidence
of organ-weight changes from chronic (2-year) studies because normal physiological changes
associated with aging and intercurrent disease contributes to interanimal variability that could
confound organ-weight interpretation fSellers etal.. 20071. as is true of the mouse liver-weight data
for RDX.
Nonneoplastic histopathological changes in the liver were not associated with RDX
exposure in the majority of experimental animal studies (Crouse etal.. 2006: Levine etal.. 1990:
Lish etal.. 1984: Levine etal.. 1983: Levine etal.. 1981a. b; Hart. 1976.1974: Martin and Hart.
19741. including 2-year oral studies in mice at doses up to 100 mg/kg-day fLish etal.. 19841 and in
rats at doses up to 40 mg/kg-day (Levine etal.. 1983). The few findings of liver lesions were
reported in studies with more limited histopathological analyses, and were not confirmed in the
studies with more complete histopathologic examination and longer exposure durations (Lish etal..
1984: Levine etal.. 1983). For example, the incidence of liver portal inflammation was increased in
female rats, but not male rats, exposed to 40 mg/kg-day in the diet for 90 days (Cholakis etal..
19801. There was an increase in the incidence of mild liver microgranulomas in female mice only
(Cholakis etal.. 1980) and kaiyomegaly of hepatocytes in male mice only exposed to
320 mg/kg-day RDX in the diet for 90 days (Cholakis etal.. 1980). Because both the rat and mouse
studies by Cholakis etal. (1980) used relatively small group sizes (n = 10/sex/group) and provided
histopathologic findings for the control and high-dose groups only, less emphasis is placed on these
findings than on those from the 2-year bioassays. Note that exposure to HMX, the primary
contaminant in several of the RDX studies, was associated with histopathological changes in the
livers of male rats fed doses >450 mg/kg-day for 13 weeks. However, similar findings were not
observed in the RDX studies, where the doses of RDX employed in the studies would have resulted
in HMX exposures of <60 mg/kg-day. The contribution of HMX exposure to the overall liver
findings in the studies of RDX toxicity is therefore expected to be negligible.
Clinical chemistry parameters, including serum ALT, AST, and ALP, showed no
treatment-related changes indicative of liver toxicity. Statistically significant changes in these
parameters in some subchronic and chronic toxicity studies in rats and mice were relatively small
(generally <50% of the control mean), were not dose related in most instances, and showed no
consistent pattern of change between sexes or across studies.
1-65
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Some subchronic and chronic oral toxicity studies in rats and mice reported dose-related
changes in serum cholesterol and triglyceride levels; however, these changes were not consistently
observed in males and females within the same study, and patterns of changes were not consistent
across studies. Specifically, serum triglyceride levels were elevated (up to 41%) in female B6C3Fi
mice exposed to RDX in the diet for 2 years, although increases were not dose related (Lish etal..
19841: male mice in the same study did not show a similar increase in triglycerides. In contrast,
serum triglycerides showed dose-related decreases in male and female F344 rats (50-62% at the
high doses) in a subchronic oral (dietary) study (Levine etal.. 1990: Levine etal.. 1981a. b). In a
chronic toxicity study by the same investigators (Levine etal.. 1983). serum triglyceride levels were
generally decreased in male and female rats (52 and 51%, respectively, at the highest dose of
40 mg/kg-day); however, triglyceride levels across the four dose groups in this study did not show
a dose-related response.
Serum cholesterol levels showed a dose-related increase (38% at the high dose of
100 mg/kg-day) in female B6C3Fi mice exposed to RDX in the diet for 2 years (Lish etal.. 1984):
however, changes in cholesterol in male mice in the same study were not dose related. Changes in
serum cholesterol in male and female F344 rats exposed to RDX in the diet for 2 years at doses up
to 40 mg/kg-day (Levine etal.. 1983). in rats exposed to RDX by gavage for 90 days at doses up to
15 mg/kg-day fCrouse et al.. 20061. and in monkeys exposed to RDX in the diet for 90 days fMartin
and Hart. 1974) were relatively small (within 38% of control mean) and were not dose related.
Integration of Liver Effects
There is limited evidence from human studies and from studies in experimental animals
that RDX may affect the liver. The observation of transient elevations of serum liver enzymes in
several human case reports of individuals who ingested unknown amounts of RDX suggests that
RDX might target the liver; however, serum liver enzymes were not elevated in a small
cross-sectional study of munition plant workers exposed to RDX. In experimental animals,
dose-related increases in liver weight were observed in some studies following subchronic oral
exposure, but liver-weight changes were not consistent across sexes within a study or across
different studies. Changes in serum chemistry were not consistent across studies and the
magnitude of change relative to concurrent controls was not indicative of liver damage.
Nonneoplastic histopathologic lesions of the liver were also not consistently associated with RDX
exposure. At this time, the available data do not support liver effects as a human hazard of RDX
exposure.
1.2.6. Other Noncancer Effects
There are some reports that RDX may induce effects on the eyes, on the cardiovascular,
musculoskeletal, immune, GI, hematological, and male reproductive systems, and on body weight.
However, there is less evidence for these effects compared to organ systems described earlier in
Section 1.2. Generally, human evidence for effects in these organ systems is limited to human case
1-66
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
reports. Evidence of effects in experimental animals is generally inconsistent across studies of
similar duration in the same species, or lacks consistent, dose-related patterns of increasing or
decreasing effect. A summary of the evidence for an association between these other noncancer
effects and RDX exposure is provided below; a more detailed discussion is provided in Appendix C,
Section C.3.2. As discussed below, the information to assess the association between RDX exposure
and toxicity for the organ systems presented below is considered inadequate.
Ocular Effects
There is no human evidence of ocular effects following exposure to RDX. In animals, the
incidence of cataracts was significantly increased in high-dose female rats (73%) relative to
controls (32%) in one chronic oral study (Levine etal.. 1983). This finding was not observed in
males in the same chronic study or in other chronic or subchronic studies in rats, mice, or monkeys
(Crouse etal.. 2006: Lish etal.. 1984: Cholakis etal.. 1980: Martin and Hart. 1974). There is
insufficient information to assess ocular toxicity following exposure to RDX.
Cardiovascular Effects
Human evidence of cardiovascular effects consists of case reports of transient arterial
hypertension, sinus tachycardia, and premature ventricular beats in male workers or men who
accidentally ingested RDX (Kuctikardali et al.. 2003: Barsotti and Crotti. 1949). In animals, evidence
is limited to inconsistent findings of changes in heart weight and a report of increased incidence of
minimal histopathological changes in a 90-day rat study at a dose that also produced 40% mortality
(Cholakis etal.. 1980). There is insufficient information to assess cardiovascular effects following
exposure to RDX.
Musculoskeletal Effects
Evidence for musculoskeletal effects in humans is limited to case reports that described
muscle twitches, soreness, and muscle injury as indicated by elevated levels of AST, creatine
phosphokinase, and myoglobinuria (Testud etal.. 2006: Kuctikardali etal.. 2003: Hett and Fichtner.
2002: Hollander and Colbach. 1969: Stone etal.. 1969: Merrill. 1968). In animal studies,
evaluations of muscle and skeletal tissues did not reveal any histopathological alterations in rats or
mice following chronic exposure or in mice, rats, or dogs following subchronic exposure. There is
insufficient information to assess musculoskeletal effects following exposure to RDX.
Immune System Effects
Increased white blood cell (WBC) counts were reported in several case reports of humans
acutely exposed to RDX (Knepshield and Stone. 1972: Hollander and Colbach. 1969: Stone etal..
1969: Merrill. 1968). In animals, there were no consistent patterns of change in WBC count or
spleen weight across the RDX database. No dose-related immune effects were observed in a 90-day
study in F344 rats that evaluated structural measures of immunotoxicity [including red blood cell
1-67
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
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]. None of the available studies included evaluation of more sensitive measures of
functional immune system changes. Therefore, there is insufficient information to assess
immunotoxicity following exposure to RDX.
Gastrointestinal Effects
Nausea, vomiting, and erosive gastroduodenitis were identified in human case reports of
RDX poisonings, generally concurrent with severe neurotoxicity fKasuske et al.. 2009: Davies et al..
2007: Kuctikardali etal.. 2003: Hett and Fichtner. 2002: Ketel and Hughes. 1972: Knepshield and
Stone. 1972: Hollander and Colbach. 1969: Stone etal.. 1969: Merrill. 1968: Kaplan etal.. 1965:
Barsotti and Crotti. 19491. There have been similar reports of vomiting in swine, dogs, and
monkeys (Musick etal.. 2010: Hart. 1974: Martin and Hart. 19741. Generally, histopathological
changes of the GI tract were not observed in RDX-exposed animals. There is insufficient
information to assess gastrointestinal toxicity following exposure to RDX.
Hematological Effects
Temporary hematological alterations, including anemia, decreased hematocrit, hematuria,
and methemoglobinemia were observed in some human case reports following acute exposure
fKasuske etal.. 2009: Kuctikardali et al.. 2003: Knepshield and Stone. 1972: Hollander and Colbach.
1969: Stone etal.. 1969: Merrill. 19681. Observations of anemia in case reports may reflect
coexposure to 2,4,6-TNT. Levine and colleagues identified that anemia resulted from exposure to
TNT in F344 rats, but not RDX (Levine etal.. 1990: Levine etal.. 1981a. b). Hematological findings
in a case-control and cross-sectional occupational study were inconsistent (West and Stafford.
1997: Hathaway and Buck. 1977): both studies used small sample sizes and were considered
low-confidence studies. In general, subchronic and chronic animal studies showed no consistent
dose-related patterns of change in hematological parameters. There is insufficient information to
assess hematological toxicity following exposure to RDX.
Reproductive Effects
Investigation of the potential effects of RDX on reproductive function is limited to a
two-generation study in rats by Cholakis etal. (1980) that also included a dominant lethal mutation
study. A reduction in number of pregnancies was reported only at a dose that also resulted in
decreased food consumption, decreased body-weight gain, and increased mortality. The limited
investigation of reproductive function in RDX-exposed rats by a single investigator provides
insufficient information to assess female reproductive toxicity following exposure to RDX.
Evidence of male reproductive toxicity comes largely from the finding of increased
incidence of testicular degeneration in male B6C3Fi mice exposed to >35 mg/kg-day RDX for
2 years in the diet compared to concurrent controls (Lish etal.. 1984). The biological significance of
1-68
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
this finding is unclear because no similar histopathological changes were observed in this study at 6
or 12 months, durations longer than the 1.4-month duration of spermatogenesis in mice, and
because of the loss of testicular function that occurs in aging rodents. The evidence fo r testicular
degeneration in mice suggested by Lish etal. (19841 was generally not supported by
histopathological findings in male reproductive organs in other studies, and changes in testes
weight across the RDX database were generally small, not dose-related, and directionally
inconsistent There is insufficient information to assess male reproductive toxicity following
exposure to RDX.
Body-Weight Effects
Changes in body-weight gain were reported in experimental animal studies involving
chronic and subchronic exposure to ingested RDX, but generally at high doses that were also
associated with elevated mortality or with severe kidney and urinary bladder toxicity in male rats
as seen in the Levine etal. (19831 study. 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. Thus, available studies of RDX provide evidence that RDX exposure causes
decreases in body-weight gain in mice and rats, but these effects appear to be secondary to effects
on other primary targets of RDX toxicity.
1.2.7. Carcinogenicity
The relationship between exposure to RDX and cancer has not been investigated in human
populations. The carcinogenicity of RDX has been examined in one oral chronic/carcinogenicity
bioassay in mice (Lish etal.. 19841 and two bioassays in rats (Levine etal.. 1983: Hart. 19761. The
2-year studies by Lish etal. (1984) and Levine etal. (1983) included comprehensive
histopathological examination of major organs, multiple dose groups and a control, and
>50 animals/dose group (plus additional interim sacrifice groups). In both studies, the maximum
tolerated dose was reached or exceeded in high-dose animals (based on decreased terminal body
weight in high-dose male and female mice of 5 and 19%, respectively, and decreased survival in
male and female rats by approximately 50 and 25%, respectively, compared to the control).19 The
earlier Hart (1976) study is largely limited by the lack of characterization of the test material and
histopathologic examination in control and high-dose groups only. A temperature spike in the
animal rooms on study Day 76 resulted in significant mortality across all dose groups and control
animals; however, there were still >80 rats/sex/group after the overheating incident and
>50 rats/sex/group at termination, and it seems unlikely that the mortality associated with the
19In high-dose mice in the l.isli etal. (19841 study, reduced survival due to acute RDX toxicity occurred during
the first 11 weeks on study at a dietary dose of 175 mg/kg-day; survival in high-dose animals was similar to
controls after 11 weeks when the dose was reduced to 100 mg/kg-day. By contrast, in high-dose rats in the
Levine etal. (19831 study, elevated mortality, particularly in males, occurred gradually over the entire period
of the study beginning after 6 months, and was attributable in large part to kidney toxicity.
1-69
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
temperature spike would have affected a tumor response in the rats. A peer review of
histopathological evaluations by the study pathologist was performed only for female mouse liver
tissues from the Lish etal. (19841 study (see discussion of the Pathology Working Group [PWG]
below). A summary of the evidence for liver and lung tumors in experimental animals from these
three bioassays is provided in Tables 1-13 and 1-14.
Liver Tumors
An increased incidence of liver tumors was observed in one chronic mouse study (Lish etal..
19841 and one chronic rat study (Levine etal.. 19831. Incidences of hepatocellular tumors are
presented in Table 1-13 and discussed in further detail below.
The incidence of hepatocellular carcinomas and the combined incidence of hepatocellular
adenomas or carcinomas showed a statistically significant positive trend with RDX dose in female,
but not male, B6C3Fi mice as compared to concurrent controls in a 2-year dietary study (Lish etal..
19841. In female B6C3Fi mice, Lish etal. (19841 observed that the liver tumor incidence in the
concurrent female control mice was relatively low (1/65), and significantly lower than the
incidence from historical controls (historical incidence data not provided by study authors). The
study authors also compared liver tumor incidence in RDX-exposed female mice to mean historical
control incidence for female mice of the same strain from National Toxicology Program (NTP)
studies conducted during the same time period [147/1,781 or 8%; range: 0-20%; (Haseman etal..
19851]20 The combined incidence of hepatocellular adenomas or carcinomas in female mice at RDX
doses >35 mg/kg-day (19% at both doses) was statistically significantly elevated when statistical
analysis was performed using NTP historical control data; limitations associated with comparisons
to historical control data originating from a different laboratory are acknowledged given
cross-study differences in diet, laboratory, pathological evaluation, and animal provider.
20Comparison of control incidences of hepatocellular adenomas or carcinomas between l.isli etal. (19841 and
Haseman etal. (19851 must be interpreted with caution because of cross-study differences in labs, diets, and
sources of animals. Specifically, the labs used by NTP and analyzed by Haseman etal. (19851 did not include
the lab contracted to perform the Lish etal. (19841 study, and it is not clear if the diet used in the Lish et al.
(19841 study was included in the diets reported in the NTP studies. Further, the NTP studies included three
different suppliers of mice; one supplier was also used in the Lish etal. (19841 study. EPA Guidelines for
Carcinogenic Risk Assessment (U.S. KPA. 2005al also note that, unless the tumor is rare, the standard for
determining statistical significance of tumor incidence is a comparison of dosed animals with the concurrent
controls.
1-70
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-13. Liver tumors observed in chronic animal bioassays
Reference and study design
Results3
Lish et al. (1984)
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 wk 11
due to excessive mortality)
Diet
2 yr
Doses
0
1.5
7.0
35
175/100b
Hepatocellular adenomas [incidence (%)]c
M
8/63
(12.7)
6/60
(10.0)
1/62*
(1.6)
7/59
(11.9)
7/27
(25.9)
F
1/65
(1.5)
1/62
(1.6)
6/64
(9.4)
6/64
(9.4)
3/31
(9.7)
Hepatocellular carcinomas [incidence (%)]c
M
13/63
(20.6)
20/60
(33.3)
16/62
(25.8)
18/59
(30.5)
6/27
(22.2)
F
0/65
(0.0)
4/62
(6.5)
3/64
(4.7)
6/64
(9.4)
3/31d
(9.7)
Hepatocellular adenoma or carcinoma combined [incidence (%)]c
M
20/63
(31.7)
26/60
(43.3)
17/62
(27.4)
25/59
(42.4)
13/27
(48.1)
F
1/65
(1.5)
5/62
(8.1)
9/64*
(14.1)
12/64*
(18.8)
6/31*d
(19.4)
PWG reanalysis of liver lesion slides from female mice (Parker et al.,
2006; Parker, 2001).e
Doses
0
1.5
7.0
35
175/100b
Hepatocellular adenomas [incidence (%)]c
F
1/67
(1.5)
3/62
(4.8)
2/63
(3.2)
8/64
(12.5)
2/31
(6.5)
Hepatocellular carcinomas [incidence (%)]c
F
0/67
(0.0)
1/62
(1.6)
3/63
(4.8)
2/64
(3.1)
2/31
(6.5)
Hepatocellular adenoma or carcinoma combined [incidence (%)]c
F
1/67
(1.5)
4/62
(6.5)
5/63
(7.9)
10/64
(15.6)
4/31d
(12.9)
1-71
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-13. Liver tumors observed in chronic animal bioassays (continued)
Reference and study design
Results3
Hart (1976)
Doses
0
1.0
3.1
10
Rats, Sprague-Dawley, 100/sex/group
Purity and particle size not specified
0,1.0, 3.1, or 10 mg/kg-d
Diet
Neoplastic nodules [incidence (%)]c
M
0/82
(0)
-
-
3/77
(3.9)
2 yr
F
1/72
(1.4)
-
-
1/81
(1.2)
Hepatocellular carcinomas [incidence (%)]c
M
1/82
(1.2)
-
-
1/77
(1.3)
F
1/72
(1.4)
-
-
l/81f
(1.2)
Neoplastic nodules or hepatocellular carcinomas combined [incidence
(%)T
M
1/82
(1.2)
-
-
4/77
(5.2)
F
2/72
(2.8)
-
-
2/81
(2.5)
1-72
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-13. Liver tumors observed in chronic animal bioassays (continued)
Reference and study design
Results3
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
Neoplastic nodules [incidence (%)]c
M
4/55
3/55
0/52
2/55
1/31
contaminant; 83-89% of particles <66 nm
(7.3)
(5.5)
(0.0)
(3.6)
(3.2)
0, 0.3,1.5, 8.0, or 40 mg/kg-d
Diet
2 yr
F
3/53
1/55
1/54
0/55
4/48
(5.6)
(1.8)
(1.9)
(0.0)
(8.3)
Hepatocellular carcinomas [incidence (%)]c
M
1/55
0/55
0/52
2/55
2/3 ld
(1.8)
(0.0)
(0.0)
(3.6)
(6.5)
F
0/53
1/55
0/54
0/55
0/48
(0.0)
(1.8)
(0.0)
(0.0)
(0.0)
Neoplastic nodules or hepatocellular carcinomas combined [incidence
(%)]'-
M
5/55
3/55
0/52
4/55
3/31
(9.1)
(5.5)
(0.0)
(7.3)
(9.7)
F
3/53
2/55
1/54
0/55
4/48
(5.6)
(3.6)
(1.9)
(0.0)
(8.3)
F = female; M = male.
Note: A dash indicates that the study authors did not measure or report a value for that dose group.
^Statistically significant difference compared to the control group (p < 0.05), identified by the study authors.
Selected percent incidences are provided in parentheses below the incidences to help illustrate patterns in the
responses.
The lower dose of 100 mg/kg-day was started in Week 11, resulting in a duration-weighted average dose of
107 mg/kg-day.
The incidences reflect the animals surviving to Month 12.
dStatistically significant trend (p < 0.05) was identified using a one-sided Cochran-Armitage trend tests performed
by EPA.
The numbers of animals at risk (i.e., the denominators) in the control group (n = 67) and 7 mg/kg-day dose group
(n = 63) as reported in the PWG reanalysis (Parker et al., 2006; Parker, 2001) differed from the numbers reported
in the original study by Lish et al. (1984) (n = 65 and 64, respectively). Further investigation of these differences by
the U.S. Army (sponsor of the mouse bioassay and subsequent PWG reevaluation) was unable to resolve the
discrepancy (email to Louis D'Amico, EPA, from Mark Johnson, U.S. Army Public Health Command, February 13,
2015).
fHart (1976) distinguishes the single high-dose carcinoma in the liver from a hepatocellular carcinoma; the
incidence of hepatocellular carcinomas in this dose group is shown as 0/81 (p. 119 of the publication).
1-73
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-14. Lung tumors observed in chronic animal bioassays
Reference and study design
Results3
Lish et al. (1984)
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 wk 11
due to excessive mortality)
Diet
2 yr
Doses
0
1.5
7.0
35
175/100b
Alveolar/bronchiolar adenomas [incidence (%)]'-
M
6/63
(9.5)
5/60
(8.3)
5/62
(8.1)
7/59
(11.9)
1/27
(3.7)
F
4/65
(6.2)
2/62
(3.2)
5/64
(7.8)
9/64
(14.1)
3/31
(9.7)
Alveolar/bronchiolar carcinomas [incidence (%)]c
M
3/63
(4.8)
6/60
(10.0)
3/62
(4.8)
7/59
(11.9)
5/27d
(18.5)
F
3/65
(4.6)
1/62
(1.6)
3/64
(4.7)
3/64
(4.7)
4/3 ld
(12.9)
Alveolar/bronchiolar adenoma or carcinoma combined [incidence
(%)f
M
9/63
(14.3)
11/60
(18.3)
8/62
(12.9)
14/59
(23.7)
6/27
(22.2)
F
7/65
(10.8)
3/62
(4.8)
8/64
(12.5)
12/64
(18.8)
7/3 ld
(22.6)
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
Alveolar/bronchiolar adenoma [incidence (%)]
M
2/83
-
-
1/77
F
0/73
-
-
0/82
No alveolar/bronchiolar carcinomas reported by study authors.
1-74
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 1-14. Lung tumors observed in chronic animal bioassays (continued)
Reference and study design
Results3
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
Alveolar/bronchiolar adenomas [incidence (%)]'-
M
1/55
0/15
1/17
0/16
1/31
F
3/53
0/7
0/8
1/10
0/48
Alveolar/bronchiolar carcinomas [incidence (%)]c
M
-
-
-
-
-
F
0/53
0/7
1/8
0/10
0/48
Alveolar/bronchiolar adenoma or carcinoma combined [incidence
(%)]c
M
-
-
-
-
-
F
3/53
0/7
1/8
1/10
0/48
F = female; M = male.
Note: A dash indicates that the study authors did not measure or report a value for that dose group.
Selected percent incidences are provided in parentheses below the incidences to help illustrate patterns in the
responses.
The lower dose of 100 mg/kg-day was started in Week 11, resulting in a duration-weighted average dose of
107 mg/kg-day.
The incidences reflect the animals surviving to Month 12.
dStatistically significant trend (p < 0.05) was identified using a one-sided Cochran-Armitage trend test performed
by EPA.
A PWG reviewed the slides of female mouse liver lesions from the Lish etal. (19841 study
(Parker etal.. 2006: Parker. 20011. Some malignant tumors were downgraded to benign status, and
several lesions initially characterized as adenomas were changed to nonneoplastic lesions based on
more recent diagnostic criteria used by the PWG (Harada et al.. 19991. There remained a
statistically significant positive trend in the combined incidence of hepatocellular adenomas or
carcinomas, consistent with the original findings of Lish etal. (1984). Because the PWG analysis
reflects more recent histopathological criteria for the grading of tumors, the incidence of
hepatocellular adenomas or carcinomas as reported by Parker etal. (20061 were considered the
more appropriate measure of liver tumor response in female mice from the Lish etal. (19841
bioassay. The PWG also offered observations about the histopathology methods used by Lish et al.
(1984) that raised some concerns about the uniformity of histologic processing (Parker et al.. 2006:
Parker. 2001). These concerns included variation in size and shape of sections, suggesting that
liver sections were not uniformly taken from the same area of the liver of all animals; only one liver
section present from most animals (two sections are commonly examined in current
carcinogenicity bioassays); and more than one section prepared for 20 mice across different
1-75
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
groups, raising some concern of sample bias but likely reflecting sections taken from visible lesions
at gross necropsy.
