UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON D.C. 20460
OFFICE OF THE ADMINISTRATOR
SCIENCE ADVISORY BOARD
September 29, 2015
EPA-SAB-15-013
The Honorable Gina McCarthy
Administrator
U.S. Environmental Protection Agency
1200 Pennsylvania Avenue, N.W.
Washington, D.C. 20460
Subject: Review of the IRIS Draft Toxicol ogical Review of Trimethylbenzenes
Dear Administrator McCarthy:
The U.S. Environmental Protection Agency's National Center for Environmental Assessment requested
a peer review of the scientific basis for the draft Toxicological Review of Trimethylbenzenes developed
for the agency's Integrated Risk Information System (IRIS). The assessment is based on a
comprehensive review of the available scientific literature on the noncancer and cancer health effects in
humans and experimental animals exposed to three trimethylbenzene isomers: 1,2,3-TMB, 1,2,4-TMB,
or 1,3,5-TMB.
In April 2011, the National Research Council released its Review of the Environmental Protection
Agency's Draft IRIS Assessment of Formaldehyde and included comments and recommendations for
improving the development of IRIS assessments in general. The Toxicological Review of
Trimethylbenzenes is one of the first IRIS assessments to address the NRC recommendations for
improving the development of IRIS assessments. The SAB was asked to review the scientific and
technical analyses used to develop reference concentrations and reference doses for the three
trimethylbenzene isomers and to comment on the agency's enhancements to the IRIS Program in
response to the NRC recommendations. The SAB Chemical Assessment Advisory Committee was
augmented with additional toxicological experts to conduct this review.
The SAB is aware that the agency is taking a phased approach to address the NRC recommendations for
several assessments that were nearly completed. For these assessments, the agency is focusing on
streamlining the documents, increasing the transparency and clarity of the assessment, and presenting
the data and information using standard tables, editing and formatting. The SAB acknowledges the
improvement in the new format for IRIS assessments and commends the agency for its progress in
addressing the NRC recommendations. The SAB recognizes that the TMB assessment was under
development prior to the NRC recommendations and only implements the first phase of the agency's
efforts to enhance the IRIS process. The SAB looks forward to reviewing future IRIS assessments with
additional enhancements. The SAB used the Toxicological Review of Trimethylbenzenes as a case study
-------
to provide advice and comments on improving IRIS toxicological assessments. Specific comments on
developing the preamble and executive summary for future assessments, as well as the TMB assessment,
are provided in the enclosed report. The SAB also found that the tables and presentation of data and
information considered are an improvement. The SAB provides specific suggestions to improve the
presentations for hazard identification and dose-response analyses. The SAB anticipates that after
several IRIS reviews are completed, the Chemical Assessment Advisory Committee will compare the
findings and recommendations from the reviews and will provide the agency, through the Chartered
SAB, advice and comments on the agency's progress to enhance IRIS assessments.
The SAB agrees with the agency that physiologically based pharmacokinetic (PBPK) modeling is an
appropriate approach to developing reference concentrations and reference doses for trimethylbenzenes.
When implementing a PBPK modeling approach the SAB strongly recommends that the EPA provide a
transparent and detailed discussion of the rationale for selecting this approach. The discussion should
include the available studies, data, and information considered by the agency, how these data were
compared and considered, and why these analyses led the agency to use a PBPK approach rather than
chemical-specific studies. In the enclosed report the SAB conducts a review of the PBPK model and
provides specific recommendations to improve the use of modeling for trimethylbenzenes. The SAB
encourages the EPA to conduct independent peer review of PBPK model and modeling results used in
assessments when the model is a new version, was previously unpublished, or is a modification of a
published model. The SAB finds that the PBPK modeling approach and extrapolating inhalation data to
an oral exposure is appropriate for the reference concentration and reference dose for 1,2,3-TMB, 1,2,4-
TMB and 1,3,5-TMB. However, the presentation of the analysis should be expanded to better describe
the inhalation and oral toxicology studies considered and the rationale for using the PBPK model.
There are inhalation and oral toxicology studies for 1,3,5-TMB and the analyses of these studies should
be expanded to develop candidate toxicity values for other endpoints (i.e., developmental and liver
toxicity) in addition to the critical effect the EPA selected. The SAB notes that the endpoints for these
studies are not the same neurotoxicological effects used in the PBPK and extrapolation from 1,2,4-TMB.
The SAB recommends that the agency derive candidate toxicity values for 1,3,5-TMB using available
toxicology studies for 1,3,5-TMB to provide a clearer explanation of its selection and the rationale for
using PBPK modeling and extrapolating inhalation data to an oral exposure. The SAB recognizes there
may be uses of candidate toxicity values in addition to selecting an overall toxicity value and strongly
supports the agency developing candidate reference concentrations and reference doses for multiple
endpoints. The Board finds that the agency needs to further clarify how candidate toxicity values should
be developed and whether they have other potential uses.
The SAB finds that the evidence for carcinogenicity of trimethylbenzenes, although limited, was well
presented by the EPA in the draft toxicological review and the SAB agrees that the EPA could not
conduct a quantitative cancer assessment for any of the TMB isomers due to the lack of appropriate
studies.
There is a limited discussion of sensitive life stages and vulnerable populations for the TMB assessment
due to lack of data on the toxicological responses in these populations. The SAB encourages the agency
to expand the description and importance of these analyses in future assessments.
Regulatory agencies are frequently required to address risks associated with short-term exposures. The
principal studies used to derive the proposed reference concentrations and reference doses for the TMBs
are subchronic in duration and the analysis needed to generate subchronic reference concentrations and
-------
reference doses has already been done. Given the usefulness of subchronic toxicity values and the small
amount of additional work needed to add them to the TMB Assessment, the SAB recommends that the
review be expanded to include the presentation of subchronic reference concentrations and reference
doses.
The SAB appreciates the opportunity to provide the EPA with advice and looks forward to the agency's
response.
Sincerely,
/Signed/
/Signed/
Peter S. Thorne
Chair
Cynthia M. Harris
Chair
SAB Chemical Assessment Advisory Committee
Augmented for the Review of the Draft IRIS
Trimethylbenzenes Assessment
Science Advisory Board
Enclosure
-------
NOTICE
This report has been written as part of the activities of the EPA Science Advisory Board (SAB), a public
advisory group providing extramural scientific information and advice to the Administrator and other
officials of the Environmental Protection Agency. The SAB is structured to provide balanced, expert
assessment of scientific matters related to problems facing the agency. This report has not been reviewed
for approval by the agency and, hence, the contents of this report do not represent the views and policies
of the Environmental Protection Agency, nor of other agencies in the Executive Branch of the Federal
government, nor does mention of trade names of commercial products constitute a recommendation for
use. Reports of the SAB are posted on the EPA website at http://www.epa.gov/sab
1
-------
U.S. Environmental Protection Agency
Science Advisory Board
Chemical Assessment Advisory Committee Augmented for
Review of the Draft IRIS Trimethylbenzene Assessment
CHAIR
Dr. Cynthia M. Harris, Director and Professor, Institute of Public Health, Florida A&M University,
Tallahassee, FL
MEMBERS
Dr. James V. Bruckner, Professor, Department of Pharmacology & Toxicology, University of Georgia,
Athens, GA
Dr. Deborah Cory-Slechta, Professor, Department of Environmental Medicine, School of Medicine
and Dentistry, University of Rochester, Rochester, NY
Dr. Helen Goeden, Senior Toxicologist, Minnesota Department of Health, St. Paul, MN
Dr. Sean Hays, President, Summit Toxicology, Allenspark, CO
Dr. Lawrence Lash, Professor, Department of Pharmacology, Wayne State University School of
Medicine, Wayne State University, Detroit, MI
Dr. Lorenz Rhomberg, Principal, Gradient, Cambridge, MA
Dr. Stephen M. Roberts, Professor, Center for Environmental and Human Toxicology, University of
Florida, Gainesville, FL
CONSULTANTS
Dr. Mitchell Cohen, Associate Professor, Department of Environmental Medicine, School of Medicine,
New York University, Tuxedo, NY
Dr. Gary Ginsberg, Toxicologist, Environmental & Occupational Health, Connecticut Department of
Public Health, Hartford, CT
Dr. Robert A. Howd, Toxicologist, ToxServices, San Jose, CA
Dr. Kannan Krishnan1, Professor and Director, Occupational and Environmental Health, Human
Toxicology Research Group, Universite de Montreal, Montreal, WI, Canada
Dr. Frederick J. Miller, Independent Consultant, Cary, NC
1 Dr. Krishnan was unable to participate on the Chemical Assessment Advisory Committee Augmented for Review of the
Draft IRIS Trimethylbenzene Assessment after the June 2014 Meeting.
ii
-------
Dr. Emanuela Taioli, Professor and Chief of Epidemiology North Shore LIJ-Hofstra School of
Medicine, Great Neck, NY
Dr. Raymond York, President, R.G. York & Associates, Manlius, NY
FEDERAL EXPERTS
Dr. Frederick A. Beland, Director, Division of Biochemical Toxicology, National Center for
Toxicological Research, U.S. Food and Drug Administration, Jefferson, AR
SCIENCE ADVISORY BOARD STAFF
Mr. Thomas Carpenter, Designated Federal Officer, U.S. Environmental Protection Agency, Science
Advisory Board (1400R), 1200 Pennsylvania Avenue, NW, Washington, DC 20460
111
-------
U.S. Environmental Protection Agency
Science Advisory Board
BOARD
CHAIR
Dr. Peter S. Thorne, Professor and Head, Department of Occupational & Environmental Health,
University of Iowa, Iowa City, IA
MEMBERS
Dr. Joseph Arvai, Max McGraw Professor of Sustainable Enterprise and Director of the Erb Institute,
School of Natural Resources & Environment, University of Michigan, MI,
Dr. Sylvie M. Brouder, Professor and Wickersham Chair of Excellence in Agricultural Research,
Department of Agronomy, Purdue University, West Lafayette, IN
Dr. Thomas Burbacher, Professor, Department of Environmental and Occupational Health Sciences,
School of Public Health, University of Washington, Seattle, WA
Dr. Ingrid Burke, Director and Wyoming Excellence Chair, Haub School and Ruckelshaus Institute of
Environment and Natural Resources, University of Wyoming, Laramie, WY
Dr. George Daston, Victor Mills Society Research Fellow, Global Product Stewardship, The Procter &
Gamble Company, Mason, OH
Dr. Costel Denson, Managing Member, Costech Technologies, LLC, Hockessin, DE
Dr. Michael Dourson, President, Toxicology Excellence for Risk Assessment, Cincinnati, OH
Dr. Joel Ducoste, Professor, Department of Civil, Construction, and Environmental Engineering,
College of Engineering, North Carolina State University, Raleigh, NC
Dr. David A. Dzombak, Hamerschlag University Professor and Department Head, Department of Civil
and Environmental Engineering, College of Engineering, Carnegie Mellon University, Pittsburgh, PA
Dr. Elaine M. Faustman, Professor and Director, Environmental and Occupational Health Sciences,
University of Washington, Seattle, WA
Dr. R. William Field, Professor, Department of Occupational and Environmental Health, and
Department of Epidemiology, College of Public Health, University of Iowa, Iowa City, IA
Dr. H. Christopher Frey, Distinguished University Professor, Department of Civil, Construction and
Environmental Engineering, College of Engineering, North Carolina State University, Raleigh, NC
Dr. Steven Hamburg, Chief Scientist, Environmental Defense Fund, Boston, MA
iv
-------
Dr. Cynthia M. Harris, Director and Professor, Institute of Public Health, Florida A&M University,
Tallahassee, FL
Dr. Robert J. Johnston, Director of the George Perkins Marsh Institute and Professor, Economics,
Clark University, Worcester, MA
Dr. Kimberly L. Jones, Professor and Chair, Department of Civil and Environmental Engineering,
Howard University, Washington, DC
Dr. Catherine Karr, Associate Professor - Pediatrics and Environmental and Occupational Health
Sciences and Director - NW Pediatric Environmental Health Specialty Unit, University of Washington,
Seattle, WA
Dr. Madhu Khanna, ACES Distinguished Professor in Environmental Economics, Department of
Agricultural and Consumer Economics, University of Illinois at Urbana-Champaign, Urbana, IL
Dr. Nancy K. Kim, Independent Consultant, Albany, NY
Dr. Francine Laden, Mark and Catherine Winkler Associate Professor of Environmental
Epidemiology, Harvard School of Public Health, and Channing Division of Network Medicine, Brigham
and Women's Hospital and Harvard Medical School, Boston, MA
Dr. Lois Lehman-McKeeman, Distinguished Research Fellow, Discovery Toxicology, Bristol-Myers
Squibb, Princeton, NJ
Dr. Cecil Lue-Hing, President, Cecil Lue-Hing & Assoc. Inc., Burr Ridge, IL
Dr. Elizabeth Matsui, Associate Professor, Pediatrics, School of Medicine, Johns Hopkins University,
Baltimore, MD
Dr. Denise Mauzerall, Professor, Woodrow Wilson School of Public and International Affairs, and
Department of Civil and Environmental Engineering, Princeton University, Princeton, NJ
Dr. Kristina D. Mena, Associate Professor, Epidemiology, Human Genetics, and Environmental
Sciences, School of Public Health, University of Texas Health Science Center at Houston, El Paso, TX
Dr. Surabi Menon, Director of Research, ClimateWorks Foundation, San Francisco, CA
Dr. James R. Mihelcic, Professor, Civil and Environmental Engineering, University of South Florida,
Tampa, FL
Dr. H. Keith Moo-Young, Chancellor, Office of Chancellor, Washington State University, Tri-Cities,
Richland, WA
Dr. Eileen Murphy, Director of Research Development, Office of Research and Economic
Development, Rutgers University, Piscataway, NJ
v
-------
Dr. James Opaluch, Professor and Chair, Department of Environmental and Natural Resource
Economics, College of the Environment and Life Sciences, University of Rhode Island, Kingston, RI
Dr. Martin Philbert, Dean and Professor, Environmental Health Sciences, School of Public Health,
University of Michigan, Ann Arbor, MI
Mr. Richard L. Poirot, Air Quality Planning Chief, Air Quality and Climate Division, Vermont
Department of Environmental Conservation, Montpelier, VT
Dr. Stephen Polasky, Fesler-Lampert Professor of Ecological/Environmental Economics, Department
of Applied Economics, University of Minnesota, St. Paul, MN
Dr. David B. Richardson, Associate Professor, Department of Epidemiology, School of Public Health,
University of North Carolina, Chapel Hill, NC
Dr. Amanda D. Rodewald, Director of Conservation Science, Cornell Lab of Ornithology and
Associate Professor, Department of Natural Resources, Cornell University, Ithaca, NY
Dr. William Schlesinger, President Emeritus, Cary Institute of Ecosystem Studies, Millbrook, NY
Dr. Gina Solomon, Deputy Secretary for Science and Health, Office of the Secretary, California
Environmental Protection Agency, Sacramento, CA
Dr. Daniel O. Stram, Professor, Department of Preventive Medicine, Division of Biostatistics,
University of Southern California, Los Angeles, CA
Dr. Paige Tolbert, Professor and Chair, Department of Environmental Health, Rollins School of Public
Health, Emory University, Atlanta, GA
Dr. Jeanne VanBriesen, Professor, Department of Civil and Environmental Engineering, Carnegie
Mellon University, Pittsburgh, PA
Dr. John Vena, Professor and Founding Chair, Department of Public Health Sciences, Medical
University of South Carolina, Charleston, SC
Dr. Elke Weber, Jerome A. Chazen Professor of International Business, Columbia Business School,
New York, NY
Dr. Charles Werth, Professor and Bettie Margaret Smith Chair in Environmental Health Engineering,
Department of Civil, Architectural and Environmental Engineering, Cockrell School of Engineering,
University of Texas at Austin, Austin, TX
Dr. Peter J. Wilcoxen, Associate Professor, Economics and Public Administration, The Maxwell
School, Syracuse University, Syracuse, NY
vi
-------
Dr. Dawn J. Wright, Chief Scientist, Environmental Systems Research Institute (Esri), Redlands, CA
SCIENCE ADVISORY BOARD STAFF
Mr. Thomas Carpenter, Designated Federal Officer, U.S. Environmental Protection Agency, Science
Advisory Board (1400R), 1200 Pennsylvania Avenue, NW, Washington, DC 20460
vii
-------
TABLE OF CONTENTS
Acronyms and Abbreviations ix
1. EXECUTIVE SUMMARY 1
2. INTRODUCTION 4
2.1. Background 4
2.2. Charge to the Science Advisory Board 5
3. RESPONSE TO CHARGE QUESTIONS 6
3.1. Enhancements to IRIS Assessments 6
3.1.1. Preamble Enhancements 6
3.1.2. Presenting Assessment Steps and Outcomes 8
3.1.3. Standardized Evaluation of Critical Studies 12
3.1.4. Addressing Public Comments on the Draft ToxicolozicalReview of Trimethylbenzenes
(May 2012) 13
3.2. Toxicological Review of Trimethylbenzenes 14
3.2.1. Executive Summary 14
3.2.2. Literature Search Strategy/Study Selection 14
3.2.3. Hazard Identification 16
3.2.4. Toxicokinetics and Pharmacokinetic Modeling 19
3.2.5. Inhalation Reference Concentration (RfC) for 1,2,4-TMB 23
3.2.6. Inhalation Reference Concentration (RfC) for 1,2,3-TMB 27
3.2.7. Inhalation Reference Concentration (RfC) for 1,3,5-TMB 29
3.2.8. Oral Reference Dose for 1,2,4-TMB 31
3.2.9. Oral Reference Dose (RID) for 1,2,3-TMB 32
3.2.10. Oral Reference Dose (RID) for 1,3,5-TMB 33
3.2.11. Carcinogenicity of 1,2,4-TMB, 1,2,3-TMB, and 1,3,5-TMB 34
3.3. Additional Recommendations 35
3.3.1. Candidate Toxicity Values 35
3.3.2. Susceptible Populations and Lifestages 36
3.3.3. Developing Subchronic RfCs and RfDs 36
REFERENCES 38
APPENDIX A: CHARGE TO THE SAB A-l
APPENDIX B: RESULTS OF REVIEW OF TRIMETHYLBENZENE PBPK MODEL
INTERNAL METRICS B-l
viii
-------
Acronyms and Abbreviations
ADME
absorption, distribution, metabolism, and excretion
BMD
benchmark dose
C-9
aromatic hydrocarbon fraction including ethyltoluenes and trimethylbenzenes
CNS
central nervous system
CYP450
cytochrome P450
EPA
U.S. Environmental Protection Agency
GD
gestational day
HEC
human equivalent concentration
IRIS
Integrated Risk Information System
Km
Michaelis-Menten constant
LOAEL
lowest-observed-adverse-effect level
MOA
mode of action
NCEA
National Center for Environmental Assessment
NOAEL
no-ob served-adverse-effect level
NRC
National Research Council
ORD
Office of Research and Development
PBPK
physiologically based pharmacokinetic
PC
partition coefficients
POD
point of departure
ppm
parts per million
RfC
reference concentration
RfD
reference dose
SAB
Science Advisory Board
SD
standard deviation
TMB
trimethylbenzene
UF
uncertainty factor
UFa
interspecies uncertainty factor
UFh
intraspecies uncertainty factor
UFs
subchronic-to-chronic uncertainty factor
UFl
LOAEL-to-NOAEL uncertainty factor
UFd
database deficiency uncertainty factor
Vmax
maximum rate of uptake/conversion
IX
-------
1. EXECUTIVE SUMMARY
The Environmental Protection Agency's (EPA) National Center for Environmental Assessment (NCEA)
requested the Science Advisory Board to conduct a peer review of the draft Toxicological Review for
Trimethylbenzenes (August 2013), hereafter referred to as the TMB assessment, developed by the
Integrated Risk Information System (IRIS) program. This assessment reviews the publicly available
studies on the three isomers of trimethylbenzene (i.e., 1,2,3-trimethylbenzene [1,2,3-TMB], 1,2,4-
trimethylbenzene [1,2,4-TMB], and 1,3,5-trimethylbenzene [1,3,5-TMB]) and identifies the adverse
health effects to characterize inhalation and oral exposure-response relationships for each isomer. This
assessment was prepared because of the presence of TMBs at Superfund sites. Of the 38 sites on the
EPA's National Priorities List that report TMB isomer contamination, 93 percent report 1,3,5-TMB
contamination, 85 percent report 1,2,4-TMB contamination, 12 percent report 1,2,3-TMB
contamination, and 17 percent report contamination by unspecified TMB isomers.