In male mice from the Lish etal. (19841 study, the incidences of hepatocellular carcinomas
in treated groups were higher than in the control, and the combined incidences of hepatocellular
adenomas or carcinomas of male mice were higher in three of four treated groups than in the
control; however, there were no statistically significant trends in either case. The incidences of
liver carcinoma in control (21%) and treated groups of male mice (22-33%) were generally within
the range for the same mouse strain reported by NTP [8-32%; (Haseman et al.. 1985)]. Similarly,
the combined incidences of liver adenoma or carcinoma in control (32%) and treated groups
(27-48%) were within the range for the same mouse strain reported by NTP [14-58%; (Haseman
etal.. 1985)].21 The PWG did not reanalyze liver tumor slides from male mice; the SAB (2017)
noted this as unusual because sections from both male and female animals are typically reevaluated
to ensure that findings in both sexes are reliable.
In the 2-year bioassay in F344 rats (Levine etal.. 1983). RDX was not associated with
dose-related increases in the incidence of nonmalignant liver tumors (neoplastic nodules) or
combined incidence of liver neoplastic nodules or carcinomas.22 However, a statistically significant
positive trend with dose was observed in the incidence of hepatocellular carcinomas in male, but
not female, F344 rats (Levine etal.. 1983). In the Levine etal. (1983) study, only a few tumors were
observed in the exposed groups of male rats (0/55, 0/52, 2/55, 2/31) relative to the control (1/55),
and inferences made from such a sparse response are uncertain. Because hepatocellular
carcinomas are rare tumors in the rat,23 some perspective is obtained by considering historical
control data. In a paper published concurrently with the Levine etal. (1983) study, NTP reported
an incidence of liver carcinomas in untreated control male F344 rats of 0.7% [12/1,719; range:
0-2%; (Haseman etal.. 1985)]. In Levine etal. (1983). the incidence of liver carcinomas in control
male rats (1/55 or 1.8%) was at the upper end of this NTP range, and the incidence in RDX-treated
male F344 rats in the highest two dose groups (3.6 and 6.4%) exceeded the NTP historical control
range. Using incidence data from NTP historical controls, the trend for carcinoma in the
RDX-treated F344 rats was statistically significant (p-value = 0.003; one-sided exact
Cochrane-Armitage trend test). It should be noted that although the NTP historical controls
(Haseman et al.. 1985) are comparable with Levine et al. (1983) in terms of the time period, they
may not be directly comparable in terms of diet, laboratory, pathological evaluation, and animal
provider. However, other historical control data sets from male F344 rats, both recent and of the
^Considerations listed in footnote 20 apply to the comparison of combined liver adenoma and carcinoma
incidence to historical controls as well.
22The incidence of neoplastic nodules of 7.3% in control male rats in Levine etal. (1983) was consistent with
the NTP historical control range of 0-12% [mean: 3.5% or 61/1,719; (Haseman et al.. 1985)].
23NTP historical control data for hepatocellular carcinomas in F344 rats as reported in Haseman etal. (1985):
12/1,719 (0.7%) in males; 3/1,766 (0.17%) in females. Historical control data for Charles River
Sprague-Dawley rats as reported in Chandra et al. (1992): 6/1,340 (0.45%) in males; 1/1,329 (0.08%) in
females.
1-76
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
time period of the Levine et al. study, indicate similar low incidences of liver carcinomas [0.36%,
(NTP. 2009): 0.31%, (Maita et al.. 1987)]. In the Levine etal. (1983) study, mortality in the highest
dose group was substantially higher than in the other dose groups during the second year leading
to uncertainty in the true cancer incidence in the high-dose group. It was not possible to estimate
mortality-adjusted incidences because no time-to-death information was available.
In a second 2-year dietary study using a different rat strain (Sprague-Dawley), the
combined incidence of hepatocellular adenomas or carcinomas was not increased with dose in rats
of either sex at doses up to 10 mg/kg-day (Hart. 1976). However, interpretation of results from this
study is limited by the comparatively lower doses employed in the study, and the recording of
effects only at the control and high-dose groups.
Lung Tumors
Lung tumors were observed in female and male B6C3Fi mice exposed to RDX in the diet for
2 years [(Lish etal.. 1984): see Table 1-14], Incidence of alveolar/bronchiolar carcinomas and the
combined incidence of alveolar/bronchiolar adenomas or carcinomas showed a statistically
significant positive trend (one-sided p-values of 0.016 and 0.009, respectively, for the
Cochran-Armitage trend test) in female mice. Incidence of alveolar/bronchiolar carcinomas in male
mice showed a statistically significant positive trend (p-value = 0.015; one-sided Cochran-Armitage
trend test). However, the combined incidence of adenomas and carcinomas was not elevated in
male mice. In such a case, NTP policy recommends analyzing the tumors both separately and in
combination (McConnell etal.. 1986). This recommendation arose out of concern that combining
benign and malignant neoplasms can result in a false negative if the chemical shows a statistically
significant increase in malignant tumors without an increase in the combined incidence. In an
addendum to the study report that included results of additional examination and sectioning of
lung specimens from the mid-dose groups in the mouse study, Lish etal. (1984) noted an increase
in the combined incidences of primary pulmonary neoplasms in males of all dose groups and in
females in the 7.0, 35, and 175/100 mg/kg-day dose groups, but regarded these neoplasms as
random and not biologically significant (rationale for this conclusion not provided).
Bioassays in rats provide no evidence of an association between RDX exposure and
induction of lung tumors. The incidence of alveolar/bronchiolar adenomas or carcinomas was not
increased in either sex of Sprague-Dawley rats exposed chronically to RDX at doses up to
10 mg/kg-day (Hart. 1976) or in F344 rats of either sex exposed chronically to RDX at doses up to
40 mg/kg-day (Levine etal.. 1983). Alveolar/bronchiolar carcinomas are rare tumors in both
species of rats, male or female (Chandra etal.. 1992: Haseman et al.. 1985).
Mechanistic Evidence
There are few mechanistic data to inform an MOA determination for either liver or lung
tumors induced by exposure to RDX.
1-77
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Largely negative findings in in vitro and in vivo genotoxicity assay for parent RDX or its
oxidative metabolites (see Appendix C, Section C.3.2) suggest that parent RDX or its oxidative
metabolites do not interact directly with deoxyribonucleic acid (DNA). In contrast, there are some
positive genotoxicity results for the /V-nitroso metabolites of RDX, specifically MNX and TNX. Trace
amounts of MNX and TNX metabolites were identified in Yucatan miniature pigs (minipigs) orally
exposed to 14C-RDX in an absorption, distribution, metabolism, and excretion study; minipigs were
chosen as the animal model for investigation of RDX metabolism because the GI tract of pigs more
closely resembles that of humans (Musick etal.. 2010: Major etal.. 2007). MNX has tested positive
in some in vitro assays, including unscheduled DNA synthesis in primary rat hepatocytes and the
mouse lymphoma forward mutation assay fSnodgrass. 19841. although MNX tested negative in the
only in vivo test performed, a mouse dominant lethal mutation test fSnodgrass. 19841. MNX was
not mutagenic in Salmonella typhimurium (strains TA98, TA100, TA1535, TA1537, and TA1538),
with or without the addition of the S9 metabolic activating mixture (Pan etal.. 2007: Snodgrass.
19841. When S. typhimurium strains TA97a and TA102, strains sensitive to frame shift and
oxidative DNA damage, were used in conjunction with elevated concentrations of the metabolizing
system (S9), MNX and TNX were mutagenic. iV-nitroso metabolites, including MNX and TNX, are
generated anaerobically and are likely a result of bacterial transformation of parent RDX in the GI
tract to various /V-nitroso derivatives (Pan etal.. 2007). Exposure to potentially mutagenic
iV-nitroso metabolites of RDX generated in the GI tract of mice may occur in the liver (and
subsequently in the systemic circulation) via enterohepatic circulation. However, as noted earlier
in pigs, the iV-nitroso metabolites of RDX have been identified only in trace amounts in urine
compared to the major metabolites, 4-nitro-2,4-diazabutanal and 4-nitro-2,4-diaza-butanamide
fMaior etal.. 20071. Thus, the contribution of the /V-nitroso metabolites to the overall carcinogenic
potential of RDX is unclear.
Aberrant expression of miRNAs was observed in the brains and livers of female B6C3Fi
mice fed 5 mg RDX/kg in the diet for 28 days [(Zhang and Pan. 2009b): dose of 0.75-1.5 mg/kg-day
estimated by Bannon etal. (2009b)]. with several oncogenic miRNAs being upregulated, while
several tumor-suppressing miRNAs were downregulated. However, the pattern of induction was
not always consistent in the livers of RDX-treated mice [e.g., miR-92a was downregulated in liver
tissue samples when it is typically upregulated in hepatocellular carcinomas; (Sweeney et al..
2012b)]. miRNAs have been associated with several cancers (Wiemer. 2007: Zhang etal.. 2007).
but the use of miRNAs as predictive of carcinogenesis has not been demonstrated (Bannon etal..
2009b). Further, it is unknown whether aberrant expression of a specific miRNA (or suite of
miRNAs) plays a role in the MOA of RDX carcinogenicity. Microarray analysis of gene expression in
male Sprague-Dawley rats after exposure to a single oral (capsule) dose of RDX revealed a general
upregulation in gene expression (predominantly genes involved in metabolism) in liver tissues
(Bannon et al.. 2009a): however, the relevance of this finding to the carcinogenicity of RDX is
unclear.
1-78
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Sweeney etal. f2012bl hypothesized a set of MOAs for the liver tumors:
• Genotoxicity mediated by either (1) RDX, (2) tissue-generated oxidative metabolites, or
(3) N-nitroso metabolites generated anaerobically in the GI tract. The key events in this
hypothesized MOA are production of DNA damage, gene mutation, formation of neoplastic
lesions, and promotion/progression of tumors. The largely negative results for genotoxicity
led Sweeney etal. f2012bl to conclude that this MOA is not plausible for RDX or its
oxidative metabolites. Although there are some positive results for the N-nitroso
metabolites, the limited evidence to support systemic uptake and distribution of
metabolites to the liver led Sweeney etal. (2012b) to conclude that this MOA is not
sufficiently plausible.
• Cell proliferation. The key events in this hypothesized MOA are Gl-tract generation of
N-nitroso metabolites, absorption, distribution to the liver, cytotoxicity (optional), and
enhanced cell proliferation, leading to preneoplastic foci that progress to hepatocellular
adenomas and carcinomas. Sweeney etal. (2012b) cited evidence of increased liver weights
in mice as consistent with cell proliferation, but noted that increased liver weights were
also observed in rats without proceeding to liver tumors. They considered this MOA
"plausible, but not particularly well supported."
In addition to the inconsistencies in the evidence identified by Sweeney etal. (2012b). EPA
notes the following evidence (or lack of evidence) that fails to support this hypothesized MOA.
(1) The absence of significant liver histopathology in mice after subchronic or chronic exposure to
RDX at doses that induced liver tumors (Lish etal.. 1984: Cholakis etal.. 1980) suggests that cellular
toxicity is not a precursor to these tumors. (2) As discussed in Section 1.2.4, changes in liver weight
showed no consistent pattern across studies or sexes and did not correlate with tumor response.
(3) No studies were available that directly measured RDX-induced cell proliferation rates. (4) No
information was available to rule out nonprecancerous causes of liver-weight increase.
In summary, the available evidence indicates that RDX is likely not mutagenic (see
Appendix C, Section C.3.2), although anaerobically derived N-nitroso metabolites have
demonstrated some genotoxic potential. While these metabolites have been measured in the
mouse (Pan etal.. 2007) and identified in minipigs (Musick etal.. 2010: Major etal.. 2007). they
have not been identified in humans, and may not be the predominant metabolites of RDX. An MOA
involving a proliferative response generated by tissue-derived oxidative metabolites of RDX has
been proposed, but is not supported by the available data. In light of limited information on
precursor events leading to the observed liver and lung tumor response in RDX-exposed rodents
and lack of toxicokinetic information on RDX metabolites, neither a cell proliferative MOA nor a
mutagenic N-nitroso metabolite MOA is supported. Thus, the MOA leading to the increased
incidence of liver and lungs tumors is not known.
1-79
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
1.3. INTEGRATION AND EVALUATION
1.3.1. Effects Other Than Cancer
Most of the evidence for the health effects of RDX comes from oral toxicity studies in
animals. The three epidemiology studies that document possible inhalation exposure are limited by
various study design deficiencies, including inability to distinguish exposure to TNT (associated
with liver and hematological system toxicity), inability to adequately characterize exposure levels,
small sample sizes, and inadequate reporting. The single animal inhalation study identified in the
literature search had deficiencies (e.g., lack of a control and incomplete exposure information) that
precluded its inclusion in this assessment (see literature search section).
The strongest evidence for a human health hazard following exposure to RDX is for nervous
system effects. Toxicity studies in multiple animal species involving chronic, subchronic, and
gestational exposures provide consistent evidence of nervous system effects following oral
exposure. Effects included dose-related increases in seizures and convulsions, as well as
observations of tremors, hyperirritability, hyper-reactivity, and other behavioral changes (Crouse
etal.. 2006: Angerhofer et al.. 1986: Levine etal.. 1983: Levine etal.. 1981a. b; Cholakis etal.. 1980:
von Oettingen et al.. 1949).
Human studies provide supporting evidence for RDX as a neurotoxicant and provide
support for the assumption that the nervous system effects observed in experimental animals are
relevant to humans. In particular, several case reports provide evidence of associations between
exposure to RDX (via ingestion, inhalation, and possibly dermal exposure) and seizures and
convulsions (Kasuske etal.. 2009: Kuctikardali etal.. 2003: Testud etal.. 1996a: Testud etal..
1996b: Woody etal.. 1986 and others, see Appendix C.2). Other nervous system effects identified in
human case reports include dizziness, headache, confusion, and hyperirritability. A cross-sectional
study described memory impairment and visual-spatial decrements in RDX-exposed workers (Ma
and Li. 1993). although confidence in these findings is relatively low because of issues with design
and reporting.
Additional support for an association between RDX exposure and nervous system effects
comes from consistent evidence of neurotoxicity across taxa, including humans, laboratory animal
species, birds, lizards, fathead minnows, and earthworms (Ouinn etal.. 2013: Garcia-Revero etal..
2011: McFarland etal.. 2009: Gogal etal.. 2003). Studies in rats demonstrate a correlation between
blood and brain concentrations of RDX and the time of seizure onset (Williams etal.. 2011: Bannon
et al.. 2009a], Additionally, the affinity of RDX for the picrotoxin convulsant site of the GABAa
channel suggests that the resulting disinhibition could lead to the onset of seizures (Williams etal..
2011).
Induction of convulsions and seizures appears to be more strongly correlated with dose
than with duration of exposure. However, there is mechanistic information that suggests repeated
binding to the receptor convulsant site of GABAa may promote a state of increased neuronal activity
1-80
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
that increases the likelihood of subsequent neurological effects fGerkin etal.. 20101. As a result,
some uncertainty remains as to whether the available mechanistic information adequately
addresses potential neurotoxicity after longer duration exposure to RDX. It is unclear if nervous
system effects progressed in severity (e.g., from subtle behavioral changes or nonconvulsive
seizures to tonic-clonic seizures) with increasing dose, as many of the studies that reported subtler
neurobehavioral changes did not provide detailed dose-response information, and most studies
were not designed to capture this information.
The nervous system effects following oral exposure to RDX were observed in humans
acutely exposed to RDX and in multiple experimental animal studies in rats, mice, monkeys, and
dogs following exposures ranging from 10 days to 2 years in duration. Notably, despite the
potential for effects on the developing nervous system based on the presumed MOA for RDX
neurotoxicity (discussed in Section 1.2.3), no studies included a thorough evaluation of potential
developmental neurotoxicity. Across the database, behavioral manifestations of seizure activity
were the most consistently observed nervous system effect associated with RDX exposure. This
most commonly included evidence of increased convulsions, as well as other related effects such as
tremors, shaking, hyperactivity, or nervousness, which were generally observed at doses that were
the same as or higher than doses that induced convulsions. Nervous system effects are a human
hazard of RDX exposure and are carried forward for consideration for dose-response analysis.
Convulsions, considered a severe adverse effect, were selected as a consistent and sensitive
endpoint representative of nervous system effects.
Evidence for urinary system toxicity is more limited than evidence for neurotoxicity. In
humans, kidney effects (including decreased urine output, blood in urine, and proteinuria) were
observed only in individuals with acute accidental exposure (ingestion and inhalation) to unknown
amounts of RDX. In experimental animal studies, histopathological changes in the kidney and
urinary bladder (medullary papillary necrosis, suppurative pyelitis, and uremic mineralization of
the kidney; luminal distention and cystitis of the urinaiy bladder) were reported in male rats
exposed to RDX in the diet following exposure durations of 1 year or longer (Levine etal.. 1983).
but not in similarly exposed female rats. Evidence for milder renal effects reported in subchronic
studies of RDX in mice, rats, and monkeys was limited and inconsistent Mice appeared to be less
sensitive than rats. Other measures of kidney effects (kidney weights and serum chemistry
parameters) did not provide consistent evidence of dose-related changes associated with RDX
exposure. In light of the dose-related increase in histopathological changes in the kidney and
urinary bladder in male rats in the Levine etal. (1983) study, and in particular the robust response
in the high-dose animals, urinary system effects are a potential human hazard of RDX exposure.
Medullary papillary necrosis was selected as an endpoint representative of kidney effects.
This histopathologic lesion was observed at higher incidence than other kidney histopathologic
lesions, was present at both the 1-year interim and 2-year final sacrifices (see Table 1-6), and
represents a severe measure of toxicity. Renal toxicity was, in fact, considered the principal cause
1-81
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
of RDX-related mortality and morbidity in male rats in the Levine etal. f!9831 2-year bioassay.
Hemorrhagic/suppurative cystitis was selected as an endpoint representative of urinary bladder
effects. Like medullary papillary necrosis of the kidney, urinary bladder cystitis is a clearly adverse
effect and was observed at both the 1-year interim and 2-year final sacrifices (see Table 1-7). A
dose-related increased incidence of luminal distention was also observed in male rats, but was not
selected as representative of urinary bladder toxicity because it is a less specific diagnosis than
cystitis and can be caused by various factors, including partial obstruction of the bladder.
Evidence for prostate toxicity is also more limited than evidence for neurotoxicity. A
dose-related, increase in the incidence of suppurative prostatitis was observed in male rats exposed
to RDX in the diet for 2 years (Levine etal.. 19831. There was also a concomitant shift from chronic
inflammation to suppurative inflammation with increasing dose of RDX starting at 1.5 mg/kg-day.
Similar types and patterns of inflammation were not observed in mice, and no other rat studies of
equivalent duration that examined the prostate were available. RDX and its interaction with GABAa
receptors, which have also been identified on the prostate (Napoleone etal.. 19901. increases
biological plausibility by providing a potential mechanism by which RDX could have effects directly
on the prostate. In their evaluation of the external review draft assessment, the SAB determined
the weight of evidence to be sufficient for identifying prostate effects as a hazard of RDX exposure
(SAB. 2017). Consistent with this determination, the incidence of suppurative prostatitis was
selected as the endpoint most representative of prostate effects.
Evidence for developmental toxicity and liver toxicity was more limited than that for the
endpoints discussed above. In animal studies, developmental effects, including offspring survival,
growth, and morphological development, were observed only at doses associated with maternal
mortality fAngerhofer etal.. 1986: Cholakis etal.. 19801. Evidence for potential hepatic effects
comes from observations of increases (generally dose related) in liver weight in some subchronic
oral animal studies (Lish etal.. 1984: Levine etal.. 1983: Levine etal.. 1981a. b; Cholakis etal.. 1980:
Hart. 1976). However, these elevations in liver weight were not consistently observed across
studies nor were they accompanied by RDX-related histopathological changes in the liver or
increases in serum liver enzymes. In addition, the interpretation of liver-weight changes in the
mouse bioassay by Lish etal. (1984) is complicated by the relatively high incidence of liver tumors
in this study. At this time, the available data do not support liver and developmental toxicity as
human hazards of RDX exposure; these effects were not considered further for dose-response
analysis and derivation of reference values.
Evidence that RDX may induce effects in other organs, including the eyes, and the
cardiovascular, musculoskeletal, immune, GI, hematological, and reproductive systems, is generally
limited to human case reports or to findings in experimental animals that were inconsistent across
studies or lacked dose-related patterns of response. Therefore, information to assess toxicity in
these organs following exposure to RDX was insufficient Treatment-related changes in body
weight or body-weight gain were generally observed at high doses in association with elevated
1-82
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
mortality or with severe kidney and urinary bladder toxicity, and thus appeared to be secondary to
effects on other primary targets of RDX toxicity. Effects on body weight and on these other organs
were not considered further for dose-response analysis and derivation of reference values.
In a number of the animal studies reporting nervous system effects, unscheduled deaths
occurred at RDX doses as low as those that induced nervous system effects (Crouse etal.. 2006:
Angerhofer etal.. 1986: Levine etal.. 1983: Levine etal.. 1981a: Cholakis etal.. 1980: von Oettingen
etal.. 1949). In a 90-day study that recorded nervous system effects and survival more thoroughly
than earlier studies, Crouse et al. (2006) reported that nearly all preterm deaths were preceded by
neurotoxic signs such as tremors and convulsions. Convulsions did not, however, necessarily lead
to early mortality; of the animals observed to have convulsed in the Crouse etal. (2006) study,
approximately 75% survived to the end of the 90-day study. Most of the earlier studies provide a
limited understanding of the association between mortality and nervous system effects because the
frequency of clinical observations was likely insufficient to observe convulsions before death. In
humans, mortality has not been reported in case reports involving workers with symptoms of
neurotoxicity exposed to RDX during manufacture or in individuals exposed acutely because of
accidental or intentional ingestion. Survival has not been specifically evaluated in studies of worker
populations exposed chronically to RDX. Ultimately, the convulsion findings, without consideration
of mortality, are sufficient to identify neurotoxic effects associated with RDX exposure as severe
and adverse.
Regarding mortality, the preference is not to use a frank health effect as severe as mortality
as the basis for a reference value. As noted in U.S. EPA (2002). a chemical may cause a variety of
effects ranging from severe—such as death—to more subtle biochemical, physiological, or
pathological changes; primary attention in assessing health risk should be given to those effects in
the lower exposure range and/or the effects most biologically appropriate for a human health risk
assessment. Where mortality occurs because of a chemical's effects on a specific organ/system
(e.g., in the case of RDX, evidence suggests some relationship between mortality and effects on the
nervous system and kidney), the preference would be to develop a quantitative assessment based
on the initial hazard and not on death. Because unscheduled deaths were observed with some
consistency across studies and, in some studies, at doses as low as those associated with
convulsions, two additional analyses of mortality data are presented in Section 2. In the first
analysis, benchmark doses (BMDs) derived using mortality data sets are compared to the BMD used
to derive the inhalation reference concentration (RfC; Section 2.1.6). As discussed in Section 1.2.1,
the relationship between convulsions and mortality is not clear and raises concerns for the
potential underreporting of convulsions. An analysis, described in Section 2.1.7, addresses the
possibility that the analyses of convulsions brought forward for dose-response analysis resulted in
an underestimate of the toxicity for RDX.
1-83
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
1.3.2. Carcinogenicity
As presented in Section 1.2.7, dietary administration of RDX induced dose-related increases
in the incidence of hepatocellular adenomas or carcinomas in male and female B6C3Fi mice (Parker
etal.. 2006: Lish etal.. 1984). In the same study, RDX also induced dose-related increases in the
incidence of alveolar/bronchiolar adenomas or carcinomas in both sexes. Some of these trends in
liver and lung were statistically significant In Fischer 344 rats, dietary administration of RDX
yielded a statistically significant trend in the incidence of hepatocellular carcinomas24 in males, but
not in females (Levine etal.. 19831. A 2-year dietary study in Sprague-Dawley rats was negative in
both sexes (Hart. 1976). although the highest dose in this study, and the only dosed group for which
pathology was examined, was somewhat lower [no increase in carcinomas at doses up to
10 mg/kg-day in Hart (1976). vs. hepatocellular carcinomas in male rats at 8 and 40 mg/kg-day in
the Levine etal. (1983) study]. The human studies are not informative.