The National Research Council (NRC), in its Review of the Environmental Protection Agency's Draft
IRIS Assessment of Formaldehyde, provided recommendations for improving the development of IRIS
assessments in general. The Draft Toxicological Review of Trimethylbenzenes is one of the first IRIS
assessments to address the NRC recommendations. For the current review, the SAB was asked to review
the scientific and technical analyses used to develop reference concentrations (RfC) and reference doses
(RfD) for the three TMB isomers and to comment on the agency's enhancements made to the IRIS
Program in response to the NRC recommendations. The SAB Chemical Assessment Advisory
Committee was augmented with additional toxicological experts to conduct the review.
Enhancements to the IRIS Program
The agency implemented a phased approach to address the NRC recommendations for several
assessments that were nearly developed. For these assessments, the agency focused on streamlining the
documents, increasing the transparency and clarity of the assessments, and better presenting the data and
information through the use of standard tables, editing and formatting. The SAB acknowledges the
improvement in the new format for IRIS assessments and commends the agency for its progress in
addressing the NRC recommendations. The SAB recognizes that the TMB assessment was under
development and implemented the first phase of the agency's efforts to enhance the IRIS process and
looks forward to reviewing future IRIS assessments with additional enhancements. The SAB used the
Toxicological Review of Trimethylbenzenes as a case study to provide advice and comments on
improving IRIS toxicological assessments by addressing the NRC recommendations. Specific comments
on developing the preamble and executive summary for future assessments, as well as the TMB
Assessment, are provided in the enclosed report. The SAB also found that the tables and presentation of
data and information considered are an improvement and provided specific suggestions to improve the
presentations for hazard identification and dose-response analyses. The SAB anticipates that after
several IRIS reviews are completed, the Chemical Assessment Advisory Committee will compare the
reviews to provide the agency, through the Chartered SAB, with advice and comments on the agency's
progress to enhance IRIS assessments.
Chronic Hazard Assessment of Trimethylbenzenes
The SAB agrees with the agency that physiologically based pharmacokinetic (PBPK) modeling is an
appropriate approach to developing RfCs and RfDs. When implementing a PBPK approach, the SAB
strongly recommends that the EPA clearly discuss the available studies, data, and information
1
-------
considered by the agency, how these were considered, and why these analyses led the agency to use a
PBPK approach rather than specific studies for the TMBs.
The SAB conducted a review of the modified Hissink PBPK model and provides specific
recommendations to improve the use of modeling for trimethylbenzenes. A review of the PBPK
modeling used to develop the TMB assessment is provided in Appendix B of this report. The SAB
report provides specific recommendations to improve the use of PBPK modeling for TMBs. Whenever
the agency uses a PBPK model that involves a new or modified PBPK model, the agency should
commission an independent peer review of the model, assumptions made in the modeling, the model's
fit to PK datasets, model predictions and the modified PBPK models application in the risk assessment.
The SAB finds that the PBPK modeling approach, which extrapolates inhalation data to an oral
exposure, is appropriate for the RfC and RfD for 1,2,3-TMB, 1,2,4-TMB and 1,3,5-TMB. However, the
SAB notes that the presentation of the analysis should be expanded to better describe the inhalation and
oral toxicology studies considered and the rationale for using the PBPK model and extrapolation
approach over the studies considered.
There are inhalation and oral toxicology studies for 1,3,5-TMB and the analyses of these studies should
be expanded to develop candidate reference toxicity values for other endpoints (i.e., developmental and
liver toxicity). The SAB notes that the endpoints for these studies are not the same neurotoxicological
effects used in the PBPK modeling and extrapolation from 1,2,4-TMB. The SAB recommends that the
agency derive candidate reference concentrations and reference doses for 1,3,5-TMB using available
toxicology studies for 1,3,5-TMB to provide a clearer explanation of the EPA's rationale for using the
PBPK modeling approach and extrapolating inhalation data to an oral exposure and compare those
results to the reference concentrations and reference doses developed for 1,3,5-TMB using the PBPK
approach extrapolating from 1,2,4-TMB.
The SAB finds that while the search strategy and rationale to select studies was clearly articulated, the
exclusion criteria and implementation of those criteria was not as transparent. The breadth of the
literature review and discussion should be expanded to include other closely related aromatic solvents
and possibly mixtures to fill gaps in the TMB database. The SAB finds that the available studies on
closely related aromatic solvents and mixtures may provide qualitative and mechanistic insights into the
toxicity of TMBs. Therefore these studies, including the C-9 fraction and white spirit studies, deserve
further discussion to transparently describe the EPA's consideration of these data.
The SAB concludes that the EPA's hazard assessment of the carcinogenicity of TMBs integrates all
available scientific evidence and agrees with the EPA that there is "inadequate information to assess the
carcinogenic potential" of TMBs. The carcinogenicity of 1,2,4-TMB has been assessed in only a single
study. The EPA found that there were a number of deficiencies concerning this bioassay and the SAB
agrees that there are not sufficient data to conduct a quantitative assessment. The SAB also notes that no
carcinogenicity bioassays have been conducted with 1,2,3-TMB or 1,3,5-TMB.
2
-------
Additional Recommendations
The SAB identified three additional topics not addressed directly in the Charge that warrant additional
consideration by the agency: (1) clarification of how the EPA considers candidate toxicity values and
their intended use, (2) an expanded discussion of sensitive life stages and vulnerable populations, and
(3) deriving the subchronic RfC and RfD for the TMB isomers.
The SAB recognizes there are uses of candidate toxicity values in addition to selecting an overall
toxicity value and strongly supports the agency developing candidate reference concentrations and
reference doses for multiple endpoints. The overall toxicity value is one that is intended to be protective
of toxicity of all types, and this is taken into consideration when selecting the uncertainty factor.
Another use of candidate RfCs/RfDs is to better understand the effects of combined or cumulative
chemical exposures and this context may change the considerations for an uncertainty factor for
databases from those used to develop the overall toxicity value. The SAB is unaware of any discussion
of this issue by EPA or clear description of how candidate RfC/RfD values are to be developed and
used. As the IRIS process moves forward, it will be important to provide much greater clarity on this
subject. The Board encourages the agency to further clarify how the candidate toxicity values should be
developed and considerations for other potential uses.
There is a limited discussion of sensitive life stages and vulnerable populations in the preamble and
main text of the assessment. The SAB recognizes that there is limited information available for TMBs.
However, the SAB encourages the agency to expand the description and importance of these analyses in
the hazard identification and dose-response for sensitive life stages and vulnerable populations in future
toxicological assessments.
The SAB notes that the agency's derivation of chronic RfCs and RfDs for the TMBs is built upon
subchronic values. The principal studies used to derive the proposed RfCs and RfDs are all subchronic
in duration, and the analysis needed to support a robust set of subchronic toxicity values has in effect
already been done. The toxic endpoints and dose-response relationships are clearly relevant for
subchronic exposure, and the same points of departure would apply to the development of a set of
subchronic RfCs and RfDs. Given the potential usefulness of these toxicity values for risk assessment,
the importance of having the values available on IRIS, and the very small amount of additional work
required to add them to the TMB Assessment, the SAB recommends that the review be expanded to
include the presentation of subchronic RfCs and RfDs for 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB.
3
-------
2. INTRODUCTION
2.1. Background
The EPA requested a peer review of the scientific basis supporting the draft Toxicological Review of
Trimethylbenzenes (1,2,3-trimethylbenzene [1,2,3-TMB], 1,2,4- trimethylbenzene [1,2,4-TMB], and
1,3,5-trimethylbenzene [1,3,5-TMB]) that will appear on the agency's online database, the Integrated
Risk Information System (IRIS). This is a new assessment; there is currently no entry in the IRIS
database for any isomer of trimethylbenzene. The SAB Chemical Assessment Advisory Committee was
augmented with additional toxicological experts to conduct this review.
IRIS is a human health effects assessment program that evaluates scientific information on effects that
may result from exposure to specific chemical substances in the environment. IRIS is prepared and
maintained by the EPA's National Center for Environmental Assessment (NCEA) within the Office of
Research and Development (ORD). Through IRIS, the EPA provides science-based human health
assessments to support the agency's regulatory activities and decisions to protect public health. IRIS
assessments contain information for chemical substances that can be used to support the first two steps
(hazard identification and dose-response assessment) of the human health risk assessment process.
When supported by available data, IRIS provides health effects information and toxicity values for
chronic health effects (including cancer and effects other than cancer). Governments and others combine
IRIS toxicity values with exposure information to characterize public health risks of chemical
substances; this information is then used to support risk management decisions designed to protect
public health.
The draft Toxicological Review of Trimethylbenzenes (August 2013), hereafter referred to as the TMB
Assessment, is based on a review of the available scientific literature on the noncancer and cancer health
effects in humans and experimental animals exposed to 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB. This
draft IRIS assessment includes the following sections:
Preamble to describe the methods used to develop IRIS assessments;
Executive Summary to concisely summarize the major conclusions of the assessment;
Literature Search Strategy and Study Selection section to describe the process for identifying and
evaluating the evidence for consideration in developing the assessment;
Hazard Identification section to systematically synthesize and integrate the available evidence of
organ/system-specific hazards; and
Dose-Response Analysis section to describe the selection of studies for consideration in
calculating toxicity values and to provide details of the analysis and methodology in deriving and
selecting toxicity values.
In addition the draft TMB Assessment includes appendices on chemical and physical properties,
toxicokinetic information, summaries of toxicity studies, and a summary of the public comments
received on the May 2012 draft. The draft assessment was developed according to guidelines for
technical reports published by the EPA and contains a qualitative characterization of the hazards for the
TMBs, including a cancer descriptor of the isomers' human carcinogenic potential, and noncancer
toxicity values, including a chronic oral RfD and a chronic inhalation RfC for all three TMB isomers. A
quantitative cancer assessment for TMBs was not conducted due to inadequate data.
4
-------
2.2. Charge to the Science Advisory Board
The draft TMB Assessment is one of the first IRIS assessments to address the NRC recommendations
for improving the development of IRIS assessments. Therefore the EPA charge for this peer review was
two-fold and requested: (1) a review of the scientific and technical analyses used to develop RfCs and
RfDs for the three TMB isomers; and (2) advice and comment on the enhancements the IRIS Program
implemented to address the NRC recommendations.
The agency asked three general questions about the agency's progress in response to the NRC
recommendations. In April 2011, the NRC released its Review of the Environmental Protection
Agency's Draft IRIS Assessment of Formaldehyde (NRC 2011). In addition to offering comments
specifically about the EPA's draft formaldehyde assessment, the NRC included comments and
recommendations for improving the development of IRIS assessments. Generally, the IRIS Program's
implementation of the NRC recommendations is following a phased approach. Phase 1 of
implementation has focused on a subset of the short-term recommendations, such as editing and
streamlining documents, increasing transparency and clarity, and using more tables, figures, and
appendices to present information and data in assessments. The Phase 1 implementation was applied to
assessments, including the draft TMB Assessment, that were nearly developed. Additional NRC
recommendations will be implemented in future IRIS assessments with input and feedback from the
public, stakeholders, and external peer review committees. This phased approach is consistent with the
NRC's Roadmap for Revision as described in Chapter 7 of the formaldehyde review report (NRC 2011).
The EPA charge to the SAB includes one question about the response to public comments on the May
2012 TMB Assessment and specific questions about the scientific and technical approaches aspect used
to develop the RfC and RfD for each of the individual isomers. Because the agency uses PBPK
modeling and extrapolation of data for 1,2,4-TMB to develop the RfC and RfD for 1,2,3-TMB and
1,3,5-TMB, the responses to questions often raised similar if not identical issues and advice for the
agency to consider. Where appropriate this report refers the reader to issues discussed in previous
sections rather than repeating the same information and advice for each isomer.
The SAB provides recommendations for three issues that were not addressed in the charge. The SAB
noted that (1) additional potential uses of the candidate toxicity values developed to select the overall
toxicity value were not well described; (2) the toxicological review provided a limited discussion of
health effects from exposure to vulnerable life stages; and (3) the approach used to develop the RfC and
RfD in this assessment built upon developing a subchronic RfC and RfD for each of the isomers and
applying uncertainty factors to arrive at a chronic value for inhalation and ingestion yet the assessment
does not provide a subchronic RfC or RfD. The SAB provides discussion and recommendations relating
to these issues after the responses to the EPA charge. Charge questions are included in italics in relevant
sections of this report and the full charge is included as Appendix A.
5
-------
3. RESPONSE TO CHARGE QUESTIONS
3.1. Enhancements to IRIS Assessments
3.1.1. Preamble Enhancements
Charge Question: NRC (2011) indicated that the introductory section of IRIS assessments needed to be
expanded to describe more fully the methods of the assessment. NRC stated that they were "not
recommending the addition of long descriptions of EPA guidelines to the introduction, but rather clear,
concise statements of criteria used to exclude, include, and advance studies for derivation of [toxicity
values]. " Please comment on whether the new Preamble provides a clear and concise description of the
guidance and methods that EPA uses in developing IRIS assessments.