This evidence leads to consideration of two hazard descriptors under the EPA's Guidelines
for Carcinogen Risk Assessment fU.S. EPA. 2005al The descriptor likely to be carcinogenic to humans
is appropriate when the evidence is "adequate to demonstrate carcinogenic potential to humans"
but does not support the descriptor carcinogenic to humans. One example from the cancer
guidelines is "an agent that has tested positive in animal experiments in more than one species, sex,
strain, site, or exposure route, with or without evidence of carcinogenicity in humans." RDX
matches the conditions of this example, having induced dose-related increases in tumors in two
species (mouse and rat), in both sexes, and at two sites (liver and lung). Liver carcinomas,
increased in male F344 rats in the Levine etal. (1983) study, are considered rare in that species.
Alternatively, the descriptor suggestive evidence of carcinogenic potential is appropriate
when the evidence raises "a concern for potential carcinogenic effects in humans" but is not
sufficient for a stronger conclusion. The incidences of alveolar/bronchiolar tumors showed a
positive trend in male and female B6C3Fi mice. Evidence of carcinogenicity in the liver from rodent
bioassays is less clear. The hepatocellular carcinoma result in male F344 rats is based on a small
number oftumors (1/55, 0/55, 0/52, 2/55, and 2/31, respectively, atO, 0.3,1.5, 8.0, and
40 mg/kg-day) that is not matched by an increase in hepatocellular neoplasms overall (5/55, 3/55,
0/52, 4/55, and 3/31, respectively), and RDX did not increase the incidence of carcinomas at any
other site in F344 or Sprague-Dawley rats of either sex. The incidence of liver tumors in female
B6C3Fi mice showed a statistically significant positive trend (Lish etal.. 1984). although the study
authors noted the relatively low tumor incidence in concurrent female control mice (1/65). The
PWG that reviewed the slides from this study (Parker etal.. 2006: Parker. 2001) confirmed the
positive trend in female mouse liver tumors, but also raised some concerns related to
^Hepatocellular carcinoma may be regarded as a rare tumor in male F344 rats. Although there is no
compilation of historical control data for the Levine laboratory, Haseman etal. (1984) reported that in NTP
studies during 1980-1983,18/2,306 (0.8%) of male F344 rats developed hepatocellular carcinomas and
78/2,306 (3.4%) developed neoplastic nodules.
1-84
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
histopathological methods and the absence of necropsy and histopathology processing records that
limited their evaluation. In male mice from this study (Lish etal.. 1984). the incidences of liver
tumors in some treated groups were higher than in the control, but trend tests were not statistically
significant. Interpretation of male mouse liver tumor incidence is complicated by the high and
variable background incidence of this tumor in the male mouse.
As discussed in Section 1.2.7, few mechanistic studies are available to inform the mode of
action by which RDX induces liver and lung tumors in rodents. The available evidence indicates
that RDX is likely not mutagenic. Anaerobically derived N-nitroso metabolites have demonstrated
some genotoxic potential. These metabolites have not been identified in humans, and their
contribution to any genotoxic potential of RDX is unknown. Precursor events leading to the
observed liver and lung tumor response in RDX-exposed rodents have not been identified.
Although characterization of the cancer MOA is not needed to determine a chemical's cancer
hazard, understanding the MOA can contribute to a cancer hazard determination. In the case of
RDX, mechanistic information is not helpful in guiding selection of a cancer descriptor.
As noted in the EPA's cancer guidelines fU.S. EPA. 2005a). choosing a hazard descriptor
cannot be reduced to a formula, as descriptors may be applicable to a variety of potential data sets
and represent points along a continuum of evidence. In the case of RDX, there are plausible
scientific arguments for more than one hazard descriptor. Overall, the considerations discussed
above, interpreted in light of the cancer guidelines, lead to the conclusion that there is suggestive
evidence of carcinogenic potential for RDX. Although the evidence includes dose-related tumor
increases in two species, two sexes, and two sites, the evidence of carcinogenicity outside the
B6C3Fi mouse is not robust, and this factor was decisive in choosing a hazard descriptor. Within
the spectrum of results covered by the descriptor suggestive evidence, the evidence for RDX is
strong. There are well-conducted studies that tested large numbers of animals at multiple dose
levels, making the cancer response suitable for dose-response analysis (Section 2).
The descriptor suggestive evidence of carcinogenic potential applies to all routes of human
exposure. Dietary administration of RDX to mice and rats induced tumors of the liver or lung, sites
beyond the point of initial contact, and human case reports have demonstrated absorption and
distribution of inhaled RDX into systemic circulation. Under the cancer guidelines, this information
provides sufficient basis to apply the cancer descriptor developed from oral studies to other
exposure routes.
1.3.3. Susceptible Populations and Life Stages for Cancer and Noncancer Outcomes
Susceptibility refers to factors such as life stage, genetics, sex, and health status that may
predispose a group of individuals to greater response to an exposure. This greater response could
be achieved either through differences in exposure to the chemical underlying RDX toxicokinetics
or differences in RDX toxicodynamics between susceptible and other populations. Little
information is available on populations that may be especially vulnerable to the toxic effects of RDX.
Reproductive and developmental toxicity studies generally did not identify effects in offspring at
1-85
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
doses below those that also caused severe maternal toxicity fAngerhofer et al.. 1986: Cholakis et al..
1980). However, the developmental importance of GABAergic systems (Kirmse etal.. 2018: Ben-
Ari. 2014: Williams et al.. 2011: Williams and Bannon. 2009: Galanopoulou. 20081 and
developmental neurotoxicity of chemicals with similar modes of action (Salari and Amani. 2017:
Marty etal.. 20001 suggest RDX may be harmful during the period of brain development. Further
raising cause for concern, seizures and seizure disorders such as epilepsy occur more frequently in
infants and children than in any other age group (many are caused by early life insults such as fever
or trauma), and research suggests that early life seizures (i.e., before the brain has fully matured)
can lead to long-lasting neurological consequences (Ronnie. 2003: Volpe. 2001: Tensen and Baram.
2000: Moshe. 2000.1987). A pilot study in rats demonstrated transfer of RDX from dam to fetus
during gestation, found RDX in milk from treated dams, and recommended further study (Hess-
Ruth etal.. 20071. Given the understanding of RDX toxicokinetics (see Section 1.1.2), it is expected
that RDX reaching the fetus or infant through either the blood or ingested milk would be readily
distributed to the brain, although specific studies have not been conducted. For these reasons, and
as noted in Section 1.2.1, the lack of developmental neurotoxicity studies was identified as a
significant data gap in understanding the nervous system effects of RDX exposure.
The primary MOA for the neurotoxic effects of RDX exposure involves RDX binding to
GABAa receptors, specifically the picrotoxin convulsant site of the GABA channel, and blocking
inhibitory GABAergic transmission, that eventually leads to the development of seizures and
related behavioral changes (see Section 1.2.1, Mechanistic Evidence). In addition to its role as the
major inhibitory neurotransmitter system in many regions of the adult brain, GABAergic signaling
plays a key role in brain development, where it contributes to a delicate equilibrium with other
signaling processes (e.g., glutamatergic) to help establish the appropriate functional connectivity of
the mature brain (Kirmse etal.. 2018: Ben-Ari. 2014). While GABAergic signaling and the overall
balance between excitation and inhibition is essential throughout brain development, which
continues through sexual maturation, a number of critical developmental processes occur
simultaneously during the perinatal period, and these coincide with prominent shifts in GABAergic
function. As a result, the perinatal period may represent a vulnerable life stage for the neurotoxic
effects of RDX exposure through GABAergic inhibition.
In the perinatal mammalian brain, GABA activity is primarily depolarizing and excitatory
(as compared to hyperpolarizing and inhibitory in the adult brain), which is presumably necessary
for its specific functions at this stage of brain development In animals, expression of chloride
cotransporters NKCC1 and KCC2 around or shortly after birth reduces intracellular CI- and
mediates a switch in GABA activity to primarily hyperpolarizing and inhibitory (Ben-Ari. 2014:
Rivera etal.. 1999). For GABAAergic signaling, the switch from depolarizing to the hyperpolarizing
phenotype in adults occurs by the end of the first postnatal month in rats, although this differs by
brain region and sex (Galanopoulou. 2008). In addition, the composition of GABAa receptors is also
subject to developmental regulation, with some subunits varying in their pattern of expression
1-86
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
during development as compared to adulthood fLuian etal.. 2005: Fritschv etal.. 1994: Laurie etal..
1992). Thus, RDX exposure during the perinatal period in humans could be impactful.
During this potentially sensitive period, excitatory GABAAergic signaling helps to regulate
the proliferation, migration, survival, and differentiation of new neurons, as well as synaptogenesis
and the development of mature neural networks (Deidda et al.. 2014: Galanopoulou. 20081.
Modulation of GABAAergic signaling at this life stage is presumably tightly controlled, as it serves to
orchestrate these processes in a region-specific manner for specific glial and neuronal subsets,
often stimulating these processes (e.g., increasing neuronal migration or survival) in some regions
while simultaneously inhibiting the same processes (e.g., decreasing neuronal migration or
survival) in other regions (Creelev. 2016: Deidda etal.. 2014: Galanopoulou. 2008: Ikonomidou et
al.. 2000). Additional concern for susceptibility during this life stage may be raised due to the
prominent role for BDNF during this time, with high expression during the first two postnatal
weeks (in rodents) before declining to adult levels, and whose role as a neurotrophic factor
includes the regulation of neuronal excitation and its sequelae (Aguado etal.. 2003). As discussed
in Section 1.2.1, molecular evidence suggests that RDX exposure in adults may impact the
expression or function of BDNF and related factors in the brain (Zhang and Pan. 2009b): the lack of
data on brain BDNF after developmental RDX exposure remains a data gap.
Alterations of GABA activity have been linked to developmental brain disorders (Kirmse et
al.. 2018). and genetic mutations causing aberrant GABAergic signaling lead to a number of seizure
disorders in infants and children (Galanopoulou. 2008). although GABAergic signaling in the
immature brain may be required for epileptogenesis (Khalilov etal.. 2003). Exposure of early
postnatal rodent hippocampus to the GABAa receptor antagonist, bicuculline, which has a similar
mode of action to RDX, increased the density of inhibitory but not excitatory synapses f Marty etal..
2000). White etal. (2008) reported that a 2.7 mg/kg subcutaneous dose of bicuculline provoked
seizures in 97% adult mice, but Salari and Amani (2017) found developmental and behavioral
impairment after a 0.3 mg/kg subcutaneous dose to neonatal mice, suggesting developmental
neurotoxicity from bicuculline is evident as low as 1/10 the convulsive dose. RDX and bicuculline
differ in their affinity for the GABAa receptor, with RDX demonstrating comparatively lower
inhibitory potency than bicuculline. However, findings from studies of bicuculline provide
suggestive evidence of perinatal susceptibility to the neurotoxicity elicited by compounds that alter
GABAergic signaling. The lack of data on how RDX exposure might impact the critical role of
GABAergic signaling during the perinatal period (and at later stages of brain development and
maturation) represents an important uncertainty.
Limited data suggest that male laboratory animals may be more susceptible to noncancer
toxicity associated with RDX exposure. In general, male animals were more sensitive to RDX
neurotoxicity than females (i.e., more convulsions; more hyperactive; greater brain-weight
changes). In the 2-year study in F344 rats (Levine etal.. 1983). RDX exposure induced severe
1-87
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
toxicity of the kidney and urinary bladder in males, but no similar effects in females, suggesting a
sex-based difference in susceptibility to RDX urinary system toxicity.
Data on the incidence of convulsions and mortality from gavage studies of RDX in the rat
provide some indication that pregnant animals may be a susceptible population. In the
developmental toxicity study by Cholakis et al. (19801. deaths were observed in pregnant F344 rats
only at a dose of 20 mg/kg-day, but convulsions were reported in a single rat at 2 mg/kg-day. In a
range-finding developmental toxicity study (Angerhofer et al.. 1986). mortality and convulsions
were reported in pregnant Sprague-Dawley rats at a dose of >40 mg/kg-day, but not at
<20 mg/kg-day, although the relatively small group sizes in this study should be noted. In the main
study by these investigators, convulsions were reported in pregnant rats only at 20 mg/kg-day, but
one death (in dose groups of 40 rats) was reported at both 2 and 6 mg/kg-day (Angerhofer etal..
19861. In comparison, increased mortality and convulsions were reported at >8 mg/kg-day in a
90-day gavage study in F344 rats (Crouse et al.. 2006). The instances of one convulsion and two
deaths in pregnant rats in the Cholakis etal. (1980) and Angerhofer et al. (1986) studies at doses of
2 or 6 mg/kg-day raise the possibility that pregnant animals may be more susceptible to the effects
of RDX; however, direct comparison between the available gavage studies in pregnant and
nonpregnant rats is uncertain because of differences in study design, including numbers of animals
tested per group, test material characteristics, and rat strain. Overall, the available information is
not considered sufficient to conclude that pregnant animals are a susceptible population.
There is limited evidence that CYP450 or similar enzymes are involved in the metabolism of
RDX (Bhushan etal.. 2003). indicating a potential for genetic polymorphisms in these metabolic
enzymes to affect susceptibility to RDX. This susceptibility may also be influenced by differential
expression of these enzymes during development. Individuals with epilepsy or other seizure
syndromes, and in particular those that have their basis in genetic mutation to GABAa receptors,
may represent another group that may be susceptible to RDX exposure. However, there is currently
no information to support predictions of how genetic polymorphisms or the presence of seizure
syndromes may affect susceptibility to RDX exposure.
1-88
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
2. DOSE-RESPONSE ANALYSIS
2.1. ORAL REFERENCE DOSE FOR EFFECTS OTHER THAN CANCER
The oral reference dose (RfD, expressed in units of mg/kg-day) is defined as an estimate
(with uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
population (including sensitive subgroups) that is likely to be without an appreciable risk of
deleterious effects during a lifetime. It can be derived from a no-observed-adverse-effect level
(NOAEL), lowest-observed-adverse-effect level (LOAEL), or the benchmark dose lower confidence
limit (BMDL), with uncertainty factors (UFs) generally applied to reflect limitations of the data
used.
2.1.1. Identification of Studies for Dose-Response Analysis of Selected Effects
As discussed in Section 1.3.1, based on findings from oral studies in experimental animals,
nervous system effects are a human hazard of RDX exposure, and urinary system (kidney and
bladder) effects are a potential human hazard of RDX exposure. There is suggestive evidence of
prostate effects associated with RDX exposure. Although animal mortality has been reported in
many of the toxicology studies conducted for RDX, it was not considered a hazard by itself or as the
basis for the derivation of a reference value (see Sections 2.1.6 and 2.1.7 for further discussion).
The effects selected to best represent each of the hazards, identified in Section 1.3.1, are
carried forward in the sections below. To identify the stronger studies for dose-response analysis,
several attributes of the studies reporting the endpoints selected for each hazard were reviewed
(i.e., study size and design, relevance of the exposure paradigm, and measurement of the endpoints
of interest). In considering the study size and design, preference was given to studies using designs
reasonably expected to have power to detect responses of suitable magnitude. Exposure paradigms
including a route of human environmental exposure (i.e., oral and inhalation) are preferred. When
developing a chronic reference value, chronic or subchronic studies are preferred over studies of
acute exposure durations. Studies with a broad exposure range and multiple exposure levels are
preferred to the extent that they can provide information about the shape of the exposure-response
relationship. Additionally, with respect to measurement of the endpoint, studies that can reliably
distinguish the presence or absence (or degree of severity) of the effect are preferred.
Human studies are generally preferred over animal studies as the basis for a reference value
when quantitative measures of exposure are reported, and the reported effects are determined to
be associated with exposure. The available epidemiological studies of worker populations exposed
to RDX examined the relationship between certain health endpoints and inhalation exposure;
however, no epidemiological studies of ingested RDX are available. Multiple case reports support
2-1
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
the identification of hazards associated with RDX exposure but are inadequate for dose-response
analysis because they do not yield incidence estimates, exposure durations are short, and
quantitative exposure information is lacking. Therefore, human studies could not be used for oral
dose-response analysis or to serve as the basis for the RfD. In the absence of human data, the
animal studies were considered for dose-response analysis.
Experimental animal studies considered for each health effect were evaluated using general
study quality considerations discussed in Section 4 of the Preamble and in the literature search
section, and the attributes described above. The rationales for selecting the strongest studies that
reported effects on the nervous system, urinary system, and prostate are summarized below.
Nervous System Effects
Convulsions, a severe adverse effect, were selected for dose-response analysis as a
consistent endpoint of nervous system effects (see Section 1.3.1 for discussion). This endpoint was
reported in seven studies fCrouse etal.. 2006: Lish etal.. 1984: Levine etal.. 1983: Levine etal..
1981a: Cholakis etal.. 1980: Martin and Hart. 1974: von Oettingen et al.. 19491. Table 2-1 provides
an overview of the information considered in the studies reporting nervous system effects (i.e.,
convulsions) evaluated for dose-response analysis.
2-2
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 2-1. Information considered for evaluation of studies that examined convulsions
Study
reference
Study design and size
Exposure paradigm
Measurement
of endpoint
Design
# of animals
Route
Duration
# of dose
groups3
Levels
(mg/kg-d)
Purity
(%)
Analytical
concentration?13
Incidence
data reported
Crouse et al.
(2006)
Toxicity study
10 rats/sex/group
Gavage
13-wk
5
4-15
99.99
Yes
Yes
Cholakis et al.
(1980)
Developmental
study
24-25 female rats/
group
Gavage
14-d
3
0.2-20
89
Yes
Yes
Martin and
Hart (1974)
Toxicity study
3 monkeys/sex/
group
Gavage
13-wk
3
0.1-10
Not specified
Not specified
Yes
Levine et al.
(1983)
Toxicity and
carcinogenicity
bioassay
75 rats/sex/group
Diet
2-yr
4
0.3-40
89-99
Yes
No
Lish et al.
(1984)
Toxicity and
carcinogenicity
bioassay
85 mice/sex/group
Diet
2-yr
4
1.5-175
89-99
Yes
No
Levine et al.
(1981a)
Toxicity study
10 rats/sex/group
Diet
13-wk
5
10-600
85
Yes
No
von Oettingen
et al. (1949)
Toxicity study
20 rats/group
Diet
13-wk
3
15-50
90-97
Not specified
No
Excluding the control group.
indicates whether study authors performed analysis to confirm the concentration of RDX in the suspension or diet administered to the animals (e.g., to
determine percentage of target concentration).
2-3
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Incidence of convulsions was reported in three studies of RDX—all involving gavage
administration: Crouse etal. (2006). Cholakis etal. (1980) (developmental toxicity study), and
Martin and Hart (1974). Qualitative findings of nervous system effects were reported in other
chronic and subchronic studies—all involving dietary administration: Lish etal. (1984). Levine et al.
(1983). Levine etal. (1981a). and von Oettingen et al. (1949). Incidence data on neurotoxic effects
of RDX were not collected in any of the dietary studies. For example, Levine etal. (1983) reported
only that convulsions and other nervous system effects were noted in rats exposed to RDX for
2 years at the highest dose (40 mg/kg-day) tested. The studies that included incidence data (i.e.,
the gavage studies) were preferred over those studies only reporting qualitative results (i.e., the
dietary studies).
The three gavage studies reporting incidence data were further considered. Crouse et al.
f20061 reported a dose-related increase in convulsions and tremors in both male and female F344
rats following a 90-day oral (gavage) exposure to RDX. This study used a test material of high
purity and six dose groups (including the control) that provided good resolution of the
dose-response curve. Cholakis etal. (1980) reported a dose-related increase in convulsions in a
developmental toxicity study in F344 rats, following a 14-day exposure to RDX on GDs 6-19.
Although this study was designed as a standard developmental toxicity study (i.e., not specifically to
examine nervous system effects), it reported information on the identity of the test material and
used three dose groups that adequately characterized the dose-response curve. Further, this study
provided evidence of nervous system effects at a relatively low dose. The study in monkeys by
Martin and Hart (1974) provides supporting evidence of nervous system effects (trembling,
shaking, ataxia, hyperactive reflexes, and convulsions); however, this study was not selected for
dose-response analysis because of small group sizes (n = 3/sex) and uncertainty in measures of
exposures (e.g., purity of the test material was not specified, and reported emesis in some animals
likely influenced the delivered dose).
Although the gavage studies reporting incidence data were preferred over four dietary
studies (Lish etal.. 1984: Levine etal.. 1983: Levine et al.. 1981a: von Oettingen etal.. 1949) that did
not provide incidence data, it is important to note that the reported neurotoxic effects in the dietary
studies were observed at dose levels higher than the doses at which effects were observed in the
gavage studies (Crouse et al.. 2006: Cholakis etal.. 1980: Martin and Hart. 1974). Given this
potential difference based on dosing method, the dietary studies were also considered for
quantitative analysis, despite the lack of incidence data, to evaluate the influence of oral dosing
method on candidate values. In the 2-year study by Levine etal. (1983). a LOAEL for nervous
system effects (convulsions, tremors, and hyper-irritability) of 40 mg/kg-day and a NOAEL of
8 mg/kg-day were identified. Other studies identified higher effect levels [i.e., 100 mg/kg-day in
the 2-year mouse study by Lish etal. (1984) and 50 mg/kg-day in the 3-month rat study by von
Oettingen et al. (1949)] and, except for Lish etal. (1984). used shorter exposure durations. The
unusual dosing regimen in the Cholakis etal. (1980) 13-week mouse study precluded identification
2-4
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
of a NOAEL and LOAEL, and the single-dose design of the 6-week dog study by von Oettingen et al.
(1949) did not allow identification of a NOAEL. As discussed in Section 1.2.1 and Table 1-3, the
technical report of the 13-week study by Levine etal. (1981a) inconsistently identified the dose
level at which convulsions occurred; therefore, a reliable NOAEL and LOAEL from this study could
not be identified.
Therefore, two gavage studies, Crouse etal. (20061 and Cholakis etal. (19801. and one
dietary study, Levine etal. (1983). were selected for dose-response analysis.
Urinary System (Kidney and Bladder) Effects
Medullary papillary necrosis and hemorrhagic/suppurative cystitis were selected for
dose-response analysis as biologically significant measures of kidney and urinary bladder effects,
respectively (see Section 1.3.1 for discussion). These histopathologic lesions of the urinary system
were primarily observed in the 2-year study by Levine etal. (1983). Levine etal. (1983) included
histopathologic examination of kidney and bladder tissues at 6-, 12-, and 24-month time points;
included four dose groups and a control group, and adequate numbers of animals per dose group
(75/sex/group, with interim sacrifice groups of 10/sex/group at 6 and 12 months); and reported
individual animal data. Other studies in rats using subchronic exposure durations or lower dose
levels did not observe similar effects on the urinary system as did Levine etal. (1983). and studies
in mice suggest that this species is less sensitive to RDX toxicity on the urinary system. Therefore,
incidence data from the 2-year dietary study by Levine etal. (1983) were selected for
dose-response analysis.
Prostate Effects
Suppurative prostatitis, as reported in male rats in the Levine etal. (1983) study, was
selected for dose-response analysis as a biologically significant measure of prostate effects (see
Section 1.3.1 for discussion). This study included histopathologic examination of the prostate at 6-,
12-, and 24-month time points; included four dose groups and a control group, and adequate
numbers of animals per dose group (75/sex/group, with interim sacrifice groups of 10/sex/group
at 6 and 12 months); and reported individual animal data. Levine etal. f!9831. the only study to
identify an increased incidence of suppurative prostatitis associated with RDX exposure, was
selected for dose-response analysis.
2.1.2. Methods of Analysis
No biologically based dose-response models are available for RDX. In this situation, the EPA
evaluates a range of dose-response models thought to be consistent with underlying biological
processes to determine how best to empirically model the dose-response relationship in the range
of the observed data. Consistent with this approach, EPA evaluated dose-response information
with the models available in EPA's Benchmark Dose Software (BMDS, Versions 2.4 and 2.5). EPA
estimated the BMD and BMDL using a benchmark response (BMR) selected for each effect. A
2-5
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
conceptual model of the analysis approach used for RDX is provided in Figure 2-1. In this
assessment, points of departure (PODs) are identified through BMD modeling (preferred) or
identification of a NOAEL, and followed by animal-to-human extrapolation using PBPK models or
the application of a dosimetric adjustment factor (DAF), depending on the data available.