To a substantial degree, the Preamble as currently written provides a concise and clear description of the
process that is followed, its steps, the places in the process where decisions or judgments are made, the
guidance that applies to making those judgments (with explanation of the main considerations and
available choices), and the process by which the results of each step feed into the next. The Preamble
certainly should be no longer; as it stands, it is near the limit of what can serve as an overview and
explanation. The Preamble is composed of three sections: The Scope (Section 1), the overall IRIS
Process for developing and reviewing assessments (Section 2), and the particulars of how an assessment
is executed (Sections 3 to 7).
The description of "Scope of the IRIS Program" (Section 1) is brief and clearly describes the IRIS
program, but the SAB notes that it lacks any overarching statement about what IRIS seeks to
accomplish, its ultimate purposes, and what its assessments are meant to represent to their users.
In view of the partial implementation of reforms to the overall process, the SAB presumes that the
Preamble will change from one assessment to the next to reflect newly adopted procedures. The SAB
recommends that it would be useful to note places where the present assessment has not yet fully
implemented changes that are already planned for application to subsequent assessments. If the
motivations (in terms of enhancing transparency, objectivity, and sound analysis) for future changes can
be borne in mind and addressed, the overall revision of the IRIS process will be smoothed. Furthermore,
assessments done before the process is complete will gain credibility and longevity.
The SAB finds that Section 2 on the IRIS Process is clear and concise in the description of the seven
steps to develop assessments and the multiple levels of review and what happens at each one. It is rather
vague, however, on the nature of the problem formulation step. The SAB recommends that it include
some discussion (without seeking to constrain the agency's further actions) about the issues needing to
be addressed, the prospects for addressing them with available data, and the uncertainties and plausible
alternative interpretations that would need to be worked through. Although the section clearly indicates
that the agency will prepare a summary record of response to peer review comments, it is not clear who
will be the ultimate EPA decision-maker. The discussion of Step 5 notes that newly published studies
that are critical to conclusions can be brought into consideration, but a more explicit reference to the
stopping rule policy (and where its details can be found) would be appropriate.
The SAB recommends that the agency take measures to ensure that the Preamble in this and future
assessments be structured so that it refers the reader to the appropriate guidance and cannot be construed
to contradict policy by over summarizing existing guidance.
6
-------
Sections 3 to 7 lay out the specific steps for conducting an assessment. A good deal of the material is a
summary of long-extant guidance, and so will be familiar to most readers. The Preamble purpose does
require setting out the processes and analyses, but perhaps this could be done somewhat more briefly.
On the other hand, the aspects that are under revision need to be flagged or expressed in more general
terms. In particular, the Preface notes that Phase 3 of the IRIS revision process is yet to happen, and this
will include review of current methods for weight-of-evidence analysis. The methods for abstracting
data, systematically considering study quality and interpretation issues, and ranking relative study
impact are also in transition.
Section 5.5 references the carcinogen classification scheme of the 2005 Carcinogenicity Assessment
guidelines. The SAB notes that the same section also cites the Integrated Science Assessment criteria for
causality (applied in the evaluation of criteria air pollutants) as "another example" and, further on in the
Preamble it is noted that the agency is investigating what descriptors to use and may use these or others.
The SAB is concerned that this may produce confusion as to what guidelines for assessment of causality
apply to the current trimethylbenzenes assessment and whether the IRIS revision process is anticipating
a revision or expansion of such guidance in the future. The SAB finds that, at this juncture, discussing
the intent of descriptors is probably more useful than recounting definitions that may or may not be used
and may or may not be seen as in keeping with the spirit of the overall revision process.
It should be clear that the Preamble itself is not guidance; it only summarizes guidance that is set out
elsewhere; an unambiguous statement to this effect should be added. This is especially critical because
being only summaries and explanations the treatment in the Preamble is less developed and is
unaccompanied by the full guidance's discussion about motivation, meaning, interpretation, and
scientific justification of the briefly described analyses, presumptions, standards, or judgments. Without
reference to the fuller treatment, the SAB is concerned that there is danger that an oversimplified version
may be mistaken for policy. In some places where existing guidance is described and explained, there is
a citation to the full guidance document, but in many spots, there is no such citation. The SAB finds that
more care in providing citations to the operative guidance documents is necessary if the Preamble is to
adhere to the distinction between established policy and explanation. Citations also give readers an
indication that there is a fuller description of the issue to be found and where to find it.
Some precepts articulated in the Preamble appear to the SAB as not consistent with existing EPA
guidance or announced policy. This raises questions about whether the agency is changing policy from
established guidance and whether such changes have been appropriately vetted, and implemented.
Several statements seem to be outside of existing guidance and are provided as examples:
p. xxii, line 67 that negative genetic toxicity studies carry less weight than positive ones;
p. xxiii, line 78 that funding source can downgrade the credibility of studies;
the assessment includes organ-specific reference values on p. xxx, line 32; and
the dismissal of specificity as an aspect of causal analysis on p. xx, line 81.
The SAB finds that all of these issues are important and should be discussed. However, they should be
supported with citations to existing policy or guidance. If these are not existing policy, then they should
be flagged as matters under discussion. In view of the incremental alterations that are expected to occur
in IRIS assessments over a series of assessment documents, it is important that any changes to the
Preamble from assessment to assessment - especially those that could be construed as altered guidance
or standards for future data interpretation - be considered carefully and called out for attention in
document reviews. In summary the SAB recommends that:
7
-------
the assessment should include some discussion about the issues needing to be addressed, the
prospects for addressing them with available data, and the uncertainties and plausible alternative
interpretations that would need to be worked through.
the document should point out places where the course of its analysis touches on aspects of the
more general IRIS review process that have not yet been implemented, but for which further
development is planned; and
the Preamble should make clear that its summary of relevant guidance does not supersede that
guidance, and it should provide adequate citation to that guidance.
3.1.2. Presenting Assessment Steps and Outcomes
Charge Question: NRC (2011) provided comments on ways to improve the presentation of steps used to
generate IRIS assessments and indicated key outcomes at each step, including systematic review of
evidence, hazard identification, and dose-response assessment. Please comment on the new IRIS
document structure and whether it will increase the ability for assessment to be more clear, concise and
easy to follow.
The objective should be to make it possible to read the document in three different modes:
1) quickly to get the main qualitative and quantitative conclusions and, in general terms, their bases;
2) somewhat more thoroughly, but still rapidly, to get a good picture of the kinds of data and
toxicity phenomena that were considered (not just those that were chosen as critical or as bases
for quantification), the main features and issues involved in the interpretation, the choices that
were made (and the nature of the main alternatives) and the main rationale for the choices; and
3) in detail, to efficiently find the particulars of study features and data, their analysis and the
detailed reasoning behind their interpretation.
In short, the reader should easily find (1) the conclusions, (2) the choices and reasoning applied when
reaching them, and (3) the fully explained justifications of choices, respectively. The SAB finds that the
structure in the TMB document does well at the first, in the form of the Executive Summary. The
leading section on "Occurrence and Health Effects" is useful as a context for the particulars that follow.
A good balance between brevity and depth is struck.
The second way of reading for the choices and reasoning has also markedly improved, though
some suggestions can be made, as discussed below.
The third way of reading the examination of particulars and the ability efficiently to find them
documented with sufficient detail is much improved from former IRIS documents. The relegation of
a lot of the details to well-structured appendices is helpful. The set of focused appendices helps a reader
to find the place where particular study aspects or analyses of issues are to be found. The organization of
the appendices and the consistency of presentation across IRIS documents are important in
making the place to find details clear. Although the general structure of the appendix entries can be
discerned, the plans for the structure and consistency have not been provided, so it will take some time
and examination of other documents following the same plan for readers to find things easily. The use of
appendices simultaneously allows presentation of more detail than may have been captured in earlier
generations of IRIS documents and also avoids cluttering the main body of the IRIS document where
interpretation and evaluation are considered. The appendix approach also frees the main document from
seeming to need to present all the details before drawing any interpretation.
8
-------
Consistent Presentation of the Studies Considered
The SAB recommends that each study should be in a consistently formatted table. The table should be in
an appropriate appendix and present the study-specific considerations that bear on evaluation of study
quality and pertinence, including shortcomings and assumptions that are needed to interpret the study's
outcomes. Consistency of format is important within each document, but it would also be a useful goal
to achieve from one IRIS assessment to another.
The SAB suggests that it would also be useful for each study to have a short overview section (also in its
appendix listing, not repeating tabulated details) of the nature of the study, its examined endpoints, and
relevant findings. The goal of the overview is to provide context for the tabulated details, so that the
details need not be read in full to gain an idea of the general nature of the study and its importance to the
assessment as a whole. This overview should not discuss interpretations.
It is clear that the intent of this structure is to free the main document to focus on choices that were made
in the analysis (selection of possible endpoints, selection of studies to represent and characterize those
endpoints, and analyses and interpretations of their bearing on human risk estimation). The challenge is
to bring the appropriate data and level of detail from the appendices into the main body, so that the
interpretations and choices can be justified and documented, without overwhelming the interpretation
discussion or leaving out potentially relevant information. Sorting this out is the essence of the
systematic review process and, though clear strides have been made, more work is left to be done. The
SAB suggests that, as the EPA develops its enhancements to the IRIS reviews, the key to this process is
to be transparent regarding both studies chosen for inclusion and those chosen for exclusion; not just
what supports an interpretation, but also what seems unexplained or even inconsistent. The results of
studies should be cited when they are consistent with a hypothesized potential for human risk and also
when they have apparently contrary results with different implications for scientifically supportable
inference about human risk impacts.
The SAB finds that the overall structure of the main report provides a good framework with sections
on literature search, hazard identification for the various candidate endpoints, and dose-response.
Describing the Literature Search
The Literature Search Strategy section is brief and focuses only on identification of pertinent studies
from the literature. The SAB is concerned that the general description of the process and the specific
implementation for TMBs may be too exclusive, missing potentially informative ancillary studies that
could help in interpretation or evaluation of those studies strictly observing toxicity outcomes of the
TMBs alone in controlled settings. The SAB recommends that, in the literature search process, a clear
distinction be drawn between, on the one hand, the primary search for sources of evidence that would be
used directly in the identification and characterization of potential hazards for TMBs as well as in
quantitative analysis of dose-response properties, and on the other hand, the secondary identification of
studies that describe effects and properties of similar chemicals or that illuminate underlying biological
processes that might be targets of toxic action. The primary literature search should be comprehensive
and subjected to an orderly process of systematic review. The secondary search is for literature that is
useful to provide context, in terms of what might be expected given the knowledge of other chemicals
and of the potential pathways of toxic action. As such, the secondary search need not be comprehensive
and could include reviews as well as original experimental studies; the aim should be to provide enough
context so that the assessment of the first set of literature can be informed by what might be expected,
given existing knowledge of similar chemicals, which can be further evaluated with respect to how these
similar cases may help to fill data gaps which exist in the TMBs primary literature.
9
-------
It should be clear that literature search is only the first step of systematic review, which needs to be
followed by evaluation of each study in terms of design, quality, shortcomings, main findings (including
both positive and negative findings), evaluation of the reliability of individual study results, and
identification of other studies, particularly on mechanisms, that could address uncertainties in the
primary database. This supports a further process of comparing results across studies to assess both the
consistency of specific effects, and also the manifestation of related effects that would be expected from
hypothesized underlying causative processes, both of which bear on the use of specific study results as
evidence regarding the existence and nature of hazards in human target populations.
Describing the Hazard Identification Steps
The individual endpoint sections of the Hazard Identification have some discussion about interpretation
across studies and evaluations of bearing and relevance, though further discussion of interpretation
rationales and consideration of alternatives would be beneficial. The SAB finds that it is the middle
section of systematic review after the studies are chosen but before the interpretation of their overall
bearing gets considered that does not have a clear home in the current document structure. As the
agency develops its approach for systematic review, including defined ways for abstracting data, judging
study quality, documenting factors bearing on interpretation and its limits, and considering the impact of
related studies, it will be important to develop the document structure that encompasses all aspects for
consideration. The SAB notes that the Preamble has a section (Section 5) on evaluation of causality,
which depends on the existence of such a documented review and evaluation process, but the present
document has no particular place where the Preamble's named considerations strength, consistency,
specificity, temporal relationship, biologic plausibility, coherence, natural experiments, and analogy
are systematically considered or documented.
The SAB recommends adding a brief summary of the main features of assessment - in this case the
pharmacokinetics and metabolism - before the section on Hazard Identification. The aim is not to
replace the fuller treatment of these issues in the appropriate appendix, but rather to set the context for
the interpretation of studies bearing on hazard, and the main presentation of pharmacokinetic details
should continue to reside in the appendix. The main text's section would note such things as extent of
absorption, rapidity of elimination, main metabolic processes, main means of clearance (and what part
of that is by metabolism), indications whether metabolic saturation or enzyme induction might play a
relevant role in toxicity studies, and any notable unusual differences between experimental animals and
humans. Again, the point would not be just to list specifics (which can remain in an appendix) but to
provide the basic insights that might bear on how toxicity data are interpreted or on the limits to such
interpretation.
A noteworthy change from earlier IRIS assessments is that the Hazard Identification section is separated
into assessments of each endpoint, with relevant data for that endpoint being reviewed within the
section. The SAB finds that discussing similar endpoints clustered together is a great improvement over
the past practice of summarizing study by study. The endpoint-by-endpoint analysis permits the
examination of consistency and sufficiency of data to draw hazard conclusions about each effect.
This being said, there are possible overarching ties among endpoints that would help in evaluation of the
hazard characterization of each (say, commonalities of dosimetry or mode-of-action) that should be
discussed in an appropriate place. It would be useful to include considerations that might indicate a
study as the critical study.
10
-------
The tabulation of studies is useful, and the dose levels and dose-specific responses are important details
to include. The hyperlinks to the detailed description of studies in the appendices helps to make those
appendices directly supportive and makes finding of relevant information more efficient. The exposure-
response arrays are useful summary devices to aid communication, though they should not be read as
meta-analysis forest plots or otherwise be used as the primary basis of conclusions. Nonetheless, they
provide a valuable overview of the data.
Describing the Dose-Response Steps
In the dose-response section, the tabulation of points of departure (PODs), health effects concentrations
(HECs), and applied uncertainty factors (UFs) is useful, allowing endpoints to be compared and the
distinction between a low POD with few UFs and a high POD with many UFs to be seen.
It represents an important advance that the Hazard Identification sections for each endpoint have specific
places for discussion of consistencies and inconsistencies among data, on the relevance of studies for
human risk evaluation, on the knowledge of mode of action (even if it must say that little is known), and
on alternative interpretations of the available data on potential causation. The format that addresses each
of these issues in an orderly way for each endpoint is important to advancing the explanation of the basis
for conclusions and enhancing transparency. However, The SAB is concerned that these interpretation
passages in each Hazard Identification section are somewhat too concise, and suggests that it would be
good to find a consistent way (perhaps more appendices) to document the arguments without unduly
distracting from the main discussion.
Presenting the Outcomes
As it stands, both the Hazard Identification and Dose-Response sections simply dive in to the first
endpoint or analysis to be considered, and then have separate sections on each. There is little overview
to prepare a reader for what is coming or to point to the parts that are critical versus those that are there
for completeness. In general, to help enable a reader to grasp the main lines of argument and only go
into detail when needed, the SAB recommends that both the Hazard Identification and the Dose-
Response sections have an initial paragraph setting out the main issues that will be considered and
indicating which considerations (to be developed in the subsequent text) are the most notable for the
larger assessment process. A parallel paragraph at the end of each of these chapters could summarize
what its contents have provided to the larger assessment process. The aim of these paragraphs would be
to make it possible to read the document in more detail than provided in the Executive Summary (which
largely documents findings) but still quickly see the deeper structure of the report and where to focus for
more information on particular aspects. That is, the initial and last paragraphs as proposed would not be
justifications of choices, but only a guide to the more detailed discussion in each section.
In summary the key recommendations are:
Each study should be in a consistently formatted table and present the study-specific
considerations of study quality and pertinence, including shortcomings and assumptions that are
needed to interpret the study's outcomes. Consistency of format is important within each
document, but it would also be a useful goal to achieve from one IRIS assessment to another.
The literature search process should be described to draw a clear distinction between the primary
search for sources of evidence and a secondary identification of studies used to describe effects
similar chemicals or studies that illuminate the underlying biological processes that might be
targets of toxic action used to fill data gaps. The primary literature search should be used directly
11
-------
in the identification and characterization of potential hazards for TMBs as well as in quantitative
analysis of dose-response properties.