BMD modeling
(or NOAEL)
> f
> '
Animal PBPK
model
Dosimetric
adjustment
factor (DAF)
Human PBPK
model
> t
Animal internal dose
Animal administered
(external) dose
Human equivalent
(external) dose (HED)
Human equivalent
(external) dose (HED)
PBPK MODELING BW3'4 SCALING
Figure 2-1. Conceptual approach to dose-response modeling for oral
exposure.
BW3/4 = body weight scaled to the % power; HED = human equivalent dose.
Nervous System Effects
Incidence data for convulsions from Crouse etal. (20061 and Cholakis etal. (19801 were
amenable to BMD modeling. For Crouse etal. (20061. statistical analysis conducted by EPA
indicated no significant difference in convulsion rates of male and female rats (exact Wald-type x2
test, accounting for dose; see Table 2-2); thus, combined incidence data from male and female rats
were used for modeling convulsion data from this study.
2-6
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 2-2. Summary of derivation of point of departures (PODs) following oral exposure to hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX)
Endpoint and reference
(exposure duration/route)
Species/ sex
Model3
BMR
BMD
(mg/kg-d)
BMDL
(mg/kg-d)
PODhed (mg/kg-d)
Administered
doseb
RDX
AUCC
RDX
Cmax^
Nervous system
Incidence of convulsions
Crouse et al. (2006)
(90-d/gavage)
Male and female
F344 rat, combined12
Multistage
3°
5% ER
5.19
2.66
0.64
1.3
1.7
Incidence of convulsions
Cholakis et al. (1980)
(GDs 6-19/gavage)
Female F344 rat
Quantal-
linear
5% ER
0.915
0.628
0.15
0.31
0.41
Incidence of convulsions
Levine et al. (1983)
(2-yr/diet)
Male and female
F344 rat
LOAEL = 40 mg/kg-d; NOAEL = 8 mg/kg-df
1.9
3.9
4.3
Urinary system (kidney and bladder)
Kidney: medullary papillary necrosis
Levine et al. (1983)
(2-yr/diet)
Male F344 rat
LOAEL = 40 mg/kg-d; NOAEL = 8 mg/kg-dg
1.9
3.9
4.3
Urinary bladder: hemorrhagic/suppurative cystitis
Levine et al. (1983)
(2-yr/diet)
Male F344 rat
Multistage
3°
10% ER
20.0
11.6
2.8
5.6
6.3
2-7
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 2-2. Summary of derivation of point of departures (PODs) following oral exposure to hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) (continued)
Endpoint and reference
(exposure duration/route)
Species/ sex
Modela
BMR
BMD
(mg/kg-d)
BMDL
(mg/kg-d)
PODhed (mg/kg-d)
Administered
doseb
RDX
AUCC
RDX
Cmax''
Prostate
Incidence of suppurative prostatitis
Levine et al. (1983)
(2-yr/diet)
Male F344 rat
LogProbit
10% ER
1.67
0.47
0.11
0.23
0.25
AUC = area under the curve; BW3/4 = body weight scaled to the % power; Cmax = peak concentration; DAF = dosimetric adjustment factor; ER = extra risk;
HED = human equivalent dose.
aFor modeling details, see Appendix D.
bPOD was converted to an HED using a standard DAF based on BW3/4. See Section 2.1.2, Methods of Analysis/Extrapolation using BW3/4 scaling for DAFs.
cPOD was converted to an HED based on the equivalence of internal RDX dose (expressed as AUC for RDX concentration in arterial blood) derived using PBPK
models. The specific ratio applied to studies that used gavage dosing (i.e., Crouse et al. (2006) and Cholakis et al. (1980)) and to studies that used continuous
(dietary) dosing (i.e., Levine et al. (1983)) was 0.487. See Appendix C, Section C.1.5 (Rat to Human Extrapolations) for more details.
dPOD was converted to an HED based on the equivalence of internal RDX dose (expressed as RDX Cmax in arterial blood, Cmax) derived using PBPK models. The
specific ratio applied to studies that used gavage dosing (i.e., Crouse et al. (2006) and Cholakis et al. (1980)) was 0.645. The ratio for studies that used
continuous (dietary) dosing (i.e., Levine et al. (1983)) was 0.540. See Appendix C, Section C.1.5 (Rat to Human Extrapolations) for more details.
eExact Wald-type x2 test for differences in convulsion incidence across sexes yielded p-value > 0.05.
fNervous system effects for male and female rats reported qualitatively; incidence of convulsions and other nervous system effects was not reported.
Therefore, available data do not support BMD modeling.
gBMD modeling was not supported for this data set; see discussion in text.
2-8
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
In general, the strong preference is to use a less severe endpoint as the basis for a
noncancer toxicity value. As discussed in the evaluation of nervous system effects (Section 1.2.1),
evidence from other seizurogenic compounds with modes of action like RDX suggests that other
generally subclinical cognitive and behavioral neurological effects are likely to occur at lower RDX
doses, although limited investigation has been conducted to establish such subclinical effects.
EPA guidelines indicate that a BMR of 5% or lower may be warranted for frank effects [such
as convulsions; (U.S. EPA. 2012a)]. EPA considered BMRs of 1 and 5% extra risk (ER) for
convulsions. A BMR of 1% ER was considered appropriate to address the severity of convulsions, a
frank effect. The use of a 1% ER BMR for convulsions in Crouse etal. (2006) resulted in
extrapolation below the range of the experimental doses. More specifically, the BMD of
3.02 mg/kg-day with a 1% BMR was below the LOAEL of 8 mg/kg-day (with a 15% response rate
for convulsions), although the BMD was not far below the dose range of 4-15 mg/kg-day used in
the study.
EPA considered the trade-off between (1) a BMR of 1% ER that addresses the severity of
convulsions, but results in extrapolation outside the experimental range and (2) a 5% ER BMR that
may be inadequate for addressing the severity of these specific outcomes, but is more consistent
with the available data. EPA selected a BMR of 5% ER, addressing the lack of incidence data for less
severe endpoints by applying the database uncertainty factor (i.e., reflecting insufficient
investigation of less severe, subclinical, nervous system effects for RDX). See Section 2.1.3,
Derivation of Candidate Values, for further discussion of the database uncertainty factor.
Because incidence data for convulsions were not provided by Levine etal. (1983). a NOAEL
was used as the point of departure (POD) for this data set rather than a BMDL.
Table 2-2 summarizes the PODs derived for each data set. More detailed BMD modeling
information is available in Appendix D; BMD and BMDL estimates for 1 and 10% ER for the selected
model (see Appendix D, Section D.1.2, Tables D-3 and D-4) are provided for comparative purposes.
Urinary System (Kidney and Bladder) Effects
Incidence data for medullary papillary necrosis in the kidney [as reported by Levine et al.
(1983)] was considered unsuitable for modeling. Aside from the lowest positive dose (which had
incidence of 1/55), only the highest dose group had a positive response (18/31 or 58%), which was
higher than a level of change considered to be minimally biologically significant (e.g., 10% ER). In
this case, because there is insufficient information to estimate the BMD (U.S. EPA. 2012a). these
data were not modeled. In the absence of sufficient information to conduct BMD modeling, a
NOAEL of 8 mg/kg-day was used as the POD for this data set (see Table 2-2).
Incidence data for hemorrhagic/suppurative cystitis in the urinary bladder as reported by
Levine etal. (1983) were amenable to BMD modeling. The BMDS models were fit to these data
using a BMR of 10% ER, under the assumption that it represents a minimally biologically significant
level of change. Table 2-2 summarizes the POD derived using data on the incidence of
2-9
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
hemorrhagic/suppurative cystitis. More detailed BMD modeling information is available in
Appendix D, Section D.1.2, Table D-6.
Prostate Effects
Incidence data on suppurative prostatitis as reported by Levine etal. (1983) were amenable
to BMD modeling. A BMR of 10% ER was applied under the assumption that it represents a
minimally biologically significant level of change. Table 2-2 summarizes the POD derived using
data on the incidence of suppurative prostatitis. More detailed BMD modeling information is
available in Appendix D, Section D.1.2, Table D-7.
Human Extrapolation
EPA guidance (U.S. EPA. 20111 describes a hierarchy of approaches for deriving human
equivalent doses (HEDs) from data in laboratory animals, with the preferred approach being PBPK
modeling. Other approaches can include using chemical-specific information in the absence of a
complete PBPK model. In lieu of either reliable, chemical-specific models or data to inform the
derivation of human equivalent oral exposures, a body-weight scaling to the % power (BW3/4)
approach is generally applied to extrapolate toxicologically equivalent doses of orally administered
agents from adult laboratory animals to adult humans.
Candidate PODs for endpoints selected from rat and mouse bioassays were expressed as
HEDs. HEDs were derived using both PBPK modeling (with alternative measures of internal dose),
and a BW3/4 scaling approach. These approaches are outlined in Figure 2-1, and the resulting
PODhed values are presented in Table 2-2.
Extrapolation Using Physiologically Based Pharmacokinetic (PBPK) Modeling.
PBPK models for RDX in rats, humans, and mice have been published (Sweeney etal..
2012a: Sweeney etal.. 2012b: Krishnan et al.. 2009) based on RDX-specific data. EPA evaluated and
further developed these models for extrapolating doses from animals to humans (see Appendix C,
Section C.1.5). In general, appropriately chosen internal dose metrics are expected to correlate
more closely with toxic responses than external doses for effects that are not occurring at the point
of contact (McLanahan et al.. 2012). Therefore, PBPK model-derived arterial blood concentration of
RDX is considered a better dose metric for extrapolating health effects than administered dose
when there is adequate confidence in the estimated value. The PBPK models for RDX were used to
estimate two dose metrics: the area under the curve (AUC) and the peak concentration (Cmax) for
RDX concentration in arterial blood. The AUC represents the average blood RDX concentration for
the exposure duration normalized to 24 hours and the Cmax represents the maximum RDX
concentration for the exposure duration.
Ideally, use of RDX concentrations in the brain would serve as the internal dose metric for
analyzing convulsion data. However, the blood concentration of RDX was preferred as the dose
metric because there is greater confidence in modeling this variable due to the substantially greater
2-10
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
number of measurements of RDX blood levels used in calibrating model parameters. Additionally,
predictions of RDX concentrations in the brain are highly correlated with predictions of RDX blood
concentrations because the model is flow-limited and no metabolism is assumed in that organ.
RDX-induction of convulsions and seizures appears to be more strongly driven by dose than
exposure duration, which might argue for use of peak blood concentration as an appropriate dose
metric; however, greater confidence was placed in model estimates of blood AUC than peak blood
concentrations because, as discussed in Appendix C, Section C.1.5, the rate constant for oral
absorption is uncertain, and peak concentrations are more sensitive to variations in this parameter
than average values. In addition, some biological support for blood AUC, rather than peak blood
concentration, comes from (1) mechanistic information on RDX binding at the picrotoxin
convulsant site of the GABA channel and (2) observations from animal studies of convulsions
occurring only after repeated exposures. [See Sections 1.2.1 and 2.1.3 (subchronic-to-chronic UF)
for further discussion.] Largely because of the greater uncertainly in peak blood concentration, and
in light of some evidence for a possible contribution of RDX exposure duration on the manifestation
of neurotoxicity, the AUC for RDX concentration in arterial blood was selected as the internal dose
metric for analyzing dose-response data for convulsions.
Tissue-specific dose metrics for kidney, bladder, and prostate were not available in the
PBPK model. Because effects in these organs were observed only after subchronic or chronic
exposure to RDX (i.e., there is no evidence that effects are associated with peak exposure) and
because greater confidence was placed in model estimates of blood AUC, AUC for RDX
concentration in arterial blood was also selected as the internal dose metric for analyzing
dose-response data for the kidney, urinary bladder, and prostate.
PODhed values based on both blood AUC and blood Cmax are presented in Table 2-2 for
completeness. As demonstrated in Table 2-2, the PODhed values derived using administered dose,
AUC, and Cmax do not differ greatly; thus, the selection of AUC is not a major determinant of the POD.
The rodent PBPK model was applied to the BMDLs generated from BMD modeling to
determine the animal internal dose, expressed as the AUC of RDX blood concentration, and
representing the cross-species toxicologically equivalent (internal) dose. The human PBPK model
was then applied to derive the corresponding HEDs (see Figure 2-1). Because the AUC is linear
with exposure level, at least in the exposure range of interest, the value of the HED would be the
same whether the rat or mouse PBPK model is applied before or after BMD modeling is performed.
Because the sequence of the calculation does not influence the results, applying the PBPK model
after BMD modeling is more efficient—BMD modeling would not have to be redone if there were
changes to the PBPK model, and it is easier to evaluate and show two dose metrics (as discussed
above). Because of the relatively high confidence in the PBPK models developed for the rat and
human, these models were used to derive reliable internal dose metrics for extrapolation. For data
sets selected from the rat bioassays, the candidate oral values were calculated assuming
2-11
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
cross-species toxicological equivalence of the AUC of RDX blood concentration derived from PBPK
modeling.
Extrapolation Using Body-Weight Scaling at % Power (BW3/4)
HEDs were also calculated using a BW3/4 scaling approach consistent with EPA guidance
(U.S. EPA. 20111. PODs (BMDLs or NOAELs) based on the RDX dose administered in the
experimental animal study were adjusted by a standard DAF derived as follows:
DAF = (BWa1/4/BWh1/4), (2-1)
where
BWa = animal body weight
BWu = human body weight
Using BWa values of 0.25 kg for rats and 0.036 kg for mice and a BWh of 70 kg for humans
(U.S. EPA. 19881. the resulting DAFs for rats and mice are 0.24 and 0.15, respectively. Applying the
DAF to the POD identified for effects in adult rats or mice yields a PODhed as follows (see Table 2-2):
POD\u \) = laboratory animal dose (mg/kg-day) x DAF (2-2)
Further details of the BMDL modeling, BMDS outputs, and graphical results for the best-fit
model for each data set included in Table 2-2 can be found in Appendix D, Section D.l. Details of the
PBPK model evaluation used for extrapolation from BMDL values can be found in Appendix C,
Section C.1.5. Table 2-2 summarizes the results of the BMD modeling and the PODhed for each data
set discussed above.
2.1.3. Derivation of Candidate Values
Under EPA's A Review of the Reference Dose and Reference Concentration Processes [(U.S.
EPA. 20021: Section 4.4.5], and as described in the Preamble, five possible areas of uncertainty and
variability were considered when determining the application of UFs to the PODs presented in
Table 2-2. An explanation follows.
An intraspecies uncertainty factor (UFh) of 10 was applied to all PODs to account for
potential differences in toxicokinetics and toxicodynamics in the absence of information on the
variability of response in the human population following oral exposure to RDX. The available
human pharmacokinetic data are not sufficient to inform human kinetic variability and derive a
chemical-specific UF for intraspecies uncertainty.
An interspecies uncertainty factor (UFa) of 3 (101/2 = 3.16, rounded to 3) was applied to all
PODs to account for uncertainty in characterizing the toxicokinetic and toxicodynamic differences
2-12
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
between rodents and humans. In the absence of chemical-specific data to quantify this uncertainty,
EPA's BW3/4 guidance (U.S. EPA. 2011) recommends using an uncertainty factor of 3. For data sets
from the rat bioassays, a PBPK model was used to convert internal doses in rats to external doses in
humans (see rationale in Section 2.1.2—Human Extrapolation). This reduces toxicokinetic
uncertainty in extrapolating from the rat to humans, but does not account for interspecies
differences due to toxicodynamics. A UFa of 3 was applied to account for this remaining
toxicodynamic uncertainty and any residual toxicokinetic uncertainty not accounted for by the
PBPK model.
A subchronic-to-chronic uncertainty factor (UFS) of 1 was applied to all PODs. Where a POD
was based on a 2-year bioassay (i.e., for urinary system [kidney/urinary bladder] and prostate), a
UFs was not necessary because the RfD was based on effects associated with chronic exposure.
Where a POD was based on studies of subchronic exposure to RDX (i.e., for nervous system effects),
EPA considered applying a UFs of either 3 or 1. Although EPA guidance recommends a default UFs
of 10 on the assumption that effects in a subchronic study would occur at an approximately 10-fold
higher concentration than in a corresponding (but absent) chronic study (U.S. EPA. 2002). the RDX
database does not support a UFs of 10. This determination is based on the MOA for nervous system
effects and the support across studies that nervous system effects are more strongly driven by dose
than duration of exposure (see Section 1.2.1). The argument for applying a UFs of 3 is based on
some remaining uncertainty regarding the potential for effects to accumulate over time. The MOA
strongly suggests that the convulsive effects of RDX are driven by the transient binding of RDX to
target GABA receptors in the brain. However, the lack of complete reversibility of the inhibited
GABAergic signaling after removal of RDX in vitro by Williams etal. (2011). as well as observations
from related chemicals suggesting that prolonged decreases in inhibitory tone might predispose
nervous system tissues to future seizurogenic events (see Section 1.2.1), introduces the possibility
that mechanisms leading to cumulative effects over time have not been adequately investigated.
The application of a UFs of 1 is supported by the findings across most studies that convulsions
occurred shortly after dosing (minutes to hours) and generally did not appear to be appreciably
influenced by duration of exposure. For example, convulsive effects were observed after a single
RDX dose of 12.5 mg/kg-day (the lowest dose tested in this experiment), and several animals
exhibited reduced seizure thresholds to other convulsants at 10 mg/kg-day [again, the lowest dose
tested; (Burdette etal.. 1988)]. These doses are comparable to the LOAEL of 8 mg/kg-day from the
90-day study by Crouse etal. (2006). Similarly, at 12 and 15 mg/kg-day in the 90-day study by
Crouse etal. (2006). neurotoxic signs were present in >80% of animals beginning on Day 0 and
continuing for the duration of the experiment. Convulsions in Crouse etal. (2006) were not
observed until 7 to 15 days of exposure at doses of 10-15 mg/kg-day, and only after 48 days of
exposure in rats receiving 8 mg/kg-day RDX (Tohnson. 2015a). In a 14-day range-finding study in
6 animals/group, Crouse etal. (2006) reported that neuromuscular signs (tremors, convulsions)
were observed at 17 mg/kg-day and above (the next lower dose was 8.5 mg/kg-day). Thus, studies
2-13
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
with comparable dosing methods reported seizure-related effects within the narrow range of
8-17 mg/kg-day, regardless of exposure duration (acute to subchronic). Data from
chronic-duration rodent studies (Lish etal.. 1984: Levine etal.. 19831 identifies a higher effective
dose range (>35 mg/kg-day) for convulsions than these acute and subchronic exposure studies
(and would result in the identification of a higher POD). However, because of cross-study
differences in methods of outcome measurement, peak internal dose from gavage administration
versus dietary administration, physical form of RDX (e.g., particle size), and dose matrix in the
dietary versus gavage preparations that could have influenced absorption rate and internal (e.g.,
peak) RDX dose, direct comparison of the effective convulsive doses from the available subchronic
and chronic studies is not appropriate. Overall, evaluation of the available evidence leaves open the
possibility for a small influence of chronic (as compared to subchronic) RDX exposure duration on
the manifestation of neurotoxicity; however, current data suggest that any such influence
specifically on convulsions would be small such that a UFs of 3 would not be warranted. Based on
the strong evidence supporting a negligible to minimal impact of exposure duration on the effective
dose for convulsions, a UFs of 1 was applied to PODs for neurotoxic effects derived from studies of
less than chronic duration.
A LOAEL-to-NOAEL uncertainty factor (UFl) of 1 was applied to all POD values because
every POD was a BMDL or a NOAEL. When the POD is a BMDL, the current approach is to address
this factor as one of the considerations in selecting a BMR for BMD modeling. In this case, the BMR
for modeled endpoints was selected under the assumption that the BMR represents a minimal,
biologically significant change for these effects.
A database uncertainty factor (UFd) of 10 was applied to all POD values. The oral toxicity
database for RDX includes subchronic and chronic toxicity studies in the rat and mouse, a
two-generation reproductive toxicity study in the rat, developmental toxicity studies in the rat and
rabbit, and subchronic studies (with study design limitations) in the dog and monkey. As discussed
below, some uncertainty is associated with characterizing RDX neurotoxicity.
EPA prefers to identify reference values based on upstream (less severe) effects that would
precede frank effects like convulsions, and uncertainty remains in understanding RDX-induced
neurotoxicity. In part, this is due to limitations in study design to assess neurotoxicity across the
RDX database; the frequency of animal observations in the available studies raises concerns that
there may be underreporting of the true incidence of convulsions, and in general the reporting of
this effect does not include a measure of the severity at the time of observation. No follow-up
studies were identified that employed more sensitive assays to assess more subtle neurotoxicity,
and the database lacks a chronic-duration study that could inform residual uncertainty regarding
the potential for chronic exposure to magnify effects (as compared to subchronic exposure). As
noted by the SAB, the convulsion endpoint in rodents does not capture the breadth of potential
human hazard, and the lack of information on more sensitive endpoints, including cognitive and
2-14
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
behavioral effects, as well as developmental neurotoxicity, is a significant data gap. Uncertainties in
the database for RDX neurotoxicity could be addressed by:
• Analysis of "seizures" using more detailed behavioral scoring methods. In the available
studies, "convulsion" or "seizure" (depending on the reporting in the study) might indicate a
range of observable behaviors in response to altered brain activity, ranging from
involuntary limb and facial twitches to tonic-clonic seizures in which animals exhibit a
sustained (seconds to hours) and widespread loss of muscle control sometimes resulting in
respiratory arrest and/or death. As there are studies where convulsions occur at the same
dose as mortality, the convulsive activity in these studies is interpreted as severe. Scoring
methods quantifying the occurrence of different behavioral aspects of the RDX-induced
convulsions, such as the Racine scale (Racine. 1972) employed in Burdette etal. (1988).
would provide a much more accurate, complete, and likely more sensitive measure of RDX
neurotoxicity.
• Additional electrophysiological measures of epileptiform activity. Well-established and
sensitive methods for evaluating brain activity exist. These measures could not only better
describe the profile of RDX-induced convulsant activity, but could also be used to identify
and quantify subconvulsive effects of RDX exposure (e.g., EEG spiking). Electrophysiological
effects of RDX in vitro and in vivo have already been characterized by Williams etal. (2011).
Additional studies building on this work, looking at the effects of different concentrations of
RDX, could potentially identify more sensitive measures of RDX neurotoxicity.
• A FOB conducted by Crouse etal. (2006) provides some limited information on
neurobehavioral effects associated with RDX exposure, yet the results of that study did not
identify notable effects associated with RDX exposure. While some components of the FOB
testing conducted by Crouse etal. f20061 would be expected to give a screening-level
evaluation of some stimuli-induced behaviors that have the potential to be related to
seizures (e.g., response to handling, touch, click, or open field), these observational
descriptions are insensitive and are expected to have missed potential subconvulsive
effects. Additional studies addressing the potential for subconvulsive behaviors resulting
from RDX exposure would be informative. For example, Burdette etal. (1988) examined
seizure susceptibility in male Long-Evans rats at gavage doses >10 mg/kg; spontaneous
seizures were already observed in this study at 12.5 mg/kg-day. Further evaluation of
seizure susceptibility at lower doses and with longer exposure durations, as well as
evaluations of potential effects of subconvulsive doses on subtler behaviors that might be
related to RDX neurotoxicity (e.g., motor, anxiety, or social behaviors; learning and memoiy
tests) may identify additional measures of RDX neurotoxicity.
• Further evaluation of potential developmental neurotoxicity associated with RDX exposure
(see Section 1.3.3 for discussion). Models for examining seizure-related behaviors during
development exist, mainly involving manipulation and analyses in preweanling rodents.
Hess-Ruth etal. (2007) reported possible transfer of RDX to offspring during gestation, as
well as the presence of RDX in the milk of dams, indicating a potential for lactational
transfer of RDX to offspring. Examination of specific developmental neurotoxicity
endpoints has not been conducted in studies of RDX toxicity. Well-conducted
developmental neurotoxicity studies could further rule out the possibility that RDX
exposure during development might result in immediate or delayed seizure activity or
2-15
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
predispose animals to developing seizures as adults. Such studies could also identify other
more sensitive indicators of toxicity.
Overall, while the RDX database adequately covers major systemic effects, including
reproductive and developmental effects, uncertainties in the adequacy of the database were
identified in characterizing the neurotoxicity hazard. Because of the possibility that additional
studies described above may lead to identification of a more sensitive endpoint or a lower POD, a
UFd of 10 was applied to all derived PODs.
Table 2-3 is a continuation of Table 2-2 and summarizes the application of UFs to each
PODhed to derive a candidate value for each data set. The candidate values presented in Table 2-3
are preliminary to deriving the organ/system-specific reference values. These candidate values are
considered individually in the selection of a representative oral reference value for a specific hazard
and subsequent overall RfD for RDX.
2-16
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 2-3. Effects and corresponding derivation of candidate values
Endpoint and reference
PODhed3
POD
type
UFa
UFh
UFl
UFs
UFd
Composite
UF
Candidate
value
(mg/kg-d)
Nervous system (rat)
Incidence of convulsions
Crouse et al. (2006)
1.3
BMDLob
3
10
1
1
10
300
4.3 x 10"3
Incidence of convulsions
Cholakis et al. (1980)
0.31
BMDLob
3
10
1
1
10
300
1.0 x 10"3
Incidence of convulsions
Levine et al. (1983)
3.9
NOAEL
3
10
1
1
10
300
1.3 x 10"2
Urinary system (kidney and bladder) (rat)
Kidney: incidence of
medullary papillary necrosis
Levine et al. (1983)
3.9
NOAEL
3
10
1
1
10
300
1.3 x 10"2
Urinary bladder: incidence
of hemorrhagic/
suppurative cystitis
Levine et al. (1983)
5.6
BMDLio
3
10
1
1
10
300
1.9 x 10"2
Prostate (rat)
Incidence of prostate
suppurative inflammation
Levine et al. (1983)
0.23
BMDLio
3
10
1
1
10
300
7.6 x 10"4
aPODHED values based on data from the rat were derived using PBPK modeling, with the HED based on
equivalence of internal RDX dose expressed as AUC for RDX concentration in arterial blood (see Section 2.1.2
and discussion of the PBPK models above and in Appendix C, Section C.1.5).
Figure 2-2 presents graphically the candidate values, UFs, and PODhed values, with each bar
corresponding to one data set described in Tables 2-2 and 2-3.
2-17
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Convulsions; Crouse et
a I. (2006)
Convulsions; Cholakis et
al. (1980)
Convulsions; Levine et
al. (1983)
Kidney - medullary
papillary necrosis;
Levine et al. (1983}
Urinary bladder -
hemorrhagic/
suppurative cystitis;
Levine et al. (1983}
Prostate - suppurative
inflammation; Levine et
al. (1983)
~ Candidate value
® PODhed
Composite UF
0.0001 0.001 0.01 0.1
mg/kg-day
10
Figure 2-2. Candidate values with corresponding point of departure (POD)
and composite uncertainty factor (UF).
2-18
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
2.1.4. Derivation of Organ/System-Specific Reference Doses
Table 2-4 distills the candidate values from Table 2-3 into a single value for each organ or
system. Organ- or system-specific reference values may be useful for cumulative risk assessments
that consider the combined effect of multiple agents acting at a common site. For example,
organ/system-specific reference values can be used to refine the hazard index (HI)25 as described in
EPA's Risk Assessment Guidance for Superfund (U.S. EPA. 1989). As noted by U.S. EPA (1989), one
limitation of the HI approach is the potential to overestimate effects when this approach is applied
to multiple chemicals that induce different types of effects or do not act by the same mode of action.
The availability of organ/system-specific reference values allows risk assessors to calculate His for
chemicals acting at a common site (i.e., and thereby more likely to induce similar effects). In
addition, derivation of multiple reference values for a chemical based on other potential health
effects can provide better characterization of the toxicity that may occur at exposures higher than
the overall reference value. Therefore, derivation of organ/system-specific reference values may be
useful to EPA program and regional offices to identify other potential health hazards above the
reference dose and to inform decisions involving multiple-chemical exposures based on a common
mode of action or common target organ.
Table 2-4. Organ/system-specific reference doses (RfDs) and overall RfD for
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
Effect
Basis
RfD
(mg/kg-d)
Study exposure
description
Confidence
Nervous system
Incidence of convulsions
(Crouse et al., 2006)
4 x 10"3
Subchronic
Medium
Urinary system
Incidence of kidney medullary papillary
necrosis
(Levine et al., 1983)
1 x 10"2
Chronic
Medium
Prostate
Incidence of suppurative prostatitis
(Levine et al., 1983)
8 x 10"4
Chronic
Low
Overall RfD
Nervous system
4 x 10"3
Subchronic
Medium
Nervous System Effects
The organ/system-specific RfD for nervous system effects was based on the incidence of
convulsions in F344 rats reported in Crouse etal. (2006). a well-conducted study that used a
99.99% pure form of RDX, five closely spaced dose groups that provided a good characterization of
the dose-response curve for convulsions, and an endpoint (convulsions) that was replicated across
25The HI is the sum of hazard quotients (HQs) for multiple chemicals and/or multiple exposure pathways,
where the HQ is derived as the ratio of the exposure level to a single chemical (e.g., in mg/kg-day or ppm in
air) to the RfD (or RfC) for that chemical (U.S. EPA. 1989). The HQ and HI are both unitless values.
2-19
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
multiple studies. Although the candidate value derived from the developmental toxicity study in
F344 rats by Cholakis etal. (1980) is approximately fourfold lower, deficiencies in the Cholakis et
al. (19801 study resulted in a candidate value with less confidence than the value derived from
Crouse etal. (20061. Crouse etal. (20061 was better designed to assess the nervous system effects
of RDX, whereas Cholakis etal. (19801 was designed as a developmental toxicity study with only
routine monitoring of clinical signs (the methods section states that "Dams were monitored daily
for toxic signs"). Crouse etal. (2006) used five dose groups (plus the control) that provided good
characterization of the dose-response curve for RDX-induced convulsions, whereas Cholakis et al.
(1980) used only three dose group (plus the control) with an order-of-magnitude dose spacing,
resulting in a less well-defined dose-response curve for this endpoint. Lack of
uniformity/homogeneity of the dosing preparation in Cholakis etal. (1980) raised concerns about
exposure quality and the potential for under- and over-dosing animals. Cholakis etal. f!9801 noted
difficulty maintaining uniform dosing suspensions, and RDX concentrations in the gavage study
ranged from 36 to 501% of target concentrations. In contrast, Crouse etal. (2006) used methods to
ensure uniform dosing suspensions; the actual RDX concentrations varied from 83 to 114% of
target concentrations, and the 114% suspension was adjusted to 100% before administration. In
light of evidence that nervous system effects are more strongly driven by dose than duration of
exposure (see Section 2.1.3), the wide deviations from the target doses in Cholakis etal. (1980)
decrease the confidence in the quantitative use of this study. Further, Crouse etal. (2006) used a
higher purity test material than did Cholakis etal. (1980) (99.99 vs. 88.6%, respectively). Finally,
the Crouse et al. (2006) study used a longer exposure duration (90 days) than did the Cholakis et al.
(1980) study (14 days) and is more representative of a chronic exposure duration. The lower
candidate value from the Cholakis etal. f 19801 developmental toxicity study could indicate that
pregnant animals are a susceptible population, which could support selection of this study as the
basis for the RfD; however, as discussed in Section 1.3.3, the available studies in pregnant and
nonpregnant rats cannot be directly compared, and the available information is not considered
sufficient to identify pregnant animals as a susceptible population.
As discussed in Section 2.1.1, the 2-year dietary study by Levine etal. (1983) was also
considered for RfD derivation because the available oral studies suggest that bolus doses of RDX
received via gavage may induce nervous system effects at doses lower than those resulting from
dietary administration (recognizing that differences in particle size and purity of the test material
may confound direct comparisons between gavage and dietary administration). Convulsion data
from Levine etal. (1983) yielded a PODhed threefold higher than the PODhed derived from Crouse et
al. (2006). The POD derived from the Levine etal. (1983) study is considered less certain than that
derived from Crouse etal. (2006). Levine etal. (1983) did not provide information on the incidence
of neurotoxic effects; thus, BMD analysis was not supported (i.e., the POD was based on a NOAEL).
As discussed in Section 1.2.1, the frequency of daily observations in the Levine etal. (1983) study
may not have been sufficient to provide an accurate measure of the occurrence of convulsions and
2-20
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
other nervous system effects, potentially leading to their underestimation. For these reasons, and
in light of the fact that data from the Levine etal. (1983) study yielded a higher POD, Levine et al.
(19831 was not used as the basis for the organ/system-specific RfD for nervous system effects.
Urinary System (Kidney and Bladder) Effects
Dose-response analysis was conducted for two data sets representing effects on the urinary
system—incidence of medullary papillary necrosis in the kidney and incidence of
hemorrhagic/suppurative cystitis in the urinary bladder, both as reported by Levine et al. (19831.
Both effects were reported primarily in high-dose male rats in this study, and both data sets yielded
similar PODhed values (3.9 and 5.6 mg/kg-day, respectively) and candidate values (1.3 x 10"2 and
1.9 x 10"2 mg/kg-day, respectively). The smaller of the two candidate values
(1.3 x 10"2 mg/kg-day) was selected as the organ/system-specific RfD for urinary system effects.
Prostate Effects
A single data set for prostate effects, specifically the incidence of suppurative prostatitis in
male F344 rats as reported in a 2-year dietary study by Levine etal. (1983). was brought forward
for quantitative analysis. The organ/system-specific RfD for prostate effects is based on this data
set
2.1.5. Selection of the Overall Reference Dose
Multiple organ/system-specific reference doses were derived for effects identified as
hazards from RDX exposure, including organ/system-specific reference doses for the nervous
system, urinary system (kidney and bladder), and prostate. There is strong support for RDX as a
nervous system toxicant, with evidence for nervous system effects, and specifically convulsions,
observed in humans and in multiple experimental animal studies, in multiple species, and following
a range of exposure durations.
The organ/system-specific RfD for nervous system effects of 4 x 10"3 mg/kg-day is smaller
than the organ/system-specific RfD for urinary system effects (1 x 10"2 mg/kg-day), suggesting that
the RfD for nervous system effects is protective of effects on both organ systems. This is consistent
with findings from animal bioassays that show RDX-related effects on the kidney and urinary
bladder at exposure levels higher than those associated with convulsions.
The organ/system-specific RfD for nervous system effects is fivefold higher than the
organ/system-specific RfD for prostate effects of 8 x 10"4 mg/kg-day. Although smaller in value,
the RfD for prostate effects was not selected as the overall RfD. Evidence for dose-related effects on
the prostate comes from a single 2-year toxicity study in male rats (Levine etal.. 1983): a second
chronic study in the rat that evaluated prostate histopathology was not available, and the 2 -year
study in mice (Lish etal.. 1984) did not identify similar patterns of prostate inflammation. There
are also uncertainties in the diagnosis of suppurative prostatitis. Levine etal. (1983) do not
provide more extensive detail on the histopathological evaluation of the prostate to account for
2-21
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
potential variation in inflammation inherent to the different lobes of the prostate. Additionally,
male rats in the high-dose group (40 mg/kg-day) were moved from group to individual housing
during Weeks 30-40 on study, due to the high incidence of fighting. As noted by the SAB, Creasy et
al. (20121 reported that fighting may cause urogenital infections in male rats. The fighting
observed by Levine etal. (19831. along with the change in housing conditions from the other
treatment groups, increases uncertainty in the response of the high-dose group.
Therefore, the organ/system-specific RfD of 4 x 10~3 mg/kg-day for nervous system effects
in the rat as reported by Crouse etal. (2006) is selected as the overall RfD for RDX given the
strength of evidence for the nervous system as a hazard of RDX exposure and the greater
confidence in the value for nervous system effects compared to urinary system and prostate effects.
The overall RfD is derived to be protective of all types of effects for a given duration of
exposure and is intended to protect the population as a whole, including potentially susceptible
populations and life stages (U.S. EPA. 2002). Decisions concerning averaging exposures over time
for comparison with the RfD should consider the types of toxicological effects and specific life
stages of concern. Fluctuations in exposure levels that result in elevated exposures during these life
stages could potentially lead to an appreciable risk, even if average levels over the full exposure
duration were less than or equal to the RfD. In the case of RDX, no specific life stage or population
has been identified as potentially susceptible.
A subchronic-to-chronic uncertainty factor (UFs) of 1 was applied to the POD for nervous
system effects in light of the MOA for nervous system effects and the support across studies that
nervous system effects (in particular convulsions) are more strongly driven by dose than duration
of exposure [see Section 2.1.3 (subchronic-to-chronic UF)]. Therefore, the chronic RfD can be
considered appropriate for assessing health risks of less-than-lifetime as well as chronic durations
of exposure.
2.1.6. Comparison with Mortality Doses Expected to be Lethal to 1% of the Animals (LD0is)
Evidence for mortality associated with RDX exposure was previously discussed at the
beginning of Section 1.2, and in particular in relation to the effects of RDX on the nervous system
(Section 1.2.1) and kidney (Section 1.2.2). EPA did not develop an RfD for mortality because EPA
generally does not develop reference values based on frank effects such as mortality; rather,
reference values are generally based on earlier (less severe) upstream events, where possible, to
protect against all adverse outcomes. Nevertheless, additional analysis of mortality data was
undertaken because some studies (see Table 2-5) identified mortality at the same RDX dose that
induced nervous system effects (Crouse etal.. 2006: Angerhofer et al.. 1986: Cholakis etal.. 1980:
von Oettingen et al.. 1949).
2-22
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 2-5. Comparison of dose levels associated with mortality and
convulsions in selected studies
Study
Doses associated with
mortality
Doses associated with
convulsions
Crouse et al. (2006)
Rats, F344,10/sex/group
0, 4, 8,10, 12, or 15 mg/kg-d
13 wks/gavage
>8 mg/kg-d
>8 mg/kg-d
von Oettingen et al. (1949)
Rats, sex/strain not specified, 20/group
0,15, 25, or 50 mg/kg-d
13 wks/diet
>25 mg/kg-d
>25 mg/kg-d
Cholakis et al. (1980)
Rats, F344, 24-25 females/group
0, 0.2, 2.0, or 20 mg/kg-d
GDs 6-19/gavage
20 mg/kg-d
Primarily 20 mg/kg-d;
1 convulsion at 2 mg/kg-d
Angerhofer et al. (1986)
Rats, Sprague-Dawley, 39-51 mated females/group
0, 2, 6, or 20 mg/kg-d
GDs 6-15/gavage
Primarily at 20 mg/kg-d, but
one death each at 2 and
6 mg/kg-d
20 mg/kg-d
A discussion of mortality evidence for RDX is presented in Appendix C, Section C.3.1, and the
relationship between mortality and nervous system effects in Sections 1.2.1 and 1.3.1. Unscheduled
deaths were observed as early as Day 8 of a 90-day gavage study fCrouse etal.. 20061 and in
developmental toxicity studies with exposure durations of 2 weeks fAngerhofer et al.. 1986:
Cholakis etal.. 1980).
Given the proximity in the dose at which mortality and nervous system effects were
observed in several studies, the dose-response relationships for mortality were compared across
studies with durations similar to those in Table 2-5 by comparing the dose expected to be lethal to
1% of the animals (LDoi) or NOAELs derived from each study. A BMR of 1% ER was used for
modeling mortality data in light of the severity of this frank effect In addition, the LDoi values and
NOAELs for mortality were compared to BMD0i for convulsions.26 For purposes of this analysis, a
BMR of 1% ER was selected for convulsions (rather than 5% ER used in the analysis to derive the
nervous system RfD) to facilitate comparison with the LDoi values for mortality.
26BMDs were compared, as opposed to BMDLs, because, as stated on p. 20 of the BMD Technical Guidance
fU.S. EPA. 2012a). "In general, it is recommended that comparisons across chemicals/studies/endpoints be
based on central estimates; this is in contrast to using lower bounds for PODs for reference values..."
2-23
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Interpretation of mortality data from chronic exposure studies in mice and rats is
complicated by other treatment-related effects and pathology regularly observed in aging animals
(e.g., kidney pathology, neoplastic lesions). Therefore, mortality data from chronic studies were not
considered in this analysis. Other studies that were less informative and not considered in this
analysis are not presented in Table 2-6.27
Of the studies in Table 2-6, dose-response analysis was conducted for all studies that
showed an increased incidence of unscheduled deaths. LDoi values are provided in Table 2-6, and
detailed modeling results are provided in Appendix D, Section D.1.2. Mortality was observed only
at the highest dose tested at Week 11 in the 2-year mouse study by Lish etal. (19841. in the
13-week rat study by von Oettingen et al. (19491. and in the two-generation reproductive and
developmental toxicity studies by Cholakis etal. (19801. In these cases, data were not amenable to
LDoi estimation, and a NOAEL (with a confidence interval [CI] on its associated response) was used
in this comparative analysis instead.
LDoi values for mortality in Table 2-6 range from 1.7 mg/kg-day (10-day gavage exposure
in pregnant rats) to 7.9 mg/kg-day (13-week dietaiy exposure in rats), with the lower values
generally from studies that administered RDX by gavage. These values may be compared to the
BMDoi for convulsions from Crouse etal. f20061 (see Appendix D, Table D-3). The BMDoi for
convulsions of 3.0 mg/kg-day is in the middle of the distribution of calculated LDoi values, and the
lowest LDoi of 1.7 mg/kg-day is within twofold of the convulsion BMDoi of 3.0 mg/kg-day.
The NOAELs from studies where mortality was observed tend to be higher than the LDoi
values. However, NOAELs are not directly comparable to BMDoi values for several reasons. CIs for
the responses characterize some statistical uncertainty for NOAELs from studies that could not be
modeled (note that the upper bound of a CI is not directly comparable to a lower bound on a
benchmark dose). The CIs suggest that comparable 1% levels for these data sets could be lower
than the NOAELs. In addition, dose spacing can affect the interpretation of NOAELs, such as that
from the Cholakis etal. (19801 developmental toxicity study because of the wide
(order-of-magnitude) spacing between doses in that study (i.e., the reported NOAEL of 2 mg/kg-day
[see Table 2-6] is 10-fold lower than the dose associated with 21% mortality [5/24 dams] at
20 mg/kg-day [see Appendix C, Table C-10]).
27The following less informative studies were not included in the analysis of early mortality:
The 13-week dietary study in the mouse by Cholakis et al. (1 9801. Mortality was observed only in the
high-dose group (257-276 mg/kg-day time-weighted average [TWA]), and the unusual dosing regimen
precluded identification of a NOAEL or LOAEL.
The 13-week dietary study in the dog by Hart (19741 and the 13-week study in the monkey by Martin and
Hart (19741. Both studies used small group sizes (3 animals/dose group), and no animals died on study
(although one high-dose monkey was euthanized).
The 6-week dietary study in the dog from the 1949 publication by von Oettingen et al. (19491. This dog study
included only one treatment group and recorded only one death.
The 30-day gavage study in the rat by MacPhail etal. (19851 The study authors did not identify
treatment-related mortality, but reporting was limited.
2-24
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 2-6. Summary of dose-response evaluation for mortality following oral
exposure to hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
Reference
(exposure duration/route)
Species/sex
Modela
BMR
LDoi
(mg/kg-d)
LDLoi
(mg/kg-d)
Diet studies
Lish et al. (1984)
(11-wk data from 2-yr
study/diet)
Male and female
B6C3Fi mouse
Not amenable to
modeling
NOAEL: 35 mg/kg-d
95% CI for response: 0-4%
Levine et al. (1981a)
(13-wk/diet)
Male and female
F344 rat,
combined
Multistage 4°
1% ER
7.9
2.2
von Oettingen et al. (1949)
(13-wk/diet)
Rats, sex/strain
not specified
Not amenable to
modeling
NOAEL: 15 mg/kg-d
95% CI for response: 0-15%
Cholakis et al. (1980)
(two-generation design/diet)
Female CD rat
Not amenable to
modeling
NOAEL: 16 mg/kg-d
95% CI for response: 0-13%
Levine et al. (1983)
(13-wk data from 2-yr
study/diet)
Male and female
F344 rat
NA
(no mortality at
highest dose tested)
NOAEL: 40 mg/kg-d
95% CI for response: 0-4%
Cholakis et al. (1980)
(13-wk/diet)
Male and female
F344 rat
NA
(no mortality at
highest dose tested)
NOAEL: 40 mg/kg-d
95% CI for response: 0-25%
Gavage studies
Crouse et al. (2006)
(90-d/gavage)
Male and female
F344 rat,
combined
Multistage 2°
1% ER
2.1
0.46
Cholakis et al. (1980)
(GDs 6-19/gavage)
Female F344 rat
Not amenable to
modeling
NOAEL: 2 mg/kg-d
95% CI for response: 0-12%
Angerhofer et al. (1986)
(GD 6-15/gavage)
Female Sprague-
Dawley rat
Multistage 3°
1% ER
1.7
0.59
Cholakis et al. (1980)
(GDs 7-29/gavage)
Female New
Zealand White
rabbit
NA
(no mortality at
highest dose tested)
NOAEL: 20 mg/kg-d
95% CI for response: 0-22%
CI = confidence interval; ER = extra risk; LD0i = dose expected to be lethal to 1% of the animals; LDLoi = lower
confidence limit on the LD0i; NA = not available.
aFor modeling details, see Appendix D, Section D.1.2, Tables D-9 to D-12.
In general, this comparison indicates that PODs derived from mortality data would be
comparable to PODs for RDX based on convulsions. Thus, the proximity of doses associated with
mortality and convulsions, as well as the potential for such effects to occur after subchronic or
shorter-term exposures, should be taken into consideration when assessing health risks from
environmental exposures to RDX.
2-25
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
2.1.7. Uncertainties in the Derivation of the Reference Dose
To derive the RfD, the UF approach (U.S. EPA. 2000.19941 was applied to a PODhed based on
nervous system effects in rats exposed to RDX for a subchronic duration. UFs were applied to the
PODhed values to account for uncertainties in extrapolating from an animal bioassay to human
exposure, the likely existence of a diverse human population of varying susceptibilities, and to
address limitations in the database. For the most part, these extrapolations are carried out with
default approaches given the lack of data to inform individual steps. One exception is the use of
PBPK modeling to perform interspecies (i.e., rat to human) extrapolation. Uncertainties associated
with the PBPK models are considered in Appendix C, Section C.1.5.
Nervous system effects have been documented in multiple studies and animal species and
strains; however, some uncertainty is associated with the incidence of reported neurological effects
in studies that employed a study design that did not monitor animals with sufficient frequency to
accurately record neurobehavioral effects, including convulsions. In the study used to derive the
RfD fCrouse et al.. 20061. Tohnson f2015al noted that convulsions were observed infrequently
outside the dosing period; more often, seizures were observed during the 2-hour (gavage) dosing
period, typically within 60-90 minutes of dosing. Similar information was not available for other
studies to assess the likelihood that observations of convulsions were missed. However, animals
were not monitored continuously during the Crouse et al. (20061 study, and investigators reported
that nearly all observed preterm deaths in rats exposed to the three higher doses were preceded by
signs of neurotoxicity. If an animal died during the study from the effects on the nervous system,
convulsions preceding death could have been missed, resulting in an underestimation of the
incidence of convulsions. Conversely, attributing all mortality to neurotoxicity (i.e., all deaths were
preceded by convulsions that may not have been observed) could result in an overestimation of the
incidence of convulsions. A dose-response analysis of the combined incidence of seizures and
mortality from Crouse etal. (2006) was conducted to evaluate the impact of these assumptions
because the true convulsion incidence would likely fall somewhere between the observed
convulsion incidence and the combined incidence of convulsions and mortality. This analysis
revealed that the PODhed of 0.24 mg/kg-day for a combined incidence of convulsions and
mortality28 was similar to the PODhed of 0.28 mg/kg-day for convulsions alone (using a BMR of 1%
ER for comparability to the analysis with mortality data), indicating that potential underestimation
of convulsion incidence in the Crouse etal. (2006) study was not likely to impact the RfD.