Add a brief summary of the main features of assessments before the section on Hazard
Identification to provide context for the interpretation of studies and the more the detailed
treatment should remain in an appendix.
Both the Hazard Identification and the Dose-Response sections should have an initial paragraph
setting out the main issues that will be considered and indicating which considerations (to be
developed in the subsequent text) are the most notable for the larger assessment process. A
parallel paragraph at the end of each of these chapters could summarize what its contents have
provided to the larger assessment process.
3.1.3. Standardized Evaluation of Critical Studies
Charge Question: NRC (2011) state that "all critical studies need to be thoroughly evaluated with
standardized approaches that are clearly formulated" and that "strengthened, more integrative, and
more transparent discussions of weight of evidence are needed. " NRC also indicated that the changes
suggested would involve a multiyear process. Please comment on EPA's success thus far in
implementing these recommendations.
The SAB finds that, in general, a great deal of progress has been made in restructuring the document to
focus the main body on documenting and explaining the interpretations, choices, and analyses, and
relegating the supporting information to appendices. The use of links to the appendices aids in using
them as support, without encumbering the flow of the main arguments. At the same time, the details of
studies are important when the study results are used in constructing arguments. It will be an ongoing
challenge to bring enough into the main text to document the reasoning, to avoid leaving important
aspects hidden in the appendices, and still to have a readable document that fully explains the choices
and conclusions made.
The SAB notes that the process of systematic review still needs development. Documentation of the
process of identifying literature has progressed, but further development is needed in establishing
standard practices for abstracting relevant data, for evaluating study quality, strengths and shortcomings,
and for integration of evidence across studies. This includes the phases of evaluating individual studies,
of comparing the results of studies of similar objective into characterizations of their joint bearing in a
way that addresses discordant results, and of the overall integration across lines of evidence to form and
justify judgments about causality and appropriate dose-response analyses.
In this development, the SAB suggests that it should be borne in mind that the process of systematic
review is not solely one of identifying the "right" or the "best" data, with the interpretation and bearing
on risk evaluation becoming clear once the right choices are made. The integration and weight-of-
evidence evaluation process requires accepting that multiple interpretations are always possible,
especially in different contexts, and that consistency of causal interpretations with available data should
be considered across all applicable studies, bearing in mind the possible role of study quality limitations
in generating apparent discordances. This process should consider how results of particular studies are to
be generalized to apply to other situations (especially to actual human exposures); it needs to account for
why other study results might disagree; and it needs to consider how other interpretations would have
different consequences for risk estimation.
12
-------
The SAB recognizes that an important challenge facing the agency is that assessments must go ahead
even as this further development proceeds and before all aspects are complete. The Board notes that a
strategy of working on the structure of the assessment, focusing on text to document the process and the
agency's choices and analytical options, is a good way to begin.
The recommendations for revision of the IRIS process come from the NRC "Roadmap" (Chapter 7 of
the Formaldehyde review) and other sources. The SAB recommends that a good principle to follow in
conducting assessments during the process of revision is to consider the reasons behind the
recommendations for change, and to make efforts to address the issues and to explain how the chosen
approaches seek to reflect the NRC recommendations, although the methods may not yet be fully
developed and agreed upon. That is, trying to address as well as one can the issues behind the
recommended methodological and procedural changes is a good way to make assessments as reformed
as they can be, and improve acceptance as the overall IRIS process continues to advance.
3.1.4. Addressing Public Comments on the Draft Toxicological Review of Trimethylbenzenes (May
2012)
Charge Question: EPA solicited public comments on the draft IRIS assessment of trimethylbenzenes
[May 2012] and has revised the assessment to respond to the scientific issues raised in the comments. A
summary of the public comments and EPA's responses are provided in Appendix F of the Supplemental
Information to the Toxicological Review of Trimethylbenzenes. Are there scientific issues that were
raised by the public as described in Appendix F that may not have been adequately addressed by EPA ?
Public comments on the draft IRIS Toxicological Review of Trimethylbenzene (as summarized in
Appendix F of the assessment) focused on the standards and transparency of the draft document and
several scientific areas: (1) why the EPA did not use the available data on C-9 mixtures (mostly TMB
isomers and ethyltoluene isomers, which according to the comments, have similar toxicological profiles)
for the IRIS evaluation; (2) why the EPA identified the critical endpoint as pain sensitivity based on a
transient latency in paw lick to a hot plate stimulus following subchronic exposure, that was not evident
after chronic exposure and was a reversible response after two-weeks post-exposure; and (3) why the
1,3,5-TMB oral gavage toxicity study (Koch Industries, 1995), the results of which were already
accepted by the EPA, was not used to reduce the identified uncertainties.
The SAB finds that Appendix F did address issues raised in public comments and that explanations were
furnished for the agency's stance on the issues and their disposition. That is, the issues were all
addressed according to the agency's judgments, and those judgments were transparently discussed.
The TMB Review Panel was divided, however, on the adequacy of the responses and the advisability of
the dispositions that were made as presented in the summary. In particular, there were a variety of views
on the role that testing of the C-9 fraction should have in the assessment, with some panelists accepting
the reasons for omission of this from the main evaluation and others feeling that these results had a role
that had not been adequately explored. The use of the C-9 fraction in the TMB assessment is further
discussed in section 3.2.3 of this report. The SAB concludes that there is value to considering the C-9
mixtures studies along with data for related alkylbenzenes (e.g., toluene, ethylbenzene, xylene, styrene)
in helping to inform gaps in the TMBs database. There was also disagreement among the TMB Panelists
related to the interpretation of the pain sensitivity data, with some members questioning whether the
document adequately examined the question of reversibility following termination of exposure, which
further bears on whether ongoing or repeated exposures to TMBs should be deemed to have
accumulating toxicity beyond effects evident in shorter-term exposure; other panel members believed
13
-------
that the data were consistent with cumulative toxicity and lack of reversibility. The full discussion of
these issues and their treatment in the TMBs assessment is covered in the responses to the charge
questions in Section 3.2 of this report.
3.2. Toxicological Review of Trimethylbenzenes
3.2.1. Executive Summary
Charge Question: The major conclusions of the assessment pertaining to the hazard identification and
dose-response analysis have been summarized in the Executive Summary. Please comment on whether
the conclusions have been clearly and sufficiently described for purposes of condensing the
Toxicological Review information into a concise summary.
The Executive Summary condenses the large amount of information presented in the draft TMB
Assessment and the Supplemental Information. Individual conclusions regarding RfC and RfD values,
as well as other relevant information (e.g., carcinogenicity) for each of the TMB isomers of concern are
clearly described. The SAB recognizes that there is always some tension to find the appropriate level of
detail to include in the Executive Summary. The Executive Summary presents somewhat detailed
information on the data used to develop the RfC and RfD for each of the three isomers and that detail
may detract from the intended purpose of brevity. As discussed in Section 3.1.1, the Executive Summary
should emphasize the major conclusions of the assessment and provide the specific details of the critical
studies in the main text of the assessment.
Recommendations to improve the Executive Summary include:
The Executive Summary should be truncated to emphasize the major conclusions. Specifically,
citations should be removed from the summary unless they are absolutely essential. Whole
sections of the Executive Summary are devoted to elaborating on "Confidence"; for example, the
last paragraphs in Sections 3 and 5 are identical except for the compound being discussed. The
SAB recommends that the EPA consider treating "Confidence" as a single, very succinct section
toward the end of the Executive Summary. Issues pertaining to the use and rationale for
assigning confidence for each isomer should be relegated to the corresponding sections in the
main text.
Another example where too much detail is provided is the middle paragraph on page xxxvi. The
text and table both describe the calculations for the RfC, even though the details are provided in
the main body of the text.
Much of Section 15 in the Executive Summary (Susceptible Populations and Lifestages) seemed
speculative. While the concepts may be correct, they were not pertinent in the executive
summary on TMBs. This section could be truncated after the first sentence, which is a clear
summary of what is known. The SAB also provides more specific comments on sensitive and
vulnerable populations in Section 3.3.1.
3.2.2. Literature Search Strategy/Study Selection
Charge Question: The process for identifying and selecting pertinent studies for consideration in
developing the assessment is detailed in the Literature Search Strategy/Study Selection section. Please
comment on whether the literature search approach, screening, evaluation, and selection of studies for
inclusion in the assessment are clearly described and supported. Please identify any additional peer-
reviewed studies from the primary literature that should be considered in the assessment of noncancer
and cancer health effects of 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB.
14
-------
The SAB finds that the search strategy was clearly articulated. The databases were clearly defined, as
were the search terms (Table LS-1). In contrast, however, the process for selecting which of the
identified studies to use for the assessment was not transparent.
A flow chart (Figure LS-1) indicates that the initial search identified approximately 4,300 papers, of
which approximately 200 were used in the draft TMB Assessment. While it was clear which papers
were used in the draft assessment, there were no means of determining which papers were excluded
from the assessment. Thus the review does not provide sufficient documentation to determine if
important papers may have been overlooked or considered and then omitted from consideration based on
EPA's criteria. As such, the SAB strongly recommends that the EPA provide citations for the 4,300
papers and group them according to the main reason why they were excluded. This could be
accomplished in several ways, ranging from an appendix at the end of the document, to a link between
the document and the Medline search that was used, to an on-line searchable data base.
The flow chart (Figure LS-1) also indicated that 65 papers were excluded "based upon manual review of
paper/abstracts." Again, there were no means of determining the identity of these papers. Furthermore,
certain papers were excluded because they were "not available in English." The SAB notes that
translation options are available and finds that this criterion for omission is unacceptable. Among the 65
papers, others were excluded because they were in vitro studies. In vitro studies are mentioned in the
assessment (e.g., Janik-Spiechowicz et al. 1998; page 1-46); thus, it was not clear why some in vitro
studies were included and others were excluded.
The SAB noted that the description of the search strategy did not mention xylenes or ethylbenzene.
Because of the close similarity of xylenes to TMBs and the very similar toxicological effects caused by
xylenes, this may have resulted in important papers being excluded, thus weakening the conclusions of
the assessment. For example, the findings of Chen et al. (1999) and Lee et al. (2005) (cited on p. 1-1)
relating painters' exposure to solvents to neurological problems have a relatively weak association to
TMBs. The SAB notes that the links in these two studies are stronger to xylene and to a mixture of
aromatic solvents including TMBs rather than the TMB isomers. For example, studies such as those of
Ruijten et al. (1994), Qian et al. (2010), Tang et al. (2011), and El Hamid Hassan et al. (2013), are
closely linked to xylene but not cited in the document. The overall association of the effects reported in
these studies in painters with exposures to aromatic solvents like the TMBs is much stronger than the
associations reported by Chen et al. (1999) and Lee et al. (2005).
Section B4 in the Supplemental Information provides details on each of the human studies. While not
directly pertinent to the search strategy, the SAB recommends the inclusion of a summary table of the
studies related to each health effect: for example, a table with the 9 studies on neurotoxicity in humans,
reporting study design, inclusion and exclusion criteria, number of subjects, and main results. This is
common practice in epidemiologic reviews and meta-analyses. The current way of presenting the study
has some advantages because it is very analytical, but it is also hard to summarize.
In summary key recommendations are:
Provide citations for the 4,300 papers and group them according to the rationales for their
exclusion. This could be accomplished in several ways, ranging from an appendix at the end of
the document, to a link between the document and the Medline search that was used, to an on-
line searchable data base.
15
-------
Include a summary table of the studies related to each health effect: for example, a table with the
9 studies on neurotoxicity in humans, reporting study design, inclusion and exclusion criteria,
number of subjects, and main results.
Additional references that should be considered by the EPA include:
Chapter 8 on Trimethylbenzenes (NRC 2013),
Health Hazards of Solvents Exposure among Workers in Paint Industry (El Hamid Hassan et al
2013)
Xylene-induced auditory dysfunction in humans (Fuente et al. 2013)
Hearing loss associated with xylene exposure in a laboratory worker. (Fuente et al. 2012)
Visual dysfunction in workers exposed to a mixture of organic solvents. (Gong et al. 2003)
Ototoxicity effects of low exposure to solvent mixture among paint manufacturing workers.
(Juarez-Perez et al. 2014)
Short latency visual evoked potentials in occupational exposure to organic solvents (Pratt et al.
2000)
Auditory brainstem response in gas station attendants (Quevedo et al. 2012)
3.2.3. Hazard Identification
Synthesis of Evidence for TMBs
Charge Question: A synthesis of the evidence for trimethylbenzene toxicity is provided in Chapter 1,
Hazard Identification. Please comment on whether the available data have been clearly and
appropriately synthesizedfor each toxicological effect. Please comment on whether the weight of
evidence for hazard identification has been clearly described and scientifically supported.
The synthesis of evidence for the three TMB isomers is nicely divided up into the various target organs
or forms of toxicity, as well as by exposure route and by human versus animal studies. The studies
chosen for review are clearly described and the summary tables and figures well supplement the text.
The tables are clear and useful, and the figures nicely summarize the available data for each effect by
each isomer. An introductory paragraph describing the section layout, including the summary tables for
each endpoint, would improve readability.
Discussion of the individual endpoints is flawed by questionable statistical statements or inferences. In
several places (pp. 1-3, 1-4, 1-5, 1-7, 1-36), the descriptions of non-statistically significant results infer
that effects have been observed. The SAB recommends that descriptions of results more closely adhere
to the rule that statistical significance provides the criterion of whether an effect has occurred (although
data trends can be cautiously noted).
The discussion of respiratory effects should be strengthened by further consideration of the relevance to
humans of the effects observed in the high-dose animal studies. While it's clear that respiratory effects
are observed and are a relevant endpoint in humans, the distinction between the high-dose animal effects
and the human effects could have been made more clearly. The limitations of the human evidence for
hematological and clinical chemistry effect, based on the uncertainties in exposures (mixture
components, doses) should be more clearly described. With regard to carcinogenicity as an endpoint, the
document clearly communicates the inadequacy of the database, including the minimal genotoxicity
database.
16
-------
The summary table (Page 1-49, Table 1-7) is very helpful in understanding the points made with regard
to toxic effects. A summary table or scheme regarding toxicokinetics and metabolism would also be
useful. Section 1.1.7, which focuses on the toxicokinetic similarities among TMB isomers, would be
improved by summarizing in a table or scheme both the similarities and differences among the isomers
in toxicokinetics and metabolism.
The synthesis has sections that summarize weight of evidence determinations for cancer and non-cancer
endpoints as well as a summary of the uncertainties raised by potentially susceptible populations. The
SAB recommends that this would also be a good place (in its own separate subsection) to describe the
major uncertainties and gaps present in the TMBs toxicological database. A synthesis of the literature
should naturally lead the reader to a summary of what is known and what is unknown for the
chemical(s) under review. This then leads to a discussion of data gaps, which in the current draft does
not appear until Page 2-11. The current discussion is brief and does not weigh the value of evidence
from related chemicals or from studies done on the C-9 mixture. Structurally related alkylbenzenes such
as toluene, xylene, ethylbenzene and styrene have similarities in neurotoxic effect and metabolic
disposition. Such information is clearly supported in the IRIS Preamble, section 3.1 (lines 11-15)
"[sjearches for information on mechanisms of toxicity are inherently specialized and may include
studies on other agents that act through related mechanisms." This is further supported in Section 5.4, p.
xxiii (lines 18-21), "Pertinent information may also come from studies of metabolites or of compounds
that are structurally similar or that act through similar mechanisms." It is therefore recommended that
additional animal and human studies on related aromatic solvents be considered in the qualitative and
mechanistic interpretations of TMB toxicity. Examples of such studies are included in comments on the
literature review. (See Section 3.2.2)
Information on mixtures containing TMBs and on compounds that are structurally related to TMBs can
be pertinent in several ways. First, the degree to which the effects seen in TMB-only studies are
consistent across related chemicals may help to evaluate the evidence for the existence and nature of
hazards, while also potentially informing MOA. Clearly, perfect consistency is not needed nor expected;
however, major discrepancies in comparably conducted studies should be noted as part of the
determination of the robustness of findings from the TMB-only studies. Where consistency of effect is
seen across related structures, studies with these related chemicals that go beyond the testing done on
TMBs may help fill data gaps, identify additional endpoints of potential concern, or better characterize
uncertainties that arise in the interpretation of TMB-only studies.
The main data gaps for the TMBs appear to be the lack of a developmental neurotoxicity study, the lack
of a multi-generational reproduction study and the lack of a chronic noncancer (neurotoxicity) study.