Some uncertainty is also associated with the influence of the method of oral dosing on the
magnitude of dose required to induce nervous system effects. As noted in Section 1.2.1, gavage
28The PODhed values were derived from data in Crouse et al. (20061 using a BMR of 1% ER and PBPK
modeling (see Section 2.1.2 and discussion of the PBPK models in Appendix C, Section C.1.5).
Calculation of PODhed-oi based on incidence of convulsions: BMDLoi = 0.569 mg/kg-day (see Appendix D.1.2,
Table D-3); converted to PODhed-oi based on AUC for RDX in arterial blood = 0.28 mg/kg-day.
Calculation of PODhed-oi based on incidence of convulsions and mortality: BMDLoi = 0.49 mg/kg-day (see
Appendix D.1.2, Table D-5); converted to PODhed-oi based on AUC for RDX in arterial blood = 0.24 mg/kg-day.
2-26
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
administration generally induced convulsions in experimental animals at lower doses than did
dietary administration, possibly due to the bolus dose resulting from gavage administration that
could lead to comparatively faster absorption and higher peak blood concentrations of RDX. To
some extent, the uncertainty associated with method of oral dosing is reflected in the threefold
difference in the candidate PODhed values derived from the Crouse etal. (20061 (gavage
administration) and Levine etal. (19831 (dietary administration) studies. A more rigorous
examination of the effect of oral dosing method cannot be performed because of the differences in
test materials and study designs used in the available gavage and dietary studies that could also
have contributed to differences in response (e.g., test article purity and particle size, number and
spacing of dose groups, exposure duration, frequency of clinical observations, and thoroughness of
the reporting of observations).
Other sources of uncertainty related to the RDX database have already been discussed at
length, namely the lack of more sensitive measures of neurotoxicity than convulsions, the lack of
studies examining the potential for RDX exposure to cause developmental neurotoxicity, and the
possibility for some increase in the incidence of neurotoxic effects with cumulative exposure (see
Sections 1.3.3 and 2.1.3). The use of a BMR of 5% addresses some of the concern for quantification
of a frank effect such as convulsions, while application of a UFd of 10 addresses limitations in the
sensitivity of the neurotoxicity measures as well as the lack of a developmental neurotoxicity study.
2.1.8. Confidence Statement
A confidence level of high, medium, or low is assigned to the study used to derive the RfD,
the overall database, and the RfD itself, as described in Section 4.3.9.2 of EPA's Methods for
Derivation of Inhalation Reference Concentrations and Application of Inhalation Dosimetry (U.S. EPA.
19941. The overall confidence in this RfD is medium. Confidence in the principal study (Crouse et
al.. 20061 is high. The study was well conducted, used 99.99% pure RDX, and had five closely
spaced dose groups that allowed characterization of dose-response curves for convulsions in the
dose range of interest. One limitation identified by the study authors was the limited ability of the
FOB to fully identify neurobehavioral effects at doses >8 mg/kg-day due to the timing of the dosing
procedure and timing of the FOB screening. Confidence in the database is medium to low. The
database includes three chronic studies in rats and mice; eight subchronic studies in rats, mice,
dogs, and monkeys; two short-term studies; and four reproductive/developmental toxicity studies
in rats and rabbits (including a two-generation reproductive study). Confidence in the database is
reduced largely because of (1) differences in test material used across studies, (2) uncertainties in
the influence of oral dosing methods, and (3) limitations in the available studies to fully
characterize potential neurological effects and developmental neurotoxicity. As discussed in
Section 2.1.7 and Appendix C, Section C.1.5, differences in test material formulation and particle
size may affect RDX absorption and subsequent toxicity, which in turn could influence the
characterization and integration of toxicity findings across studies. The available evidence also
suggests that bolus dosing of RDX that results from gavage administration induces neurotoxicity at
2-27
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
doses lower than administration in the diet, although a rigorous examination of these differences
cannot be performed with the available database. To the extent that a bolus dose of RDX, with
associated high peak blood concentrations, may not represent likely human exposure, the use of
toxicity data from a gavage (bolus dosing) study may introduce uncertainty in the RfD. Finally, as
noted in Section 1.2.1 and 1.3.3, the convulsion incidence endpoint in rodents does not reflect the
spectrum of potential human hazard; the lack of information on developmental neurotoxicity, as
well as more sensitive cognitive and behavioral effects, introduces uncertainty into the derived RfD.
Reflecting high confidence in the principal study and medium to low confidence in the database,
overall confidence in the RfD is medium.
2.1.9. Previous Integrated Risk Information System (IRIS) Assessment
The previous RfD for RDX, posted to the IRIS database in 1988, was based on a 2-year rat
feeding study by Levine etal. (19831. The no-observed-effect level of 0.3 mg/kg-day based on
suppurative inflammation of the prostate in male F344 rats from this study was identified as the
POD. An RfD of 3 x 10"3 mg/kg-day was derived following application of an overall UF of 100
(UFa=10,UFh = 10).
2.2. INHALATION REFERENCE CONCENTRATION FOR EFFECTS OTHER
THAN CANCER
The RfC (expressed in units of mg/m3) is defined as an estimate (with uncertainty spanning
perhaps an order of magnitude) of a continuous inhalation exposure to the human population
(including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects
during a lifetime. It can be derived from a NOAEL, LOAEL, or the 95% lower bound on the
benchmark concentration, with UFs generally applied to reflect limitations of the data used.
As discussed in Section 1.3.1, the available inhalation literature does not support
characterization of the health hazards specifically associated with chronic inhalation exposure to
RDX, nor do the studies support quantitative dose-response analysis. Of the available human
epidemiological studies of RDX fWest and Stafford. 1997: Ma and Li. 1993: Hathaway and Buck.
1977). none provided data that could be used for dose-response analysis. The studies by Ma and Li
(1993) of neurobehavioral effects in Chinese workers and West and Stafford (1997) of
hematological abnormalities in ordnance factory workers had numerous methodological
limitations that preclude their use for quantitative analysis (see Literature Search Strategy | Study
Selection and Evaluation). The study by Hathaway and Buck (1977) found no evidence of adverse
health effects in munition plant workers (based on evaluation of liver function, renal function, and
hematology), and therefore does not identify a POD at which there would be an effect from which to
derive an RfC. Multiple case reports provide some evidence of effects in humans associated with
acute exposure to RDX; however, while case reports can support the identification of hazards
associated with RDX exposure, data from case reports are inadequate for dose-response analysis
2-28
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
and subsequent derivation of a chronic reference value because of short exposure durations and
incomplete or missing quantitative exposure information.
As discussed in Literature Search Strategy | Study Selection and Evaluation, a single
experimental animal study involving inhalation exposure was identified in the Defense Technical
Information Center database; the study is not publicly available. However, the study would not
have provided useful data on responses to inhaled RDX because it was limited by small numbers of
animals tested, lack of controls, and incomplete reporting of exposure levels.
Therefore, the available health effects literature does not support the derivation of an RfC
for RDX. While inhalation absorption of RDX particulates is a plausible route of exposure, there are
no toxicokinetic studies of RDX inhalation absorption to support an inhalation model. Therefore, a
PBPK model for inhaled RDX was not developed to support route-to-route extrapolation from the
RfD.
2.2.1. Previous Integrated Risk Information System (IRIS) Assessment
An RfC for RDX was not previously derived under the IRIS Program.
2.3. ORAL SLOPE FACTOR FOR CANCER
The oral slope factor (OSF) is a plausible upper bound on the estimate of risk per
mg/kg-day of oral exposure. The OSF can be multiplied by an estimate of lifetime exposure (in
mg/kg-day) to estimate the lifetime cancer risk.
2.3.1. Analysis of Carcinogenicity Data
As noted in Section 1.3.2, there is "suggestive evidence of carcinogenic potential" for RDX.
The Guidelines for Carcinogen Risk Assessment (U.S. EPA. 2005a) state:
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.
In the case of RDX, there are well-conducted studies that tested large numbers of animals at
multiple dose levels fLish etal.. 1984: Levine etal.. 19831. making the cancer response suitable for
dose-response analysis. Considering the data from these studies, along with the uncertainty
associated with the suggestive nature of the weight of evidence, quantitative analysis of the tumor
data may be useful for providing a sense of the magnitude of potential carcinogenic risk.
The incidences of liver and lung tumors in female mice from the study by Lish etal. (19841
were selected for quantitative dose-response analysis. The study by Lish etal. (19841 (1) included
2-29
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
comprehensive histopathological examination of major organs, (2) contained four dose groups and
a control, (3) used adequate numbers of animals per dose group (85/sex/group, with interim
sacrifice groups of 10/sex/group at 6 and 12 months) and a sufficient overall exposure duration
(2 years), and (4) contained adequately reported methods and results (including individual animal
data). Female mouse liver tissues from the original unpublished study by Lish etal. (1984) were
reevaluated by a PWG (Parker etal.. 2006) in order to apply more up-to-date histopathological
criteria established by Haradaetal. (1999). The updated liver tumor incidences from the PWG
reanalysis of Lish etal. (1984) were used for quantitative dose-response analysis.
In the case of both liver and lung tumors, benign and malignant tumors (i.e., adenomas and
carcinomas) were combined for dose-response analysis because benign and malignant tumors in
both organs develop from the same cell line and there is evidence for progression from benign to
the malignant stage fU.S. EPA. 2005a: McConnell etal.. 19861. In addition, the highest dose group
was excluded from the analyses because of the death of almost half the animals in that group from
overdosing. As a group, mice that survived exposure to 175 mg/kg-day RDX for 11 weeks may not
have constituted an unbiased representation of the population of animals exposed to the final high
dose of 100 mg/kg-day from Week 11 to study termination at 2 years. Sensitivity of the surviving
animals to RDX may have differed from the larger group of animals on study, and if so, to an
unknown degree. Therefore, this group was excluded because its tumor rates may not have been
representative of the population tumor rate at this dose. Female mouse liver and lung tumor
incidences from the Lish etal. (1984) study are summarized in Appendix D, Table D-13.
The incidence of hepatocellular carcinomas in male F344 rats from the study by Levine et al.
(1983) and the incidence of alveolar/bronchiolar carcinomas in male B6C3Fi mice from the study
by Lish etal. f19841 were also considered for quantitative dose-response analysis. Both studies
were well conducted, using similar study designs (described above). In both instances, the
response was less robust than the response observed in female mice from the Lish etal. (1984)
study. The hepatocellular carcinoma result in male F344 rats is based on a small number of tumors
(1/55, 0/55, 0/52, 2/55, and 2/31, respectively, at 0, 0.3,1.5, 8.0, and 40 mg/kg-day), and
inferences made from such a sparse response are uncertain. There was no increased trend in
hepatocellular adenomas and carcinomas combined. The alveolar/bronchiolar carcinomas in male
B6C3Fi mice showed a positive trend; however, a positive trend was not observed when the
incidence of adenomas and carcinomas was combined. Modeling results are provided in
Appendix D, Section D.2.3 for comparison.
2.3.2. Dose-Response Analysis—Adjustments and Extrapolation Methods
The EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA. 2005a) recommend that the
method used to characterize and quantify cancer risk from a chemical be determined by what is
known about the MOA of the carcinogen and the shape of the cancer dose-response curve. The
linear approach is recommended when there are MOA data to indicate that the dose-response curve
is expected to have a linear component below the POD or when the weight-of-evidence evaluation
2-30
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
of all available data are insufficient to establish the MOA for a tumor site fU.S. EPA. 2005al In the
case of RDX, the mode of carcinogenic action for hepatocellular and alveolar/bronchiolar tumors is
unknown (see discussion of Mechanistic Evidence in Section 1.2.7). Therefore, a linear low-dose
extrapolation approach was used to estimate human carcinogenic risk associated with RDX
exposure.
The survival curves were compared across dose groups in each study to determine whether
time of death should be incorporated in the dose-response analysis of tumors. For female mice in
Lish etal. (1984). the survival curves were determined to be similar across dose groups after
excluding the high-dose group (log-rank test, p-value > 0.10); therefore, a time-to-tumor analysis
was not necessary for this study. Tumor incidence was modeled using the multistage-cancer
models in BMDS (Versions 2.4 and 2.5). A standard BMR of 10% ER was applied to both tumor sites
in the mouse.
Given the finding of an association between RDX exposure in the female mouse and
increased tumor incidence at two tumor sites, basing the OSF on only one tumor site could
potentially underestimate the carcinogenic potential of RDX. Therefore, an analysis that combines
the results from the mouse liver and lung tumor incidence is preferred. The MS-COMBO procedure
(BMDS, Version 2.5) extends the multistage-cancer models to the case with multiple tumors
assuming independence between tumor types. There is no known biological relationship between
liver and lung tumors in RDX-exposed mice, and therefore, as noted by the NRC (1994). this
assumption of independence is not considered likely to produce substantial error in risk estimates.
The procedure derives a maximum likelihood estimate of the combined risk at a 95% confidence
level based on the parameter values obtained for the individual tumor multistage model fits.
Additional details on the MS-COMBO procedure are provided in Appendix D, Section D.2.1.
In addition, a sensitivity analysis was conducted as recommended by the SAB in its
evaluation of the external review draft of the RDX assessment (SAB. 2017). The SAB recommended
this analysis 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. The sensitivity analysis is provided in Appendix D, Section D.2.4.
EPA's preferred approach for extrapolating results from animal studies to humans is
toxicokinetic modeling. As described in Appendix C, Section C.1.5, PBPK models for RDX in mice
and humans published by Sweeney etal. (2012b) were evaluated and further developed by EPA.
Consideration was given to whether the available toxicokinetic information supported using an
internal dose metric derived by PBPK modeling. The available mechanistic data (Section 1.2.7)
point to some evidence, although not conclusive, that RDX-generated metabolites may be
implicated in the observed tumorigenicity in the female mouse. However, there are no data on the
toxicokinetics of RDX metabolites, and metabolism in the liver is the only route of elimination of
RDX in the PBPK model. In this case, as is to be expected from mass balance principles, the PBPK
modeling provides no further information; the HED obtained from the model-estimated amount of
2-31
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
total RDX metabolites scaled by BW3/4 was equal to that calculated using administered dose scaled
by BW3/4.
In addition to the lack of data on metabolism, other major uncertainties were identified in
the mouse PBPK modeling. The mouse model was based on fitting both the absorption and
metabolic rate constants to a single set of blood concentration measurements. In this study, the
lowest dose that resulted in a detectable level of RDX in blood was 35 mg/kg, a dose high enough to
manifest some toxicity in the chronic mouse bioassay. At the 4-hour time point in this study,
measurement of blood RDX was based on results from only one of six exposed mice [the five other
data points were nondetects, excluded as an outlier, or not collected because of death; (Sweeney et
al.. 2012b)]. The type of additional data that increased confidence in the rat and human models
(e.g., in vitro measurements of RDX metabolism and RDX elimination data) were not available for
mice. Consequently, confidence in the mouse model parameter values and in the calibration of the
mouse PBPK model is low. Further, no data were available to characterize the fraction of RDX
metabolized in the mouse; this is problematic considering there is evidence indicating that the role
of metabolism in RDX toxicity differs across species (e.g., mice may have more efficient or higher
expression of the CYP450 enzymes). Given the high sensitivity of the model to the metabolic rate
constant, the uncertainty in mouse toxicokinetics significantly decreases confidence in using the
mouse PBPK model for predicting mouse blood RDX concentrations (see Summary of Confidence in
PBPK Models for RDX in Appendix C, Section C.1.5 for further discussion of confidence in the mouse
model). In light of insufficient toxicokinetic information to identify a supported internal dose
metric and model uncertainties, the PBPK model developed for the mouse was not used. Consistent
with the EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA. 2005a). the approach used to
calculate an HED from the mouse tumors, in the absence of a suitable PBPK model, was adjustment
of the administered dose by allometric scaling to achieve toxicological equivalence across species.
As discussed in Section 2.1.1, the administered dose in animals was converted to an HED on
the basis of BW3/4 (U.S. EPA. 1992). This was accomplished by multiplying administered dose by
(animal body weight in kg/human body weight in kg)1/4 (U.S. EPA. 1992). where the body weight
for the mouse is 0.036 kg and the reference body weight for humans is 70 kg (U.S. EPA. 1988).
Details of the BMD modeling can be found in Appendix D, Section D.2.
2.3.3. Derivation of the Oral Slope Factor
The lifetime cancer OSF for humans is defined as the slope of the line from the BMR (10%
ER) at the BMDL (expressed as the HED) to the estimated control response at zero
(OSF = 0.1/BMDLio-hed). This slope, a 95% upper confidence limit on the true slope, represents a
plausible upper bound on the true slope or risk per unit dose. The PODs estimated for each mouse
tumor site are summarized in Table 2-7. Using linear extrapolation from the BMDLio-hed, human
equivalent OSFs were derived for each tumor site individually and both sites combined and are
listed in Table 2-7.
2-32
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 2-7. Model predictions and oral slope factors (OSFs) for hepatocellular
and alveolar/bronchiolar adenomas or carcinomas in female B6C3Fi mice
administered hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) in the diet for
2 years (Lish et al.. 1984)
Tumor type
Selected
model3
BMR
BMD
(mg/kg-d)
BMDL
(mg/kg-d)
BMDl0-HEDb
(mg/kg-d)
POD =
BMDLio-hedc
(mg/kg-d)
OSFd
(mg/kg-d)"1
Hepatocellular
adenomas or
carcinomas'2
Multistage 1°
10% ER
25.5
14.2
3.81
2.12
0.047
Alveolar/
bronchiolar
adenomas or
carcinomas
Multistage 1°
10% ER
29.9
14.9
4.47
2.23
0.045
Liver + lung
tumors
Multistage 1°
(MS-COMBO)
10% ER
13.8f
8.53f
2.06
1.28
0.078
aThe highest dose was dropped prior to analysis (see Section 2.3.1).
"BMDio-hed = BMDio x (BWa^/BWh1/4), where BWa = 0.036 kg, and BWh = 70 kg.
cBMDLio-hed = BMDLio x (BWa1/4/BWh1/4), where BWa = 0.036 kg, and BWh = 70 kg.
dOSF = BMR/BMDLio-hed, where BMR = 0.1 (10% ER).
incidences of female mouse liver tumors from Lish et al. (1984) are those reported in the PWG reevaluation (Parker et
al.. 2006).
fData for hepatocellular adenomas and carcinomas and for liver and lung tumors combined were remodeled using the
original sample sizes provided in Lish et al. (1984), which were slightly different for two groups than those reported in
Parker et al. (2006). The resulting BMDs and BMDLs from the remodeling were 25.7 and 14.3 mg/kg-day, respectively,
for hepatocellular adenomas and carcinomas and 13.8 and 8.56 mg/kg-day, respectively, for liver and lung tumors
combined. See Appendix D, Table D-16 and the subsequent MS-COMBO results for details.
An OSF was derived from the BMDLio-hed based on a significantly increased trend in the
incidence of hepatocellular and alveolar/bronchiolar adenomas or carcinomas in female B6C3Fi
mice (i.e., the Liver + Lung BMDLio-hed from MS-COMBO). The OSF of 0.08 (mg/kg-day)-1 is
calculated by dividing the BMR (10% ER) by the Liver + Lung BMDLio-hed and represents an upper
bound on cancer risk per unit dose associated with a continuous lifetime exposure:
OSF = 0.10 4- (Liver + Lung) BMDLio-hed = 0.10 4 1.28 mg/kg-day (2-3)
= 7.8 x 10"2 (mg/kg-day)-1
= 8 x 10"2 (mg/kg-day)-1 (rounded to one significant figure)
The slope of the linear extrapolation from the central estimate of exposure associated with
10% extra cancer risk (BMDio-hed) from the same data sets is given by:
2-33
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Slope of the linear extrapolation from the central estimate (2-4)
= 0.10 -7- (Liver + Lung) BMDw-hed = 0.10 -f- 2.06 mg/kg-day
= 4.9 x 10"2 (mg/kg-day)-1
= 5 x 10"2 (mg/kg-day)-1 (rounded to one significant figure)
The OSF for RDX should not be used with exposures exceeding the POD (1.28 mg/kg-day),
because above this level, the fitted dose-response model better characterizes what is known about
the carcinogenicity of RDX.
2.3.4. Uncertainties in the Derivation of the Oral Slope Factor
Several uncertainties underlie the cancer unit risk for RDX. Table 2-8 summarizes the
impact on the assessment of issues such as the use of models and extrapolation approaches,
particularly those underlying the Guidelines for Carcinogen Risk Assessment (U.S. EPA. 2005a). the
effect of reasonable alternatives, the approach selected, and its justification.
2-34
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 2-8. Summary of uncertainty in the derivation of the cancer risk value
for hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX)
Consideration and impact on
cancer risk value
Decision
Justification
Selection of study
The cancer bioassay in the rat
(Levine et al., 1983) would
provide a lower estimate of the
OSF
Lish et al. (1984) as
principal oral study to
derive the human
cancer risk estimate
Lish et al. (1984) was a well-conducted study;
five dose levels (including control) used, with a
sufficient number of animals per dose group (at
terminal sacrifice, n = 62-65 female mice/dose group
for all groups besides the highest dose group). Tumor
data from the mouse provided a stronger basis for
estimating the OSF than rat data. Confidence in the
OSF based on rat data was low because of the small
numbers of tumors.
Species/sex
Use of data sets from the male
mouse or male rat would provide
a lower OSF
OSF based on tumors
in female B6C3Fi
mouse
It is assumed that a positive tumor response in animal
cancer studies indicates that the agent can have
carcinogenic potential in humans in the absence of
data indicating that animal tumors are not relevant to
humans (U.S. EPA, 2005a). As there are no data to
inform whether the response in any given
experimental animal species or sex would be most
relevant for extrapolating to humans, tumor data
from the most sensitive species and sex were
selected as the basis for the OSF. Other data sets
would provide smaller OSF values, and are not
considered any more or less relevant to humans than
data from the female mouse (i.e., 0.017 per
mg/kg-day based on hepatocellular carcinomas in
male F344 rats, and 0.027 per mg/kg-day based on
alveolar/bronchiolar carcinomas in male B6C3Fi mice;
see Appendix D, Section D.2.3).
Combined tumor types
Human risk would be
underestimated if OSF was based
on analysis using only a single
tumor type
OSF based on liver and
lung tumors in female
B6C3Fi mouse
Basing the OSF on one tumor site could potentially
underestimate the carcinogenic potential of RDX, so
an analysis that included data from the two tumor
sites was chosen to calculate the combined risk (see
Appendix D, Section D.2.1). Because there is no
known biological dependence between the liver and
lung tumors, independence between the two tumor
sites was assumed. NRC (1994) considered the
assumption of independence in incidence between
tumor types to be reasonable when no evidence
exists to the contrary.
2-35
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 2-8. Summary of uncertainty in the derivation of the cancer risk value
for hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) (continued)
Consideration and impact on
cancer risk value
Decision
Justification
Selection of dose metric
PBPK models are available for the
rat, mouse, and human, and
using an appropriate internal
metric can T* accuracy in human
extrapolation
Mouse liver and lung
tumors: administered
dose used
EPA evaluated a published PBPK model in the mouse
(Sweeney et al., 2012b); maior uncertainties
associated with limited toxicokinetic data in the
mouse and unknown differences in metabolism
across species were identified. Although EPA's
preferred approach for extrapolating results from
animal studies to humans is toxicokinetic modeling,
the uncertainties associated with use of the mouse
PBPK model for RDX were considered higher than use
of administered dose.
Cross-species scaling
Alternatives could 4^ or T* OSF
(e.g., 3.5-fold 4/ [scaling by body
weight] or T* 2-fold [scaling by
BW2/3])
BW3/4 scaling (default
approach)
There are no data to support alternatives. Because
the dose metric was not an AUC, BW3/4 scaling was
used to calculate equivalent cumulative exposures for
estimating equivalent human risks. While the true
human correspondence is unknown, this overall
approach is not expected to over- or under-estimate
human equivalent risks.