The EPA could potentially utilize data from these analogous alkylbenzenes to inform these data gaps
and determine whether the database uncertainty factor (now 3 fold) and the subchronic to chronic
uncertainty factor (now 10 fold) should be modified on this basis. Further, discussion of the existing C-
9 mixtures studies should be brought into the main document at this juncture by describing their
strengths and weaknesses and relevance to the setting of RfDs/RfCs for individual TMB isomers, with
particular emphasis on whether they provide evidence to inform the aforementioned data gaps. For
example, regarding the developmental neurotoxicity data gap, a Hungarian study (Lehotsky et al. 1985)
did test a C-9 mixture containing trimethylbenzenes (Aromatol) for developmental neurotoxicity in rats.
That study had minimal reporting of results, simply stating that there were no effects of Aromatol on
dams or offspring at any time point (Lehotsky et al. 1985). This is in spite of the fact that the high dose
of Aromatol was 2000 mg/m3, a dose that one would expect to have a neurotoxic effect in dams during
and after exposure, based upon results of other testing. The lack of any toxicity in dams or offspring,
17
-------
combined with the lack of reporting of any data (including Aromatol treatment group neurological
testing or Aromatol composition) and the fact that it was a mixture and not a specific TMB, makes this
study of limited utility for filling the developmental neurotoxicity data gap. Thus, chemical relevance
and study quality need to be considered when bringing in other chemicals or mixtures to help fill data
gaps.
If data for individual alkylbenzenes (toluene, ethylbenzene, xylene, styrene) or from the C-9 mixtures
studies are used by the EPA to modify the assignment of uncertainty factors, a comparative discussion
would be needed to describe similarities and differences between the surrogate chemical (or mixture)
and the TMBs and whether the extrapolation across chemicals reduces overall uncertainty or merely
reduces uncertainty in one area (e.g., subchronic to chronic extrapolation) only to add back a different
type of uncertainty (extrapolation across chemicals). The EPA's use of surrogate chemicals (toluene,
Page 1-19) to draw inferences regarding TMB mode of action is appropriate and as suggested below,
early life toxicokinetic data with toluene may be useful to decrease the uncertainty associated with early
life exposures. There may be additional value to review these surrogate chemicals and mixtures, not
only from the perspective of filling TMB data gaps, but to identify other effects of potential concern that
have not received adequate attention in TMB studies. This can include consideration of the human
clinical toxicology and epidemiology studies that may exist for these related chemicals.
The testing of the C-9 fraction has interpretative issues but is relevant to the TMB isomers under review
because this mixture, as tested, was about half TMBs; therefore, much of the observed effects could
have been due to the TMBs. The various C-9 components of the mixture may have created competition
for metabolic clearance that could have increased duration of exposure to the TMBs. However, certain
components may have induced metabolic clearance enzymes or competed for distributional pathways
(e.g., uptake into the CNS) which might have decreased the response. Thus, the minimal observed
toxicity in several C-9 studies (e.g., Douglas et al. 1993) provides relevant information for the
evaluation of individual TMB isomers and the agency should carefully explain its reasoning of the role
of C-9 studies in its final evaluation of TMBs. Relevance increases to the extent that application of the
IRIS assessment for TMBs may be used in certain settings for the evaluation of exposure to mixtures
containing TMBs and related alkylbenzenes. Thus as stated above the C-9 mixtures studies have
relevance and need further discussion. Within this context, data from additional mixture studies may
provide further perspectives on this question, as reviewed, for example, by Richie et al. (2001).
Noncancer Health Effects
Charge Question: Does EPA's hazard assessment of noncancer human health effects of
trimethylbenzenes clearly integrate the available scientific evidence (i.e., human, experimental animal,
and mechanistic evidence) to support the conclusions that trimethylbenzenes pose potential hazards to
the nervous system, respiratory system, the developing fetus, and the circulatory system (i.e., blood)?
Hazard assessment results in the identification of the potential adverse health effects attributable to a
specific environmental agent, the mechanisms by which agents exert their toxic effects, and the
associated doses, route, duration, and timing of exposure. Section 1.2.1 (Weight of Evidence for Effects
Other Than Cancer) contains a summary description of the toxicological evidence of effects of the
TMBs on the nervous, respiratory, circulatory and developmental systems. The section, however, does
not adequately describe the limitations and uncertainties within the database or how the results of the
hazard assessment will be utilized in the subsequent dose response evaluation. The SAB recommends
that Section 1.2.1 be revised to include the following:
18
-------
A short summary of the toxicokinetic similarities and differences among the three isomers early
in the section to provide context to the subsequent effect summaries.
A short summary of the neurological effects database limitations and accompanying
uncertainties such as lack of subchronic data for some isomers, lack of chronic data for all
isomers, questions of reversibility and lack of mechanistic data. The SAB notes that summaries
for the respiratory, hematological and development effects already make these distinctions.
Statement(s) regarding the confidence in the hazard identification results given the limitations of
the available database. This statement(s) should address the question: based on the sensitivity of
endpoints assessed in the limited database, lack of mechanistic information and effects observed
with similar compounds but not assessed for TMBs, what is the confidence that the hazards (i.e.,
sensitive health endpoints) have been adequately identified?
Inclusion of a concluding paragraph(s) which states how the results of the hazard identification
(e.g., the effects on the nervous system, respiratory system, the hematological system, and
developing fetus) will be utilized in the subsequent dose-response evaluation as well as
describing the relative importance of the different health effects.
Carcinogenicity
Charge Question: Does EPA's hazard assessment of the carcinogenicity of trimethylbenzenes clearly
integrate the available scientific evidence to support the conclusions that under EPA's Guidelines for
Carcinogen Risk Assessment (U.S. EPA, 2005), there is "inadequate information to assess the
carcinogenic potential" of trimethylbenzenes?
As noted in the detailed response to the charge question on carcinogenicity (See section 3.2.11), 1,2,4-
TMB has been assessed in only one study. The EPA found that there were a number of deficiencies
concerning this bioassay and the SAB agrees with the agency's finding. The EPA also noted that no
carcinogenicity bioassays have been conducted with 1,2,3-TMB or 1,3,5-TMB. As such, the SAB
concludes that the EPA's hazard assessment of the carcinogenicity of the TMBs did integrate all
available scientific evidence and agrees with the EPA that there is "inadequate information to assess the
carcinogenic potential" of trimethylbenzenes. The SAB provided recommendations to use data for
analogous compounds qualitatively (see section 3.2.3 Hazard Identification) to fill data gaps and
uncertainty factors. The SAB found that the available carcinogenic data for the TMB isomers is
inadequate to assess quantitatively as discussed later in this report.
3.2.4. Toxicokinetics and Pharmacokinetic Modeling
Charge Question: Data characterizing the toxicokinetics of 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB
following inhalation and oral exposures in humans and experimental animals support the use of
physiologically-based pharmacokinetic (PBPK) models for 1,2,4-TMB. For the purposes of this
assessment, the Hissink et al. (2007)model, originally describing 1,2,4-TMB toxicokinetics following
exposure to white spirit (a complex mixture of volatile organic compounds), was modified by EPA to
calculate internal dose metrics following exposure to 1,2,4-TMB alone for the derivation of an
inhalation RfC for 1,2,4-TMB. Additionally, the model was further modified by the addition of an oral
route of exposure for use in a route-to-route extrapolation for the derivation of an oral RfD for 1,2,4-
19
-------
TMB.
Please comment on whether the selectedPBPK model (Hissink et al. 2007) with EPA 's modifications
adequately describe the toxicokinetics of 1,2,4-TMB (Appendix B [of the TMB Assessment]). Was the
PBPK modeling appropriately utilized and clearly described? Are the model assumptions and
parameters scientifically supported and clearly described? Are the uncertainties in the model structure
adequately characterized and discussed?
The SAB finds that the selected model did an adequate job of simulating the time-course of TMB in the
blood of human subjects during and following acute inhalation exposures. There was excellent
agreement between predicted and measured blood TMB levels, both during and following 4-hour
exposures, for the subjects of Hissink et al. (2007) inhaling 100 ppm white spirit. All three of these
subjects regularly consumed alcohol, which would induce cytochrome P4502E1 and enhance TMB
metabolism. The model modestly, but consistently underpredicted blood levels in volunteers inhaling 30
ppm TMB for 8 hours (Kostrezewski et al. 1997). The model also consistently underpredicted blood
levels in persons inhaling 2 or 25 ppm TMB for 2 hours (Jarnberg et al. 1996, 1997, 1998), but to a
larger degree. Agreement was better at the lower exposure level. These subjects exercised during
exposure, which would increase their systemic uptake of TMB. Post-exposure blood levels were well
predicted for all human data sets.
In most instances, the model over-predicted blood TMB levels in rats subjected to single exposures to
white spirit (Hissink et al. 2007) and TMB (Swiercz et al. 2003). The differences between predicted and
empirical levels typically increased from IV2- to 2-fold at lower inhaled concentrations to 4- to 6-fold at
>100 ppm. The accuracy of predictions of brain levels was similar to those for blood. The model
reasonably simulated blood and brain levels in rats after repeated TMB exposures. Again, the disparity
between simulated and empirical data increased with increasing vapor concentration. With the repeated
exposure data of Swiercz et al. (2003), there were ~2- and 3-fold differences for the 25 and 50 ppm
exposures, respectively. Differences in brain levels after 606 hours were somewhat greater. There was
more disparity (4- to 5-fold) for blood and brain levels in the rats of Zahlsen et al. (1992) inhaling 100
ppm TMB for 3 days.
The poor model prediction for inhaled concentrations >100 ppm in rats is acknowledged by the EPA
authors. Nevertheless, they use the model to provide simulations for exposures outside its application
domain. This is necessitated by the fact that the 100-ppm dose is in the middle of the rat dose-response
range used for benchmark dose modeling. Over-predicting rat dosimetry in this range thus has the
potential to influence the results of dose-response modeling and extrapolation of potency to humans.
Marked over-prediction of high-dose data necessitated omission of the highest dose for benchmark dose
(BMD) modeling.
The EPA has two options for alleviating this issue; refine the rat PBPK model to improve fits or conduct
BMD modeling using inhaled concentration first to identify the point of departure (POD), then use the
rat and human PBPK models to determine the human equivalent concentration. Refining the PBPK
model may require recalibration of some type, such as the addition of a first-order metabolic pathway
consistent with the PBPK model of Jarnberg and Johanson (1999), or changing hepatic blood flow to 25
percent instead of 17 percent of cardiac output, which is a more common physiologic parameter value.
Alternatively, the EPA could conduct BMD modeling of the Korsak and Rydzynski (1996) data using
air TMB concentration as the dose metric to derive the POD. Subsequently, the PBPK model would be
used to convert the POD to the weekly average blood concentration. This alternative approach yields a
20
-------
BMD of 84 mg/m3 (17ppm), which would be predicted by the PBPK model to yield a blood
concentration of 0.087mg/L in rats. The result is identical to the values derived by the EPA, suggesting
that the approach of dropping the high-dose group used by the EPA does not introduce any bias. EPA
can use this alternative approach to support their BMD modeling approach.
The SAB conducted a quality control quality assurance review and confirmed the model simulations
presented in Appendix B of the IRIS document draft. Although a couple minor technical issues were
identified, no fundamental flaws or issues were found. This review is provided in Appendix B of this
report.
The EPA's assumptions, in modifying the Hissink et al. (2007) model to predict the kinetics of inhaled
TMB for repeated exposure scenarios, were reasonable and appropriate. The major caveats, however,
were not identified up-front on page B-20 (e.g. that the original model and its parameters were for TMB
and white spirit, lack of parameters for the oral route, lack of parameters for pregnancy). The SAB
recommends that the EPA expand the explanation and justification for the modifications of model
parameters. Specifically, the discussion of the input parameters (e.g., human tissue:blood partition
coefficients (PCs), cardiac output, liver blood flow) should be justified. Additionally the use of scaled-
up rat Vmax values, when human values were available, requires further explanation. Metabolic constants
could be questioned, as they summarily reflect the rate of TMB metabolism during mixed exposures to
white spirits, rather than exposure to TMB alone. The use of a liver blood flow of 17.5 percent of
cardiac output should be justified, as it differs substantially from the traditional value of 25 percent. The
EPA did not attempt any re-estimation or adjustment of parameters for chronic exposure (e.g., enzyme
induction, dose-dependency, growth dilution). Results of sensitivity analyses can be used to indicate
whether the choice of liver blood flow substantially impacts the model predictions and thus warrants
revisiting. It was noted that human tissue:blood PCs used in modeling were twice those for rats.
Meuhlenberg and Vijverberg (2000) estimated human brain:blood, fat:blood and kidney:blood PCs that
were higher for rats than for humans. It was suggested that first order and saturable metabolism be
incorporated into the model, and the model run to explore the impact of the change.
The SAB did not find a specific discussion of the uncertainties in the model's structure. While these
uncertainties may be implicitly included in the uncertainties discussion, they should be specifically
discussed in reference to the PBPK model.
One TMB panel member noted that there is a published human PBPK model (Jarnberg and Johanson
1999). The EPA requested the model through email and was unable to obtain the model. The model is
for TMB alone, and avoids the complications and uncertainties of: (1) concurrent exposure to other
components in white spirit; and (2) species-to-species extrapolations. Empirical human kinetic data are
available from the same laboratory for model parameterization and validation. Human neurobehavioral
data are also available in the literature from other research groups. The results of these studies identify
human NOAELs/LOAELs for acute irritation and central nervous system (CNS) effects by TMB and
white spirit. The SAB notes that EPA policy is to use and consider human data and validated human
models when available. Because the EPA could not obtain Jarnberg and Johanson model, the SAB has
provided recommendations to improve the use of the Hissink model and encourages the EPA to, at a
minimum, be more transparent in its discussion of available models and model selection in this and
future assessments.
The SAB recommends the EPA conduct BMD modeling of the Korsak and Rydzynski (1999) data using
air TMB concentration as the dose metric to derive the POD and subsequently use the PBPK model to
21
-------
convert the POD to the weekly average blood TMB concentration. This can be done to either replace the
EPA's current approach or offered as support of the EPA's approach (i.e., to demonstrate the same
answer results from either approach).
Charge Question: The internal dose metric selected for use in the derivation of the RfC and RfD for
1,2,4-TMB was the steady-state weekly average venous blood concentration (mg/L) of 1,2,4-TMB for
rats exposed for 6 h/day, 5 days/week. Please comment on whether the selection of this dose metric is
scientifically supported and clearly described. If a different dose metric is recommendedfor deriving the
RfC, please identify this metric and provide scientific support for this choice. Are the uncertainties in the
selected dose metric adequately characterized and discussed?
The use of any dose metric should be guided by the MO A of the chemical being examined. For the
TMBs, there is a paucity of information on their MO A, and the agency has inferred the mode of action
to be similar to that for chemicals such as toluene. Given the uncertainties in the MO A, the SAB finds
that the selection of the internal dose metric of the venous blood concentration averaged over a week of
exposure is reasonable.
In the absence of knowing the MO A, the area under the dose-response curve can be used to estimate the
average venous blood concentration as a viable dose metric. Given that the critical effects upon which
the RfC is being determined are neurological and, therefore, are extrapulmonary effects due to inhalation
of the TMBs, the selection of the internal dose metric comes down to either the maximum venous
concentration or the steady-state weekly average venous blood concentration. While there are acute
effects of 1,2,4-TMB that might bring into play the maximum blood concentration, there were also
effects with 90 days of exposure.
Clarification is needed on how the average weekly venous concentration was determined. This is
because the longer phase half-life of the TMB isomers indicates that an exposure period longer than a
week is required for blood levels to achieve a steady state. In addition, the experimental data for both
rats and humans show that steady state is not achieved with only a single week of exposure. Executing
the PBPK model over a 4-week period shows that the average blood levels are still continuing to rise
slightly. The model should be run long enough to come to a weekly steady state and then the associated
venous blood concentration used as the internal dose metric.
The multiple tissues of interest for derivation of an RfC are primarily extrapulmonary tissues. However,
the agency has a goal to establish RfCs for multiple endpoints beyond the critical effect endpoint
currently being addressed. If an effect in the respiratory tract is established such as a change in bronchial
alveolar lavage fluid composition and an RfC is to be determined, then the appropriate dose metric
would be based on the mass deposited per unit surface area of the lung rather than on the average venous
blood concentration. A mass per unit lung surface area dose metric enables species with significantly
different lung sizes than humans to be used in the derivation of the RfC.
Using the PBPK model-estimated internal dose metrics as the dose inputs for BMD modeling required
the agency to drop the high dose exposures from all modeling efforts because the venous blood dose
metrics consistently over-predicted experimental results for high exposures. This overestimation may be
due in part to the agency using minute ventilation as the driver function for internal dose rather than
decomposing minute ventilation into its two components, namely tidal volume and breathing frequency.
While the exposure level is high and that may lead to a 50 percent reduction in respiratory rate,
respiratory irritants such as the TMBs cause subtle shifts in the breathing pattern while maintaining the
same overall minute ventilation. Shallower breathing leads to a shift upward in the respiratory tract for
22
-------
the site of deposition. In addition, the PBPK modeling for humans did not take into account the periods
of exercise the subjects underwent, which may explain the model's greater deviations from experimental
results at high exposure levels. While the high doses would not need to be dropped if the agency added
an exponential rising model to their suite of models to be fit, the SAB notes that external air can be used
as the dose metric and then the PBPK model used to back-calculate the appropriate venous blood levels,
arriving at the same result that the agency obtained. If the SAB's suggestions for improvements in the
PBPK model do not lead to a better agreement with the high dose exposures, the agency would be well
advised to include the external air dose metric and corresponding venous blood back-calculations.
While uncertainties concerning model parameters, potential for kinetic changes with repeated exposures,
and model estimates of internal dose are discussed, the uncertainties in the selected dose metric (weekly
average venous blood concentration) are not adequately characterized or discussed in the TMB
assessment.
3.2.5. Inhalation Reference Concentration (RfC) for 1,2,4-TMB
Charge Question: A 90-day inhalation toxicity study of 1,2,4-TMB in male rats (KorsakandRydzynski,
1996) was selected as the basis for the derivation of the RfC. Please comment on whether the selection
of this study is scientifically supported and clearly described. If a different study is recommended as the
basis for the RfC, please identify this study and provide scientific support for this choice.
The SAB generally agrees with the choice of the Korsak and Rydzynski (1996) study as the basis for
derivation of the RfC for 1,2,4-TMB. The study utilized a 90-day exposure period and, thus, the longest
duration exposure study available in the literature; in addition, it included multiple exposure levels. It
was well-conducted and utilized adequate sample sizes of rats. In addition, it was based on widely-used
behavioral assays. An examination of the study indicates these behavioral studies were carefully carried
out and data from control animals were consistent with previously published observations.
Clarification of this choice, however, could be significantly improved in the document in several ways:
The rationale for the choice of Korsak and Rydzynski (1996) is not specifically described and the
reasons for its choice over other studies, e.g., the 4-week exposure studies, need to be more clearly
stated.
As currently written, there is confusion over chronicity of exposure vs. effects. It would be helpful
to modify the terminology particularly related to outcome measures, perhaps as acute effects vs.
long-term effects/irreversible effects and retain the use of the word chronic/sub chronic etc. to
descriptions of statements related specifically to exposure.
Separate the write-up into sections that specifically elaborate on the acute effects and provide a
separate section related to effects observed post-exposure. Given the commonality of even the
trends in data across these studies, some mention of the biological significance in the absence of
statistical significance (a = 0.05 as an arbitrarily chosen value) should be mentioned.
The text, where applicable, could include additional qualifications as to "reversibility of effects" at
the 2-week post-exposure time point. This assessment of reversible effects of failures on the
rotarod is based on the finding of lack of statistical difference between treated and control groups
at one week post-exposure following a 13-week exposure period for one of two isomers. Some
TMB Panel members felt that this was sufficient evidence for reversibility. For other members,
however, this did not provide sufficient evidence. Specifically, this interpretation of a reversal
relied on a reduction from 40 percent rotarod failure during the final week of exposure compared
to 35 percent one week post-exposure, as compared to 0 percent rates for controls. There was no
23
-------
such statistical reversal for the other isomer, and for both isomers, the magnitude of the reduction
post exposure was minimal. Further, it was not clear that the statistical analyses of these data
incorporated a repeated measures component that would be required by the experimental design.
Thus, while a case was stated for a statistically significant reversal, several TMB Panel members
felt that it was not consistent across nor did it appear to be biologically meaningful.
It was recommended that the EPA re-calculate the RfC as if the study were subchronic (i.e., UF
converts to 1 from 3) and report this value as well.
Include more specific mention of the potential cumulative neurotoxicity that is suggested by the
repeated measurement finding of rotarod performance failures across the course of exposure.
Include more specific descriptions of the similarity of the animal behavioral endpoints to what has
been observed in humans.
Charge Question: Decreased pain sensitivity (measured as an increased latency to pawlick response
after a hotplate test) in male Wistar rats was concluded by EPA to be an adverse effect on the nervous
system and was selected as the critical effect for the derivation of the RfC. Please comment on whether
the selection and characterization of this critical effect is scientifically supported and clearly described.
If a different endpoint(s) is recommended as the critical effect(s) for deriving the RfC, please identify
this effect and provide scientific support for this choice.
The SAB agrees that the reduction in pain sensitivity as indicated by an increased latency to pawlick
response in a hotplate test is a valid adverse nervous system effect and appropriately selected as a
critical effect for the derivation of the RfC. This effect was variously seen in response to short-term, 4-
week, and 90-day studies. The associated U-shaped dose-effect curves seen with these isomers,
moreover, are highly consistent with the effects of various other pharmacological agents (e.g., opioids)
on this response and likely reflective of the mechanisms by which these isomers act. This assay is
widely used in the behavioral pharmacology literature and particularly in the study of pain nociception
and opioid pharmacology.
The SAB agrees that the observation of prolonged latency in the hot plate test 24 hour post-footshock
delivery that was observed in studies by Gralewicz and colleagues (1997, 2001) also constitutes an
adverse effect. The administration of footshock immediately after the hotplate test trial essentially
maximizes the capabilities of the nervous system and, thus, provides a type of nervous system probe that
then unmasks a prolonged latency to a hotplate stimulus 24 hours later. It shows that when the nervous
system is maximally stressed, it cannot respond/recover in a normal timeframe.
In addition to the recommendations above for the document related to the nervous system effects, this
section could also benefit from some additional description of the hotplate procedures, including the
rationale/approach for using the footshock intervention in the post-exposure behavioral assessments
carried out after the 4-week exposures.
Charge Question: In order to characterize the observed dose-response relationship comprehensively,
benchmark dose (BMD) modeling was used in conjunction with dosimetric adjustments for calculating
the human equivalent concentration (HEC) from a rat and human PBPK model (Hissink et al., 2007) to
identify the point of departure (POD) for derivation of the RfC. Please comment on whether this
approach is scientifically supportedfor the available data, and clearly described.
a. Has the modeling been appropriately conducted and clearly described, based on EPA 's
Benchmark Dose Technical Guidance (U.S. EPA, 2012)?
b. Has the choice of the benchmark response (BMR) for use in deriving the POD (i.e., a BMR
24
-------
equal to 1 standard deviation change in the control mean for the latency to paw lick response)
been supported and clearly described?
EPA's decision to omit the high dose group from the Korsak and Rydzynski (1999) study before BMD
modeling is an initial concern. However, an analysis conducted on BMD modeling on the same dataset
using air concentration as the dose metric results in the same POD air concentration as BMD modeling
based on internal dose and using the low and mid-dose groups. As a result, the SAB agrees that the
overall results for the POD generated by the EPA are adequate but strongly suggests that the agency
provide a more robust explanation of any analyses. The SAB also considered Appendix C-2 in the TMB
Assessment as inappropriate and recommends deleting it. If the EPA is so inclined, they could replace it
with the BMD analysis using air concentration as the dose metric.
The SAB recommends that the EPA provide better justification for applying the "one standard
deviation" from the mean of the control group for the neurotoxicological endpoint than using the agency
default value. The EPA should also provide better explanation of the issues associated with the
homogeneity of variance across dose groups in the Korsak and Rydzynski (1999) study, its implications
for BMD modeling, and how the EPA addressed this in their BMD modeling.
Charge Question: Please comment on the rationale for the selection of the uncertainty factors (UFs)
applied to the POD for the derivation of the RfC for 1,2,4-TMB. Are the UFs appropriate based on the
recommendations described in Section 4.4.5 of A Review of the Reference Dose and Reference
Concentration Process (U.S. EPA, 2002), and clearly described? If changes to the selected UFs are
proposed, please identify and provide scientific support for the proposed changes.
Consistent with guidance provided in A Review of the Reference Dose and Reference Concentration
Process (U.S. EPA, 2002), five possible areas of variability and uncertainty were considered by the EPA
in deriving the proposed RfC for 1,2,4-TMB. This consideration is reflected in choices regarding five
specific uncertainty factors, namely:
UFA - an interspecies uncertainty factor;
UFH - an intraspecies uncertainty factor;
UFL - a LOAEL (lowest observed adverse effect level) to NOAEL (no observed adverse effect
level) uncertainty factor;
UFS - a subchronic to chronic uncertainty factor; and
UFD - a database uncertainty factor.
In responding to this charge question, the SAB evaluated the choice and rationale for each of these UFs,
reaching the following conclusions.
UFa. The SAB agrees with the UFa of 3 and its rationale. The default UFa of 10 can be divided into two
half-log UF components of 3 each to account for species differences in toxicokinetics and
toxicodynamics, respectively. In developing the RfC for 1,2,4-TMB, the EPA used PBPK modeling to
convert estimated internal doses in rats in toxicity studies of 1,2,4-TMB to corresponding applied doses
in humans. PBPK modeling substantially reduces uncertainty associated with extrapolating animal
exposures to humans based upon toxicokinetic differences, justifying elimination of one of the half-log
components of the default UFa of 10 (U.S. EPA, 2002). Uncertainty regarding possible toxicodynamic
differences among species, i.e., different sensitivity to toxicity at equivalent internal doses, remains,
justifying keeping the other half-log component of 3.
25
-------
UFh. The SAB agrees with the UFh of 10 and its rationale, although one TMB Review Panel member
thought that a UFh of 3 would be adequate. This UF is intended to account for potential differences
among individuals in susceptibility to toxicity. The EPA concluded that no information on potential
variability in human susceptibility to 1,2,4-TMB toxicity exists with which to justify using a value other
than the default of 10. It was noted during discussion that numerous clinical studies have demonstrated
that humans, including pediatric and geriatric patients, differ by only about 2-fold in their
susceptibility/sensitivity to inhaled lipophilic anesthetics (e.g., chloroform, halothane), indicating to one
Panel member that a UFh of 3 would be scientifically defensible given the neurotoxicity endpoint used
to establish the POD. Other TMB Panel members disagreed, stating that the mode of action of
neurotoxicity of 1,2,4-TMB is unknown and that the actions of general anesthetics may have little or no
bearing on variability in TMB susceptibility. In their opinion, the full UFh of 10 is warranted.
UFl. The SAB agrees with the EPA's choices for UFl values, i.e., a UFl of 1 for all endpoints except
increased bronchoalveolar lung cells, for which a UFl of 10 was selected. However, the SAB suggests
that the justification for the UFl be strengthened. This UF is intended to be used when the POD is a
LOAEL rather than a NOAEL. In conducting BMD modeling, a BMD equal to one standard deviation
change in the control mean for modeled endpoints was selected. The document would be improved by
adding an explanation of the reasoning for selection of one standard deviation (versus one-half standard
deviation) along with a clearer discussion of why this is expected to lead to a POD for which a UFl of 1
is appropriate.
UFs. The SAB agrees with the UFs of 3, although one TMB Panel member thought that a UFs of 10
would be more appropriate. When the data used to generate a chronic RfC are from subchronic studies, a
UFs is used to address uncertainty around whether longer exposures might lead to effects at lower doses.
The EPA justified using less than a full default factor of 10 for this UF stating:
A full subchronic to chronic uncertainty factor of 10 was not applied in this case
as there was evidence of reversibility of not only neurotoxic effects, but also
hematological effects in rats exposed to 1,2,4-TMB for subchronic durations.
Also, the respiratory effects appeared to be inflammatory in nature. Although
reversibility was not investigated for these endpoints, it is possible that adaptive
mechanisms may alleviate these effects following termination of exposure.
Most of the TMB Panel were satisfied with this justification, but some members of the TMB Panel
disputed the evidence for reversibility of effects. In addition several TMB Panel members noted that
reversibility following cessation of exposure was irrelevant since the chronic RfC is applicable to
lifetime of exposure - i.e., there is no post exposure period. The discussion regarding reversibility of
neurotoxic effects is presented in response to the RfC for 1,2,4-TMB (see Section 3.2.5). The TMB
Review Panel discussed that some hematologic effects considered by the EPA appeared to resolve when
exposure ceased, but other effects did not resolve, and that inflammatory pulmonary effects can lead to
persistent injury. The SAB notes that factors other than reversibility could contribute to selection of a
UFs less than 10, such as evidence from PBPK modeling that 1,2,4-TMB does not accumulate in the
body over time and empirical evidence that the POD does not appear to decrease when results from
subchronic studies are compared with studies of shorter duration. One TMB Review Panel member
thought that none of these considerations had sufficient merit to justify using less than the full default
UFs of 10.
UFd. The TMB Panel was divided on whether the UFd should be 3, as selected by the agency, or 10.
The purpose of this UF is to account for overall deficiencies in the database of studies available to assess
potential toxicity. The EPA cited strengths in the database in terms of availability of information on
26
-------
multiple organ systems from three well-designed subchronic toxicity studies in justifying not using the
full default factor of 10. In retaining a half-log factor of 3, the EPA noted the absence of a multi-
generation reproductive/developmental toxicity as a weakness in the database, and specifically concern
for the absence of a developmental neurotoxicity study for 1,2,4-TMB given the importance of
neurotoxicity in establishing the RfC. Among those who agreed with a UFd of 3, some found the
justification provided by the EPA to be satisfactory, while others thought that toxicity data available for
C-9 mixtures should contribute to the rationale to lower the value from the default of 10. Others
disagreed with including C-9 mixture data as relevant to the database UF. (See Section 3.2.3). Panel
members who thought that the UFd should be 10 cited various reasons, including the absence of data in
other species and the absence of a multi-generational reproductive study, as well as the opinion that the
absence of a developmental neurotoxicity study alone warranted a full factor 10. One TMB Panel
member pointed out that analogy with toluene suggests that the perinatal exposure could lead to
neurodevelopmental effects at doses 10-fold lower than the NOAEL for effects in adults. An additional
point made by another Panel member was that because the RfCs for all of the isomers are being set at
the same value, whereas the database is severely limited for the 1,2,3- and 1,3,5-TMB isomers and the
latter two compounds deserve a UFd of 10. Therefore, for consistency, a factor of 10 should be used for
all the isomers.
3.2.6. Inhalation Reference Concentration (RfC) for 1,2,3-TMB
Charge Question: A 90-day inhalation toxicity study of 1,2,3-TMB in male rats (KorsakandRydzynski,
1996) was selected as the basis for the derivation of the RfC. Please comment on whether the selection
of this study is scientifically supported and clearly described. If a different study is recommended as the
basis for the RfC, please identify this study and provide scientific support for this choice.
The SAB agrees with the EPA's conclusion not to base the RfC derivation for 1,2,3-TMB on isomer-
specific data. The justification for this conclusion is supported and clearly described. The SAB is not
aware of chronic or subchronic studies that could be used to support an RfC derivation for 1,2,3-TMB
with neurotoxicity as the critical endpoint, similar to the Korsak and Rydynski (1996) study used to
develop the 1,2,4-TMB RfC. As with 1,2,4-TMB, the SAB finds that the clarification of this choice,
however, could be greatly improved by expanding the assessment on the same points discussed for
1,2,4-TMB (see section 3.2.5)
27
-------
Charge Question: Decreased pain sensitivity (measured as an increased latency to pawlick response
after a hotplate test) in male Wistar rats was concluded by EPA to be an adverse effect on the nervous
system and was selected as the critical effect for the derivation of the RfC. Please comment on whether
the selection and characterization of this critical effect is scientifically supported and clearly described.
If a different endpoint(s) is recommended as the critical effect(s) for deriving the RfC, please identify
this effect and provide scientific support for this choice.
The SAB agrees that reduction in pain sensitivity as indicated by an increased latency to pawlick
response in a hotplate test was a valid adverse nervous system effect and appropriately selected as a
critical effect for RfC derivation of 1,2,3-TMB. The SAB notes that the agency appropriately uses the
same rationale to derive the RfC for 1,2,4-TMB and uses this information. A detailed response is in
Section 3.2.5 and the SAB refers the reader to that section rather than reiterate the response to the
Charge question.
Charge Question: In order to characterize the observed dose-response relationship comprehensively,
benchmark dose (BMD) modeling was used in conjunction with default dosimetric adjustments (U.S.
EPA, 1994b) for calculating the human equivalent concentration (HEC) to identify the point of
departure (POD) for derivation of the RfC. Please comment on whether this approach is scientifically
supportedfor the available data, and clearly described.
a. Has the modeling been appropriately conducted and clearly described, based on EPA 's
Benchmark Dose Technical Guidance (U.S. EPA, 2012)7
b. Has the choice of the benchmark response (BMR) for use in deriving the POD (i.e., a BMR
equal to a 1 standard deviation change in the control mean for the latency to pawlick response)
been supported and clearly described?