BMD model uncertainty
Alternative models could 4^ or T*
OSF
Use multistage model
to derive a BMD and
BMDL for combined
tumor incidence
No biologically based models for RDX are available,
and there is no a priori basis for selecting a model
other than the multistage. The multistage model has
biological support because it allows for the statistical
plausibility of low-dose linearity (see Appendix D,
Section D.2.4), and is the model most consistently
used in EPA cancer assessments (Gehlhaus et al.,
2011). A sensitivity analysis using multistage and
nonmultistage models, with the highest dose
dropped, revealed that the multistage models were
the best-fitting or near best-fitting models for both
liver and lung tumors.
Low-dose extrapolation approach
4/ cancer risk would be expected
with the application of nonlinear
extrapolation
Linear extrapolation
from the POD
Where the available information is insufficient to
establish the MOA for tumors at a given site, linear
extrapolation is recommended because this
extrapolation approach is generally considered to be
health protective (U.S. EPA, 2005a). Because the
MOA for RDX-induced liver and lung tumors has not
been established, linear low-dose extrapolation was
applied, consistent with EPA guidance.
Statistical uncertainty at the POD
4/ OSF by 1.6-fold if BMD used as
the POD rather than the BMDL
BMDL (default
approach for
calculating plausible
upper bound OSF)
Lower bound is 95% CI on administered exposure at
10% ER of liver and lung tumors.
2-36
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Table 2-8. Summary of uncertainty in the derivation of the cancer risk value
for hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) (continued)
Consideration and impact on
cancer risk value
Decision
Justification
Sensitive subpopulations
T* OSF to an unknown extent
Considered
qualitatively
There is little information on whether some
subpopulations may have different sensitivities to the
potential carcinogenicity of RDX (i.e., because of
variability in toxicokinetics or toxicodynamics for
RDX). The mode of carcinogenic action for liver and
lung tumors in experimental animals is unknown, and
little information is available on RDX metabolites or
variation in metabolic rates that could be used to
evaluate human variability in cancer response to RDX.
Historical control
OSF changes no more than
twofold if mean historical control
tumor rates (from NTP) used
rather than concurrent control
rates
Concurrent control
rate used in BMD
modeling and to drive
OSF
The concurrent control liver tumor rate (1.5%) was at
the low end of the range (0-20%) for historical
controls from NTP studies (Haseman et al., 1985).
Concurrent control is generally preferred to historical
control in BMD modeling, especially where historical
control data come from a different laboratory. See
Appendix D, Section D.2.4, and Table D-29.
BW2/3 = body-weight scaling to the 2/3 power.
2.3.5. Previous Integrated Risk Information System (IRIS) Assessment
The previous cancer assessment for RDX was posted to the IRIS database in 1990. The OSF
in the previous cancer assessment was based on the bioassay by Lish etal. (19841 and analysis of
data for hepatocellular adenomas or carcinomas in female mice. An OSF of 1.1 x 10"1 (mg/kg-day)-1
was derived using a linearized multistage procedure (extra risk) and scaling by body weight to the
2/3 power for cross-species extrapolation. In addition, the previous assessment dropped the
high-dose group because the dose was reduced at Week 11 to address high mortality.
2.4. INHALATION UNIT RISK FOR CANCER
The carcinogenicity assessment provides information on the carcinogenic hazard potential
of the substance in question, and quantitative estimates of risk from oral and inhalation exposure
may be derived. Quantitative risk estimates may be derived from the application of a low-dose
extrapolation procedure. If derived, the inhalation unit risk (IUR) is a plausible upper bound on the
estimate of risk per |J.g/m3 air breathed.
An IUR value was not calculated because inhalation carcinogenicity data for RDX are not
available. While inhalation absorption of RDX particulates is a plausible route of exposure, there
are no toxicokinetic studies of RDX inhalation absorption to support an inhalation model.
2-37
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Therefore, a PBPK model for inhaled RDX was not developed to support route-to-route
extrapolation of an IUR from the OSF.
2.5. APPLICATION OF AGE-DEPENDENT ADJUSTMENT FACTORS
As discussed in the Supplemental Guidance for Assessing Susceptibility from Early-Life
Exposure to Carcinogens fU.S. EPA. 2005b). either default or chemical-specific age-dependent
adjustment factors (ADAFs) are recommended to account for early life exposure to carcinogens that
act through a mutagenic MOA. Because no chemical-specific data on life stage susceptibility for
RDX carcinogenicity are available, and because the MOA for RDX carcinogenicity is not known (see
Section 1.2.7), application of ADAFs is not recommended.
2-38
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
REFERENCES
Aguado. F: Carmona. MA: Pozas. E: Aguilo. A: Martinez-Guijarro. FT: Alcantara. S: Borrell. V: Yuste. R:
Ibanez. CF: Soriano. E. (2003). BDNF regulates spontaneous correlated activity at early
developmental stages by increasing synaptogenesis and expression of the K+/C1- co-
transporter KCC2. Development 130: 1267-1280. http://dx.d0i.0rg/l0.1242/dev.00351.
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.
ATSDR (Agency for Toxic Substances and Disease Registry). (1992). Health assessment for
Nebraska Army Ordnance Plant (former), Mead, Saunders County, Nebraska, Region 7.
CERCLIS No. NE6211890011. (PB92177815). Atlanta, GA: U.S. Department of Health and
Human Services, Public Health Service.
ATSDR (Agency for Toxic Substances and Disease Registry). (1993). Health Assessment for Milan
Army Ammunition Plant, Milan, Carroll and Gibson Counties, Tennessee, Region 4. CERCLIS
No. TN0210020582. Atlanta, GA: U.S. Department of Health and Human Services, Public
Health Service.
ATSDR (Agency for Toxic Substances and Disease Registry). (1996). Symptom and disease
prevalence with biomarkers health study, Cornhusker Army Ammunition Plant Hall County,
Nebraska. (ATSDRHS9672). Atlanta, GA: U.S. Department of Health and Human Services,
Public Health Service.
https://ntrl.ntis.gov/NTRL/dashboard/searchResults.xhtml?search0uery=PB96187760.
ATSDR (Agency for Toxic Substances and Disease Registry). (1999). Public health assessment for
Iowa Army Ammunition Plant, Middletown, Des Moines County, Iowa. CERCLIS No.
IA7213820445. Atlanta, GA: U.S. Department of Health and Human Services, Public Health
Service.
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.
Bailey. SA: Zidell. RH: Perry. RW. (2004). Relationships between organ weight and body/brain
weight in the rat: What is the best analytical endpoint? Toxicol Pathol 32: 448-466.
http://dx.d0i.0rg/l 0.1080/01926230490465874.
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. (2009a). 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. EI: 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.toxicology.org/pubs/docs/Tox/2006Tox.pdf.
R-l
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Bannon. PI: Tohnson. M: Williams. L: Adams. V: Perkins. E: Gust. K: Gong. P. (2009b). RDX and
miRNA expression in B6C3F1 mice [Letter], Environ Health Perspect 117: A98.
http://dx.doi.org/10.1289/ehp.080Q276.
Barsotti. M: Crotti. G. (1949). [Attacchi epileptici come manifestazione di intossicazione
professionale da trimetilen-trinitroamina (T4)]. Med Lav 40: 107-112.
Bart. 1C: ludd. LL: Hoffman. KE: Wilkins. AM: Kusterbeck. AW. (1997). Application of a portable
immunosensor to detect the explosives TNT and RDX in groundwater samples. Environ Sci
Technol 31: 1505-1511. http://dx.doi.org/10.1021/es9607771.
Beller. HR: Tiemeier. K. (2002). Use of liquid chromatography/tandem mass spectrometry to detect
distinctive indicators of in situ RDX transformation in contaminated groundwater. Environ
Sci Technol 36: 2060-2066. http: //dx.doi.org/l 0.1021 /esOl 57696.
Ben-Ari. Y. (2014). The GABA excitatory/inhibitory developmental sequence: a personal journey
[Review], Neuroscience 279: 187-219.
http://dx.doi.Org/10.1016/i.neuroscience.2014.08.001.
Bergeret. M: Khrestchatiskv. M: Tremblav. E: Bernard. A: Gregoire. A: Chanv. C. (1998). GABA
modulates cytotoxicity of immunocompetent cells expressing GABAA receptor subunits.
Biomed Pharmacother 52: 214-219. http: //dx.d0i.0rg/l 0.1016/S0753-3322f98180019-X.
Berry. AP: Arbuckle. IB: Nicol. 1. (1983). Cyclonite poisoning in a dog. Vet Rec 113: 449.
http: / /dx. doi. org /10.113 6 /vr. 113.19.449.
Best. EP: Sprecher. SL: Larson. SL: Fredrickson. HL: Bader. DF. (1999a). Environmental behavior of
explosives in groundwater from the Milan Army Ammunition Plant in aquatic and wetland
plant treatments. Removal, mass balances and fate in groundwater of TNT and RDX.
Chemosphere 38: 3383-3396. http: //dx.doi.org/10.1016/S0045-6535r98100550-5.
Best. EP: Sprecher. SL: Larson. SL: Fredrickson. HL: Bader. DF. (1999b). Environmental behavior of
explosives in groundwater from the Milan Army Ammunition Plant in aquatic and wetland
plant treatments. Uptake and fate of TNT and RDX in plants. Chemosphere 39: 2057-2072.
http://dx.doi.org/10.1016/S0045-6535r99100117-4.
Bhat. R: Axtell. R: Mitra. A: Miranda. M: Lock. C: Tsien. RW: Steinman. L. (2010). Inhibitory role for
GABA in autoimmune inflammation. Proc Natl Acad Sci USA 107: 2580-2585.
http://dx.doi.org/10.1073/pnas.0915139107.
Bhushan. B: Trott. S: Spain. TC: Halasz. A: Paquet. L: Hawari. 1. (2003). Biotransformation of
hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) by a rabbit liver cytochrome P450: Insight
into the mechanism of RDX biodegradation by Rhodococcus sp. strain DN22. Appl Environ
Microbiol 69: 1347-1351. http://dx.doi.Org/10.1128/AEM.69.3.1347-1351.2003.
Bingham. E: Cohrssen. B: Powell. CH. (2001). Patty's toxicology. In E Bingham; B Cohrssen; CH
Powell (Eds.), Patty's toxicology vol 6: Ketones, alcohols, esters, epoxy compounds, organic
peroxides (5 ed.). Hoboken, NJ: John Wiley & Sons. http: //dx.doi.org/10.1002 /Q471125474.
Bishop. RW: Kennedy. TL: Podolak. GE: Rvea. TL. Tr. (1988). A field evaluation of air sampling
methods for TNT and RDX. Am Ind Hyg Assoc J 49: 635-638.
http://dx.doi.org/10.1080/15298668891380376.
Bloomquist. TR. (1992). Intrinsic lethality of chloride-channel-directed insecticides and convulsants
in mammals. Toxicol Lett 60: 289-298. http: //dx.doi.org/10.1016/0378-4274r9219Q287-T.
R-2
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Boileau. T: Fauquignon. C: Hueber. B: Meyer. H. (2003). Explosives. In Ullmann's Encyclopedia of
Industrial Chemistry. Weinheim, Germany: Wiley-VCH Verlag Gmbh & Co. Kgaa.
http://dx.doi.Org/10.1002/14356007.al0 143.
Bosland. MC. (1992). Lesions in the male accessory sex glands and penis. In Pathobiology of the
Aging Rat, vol 1. Washington: ILSI Press.
Brar. R: Singh. I: Kaur. T: Arora. S: Singh. A. (2014). Role of GABAergic activity of sodium valproate
against ischemia-reperfusion-induced acute kidney injury in rats. Naunyn Schmiedebergs
Arch Pharmacol 387: 143-151. http://dx.doi.Org/10.1007/s00210-013-0928-2.
Bruchim. Y: Saragusty. I: Weisman. A: Sternheim. D. (2005). Cyclonite (RDX) intoxication in a police
working dog. Vet Rec 157: 354-356. http://dx.doi.org/10.1136/vr.157.12.354.
Burdette. LT: Cook. LL: Dyer. RS. (1988). Convulsant properties of cyclotrimethylenetrinitramine
(RDX): Spontaneous, audiogenic, and amygdaloid kindled seizure activity. Toxicol Appl
Pharmacol 92: 436-444. http://dx.doi.org/10.1016/0041-008XC88190183-4.
Chandra. M: Riley. MG: Tohnson. DE. (1992). Spontaneous neoplasms in aged Sprague-Dawley rats.
Arch Toxicol 66: 496-502. http://dx.doi.org/10.1007/BF01970675.
Cholakis. TM: Wong. LCK: Van Goethem. PL: Minor. I: 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.
Collier. ZA: Gust. KA: Gonzalez-Morales. B: Gong. P: Wilbanks. MS: Linkov. I: Perkins. El. (2016). A
weight of evidence assessment approach for adverse outcome pathways. Regul Toxicol
Pharmacol 75: 46-57. http://dx.doi.org/10.1016/i.vrtph.2015.12.014.
Cowin. PA: Gold. E: Aleksova. 1: O'Brvan. MK: Foster. PMD: Scott. HS: Risbridger. GP. (2010).
Vinclozolin exposure in utero induces postpubertal prostatitis and reduces sperm
production via a reversible hormone-regulated mechanism. Endocrinology 151: 783-792.
http://dx.doi.org/10.1210/en.2009-Q982.
Craig. EA: Yan. Z: Zhao. 01. (2014). The relationship between chemical-induced kidney weight
increases and kidney histopathology in rats. J Appl Toxicol 35: 729-736.
http://dx.doi.org/10.1002/iat3036.
Creasy. D: Bube. A: de Rijk. E: Kandori. H: Kuwahara. M: Masson. R: Nolte. T: Reams. R: Regan. K:
Rehm. S: Rogerson. P: Whitney. K. (2012). Proliferative and nonproliferative lesions of the
rat and mouse male reproductive system [Review], Toxicol Pathol 40: 40S-121S.
http://dx.doi.org/10.1177/0192623312454337.
Creelev. CE. (2016). From drug-induced developmental neuroapoptosis to pediatric anesthetic
neurotoxicity-where are we now? [Review], Brain Sci 6: 32.
http://dx.doi.org/10.3390/brainsci6030032.
Crouse. LCB: Michie. MW: Major. 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 /1050903.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.
R-3
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
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/10.1080/15563650601118Q44.
Deidda. G: Bozarth. IF: Cancedda. L. (2014). Modulation of GABAergic transmission in development
and neurodevelopmental disorders: investigating physiology and pathology to gain
therapeutic perspectives [Review], Front Cell Neurosci 8: 119.
http://dx.doi.org/10.3389/fncel.2014.00119.
Deng. Y: Ai. 1: Guan. X: Wang. Z: Yan. B: Zhang. D: Liu. C: Wilbanks. MS: Escalon. BL: Meyers. SA:
Yang. M0: Perkins. El. (2014). MicroRNA and messenger RNA profiling reveals new
biomarkers and mechanisms for RDX induced neurotoxicity. BMC Genomics 15 Suppl 11:
SI. http://dx.doi.org/10.1186/1471-2164-15-Sll-Sl.
Erdo. SL. (1985). Peripheral GABAergic mechanisms. Trends Pharmacol Sci 6: 205-208.
http://dx.d0i.0rg/l 0.1016/0165-6147(85190096-3.
Erdo. SL: Dobo. E: Parducz. A: Wolff. TR. (1991). Releasable GABA in tubular epithelium of rat
kidney. Experientia 47: 227-229. http://dx.doi.org/10.1007/BF01958142.
Fellows. RT: 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.d0i.0rg/l 0.1897/05-446R1.1.
Fowler. CI: Griffiths. D: de Groat. WC. (2008). The neural control of micturition [Review], Nat Rev
Neurosci 9: 453-466. http://dx.doi.org/10.1038/nrn2401.
Fritschv. TM: Pavsan. 1: Enna. A: Mohler. H. (1994). Switch in the expression of rat GABAA-receptor
subtypes during postnatal development: An immunohistochemical study. J Neurosci 14:
5302-5324.
Fuller. ME: Perreault. N: Hawari. T. (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.llll/i.1472-765X.2010.02897.x.
Funk. SB: Roberts. DT: Crawford. PL: Crawford. RL. (1993). Initial-phase optimization for
bioremediation of munition compound-contaminated soils. Appl Environ Microbiol 59:
2171-2177.
Gadagbui. B: Patterson. 1: Rak. A: Kutzman. RS: Reddv. G: Tohnson. MS. (2012). Development of a
relative source contribution factor for drinking water criteria: The case of hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX). Hum Ecol Risk Assess 18: 338-354.
http://dx.doi.Org/10.1080/10807039.2012.650588.
Galanopoulou. AS. (2008). GABA(A) receptors in normal development and seizures: friends or foes?
Curr Neuropharmacol 6: 1-20. http://dx.doi.org/10.2174/1570159Q8783769653.
Garcia-Revero. N: Habib. T: Pirooznia. M: Gust. KA: Gong. P: Warner. C: Wilbanks. M: Perkins. E.
(2011). Conserved toxic responses across divergent phylogenetic lineages: A meta-analysis
of the neurotoxic effects of RDX among multiple species using toxicogenomics.
Ecotoxicology 20: 580-594. http: //dx.doi.org/10.1007/slQ646-011-0623-3.
Gatebeck. L: Aronsson. A: Tohansson. B: Stromberg. L. (1987). Stress stimuli-induced
histopathological changes in the prostate: An experimental study in the rat. Prostate 11: 69-
76. http://dx.doi.Org/10.1002/pros.2990110109.
R-4
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Gehlhaus. MW. Ill: Gift. TS: Hogan. KA: Kopvlev. L: Schlosser. PM: Kadrv. AR. (2011). Approaches to
cancer assessment in EPA's Integrated Risk Information System [Review], Toxicol Appl
Pharmacol 254: 170-180. http://dx.doi.Org/10.1016/i.taap.2010.10.019.
Gerkin. RC: Clem. RL: Shruti. S: Kass. RE: Barth. AL. (2010). Cortical up state activity is enhanced
after seizures: A quantitative analysis. J Clin Neurophysiol 27: 425-432.
http://dx.d0i.0rg/l 0.1097/WNP.0h013e3181fdf8bd.
Gilbert. ME. (1994). The phenomenology of limbic kindling [Review], Toxicol Ind Health 10: 343-
358.
Godeiohann. M: Preiss. A: Levsen. K: Wollin. K. -M: Mtigge. C. (1998). Determination of polar organic
pollutants in aqueous samples of former ammunition sites in lower Saxony by means of
HPLC/photodiode array detection (HPLC/PDA) and proton nuclear magnetic resonance
spectroscopy (1H-NMR). Acta Hydrochim Hydrobiol 26: 330-337.
http://dx.doi.org/10.1002/rSTCni521-401Xri9981 n26:6<330::ATD-AHEH330>3.0.CQ:2-T.
Gogal. OM. Tr: Tohnson. MS: Larsen. CT: Prater. MR: Duncan. RB: Ward. PL: Lee. RB: Salice. CI:
Tortner. B: Holladav. SD. (2003). Dietary oral exposure to l,3,5-trinitro-l,3,5-triazine in the
northern bobwhite (Colinus virginianus). Environ Toxicol Chem 22: 381-387.
http://dx.doi.Org/10.1002/etc.5620220220.
Goldberg. DT: Green. ST: Nathwani. D: McMenamin. I: Hamlet. N: Kennedy. PH. (1992). RDX
intoxication causing seizures and a widespread petechial rash mimicking
meningococcaemia. J R Soc Med 85: 181.
Gong. P: Hong. H: Perkins. EI. (2015). Ionotropic GABA receptor antagonism-induced adverse
outcome pathways for potential neurotoxicity biomarkers [Review], Biomark Med 9: 1225-
1239. http://dx.doi.org/10.2217/bmm.15.58.
Goodman. LS: Gilman. AG: Hardman. TG: Limbird. LE. (1996). Goodman & Gilman's the
phamacological basis of therapeutics. In JG Hardman; LE Limbird (Eds.), Goodman &
Gilman's the Phamacological Basis of Therapeutics (9th ed. ed.). New York: McGraw-Hill.
Guo. L: Xu. H: Chen. Y: Chang. Y. (1985). Distribution and metabolism of tritium-labeled hexogen in
white mice. Chinese Journal of Industrial Hygiene and Occupational Diseases 3: 335-339.
Halasz. A: Manno. D: Perreault. NN: Sabbadin. F: Bruce. NC: Hawari. 1. (2012). Biodegradation of
RDX nitroso products MNX and TNX by cytochrome P450 XplA. Environ Sci Technol 46:
7245-7251. http://dx.doi.org/10.1021/es3011964.
Harada. T: Enomoto. A: Boorman. GA: Maronpot. RR. (1999). Liver and gallbladder. In R Maronpot;
G Boorman; B Gaul (Eds.), Pathology of the Mouse (pp. 119-183). Vienna, IL: Cache River
Press, http: //www.cacheriverpress.com/books/pathmouse.htm.
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 RDX and TNT in dogs. Final report. (A717530). Kensington,
MD: Litton Bionetics, Inc. http://www.dtic.mil/dtic/tr/fulltext/u2/a035717.pdf.
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/a040161.pdf.
Harvey. SD: Fellows. RT: Cataldo. DA: Bean. RM. (1991). Fate of the explosive hexahydro-1,3,5-
trinitro-l,3,5-triazine (RDX) in soil and bioaccumulation in bush bean hydroponic plants.
Environ Toxicol Chem 10: 845-855. http://dx.doi.org/10.1002/etc.5620100701.
R-5
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Haseman. TK: Huff. T: Boorman. GA. (1984). Use of historical control data in carcinogenicity studies
in rodents. Toxicol Pathol 12: 126-135. http://dx.doi.org/10.1177/019262338401200203.
Haseman. TK: 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/jnci/75.5.975.
Haskell Laboratories. (1942). Initial submission: Toxicity of RDX (cyclotrimethylenetrinitramine)
with cover letter dated 101592 [TSCA Submission], (88920009850). Wilmington, DE:
DuPont Chemical Company.
https://ntrl.ntis.gov/NTRL/dashboard/searchResults.xhtml?search0uery=OTS0571501.
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/jrsm.95.5.251.
Hollander. AI: Colbach. EM. (1969). Composition C-4 induced seizures: A report of five cases. Mil
Med 134: 1529-1530.
Igawa. Y: Mattiasson. A: Andersson. K. -E. (1993). Effects of GABA-receptor stimulation and
blockade on micturition in normal rats and rats with bladder outflow obstruction. J Urol
150: 537-542.
Ikonomidou. C: Bittigau. P: Ishimaru. MI: Wozniak. DF: Koch. C: Genz. K: Price. MT: Stefovska. V:
Horster. F: Tenkova. T: Dikranian. K: Olnev. TW. (2000). Ethanol-induced apoptotic
neurodegeneration and fetal alcohol syndrome. Science 287: 1056-1060.
http://dx.doi.org/10.1126/science.287.5455.1056.
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/j.taap.2012.ll.022.
Teffervs. TGR: Tiruska. P: de Curtis. M: Avoli. M. (2012). Limbic network synchronization and
temporal lobe epilepsy. In JL Noebel; HH Jasper (Eds.), Jasper's Basic Mechanisms of the
Epilepsies (4th ed ed.). New York, NY: Oxford University Press.
http://www.ncbi.nlm.nih.gov/books/NBK98158/.
Tensen. FE: Baram. TZ. (2000). Developmental seizures induced by common early-life insults: short-
and long-term effects on seizure susceptibility [Review], Ment Retard Dev Disabil Res Rev 6:
253-257. http://dx.doi.org/10.1002/1098-2779r200016:4<:253::AID-
MRDD4>:3.0.CQ:2-P.
Tohnson. MS. (2015a). Memorandum for National Center for Environmental Assessment (8101R/Dr.
Louis D'Amico), U.S. EPA - Office of Research 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
R-6
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
seizure relative to mortality events. (MCHB-IP-T). Aberdeen Proving Ground, MD:
Department of the Army, US Army Institute of Public Health.
Tohnson. MS. (2015b). Memorandum for Office of Research and Development (NCEA/Dr. Louis
DAmico), National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Two Potomac Yard, 2733 South Crystal Drive, Arlington, VA 22202. Subject: Lack of
evidence for GABAa receptor-mediated MOA of RDX-induced prostatitis. Aberdeen Proving
Ground, MD: Department of the Army, U.S. Army Institute of Public Health.