The SAB response to this charge question deals with the same issues as charge question for 1,2,4-TMB
and did not identify any issues specific to 1,2,3-TMB and refers the reader the section 3.2.5.
Charge Question: Please comment on the rationale for the selection of the uncertainty factors (UFs)
applied to the POD for the derivation of the RfC for 1,2,3-TMB. Are the UFs appropriate based on the
recommendations described in Section 4.4.5 of A Review of the Reference Dose and Reference
Concentration Process (U.S. EPA, 2002), and clearly described? If changes to the selected UFs are
proposed, please identify and provide scientific support for the proposed changes.
The SAB notes that the uncertainty factor values selected by the EPA for 1,2,3-TMB are identical to
those selected for 1,2,4-TMB, and that the justifications are essentially the same. Thus, the SAB
response to this charge question and recommendation are the same as the response to Charge Question
for 1,2,4-TMB and refers the reader to Section 3.2.5.
28
-------
3.2.7. Inhalation Reference Concentration (RfC) for 1,3,5-TMB
Charge Question: One developmental toxicity study (Saillenfait et al. 2005) following inhalation
exposure to 1,3,5-TMB was identified in the literature and was considered as a potential principal study
for the derivation of the RfC for 1,3,5-TMB. However, the candidate RfC derivedfor 1,3,5-TMB based
on this study (and the critical effect of decreased maternal weight gain) was 20-fold higher than the RfC
derivedfor 1,2,4-TMB (based on decreased pain sensitivity). Given the available toxicological database
for 1,2,4-TMB and 1,3,5-TMB, there are several important similarities in the two isomers' neurotoxicity
that support an RfC for 1,3,5-TMB that is not substantially different than the RfC derivedfor 1,2,4-TMB.
Additionally, the available toxicokinetic database for the two chemicals indicates that internal dose
metrics would be comparable. Thus, EPA concluded that deriving such disparate RfCs for these two
isomers was not scientifically supported. Rather, EPA concluded that given the similarities in
toxicokinetics and toxicity between the two isomers, there was sufficient evidence to support adopting
the RfC for 1,2,4-TMB as the RfC for 1,3,5-TMB.
Please comment on EPA 's conclusion to not base the RfC derivation for 1,3,5-TMB on isomer-specific
data. Is the scientific justification for not deriving an RfC based on the available data for 1,3,5-TMB
supported and has been clearly described?
The SAB agrees with the EPA conclusion not to base the RfC derivation for 1,3,5-TMB on isomer-
specific data. The justification for this conclusion is supported and clearly described. The SAB is not
aware of chronic or subchronic studies that could be used to support an RfC derivation for 1,3,5-TMB
with neurotoxicity as the critical endpoint, similar to the Korsak and Rydynski (1996) study used to
develop the 1,2,4-TMB RfC. The candidate RfCs for 1,3,5-TMB, based on maternal and fetal toxicity
from the study of Saillenfait et al. (2005) are presented by EPA, but were not chosen as the overall RfC.
Although the SAB takes issue with the PODs selected by EPA in their analysis of the Saillenfait et al.
study, as discussed below, it nevertheless agrees with the decision not to use this study to derive the
overall RfC for 1,3,5-TMB. The SAB concurs with EPA that the best approach under the circumstances
is to adopt the RfC for 1,2,4-TMB, based on decreased pain sensitivity, as the overall RfC for 1,3,5-
TMB.
The SAB provides the following recommendations to develop a candidate RfC based on Saillenfait et al.
(2005). This study was well-conducted and followed the appropriate European Union guidelines and
experimental methods for an inhalation developmental toxicity study (i.e., animal model and strain;
exposure chamber generation; five concentration groups; atmosphere sampling and analysis; group
sizes; maternal and fetal evaluations; and, statistical data analyses). The SAB acknowledges that the
Saillenfait study has two major limitations: (1) no neurotoxic endpoints were collected (decreased pain
sensitivity had been determined by the EPA as the critical effect for the other two TMB isomers because
it was observed following inhalation exposures in multiple rat studies); and, (2) the exposure period was
short (GD 6-15; only 10 days). Nevertheless, the SAB recommends that the EPA revise the calculations
for the fetal and maternal endpoint-based candidate RfCs.
Saillenfait et al. (2005) selected 100 ppm (492 mg/m3) for the maternal NOAEL for 1,3,5-TMB with
300 ppm (1476 mg/m3) as the maternal LOAEL based on decreased maternal weight gain and food
intake. The developmental NOAEL in the study was 300 ppm (1476 mg/m3) and the developmental
LOAEL was 600 ppm (2952 mg/m3) based on decreased mean male fetal body weights.
In the draft TMB Assessment, the EPA set the maternal NOAEL at 300 ppm (1476 mg/m3) and the
maternal LOAEL at 600 ppm (2952 mg/m3) based on decreased corrected body weight gain and higher
29
-------
exposure levels than Saillenfait et al. The SAB finds that this is not a correct interpretation of a maternal
NOAEL for the Saillenfait et al. paper. Decreased corrected body weight gain was measured only at one
time point (C-section) one day after cessation of exposure. Statistically significant decreased maternal
weights were observed at gestational days (GDs) 13-21 when the fetuses would be contributing far less
to the mother's weight and at GDs 6-21 (entire treatment period). Reduced maternal body weights
correspond exactly with the statistically significant decreased food consumption values recorded at GDs
6-13, 13-21 and 6-21 (entire treatment period). An evaluation of statistical methods used in the
Saillenfait et al. study may also be appropriate.
The SAB recommends that EPA use decreased maternal body weight gain data from GDs 6-13 and 6-21
as the basis of the maternal endpoint POD and candidate RfC rather than corrected maternal weight gain
data. If BMD modeling is unsuccessful, the SAB recommends that EPA use the maternal NOAEL of
492 mg/m3 as the POD.
Section 2.3.2 of the TMB Assessment [Methods of Analysis for 1,3,5-TMB)] incorrectly identifies 2,974
mg/m3 as the NOAEL) for the developmental endpoint (decreased male fetal body weight). The SAB
recommends using the NOAEL of 1476 mg/m3 as the POD for derivation of a developmental endpoint
RfC. The SAB also suggests that EPA consider increasing the UFd from 3 to 10, to address the lack of
neurodevelopmental testing, in the derivation of the developmental candidate RfC. The SAB notes that
this approach may not fully address neurological effects which serve as the basis for the other isomers.
However, the revised developmental endpoint RfC calculation will be based on a more appropriate POD
and improve the justification for using the extrapolation from the lower neurological-based RfC from
1,2,4-TMB.
In addition to the above analysis and considerations, the SAB noted that there are minor errors in the
description of the 1,3,5-TMB inhalation data. In Section 2.3.1. (Identification of Studies and Effects
Other Than Cancer for 1,3,5-TMB), there were errors in Table 2-12 that need to be addressed:
The female fetal body weight average for the 100 ppm (492 mg/m3) group should be 5.47 ± 0.21
and not 5.74 ± 0.21 (it is correct in other tables of the document).
The level of significance for decreased maternal body weight gain for the 600 ppm (2,952
mg/m3) group should have two (**) and not one (*) asterisk to indicate p < 0.01.
The table also states with a footnote (b) that numbers of live fetuses was not explicitly reported.
However, Saillenfait at al. (2005) did report them in Table 3 of their manuscript. The total
numbers of fetuses were 297, 314, 282, 217, and 236, for the control and exposure groups,
respectively, and should be included in Tables 2-2 and 2-12 of the draft TMB Review document.
Charge Question: Please comment on whether EPA 's approach to developing the RfC for 1,3,5-TMB is
scientifically supportedfor the available data and clearly described.
The SAB acknowledges that the agency's approach to developing the overall RfC (based on
neurological effects) for 1,3,5-TMB based on a structurally and toxicologically related isomer is
scientifically appropriate. However, the SAB recommends that the agency strengthen the justification
for using this approach for 1,3,5-TMB by: 1) following the recommendations provided above regarding
recalculating the maternal and developmental-based candidate RfCs from Sallenfait et al. (2005); and 2)
discussing the differences as well as similarities in physical and toxicological parameters (i.e., Henry's
Law constant and toxicokinetics) for 1,3,5-TMB as compared with the other isomers.
30
-------
3.2.8. Oral Reference Dose for 1,2,4-TMB
Charge Question: The oral database for 1,2,4-TMB was considered inadequate for derivation of an
RfD. However, available evidence demonstrates similar qualitative profiles of metabolism and patterns
ofparent compound distribution across exposure routes (i.e., oral and inhalation). Furthermore, there is
no evidence that would suggest the toxicity profiles would differ to a substantial degree between oral
and inhalation exposures. Therefore, route-to-route extrapolation, from inhalation to oral, using the
modified Hissink et al. (2007) PBPK model was used to derive a chronic oral RfD for 1,2,4-TMB. In
order to perform the route-to-route extrapolation, an oral component was added to the model, assuming
a constant infusion rate into the liver. Specifically, in the absence of isomer-specific information, an
assumption was made that 100% of the ingested 1,2,4-TMB would be absorbed by constant infusion of
the oral dose into the liver compartment. The contribution of first-pass metabolism was also evaluated.
Please comment on whether EPA 's conclusion that the oral database for 1,2,4-TMB is inadequate for
derivation of an RfD is scientifically supported and clearly described. Please comment on whether oral
data are available to support the derivation of an RfD for 1,2,4-TMB. If so, please identify these data.
The SAB agrees that the primary toxicological endpoints for 1,2,4-TMB (neurotoxicity, hematotoxicity)
can be extrapolated across dose routes from the inhalation data with the assistance of PBPK modeling.
There is ample precedent with IRIS assessments to use this approach to derive a reference value for a
chemical with missing data by a particular dose route.
The SAB is not aware of adequate repeat dose studies for 1,2,4-TMB via the oral dose route. The
available acute exposure studies offer limited support in developing an RfD. The SAB recognizes that
this represents a data gap. One potential way to fill this data gap is to use oral data for a closely related
TMB. There are subchronic gavage toxicology data available for 1,3,5-TMB (Koch Industries 1995;
Adenuga et al. 2014). The EPA chose not to use the Koch Industries (1995) study for derivation of a
RfD, because it did not assess the potential for neurological effects. The EPA should consider deriving
RfD(s) for endpoints developed in the Koch Industries (1995) and Adenuga et al. (2014), such as liver
and kidney weight changes, which were not seen in inhalation studies. This would be consistent with the
EPA's goal to derive RfDs for multiple endpoints. Such oral RfDs for 1,3,5-TMB could then be
considered for extrapolation to the other TMB isomers. Upon doing so, the EPA can consider the
appropriateness of applying a database uncertainty factor to the oral point of departure to compensate for
the data gap of not having an oral neurotoxicity endpoint in the current approach. This option is
commonly utilized for derivation of RfDs in these situations. By comparing the RfDs generated from the
oral studies and from the extrapolation from the RfC through using route-to-route extrapolation, the
EPA can provide a clear explanation for why the use of the PBPK route-to-route based RfD for 1,2,4-
TMB may be preferable to application of a database uncertainty factor to an oral POD.
The SAB notes there were limitations in the Koch Industries study (primarily that it didn't involve
neurotoxicity endpoints) and the study does involve an extrapolation across congeners. Presented with
those limitations, the Koch Industries study does not provide a superior alternative to the PBPK
approach for dose route extrapolation that the EPA implemented. As discussed in Section 3.2.10, the
Koch Industries study may provide a means to derive RfDs for several additional endpoints (e.g., liver,
kidney) for 1,3,5-TMB. The EPA can consider such additional RfDs and whether they are potentially
useful for 1,2,4-TMB based upon extrapolation across congeners.
Charge Question: A route-to-route extrapolation from inhalation to oral exposure using the modified
Hissink et al. (2007) PBPK model has been used to derive an oral RfD for 1,2,4-TMB. Please comment
31
-------
on whether the PBPK modeling been appropriately utilized and clearly described. Are the model
assumptions and parameters scientifically supported and clearly described? Are the uncertainties in the
model structure adequately characterized and discussed? Please comment on whether this approach is
scientifically supported and clearly described in the document.
The EPA adapted the modified Hissink et al. (2007) model for dose route extrapolation of internal dose
by adding an oral delivery component (continuous gastric infusion, instantaneous and complete
absorption). The Hissink et al. (2007) inhalation human model is a reasonable starting point as it
simulated the available human toxicokinetic data fairly well. While the incorporation of the oral dose
route is simplistic, it is acceptable for the current purposes in that the dose metric used for dose response
modeling (parent compound average weekly venous concentration) is not sensitive to peaks and valleys
of a more normal oral intake pattern. A constant infusion averages out the exposure over the course of
the day, thus creating an average venous concentration that is compatible with the dose metric without
further calculation. Overall, the modified Hissink et al. (2007) model adapted for the oral route is likely
to adequately predict human oral exposures and be useful for dose-response modeling and the derivation
of the RfD.
Charge Question: Please comment on the rationale for the selection of the uncertainty factors (UFs)
applied to the POD for the derivation of the RfD for 1,2,4-TMB. Are the UFs appropriate based on the
recommendations described in Section 4.4.5 of A Review of the Reference Dose and Reference
Concentration Processes, and clearly described? If changes to the selected UFs are proposed, please
identify and provide scientific support for the proposed changes.
The SAB agrees with the uncertainty factors selected in the development of the oral RfD for 1,2,4-TMB,
but recommends that the discussion of uncertainty be strengthened with respect to bioavailability
assumptions. As discussed in the previous response, the oral RfD for 1,2,4-TMB was derived by
incorporating an oral intake component into the PBPK model for 1,2,4-TMB to obtain a human
equivalent oral dose POD. The EPA used the same uncertainty factors for the oral RfD as were used in
the development of the inhalation RfC. Given that the oral RfD was based upon the same endpoint and
derived from the same study as the RfC, the SAB agrees that it is logical to use the same uncertainty
factors. Thus, the comments and recommendations regarding uncertainty factors are applicable to this
charge question as well (see Section 3.2.5). There was discussion regarding whether there is additional
uncertainty associated with incorporation of the oral intake component in the PBPK model, and
specifically regarding assumptions made with that component regarding oral absorption of 1,2,4-TMB
and first-pass metabolism. Unlike modeling of internal concentrations from inhalation exposure that can
be verified with existing experimental data, there are no data with which to assess model predictions of
internal doses following oral 1,2,4-TMB exposures. The SAB does not consider this additional
uncertainty sufficient to increase the composite UF for the oral RfD, largely because the nature of the
uncertainty (possible lower absorption by the oral route), would add extra health protection. The SAB
recommends that the potential uncertainties associated with oral bioavailability of 1,2,4-TMB be
discussed more clearly in the document.
3.2.9. Oral Reference Dose (RfD) for 1,2,3-TMB
Charge Question: The oral database for 1,2,3-TMB was considered to be inadequate for derivation of
an RfD. Based on the similarities in chemical properties, toxicokinetics, and toxicity profiles between
the 1,2,4-TMB and 1,2,3-TMB isomers, EPA concluded that there was sufficient evidence to support
adopting the 1,2,4-TMB RfD as the RfD for 1,2,3-TMB.
32
-------
Please comment on whether EPA 's conclusion that the oral database for 1,2,3-TMB is inadequate for
derivation of an RfD is scientifically supported and clearly described. Please comment on whether oral
data are available to support the derivation of an RfD for 1,2,3-TMB. If so, please identify these data.
Please comment on whether EPA 's approach to developing the RfD for 1,2,3-TMB is scientifically
supported and clearly described.
The SAB is not aware of adequate repeat dose studies for 1,2,3-TMB via the oral dose route. The
available acute exposure studies offer limited support in developing an RfD.
The SAB agrees that the primary toxicological endpoints used for 1,2,4-TMB (neurotoxicity,
hematotoxicity) and extrapolated across dose routes from the inhalation data with the assistance of
PBPK modeling are appropriate for 1,2,3-TMB. There is ample precedent within the IRIS system for
this approach to derive a reference value for a chemical with missing data by a particular dose route.
The SAB notes that the agency appropriately uses the same rationale to derive the RfD for 1,2,4-TMB.
A detailed response is in Section 3.2.8 and the SAB refers the reader to that section rather than reiterate
the response to the charge question.
3.2.10. Oral Reference Dose (RfD) for 1,3,5-TMB.
Charge Question: The oral database for 1,3,5-TMB was considered to be inadequate for derivation of
an RfD. EPA concluded that given the similarities in the chemical properties, toxicokinetics, and toxicity
profiles between the two isomers, there was sufficient evidence to support adopting the RfD for 1,2,4-
TMB as the RfD for 1,3,5-TMB.
Please comment on whether EPA 's conclusion that the oral database for 1,3,5-TMB is inadequate for
derivation of an RfD is scientifically supported and clearly described.