Tohnson. MS: Ouinn. MI: Bazar. MA: Gust. KA: Escalon. BL: Perkins. El. (2007). Subacute toxicity of
oral 2,6-dinitrotoluene and l,3,5-trinitro-l,3,5-triazine (RDX) exposure to the northern
bobwhite (Colinus virginianus). Environ Toxicol Chem 26: 1481-1487.
http://dx.doi.Org/10.1897/06-525.l.
Kalueff. AY. (2007). Mapping convulsants' binding to the GABA-A receptor chloride ionophore: A
proposed model for channel binding sites. Neurochem Int 50: 61-68.
http: / /dx.doi. or g/10.1016/i .neuint 2006.07.004.
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/j.jen.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.
Khalilov. I: Le Van Ouven. M: Gozlan. H: Ben-Ari. Y.. (2003). Epileptogenic actions of GABA and fast
oscillations in the developing hippocampus. Neuron 48: 787-796.
http://dx.doi.Org/10.1016/j.neuron.2005.09.026.
Kirmse. K: Hiibner. CA: Isbrandt. D: Witte. OW: Holthoff. K. (2018). GABAergic transmission during
brain development: Multiple effects at multiple stages. Neuroscientist 24: 36-53.
http://dx.doi.org/10.1177/1073858417701382.
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.
Kontani. H: Kawabata. Y: Koshiura. R. (1987). In vivo effects of gamma-aminobutyric acid on the
urinary bladder contraction accompanying micturition. Jpn J Pharmacol 45: 45-53.
http: / /dx. do i. org/10.12 5 4 /i ip. 45.45.
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.d0i.0rg/l 0.1002/iatl 455.
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.
Laurie. DT: Wisden. W: Seeburg. PH. (1992). The distribution of thirteen GABAA receptor subunit
mRNAs in the rat brain. III. Embryonic and postnatal development. J Neurosci 12: 4151-
4172.
R-7
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Levine. BS: Furedi. EM: Gordon. DE: Barklev. TT: 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.
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.d0i.0rg/l 0.1016/0378-4274(81190108-9.
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.
Lish. PM: Levine. BS: Furedi. EM: Sagartz. TM: Rac. VS. (1984). 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). FortDetrick,
Frederick, MD: U.S. Army Medical Research and Development Command.
Luian. R: Shigemoto. R: Lopez-Bendito. G. (2005). Glutamate and GABA receptor signalling in the
developing brain. Neuroscience 130: 567-580.
http://dx.doi.Org/10.1016/j.neuroscience.2004.09.042.
Ma. B: Li. H. (1993). Neurobehavioral effects ofhexogen 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/al68266.pdf.
Maita. K: Hirano. M: Harada. T: Mitsumori. K: Yoshida. A: Takahashi. K: Nakashima. N: Kitazawa. T:
Enomoto. A: Inui. K: Shirasu. Y. (1987). Spontaneous tumors in F344/DuCrj rats from 12
control groups of chronic and oncogenicity studies. J Toxicol Sci 12: 111-126.
Major. 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.
Marty. S: Wehrle. R: Sotelo. C. (2000). Neuronal activity and brain-derived neurotrophic factor
regulate the density of inhibitory synapses in organotypic slice cultures of postnatal
hippocampus. J Neurosci 20: 8087-8095.
Maverhofer. A: Hohne-Zell. B: Gamel-Didelon. K: Tung. H: Redecker. P: Grube. D: Urbanski. HF:
Gasnier. B: Fritschv. T. -M: Gratzl. M. (2001). Gamma-aminobutyric acid (GABA): A para-
R-8
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
and/or autocrine hormone in the pituitary. FASEB J 15: 1089-1091.
http://dx.doi.org/10.1096/fi.00-Q546fie.
McConnell. EE: Solleveld. HA: Swenberg. TA: Boorman. GA. (1986). Guidelines for combining
neoplasms for evaluation of rodent carcinogenesis studies. J Natl Cancer Inst 76: 283-289.
http://dx.doi.Org/10.1093/jnci/76.2.283.
McCormick. NG: Cornell. TH: Kaplan. AM. (1981). Biodegradation of hexahydro-1,3,5-trinitro-1,3,5-
triazine. Appl Environ Microbiol 42: 817-823.
McDonough. TH. Tr: Shih. T. -M. (1997). Neuropharmacological mechanisms of nerve agent-induced
seizure and neuropathology [Review], Neurosci Biobehav Rev 21: 559-579.
http://dx.doi.Org/10.1016/S0149-7634f96100050-4.
McFarland. CA: Ouinn. Ml. Tr: Bazar. MA: Talent. LG: Tohnson. MS. (2009). Toxic effects of oral
hexahydro-l,3,5-trinitro-l,3,5-triazine in the Western fence lizard (Sceloporus
occidentalis). Environ Toxicol Chem 28: 1043-1050. http://dx.doi.Org/10.1897/08-419.l.
McLanahan. ED: El-Masri. HA: Sweeney. LM: Kopvlev. LY: Clewell. HT: Wambaugh. TF: Schlosser. PM.
(2012). Physiologically based pharmacokinetic model use in risk assessment-Why being
published is not enough. Toxicol Sci 126: 5-15. http: //dx.doi.org/10.1093 /toxsci/kfr295.
McNamara. BP: Averill. HP: Owens. El: Callahan. IF: Fairchild. DG: Ciuchta. HP: Rengstorff. RH:
Biskup. RK. (1974). The toxicology of cyclotrimethylenetrinitramine (RDX) and
cyclotetramethylenetetranitramine (HMX) solutions in dimethylsulfoxide (DMSO),
cyclohexanone, and acetone (pp. 112). (AD780010. CBRNIAC-CB-148341. EB-TR-73040.
CPIAC-1975-0806). Aberdeen Proving Ground, MD: Edgewood Arsenal, Dept of the Army.
http://www.dtic.mil/dtic/tr/fulltext/u2/780010.pdf.
Meldrum. BS: Rogawski. MA. (2007). Molecular targets for antiepileptic drug development
[Review], Neurotherapeutics 4: 18-61. http://dx.doi.Org/10.1016/j.nurt.2006.ll.010.
Merrill. SL. (1968). Ingestion of an explosive material, composition C-4: A report of two cases.
USARVMed Bull 40: 5-11.
Meyer. SA: Marchand. AT: Hight. TL: 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-1,3,5-triazine (RDX) [Abstract], J Appl Toxicol 25: 427-
434. http: //dx.doi.Org/10.1002/jat.1090.
Mohler. H. (2006). GABA(A) receptor diversity and pharmacology [Review], Cell Tissue Res 326:
505-516. http://dx.doi.Org/10.1007/s00441-006-0284-3.
Moshe. SL. (1987). Epileptogenesis and the immature brain. Epilepsia 28: S3-S14.
Moshe. SL. (2000). Seizures early in life [Review], Neurology 55: S15-20; discussion S54-18.
MTA (Midwest Teratology Association). (1992). Historical control project 1988-1992: Skeletal
findings on Sprague-Dawley CD rats and New Zealand white rabbits. Natural delivery and
postnatal data: Sprague-Dawley CD rats. Ann Arbor, MI: Parke-Davis Pharmaceutical
Research.
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/GetTRDoc?AD=ADA526472.
R-9
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Napoleone. P: Bronzetti. E: Cavallotti. C: Amenta. F. (1990). Predominant epithelial localization of
type A gamma-aminobutyric acid receptor sites within rat seminal vesicles and prostate
glands. Pharmacology 41: 49-56. http: //dx.doi.org/10.1159/00Q138698.
NRC (National Research Council). (1994). Science and judgment in risk assessment In Science and
Judgment in Risk Assessment (pp. 672). Washington, DC: National Academy Press.
http://dx.doi.org/10.17226/2125.
NRC (National Research Council). (2011). Review of the Environmental Protection Agency's draft
IRIS assessment of formaldehyde (pp. 1-194). Washington, DC: The National Academies
Press, http:/ /dx.doi.org/10.17226/13142.
NTP (National Toxicology Program). (2009). NTP historical controls report: All routes and vehicles
: Rats [NTP], http://ntp.niehs.nih.gov/ntp/historical controls/ntp2000 2009/2009-05-14-
hist-ratsallroutes.pdf.
Ong. I: Kerr. PI. (1990). GAB A-receptors in peripheral tissues [Review], Life Sci 46: 1489-1501.
http://dx.doi.org/10.1016/0024-3205r90190421-M.
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/S0014-827Xf03100069-7.
Pan. X: Zhang. B: Smith. IN: San Francisco. M: Anderson. TA: Cobb. GP. (2007). 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.
Parker. G. (2001). Attachment 1: Pathology Working Group. Chairperson's report. Reevaluation:
Twenty-four month chronic toxicity/carcinogenicity study of hexahydro-1,3,5-trinitro-
1,3.5-triazine (RDX) in the B6C3F1 hybrid mouse. Research Triangle Park, NC: National
Institute of Environmental Health Sciences.
Parker. GA: Grabau. 1. (1987). Chronic prostatitis, rat In Genital System. Berlin, Germany: Springer
Berlin Heidelberg, http: //dx.doi.org/10.1007/978-3-642-72550-0 44.
Parker. GA: Reddv. G: Major. 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.
Ouinn. Ml: Hanna. TL: Shiflett. AA: McFarland. CA: Cook. ME: Tohnson. MS: Gust. KA: Perkins. El.
(2013). Interspecific effects of 4A-DNT (4-amino-2,6-dinitrotoluene) and RDX (1,3,5-
trinitro-l,3,5-triazine) in Japanese quail, Northern bobwhite, and Zebra finch. Ecotoxicology
22: 231-239. http://dx.doi.org/10.1007/sl0646-012-1019-8.
Racine. RT. (1972). Modification of seizure activity by electrical stimulation. II. Motor seizure.
Electroencephalogr Clin Neurophysiol 32: 281-294. http:/ /dx.doi.org/10.1016/0013-
4694172190177-0.
Reddv. G: Allen. NA: Major. 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.
Reifenrath. WG: Kammen. HO: Reddv. G: Maior. MA: Leach. GT. (2008). Interaction of hydration,
aging, and carbon content of soil on the evaporation and skin bioavailability of munition
R-10
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
contaminants. J Toxicol Environ Health A 71: 486-494.
http://dx.doi.Org/10.1080/15287390801906956.
Reves-Garcia. MG: Hernandez-Hernandez. F: Hernandez-Tellez. B: Garcia-Tamavo. F. (2007).
GABA(A) receptor subunits RNA expression in mice peritoneal macrophages modulate their
IL-6/IL-12 production. J Neuroimmunol 188: 64-68.
http://dx.doi.Org/10.1016/j.jneuroim.2007.05.013.
Rivera. C: Voipio. 1: Payne. TA: Ruusuvuori. E: Lahtinen. H: Lamsa. K: Pirvola. U: Saarma. M: Kaila. K.
(1999). The K+/C1- co-transporter KCC2 renders GABA hyperpolarizing during neuronal
maturation. Nature 397: 251-255. http://dx.doi.org/10.1038/16697.
Robinette. CL. (1988). Sex-hormone-induced inflammation and fibromuscular proliferation in the
rat lateral prostate. Prostate 12: 271-286. http: //dx.doi.org/10.1002 /pros.2990120310.
Ronnie. 1. ,M. (2003). Neonatal seizures and their treatment. Curr Opin Neurol 16: 177-181.
SAB (Science Advisory Board). (2017). Letter to E. Scott Pruitt re: Review of EPAs draft assessment
entitled toxicological review of hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) (September
2016). Available online at (accessed
Salari. AA: Amani. M. (2017). Neonatal blockade of GABA-A receptors alters behavioral and
physiological phenotypes in adult mice. Int J Dev Neurosci 57: 62-71.
http://dx.doi.Org/10.1016/j.ijdevneu.2017.01.007.
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/10.1016/0041-008Xf77190144-2.
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 -008Xf78190147-3.
Sellers. RS: Mortan. D: Michael. B: Roome. N: Tohnson. IK: Yano. BL: Perry. R: Schafer. K. (2007).
Society of toxicologic pathology position paper: Organ weight recommendations for
toxicology studies [Review], Toxicol Pathol 35: 751-755.
http://dx.doi.Org/10.1080/01926230701595300.
Sikka. HC: Bannerjee. S: Pack. El: Appleton. HT. (1980). Environmental fate of RDX and TNT (pp. 1-
142). (TR 81-538). Fort Detrick, MD: U.S. Army Medical Research and Materiel Command.
Simini. M: Checkai. RT. (1996). Uptake of RDX and TNT in crop plants irrigated with contaminated
water [Abstract], Phytopathology 86: S86.
Smith. IN: Pan. X: Gentles. A: Smith. EE: Cox. SB: Cobb. GP. (2006). Reproductive effects of
hexahydro-1,3,5-trinitroso-1,3,5-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 351.
Soto-Otero. R: Mendez-Alvarez. E: Sierra-Paredes. G: Galan-Valiente. 1: Aguilar-Veiga. E: Sierra-
Marcuno. G.. (1989). Simultaneous determination of the two components of picrotoxin in
R-ll
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
serum by reversed-phase high-performance liquid chromatography with application to a
pharmacokinetic study in rats. J Pharm Biomed Anal 7: 369-375.
http://dx.doi.org/10.1016/0731-7085r89180104-9.
Spanggord. Rl: Mill. T: Chou. T. -W: Mabev. WR: Smith. 1H: Lee. S. (1980). Environmental fate studies
on certain munition wastewater constituents. Final report, phase II Laboratory studies.
Menlo Park, CA: SRI International, http://www.dtic.mil/cgi-
bin/GetTRDoc?AD=ADA099256&Location=U2&doc=GetTRDoc.pdf.
Steuckart. C: Berger-Preiss. E: Levsen. K. (1994). Determination of explosives and their
biodegradation products in contaminated soil and water from former ammunition plants by
automated multiple development high-performance thin-layer chromatography. Anal Chem
66: 2570-2577. http://dx.doi.org/10.1021/ac00087a023.
Stoker. TE: Robinette. CL: Britt. BH: Laws. SC: Cooper. RL. (1999a). Prepubertal exposure to
compounds that increase prolactin secretion in the male rat: Effects on the adult prostate.
Biol Reprod 61: 1636-1643. http://dx.doi.Org/10.1095/biolreprod61.6.1636.
Stoker. TE: Robinette. CL: Cooper. RL. (1999b). Perinatal exposure to estrogenic compounds and
the subsequent effects on the prostate of the adult rat: Evaluation of inflammation in the
ventral and lateral lobes. Reprod Toxicol 13: 463-472. http: //dx.doi.org/10.1016/S0890-
6238f99100049-0.
Stone. E: Coote. TH: Allard. I: Lovick. TA. (2011). GABAergic control of micturition within the
periaqueductal grey matter of the male rat J Physiol 589: 2065-2078.
http://dx.doi.org/10.1113/iphysiol.2010.202614.
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.
Sunol. C: Tusell. TM: Gelpi. E: Rodriguez-Farre. E. (1989). GABAergic modulation of lindane (gamma-
hexachlorocyclohexane)-induced seizures. Toxicol Appl Pharmacol 100: 1-8.
http://dx.doi.org/10.1016/0041-008Xr89190086-0.
Suwa. T: Nvska. A: Peckham. TC: Hailev. TR: Mahler. IF: Haseman. IK: Maronpot. RR. (2001). A
retrospective analysis of background lesions and tissue accountability for male accessory
sex organs in Fischer-344 rats. Toxicol Pathol 29: 467-478.
http: //dx.doi.org/10.1080/01926230152500086.
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
based pharmacokinetic (PBPK) modeling [Review], Regul Toxicol Pharmacol 62: 107-114.
http://dx.doi.Org/10.1016/j.yrtph.2011.12.007.
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.Org/10.1016/i.yrtph.2012.07.005.
Tangbanluekal. L: Robinette. CL. (1993). Prolactin mediates estradiol-induced inflammation in the
lateral prostate of Wistar rats. Endocrinology 132: 2407-2416.
http://dx.doi.Org/10.1210/endo.132.6.8504745.
Testud. F: Descotes. 1: Le Meur. B. (2006). Acute occupational poisoning by octogen: First case
report. Clin Toxicol 44: 189-190.
R-12
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Testud. F: Glanclaude. T. -M: Descotes. T. (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.
Tian. 1: Chau. C: Hales. TG: Kaufman. PL. (1999). GABA(A) receptors mediate inhibition of T cell
responses. J Neuroimmunol 96: 21-28. http: //dx.doi.org/10.1016/S0165-5728(98)00264-
1.
Tian. 1: Lu. Y: Zhang. H: Chau. CH: Dang. HN: Kaufman. PL. (2004). Gamma-aminobutyric acid
inhibits T cell autoimmunity and the development of inflammatoiy responses in a mouse
type 1 diabetes model. J Immunol 173: 5298-5304.
http://dx.doi.Org/10.4049/iimmunol.173.8.5298.
U.S. EPA (U.S. Environmental Protection Agency). (1988). Recommendations for and documentation
of biological values for use in risk assessment (pp. 1-395). (EPA/600/6-87/008). Cincinnati,
OH: U.S. Environmental Protection Agency, Office of Research and Development, Office of
Health and Environmental Assessment.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=34855.
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). (1991). Guidelines for developmental toxicity risk
assessment (pp. 1-71). (EPA/600/FR-91/001). Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=23162.
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). (1994). Methods for derivation of inhalation
reference concentrations and application of inhalation dosimetry [EPA Report],
(EPA/600/8-90/066F). Research Triangle Park, NC: U.S. Environmental Protection Agency,
Office of Research and Development, Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office.
U.S. EPA (U.S. Environmental Protection Agency). (1996). Guidelines for reproductive toxicity risk
assessment. Fed Reg 61: 56274-56322.
U.S. EPA (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk
assessment [EPA Report] (pp. 1-89). (EPA/630/R-95/001F). Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment Forum.
http://www.epa.gov/risk/guidelines-neurotoxicity-risk-assessment.
U.S. EPA (U.S. Environmental Protection Agency). (2000). Science policy council handbook: Risk
characterization [EPA Report] (pp. 1-189). (EPA/100/B-00/002). Washington, D.C.: U.S.
Environmental Protection Agency, Science Policy Council, https://www.epa.gov/risk/risk-
characterization-handbook.
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:
R-13
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
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). (2005a). 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- car cinogen-risk- assessment.
U.S. EPA (U.S. Environmental Protection Agency). (2005b). Supplemental guidance for assessing
susceptibility from early-life exposure to carcinogens [EPA Report], (EPA/630/R-03/003F).
Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
https://www3.epa.gov/airtoxics/childrens supplement final.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2006). Approaches for the application of
physiologically based pharmacokinetic (PBPK) models and supporting data in risk
assessment (Final Report) [EPA Report] (pp. 1-123). (EPA/600/R-05/043F). Washington,
DC: U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=15 7668.
U.S. EPA (U.S. Environmental Protection Agency). (2011). Recommended use of body weight 3/4 as
the default method in derivation of the oral reference dose (pp. 1-50).
(EPA/100/R11/0001). Washington, DC: U.S. Environmental Protection Agency, Risk
Assessment Forum, Office of the Science Advisor.
https://www.epa.gov/risk/recommended-use-body-weight-34-default- method-derivation-
oral- reference-dose.
U.S. EPA (U.S. Environmental Protection Agency). (2012a). 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). (2012b). EPA Federal facilities forum issue paper:
Site characterization for munitions constituents [EPA Report] (pp. 1-170). (EPA-505-S-11-
001). Washington, D.C. http://www2.epa.gov/fedfac/epa-federal-facilities-forum-issue-
paper-site-characterization-munitions-constituents.
U.S. EPA (U.S. Environmental Protection Agency). (2012c). Peer review summary report. External
letter peer review of study reports on l,3,5-trinitro-l,3,5-hexahydrotriazine (RDX).
Washington, DC.
U.S. EPA (U.S. Environmental Protection Agency). (2014). Guidance for applying quantitative data to
develop data-derived extrapolation factors for interspecies and intraspecies extrapolation
[EPA Report], (EPA/100/R-14/002F). Washington, DC: Risk Assessment Forum, Office of
the Science Advisor, https: //www.epa.gov/sites/production/files/2015-
01/documents/ddef-final.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2017). External peer review of study report on
l,3,5-trinitro-l,3,5-hexahydrotriazine (RDX) (CAS no. 121-82-4). (Task Order 107 -
Contract EP-C-12-029). Washington, DC.
Vale. C: Fonfria. E: Bujons. 1: Messeguer. A: Rodriguez-Farre. E: Sunol. C. (2003). The organochlorine
pesticides gamma-hexachlorocyclohexane (lindane), alpha-endosulfan and dieldrin
differentially interact with GABA(A) and and glycine-gated chloride channels in primary
cultures cells of cerebellar granule. Neuroscience 117: 397-403.
http://dx.doi.org/10.1016/S0306-4522r02100875-8.
R-14
-------
Toxicological Review ofHexahydro-l,3,5-trinitro-l,3,5-triazine
Volpe. T. .T. (2001). Neonatal seizures. In Neurology of the Newborn. Philadelphia: WB Saunders.
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.d0i.0rg/l 0.1016/S0145-2126f97100042-8.
White. HS: Franklin. MR: Kupferberg. HI: Schmutz. M: Stables. TP: Wolf. HH. (2008). The
anticonvulsant profile of rufinamide (CGP 33101) in rodent seizure models. Epilepsia 49:
1213-1220. http://dx.doi.Org/10.llll/i.1528-1167.2008.01552.x.
Wiemer. EAC. (2007). The role of microRNAs in cancer: No small matter [Review], Eur J Cancer 43:
1529-1544. http://dx.doi.0rg/lO.lOl6/i.eica.2OO7.O4.OO2.
Williams. LR: Aroniadou-Anderjaska. V: Oashu. F: Finne. H: Pidoplichko. V: Bannon. PI: Braga. MFM.
(2011). RDX binds to the GABA(A) receptor-convulsantsite and blocks GABA(A) receptor-
mediated currents in the amygdala: A mechanism for RDX-induced seizures. Environ Health
Perspect 119: 357-363. http://dx.doi.org/10.1289/ehp.1002588.
Williams. LR: Bannon. PI. (2009). Mechanism of RDX-induced seizures in rats (Toxicology Study No.
87-XE-0BT9-09). (ADA511072. USACHPPM-TSN-87-XE-0BT9-09). Aberdeen Proving
Ground, MD: U.S. Army Center for Health Promotion and Preventive Medicine, Health Effects
Research Program.
http: / /www, dtic. mil /do cs /citations /AD A511072.
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/10.3109/15563658608992595.
Yalkowskv. SH: He. Y. (2003). Entry #172 for cyclonite (RDX). In Handbook of Aqueous Solubility
Data (1st ed.). Boca Raton, FL: CRC Press, http: //www.worldcat.org/title/handbook-of-
aqueous-solubility-data/.
Yoshimura. N: de Groat. WC. (1997). Neural control of the lower urinary tract [Review], Int J Urol 4:
111-125. http://dx.doi.Org/10.llll/i.1442-2042.1997.tb00156.x.
Zemkova. HW: Bjelobaba. I: Tomic. M: Zemkova. H: Stojilkovic. SS. (2008). Molecular,
pharmacological and functional properties of GABA(A) receptors in anterior pituitary cells. J
Physiol 586: 3097-3111. http://dx.doi.org/10.1113/iphysiol.2008.153148.
Zhang. B: Pan. X. (2009a). RDX and miRNA expression: Zhang and Pan respond. Environ Health
Perspect 117: A98-A99. http: //dx.doi.org/10.1289/ehp.080Q276R.
Zhang. B: Pan. X. (2009b). RDX induces aberrant expression of microRNAs in mouse brain and liver.
Environ Health Perspect 117: 231-240. http://dx.doi.org/10.1289/ehp.11841.
Zhang. B: Pan. X: Cobb. GP: Anderson. TA. (2007). MicroRNAs as oncogenes and tumor suppressors
[Review], Dev Biol 302: 1-12. http://dx.doi.Org/10.1016/i.ydbio.2006.08.028.
R-15
------- |