Please comment on whether oral data are available to support the derivation of an RfD for 1,3,5-TMB.
If so, please identify these data.
The SAB agrees with the EPA's approach to extrapolating the RfD of 1,2,4-TMB to 1,3,5-TMB. The
SAB is aware of an isomer specific study (Koch Industries 1995) and the recently released data on 1,3,5-
TMB (Adenuga et al. 2014) provided by public commenters.
The Koch Industries study (1995) was the only isomer-specific and route-specific study available in the
peer-reviewed literature for oral exposure to 1,3,5-TMB when the TMB Assessment was drafted in
2013. The EPA chose not to use this study for derivation of an RfD because it did not assess the
potential for neurological effects and "presented limited toxicological information." (see Appendix F of
the draft TMB Toxicological Review). Although the rationale for this decision is clearly described, the
SAB disagrees and considers the Koch Industries study suitable for development of one or more
candidate RfDs for 1,3,5-TMB. The Koch Industries study of 1,3,5-TMB toxicity after subchronic (90-
day) gavage treatment was consistent with good laboratory practices and requirements and, when
submitted for an EPA Office of Water test rule, was peer reviewed by three senior scientists (Versar
2013). Although the study does not include neurological endpoints, it does provide information on
toxicity to other organ systems such as liver and kidney. In the opinion of the SAB, this study is suitable
for providing candidate RfDs for one or more endpoints in the same way that, for example, candidate
RfC values based upon a variety of endpoints were developed and presented for 1,2,4-TMB (see Table
2-4 of the draft TMB Toxicological Review). In view of the principle of toxicological equivalence
among TMB isomers for the purposes of development of toxicity values proposed by the EPA and
33
-------
accepted by the SAB (see response to charge questions in Section 3.2.7), these candidate RfDs could
also be considered for 1,2,3- and 1,2,4-TMB as well.
Given the importance of neurotoxicity as a critical endpoint for inhalation exposure to TMB isomers,
there should be confidence that any value selected as the RfD for 1,3-5-TMB is adequately protective of
this type of effect. In order to produce an RfD protective of neurotoxicity using PODs from the Koch
Industries study, a large UFd (e.g., 10) could be used to account for the absence of isomer- and route-
specific neurotoxicity data. However, in the opinion of the SAB, there is stronger scientific support for
use of PBPK-extrapolated RfD for 1,2,4-TMB based on a neurotoxic endpoint as the overall RfD for
1,3,5-TMB. Thus, while the SAB recommends use of the Koch Industries data and Adenuga et al.
(2014) to develop candidate RfDs for comparison purposes, it agrees with the overall RfD for 1,3,5-
TMB as proposed by EPA.
3.2.11. Carcinogenicity of 1,2,4-TMB, 1,2,3-TMB, and 1,3,5-TMB
Charge Question: The draft Toxicological Review of Trimethylbenzenes did not conduct a quantitative
cancer assessment for any isomer due to the lack of available studies. Please comment on whether data
are available to support the derivation of a quantitative cancer risk estimate.
The SAB finds that the evidence for carcinogenicity of trimethylbenzenes is limited and that this fact
was well presented by the EPA in the draft toxicological review.
The carcinogenicity of 1,2,4-TMB has been assessed in a single study (Maltoni et al. 1997), in which a
single dose level was administered to rats for two years. The SAB determined that this study had a
number of shortcomings. For example, it is unclear how the dose was selected, and only one dose was
used so nothing can be said about dose-response. The dosing schedule was quite unusual and the authors
stated that a more frequent schedule (i.e., 5 or 6 days per week) would have resulted in unacceptable
toxicity. Survival was affected by treatment, but quantitative data and statistical analyses were not
presented. Body weights were collected, but the data were not reported. The only remarkable finding
from the study was neuroesthesioepitheliomas, a tumor arising from the olfactory neuroepithelium,
which occurred in treated but not control animals. No statistical analyses were presented in the paper but
a Fishers Exact test conducted by the EPA indicated that the result was not significant. Nonetheless,
these tumors are very rare in rats and it is noteworthy that in the same study ethylbenzene also induced
neuroesthesioepitheliomas. Carcinogenicity bioassays do not appear to have been conducted with 1,2,3-
TMB or 1,3,5-TMB.
Trimethylbenzenes do not appear to be genotoxic when assessed in a standard battery of genotoxicity
assays. The one exception was 1,2,3-TMB in the Ames assay in the absence of S9. The SAB concluded
that the significance of the finding was uncertain because it was not clear what mechanism could lead to
such a response.
The SAB is not aware of any human studies on carcinogenicity of TMBs, and notes that a number of
biomarker studies and their association with cancer of various sites have been published. These
biomarker studies should be reviewed and included. Some examples are:
Solid phase microextraction, mass spectrometry and metabolomic approaches for detection of
potential urinary cancer biomarkersa powerful strategy for breast cancer diagnosis. (Silva et al.
2012)
34
-------
Investigation of urinary volatile organic metabolites as potential cancer biomarkers by solid-
phase microextraction in combination with gas chromatography-mass spectrometry. (Silva et al.
2011)
Cellular responses after exposure of lung cell cultures to secondary organic aerosol particles.
(Gaschen et al. 2010)
Based upon the deficiencies of the Maltoni et al. (1997) study, the lack of bioassays with 1,2,3-TMB and
1,3,5-TMB, and the lack of human studies, the SAB agrees that the EPA could not conduct a
quantitative cancer assessment for any isomer due to the lack of appropriate studies.
3.3. Additional Recommendations
The SAB identified three additional topics not addressed directly in the Charge that warrant additional
consideration by the agency: (1) A clarification of how the EPA considers candidate toxicity values and
their intended use, (2) an expanded discussion of sensitive life stages and vulnerable populations, and
(3) deriving subchronic RfC and RfD for the TMB isomers.
3.3.1. Candidate Toxicity Values
Section 7.6 of the preamble describes how IRIS assessments derive candidate values for each suitable
data set and effect that is credibly associated with an agent. These results are arrayed, using common
dose metrics, to show where effects occur across a range of exposures using guidance on methods to
derive RfCs and RfDs. The assessment process develops an organ- or system-specific reference value
for each organ or system affected by the agent and selects an overall reference dose and an overall
reference concentration for the agent to represent lifetime human exposure levels where effects are not
anticipated to occur. Providing these organ/system-specific candidate reference values, IRIS assessments
may facilitate subsequent risk assessments that consider the combined effect of multiple agents acting at
a common site or through common mechanisms.
In considering the merits and use of candidate toxicity values, the SAB encountered an issue where
further clarification by EPA is strongly encouraged. Interest by the EPA in developing PODs and
RfCs/RfDs for multiple endpoints in new IRIS profiles is noted. As shown in this toxicological review,
one of the uses of RfCs/RfDs for various endpoints is as candidates for selection as the overall toxicity
value. The overall toxicity value is one that is intended to be protective of toxicity of all types, and this
is taken into consideration when selecting the UFd. Another use of candidate RfCs/RfDs is to better
understand the effects of combined chemical exposures. Risks from combined or cumulative exposures
to chemicals is generally of greatest concern when the chemicals affect the same targets organs. While
an overall RfC or RfD is based upon one effect chosen as the critical effect, that chemical may produce
other types of toxicity at doses that are only marginally higher than the selected overall toxicity value.
To illustrate the problem, consider the situation in which individuals are exposed to three chemicals,
each with an RfC based upon a different endpoint, but all have the potential to affect the liver. For the
risk assessor, the combined effect of the three chemicals on the liver may be greater concern than the
effects of the individual chemicals on other organ systems. In order to evaluate the risk of liver injury
from combined exposure, the risk assessor needs a liver RfC for each compound. Conceivably this
information could come from candidate RfCs for the chemicals, if available for the liver, but there is a
difference in the way that an RfC for this use would be developed versus an RfC suitable for selection as
the overall RfC. The difference is in the way that the UFd is selected - on one hand to insure that the
RfC is protective against all forms of toxicity and on the other that it is reliably protective of toxicity to a
specific target organ. Conceivably, the UFd values selected for those two purposes, and the resulting
35
-------
RfC/RfD values, could be quite different. The SAB is unaware of any discussion of this issue by EPA or
clear description of how candidate RfC/RfD values are to be developed and used. As the IRIS process
moves forward, it will be important to provide much greater clarity on this subject.
3.3.2. Susceptible Populations and Lifestages
The draft TMB Assessment provides one paragraph on this subject, spanning pages 1-54 to 1-55. It
correctly identifies various types of immaturity (metabolism, renal clearance) as potentially leading to
greater vulnerability in early life. However, this section could provide a better outline of the kinds of
information needed to understand the potential vulnerabilities in early life, including key aspects of
TMB mode of action and key developmental features.
Regarding mode of action, it is important to know:
whether it is the parent compound or metabolites (or both) that contribute to toxic effect;
which metabolic systems are responsible for removing the parent compound and creating
important metabolites; and
what is the role of distributional phenomena (e.g., uptake into brain; partitioning into fat) and
other clearance mechanisms in determining chemical fate and access to target sites.
Based upon the available MOA information, the developmental factors which may influence
toxicokinetics can be discussed. For TMBs the draft document assumes that the parent compound is
responsible for toxicity with modeling assuming that a saturable Phase I oxidative Cytochrome P450
(CYP) process is responsible for decreasing parent compound levels in venous blood. This section
should state whether it is known which CYP(s) are responsible for TMB saturable metabolism as
different CYPs have different developmental patterns. Analogy may be drawn with other alkylbenzenes
which do have toxicokinetic modeling data in early life such as toluene. Toluene has already been
referred to in the mode of action section of the document; it is also neurotoxic and its mode of action is
based upon parent compound with the level getting to the brain determined by saturable CYP
metabolism. If the EPA determines these parallels to provide a useful analogy, then early life modeling
papers for toluene by Pelekis et al. (2001) and Nong et al.( 2006) may useful for describing the degree of
toxicokinetic uncertainty presented by early life stage exposure to TMBs.
Some discussion is warranted concerning what is known about early life vulnerability to aromatic
solvent neurotoxicity. Several studies are available suggesting a vulnerable window of brain
development in mice to the neurotoxic effects of toluene (Win-Shwe et al. 2010, 2012). The USEPA
should evaluate this evidence relative to other developmental neurotoxicity studies that may be available
for toluene and other related alkylbenzenes to determine whether this data gap represents a large
uncertainty.
This section should conclude with a statement as to whether any specific data exist for TMBs that would
show the extent of early life vulnerability based upon toxicokinetic and toxicodynamic considerations
and the degree to which such data for related alkylbenzenes helps to fill these data gaps.
3.3.3. Developing Subchronic RfCs and RfDs
In addition to responding to the charge questions related to development of chronic toxicity values for
1,2,4-, 1,2,3-, and 1,3,5-TMB, the TMB Review Panel discussed using the analysis presented in the
TMB Assessment to support development of subchronic toxicity values (i.e., subchronic RfCs and oral
RfDs) for these chemicals.
36
-------
The EPA and other environmental regulatory agencies are frequently required to address the risks
associated with exposures lasting less than a lifetime. Because the toxic endpoint(s) of concern for a
given chemical, as well as threshold doses or concentrations for toxicity, can change with exposure
duration, the toxicity value used in risk assessment should be matched to the extent possible to the
length of exposure associated with the scenario of interest. Recognizing the need for toxicity values for
less-than-lifetime exposures, the EPA Risk Assessment Forum recommended that the agency develop
such values and incorporate them into the IRIS database (U.S. EPA, 2002).
In the case of the TMBs, the principal studies used to create the proposed RfCs and RfDs are all
subchronic in duration, and the analysis needed to support a robust set of subchronic toxicity values has
in effect already been done for these chemicals. The SAB acknowledges that the derivation of
subchronic RfCs and RfDs may not always be appropriate. However, the toxic endpoints and dose-
response relationships for the TMBs in the draft report are clearly relevant for subchronic exposure, and
the same PODs and the same uncertainty factors except UFs, which is used to generate a chronic
toxicity value from subchronic study data would apply to the development of a set of subchronic
RfCs and RfDs.
Given the potential usefulness of these toxicity values for risk assessment, the importance of having the
values available on IRIS, and the very small amount of additional work required to add them to the
TMB Assessment, the SAB suggests that the EPA consider including subchronic RfCs and RfDs for
1,2,4-TMB, 1,2,3-TMB, and 1,3,5-TMB. These values would be calculated using the same inputs as for
the chronic toxicity values, but omitting the UFS. The SAB anticipates that incorporation of these values
will require minimal edits to existing tables and text.
37
-------
REFERENCES
Adenuga, D., J.C. Carrillo, and R.H. McKee. 2014. The sub-chronic oral toxicity of 1,3,5-
trimethylbenzene in Sprague-Dawley rats. Regulatory Toxicology and Pharmacology 68:144-
143.
Chen, R., F. Dick, and A. Seaton. 1999. Health effects of solvent exposure among dockyard painters:
Mortality and neuropsychological symptoms. Occup Environ Med 56: 383-387.
http://dx.doi.org/10.1136/oem.56.6.383 doi: 10.1093/annhyg/meq021. Epub 2010 Mar 30.
Douglas, J.F., R.H. McKee, S.Z. Cagen, S.L. Schmitt, P.W. Beatty, M.S. Swanson, C.A. Schreiner, C.E.
Ulrich, B.Y. Cockrell. 1993. A neurotoxicity assessment of high flash aromatic naphtha. Toxicol
IndHealth9\ 1047-1058.
El Hamid Hassan A.A., S.A. El Moez Elnagar, I.M. El Tayeb, and S.A. El Halim Bolbol. 2013. Health
Hazards of Solvents Exposure among Workers in Paint Industry. Open Journal of Safety Science
and Technology 3:87-95. (http ://www.scirp.org/journal/oi sst)
Fuente A., B. McPherson, and F. Cardemil. 2013. Xylene-induced auditory dysfunction in humans. Ear
Hear. 34(5):651-60.
Fuente A., B. McPherson, and L.J. Hood. 2012. Hearing loss associated with xylene exposure in a
laboratory worker. JAm AcadAudiol. 23(10):824-30.
Gaschen A., D. Lang, M. Kalberer, M. Savi, T. Geiser, A. Gazdhar, C.M. Lehr, M. Bur, J. Dommen, U.
Baltensperger, and M. Geiser. 2010. Cellular responses after exposure of lung cell cultures to
secondary organic aerosol particles. Environ Sci Technol. 44(4): 1424-30. doi:
10.1021/es902261m.
Gong Y, R. Kishi, S. Kasai, Y. Katakura, K. Fujiwara, T. Umemura , T. Kondo, T. Sato, F. Sata, E.
Tsukishima, S Tozaki, T. Kawai, Y. Miyama. 2003. Visual dysfunction in workers exposed to a
mixture of organic solvents. Neurotoxicology. 24(4-5):703-10.
Gralewicz, S. and D. Wiaderna. 2001. Behavioral effects following subacute inhalation exposure to m-
xylene or trimethylbenzene in the rat: A comparative study. Neurotoxicology 22: 79-89.
http://dx.doi.org/10.1016/S0161-813 X( 00)00003 -6
Gralewicz, S, D. Wiaderna,T. Tomas; K. Rydzynski, K. 1997. Behavioral changes following 4-week
inhalation exposure to pseudocumene (1,2,4-trimethylbenzene) in the rat. Neurotoxicol Teratol
19: 327-333. http://dx.doi.org/10.1016/S0892-03 62(97)00001 -9
Hissink, A.M.; J. Kriise, B.M. Kulig,, M. Verwei,H. Muijser,F. Salmon, L.H. Leenheers, D.E. Owen,,
J.H. Lammers,, A.P. Freidig, R.H. McKee,. 2007. Model studies for evaluating the
neurobehavioral effects of complex hydrocarbon solvents III. PBPK modeling of white spirit
constituents as a tool for integrating animal and human test data. Neurotoxicology 28: 751-760.
http://dx.doi.org/10.1016/i.neuro.2007.03.005
Janik-Spiechowicz, E; K. Wyszynska,and E. Dziubaltowska, E. 1998. Genotoxicity evaluation of
trimethylbenzenes. Mutat Res Genet Toxicol Environ Mutagen 412: 299-305.
http://dx.doi. org/10.1016/S1383-5718(97)00202-7
Jarnberg J, G. Johanson , A. Lof, and B. Stahlbom. 1998. Toxicokinetics of 1,2,4-trimethylbenzene in
humans exposed to vapours of white spirit: comparison with exposure to 1,2,4-trimethylbenzene
alone. Arch Toxicol. 72(8):483-91.
38
-------
Jarnberg J, G. Johanson, A. Lof, and B. Stahlbom. 1997. Inhalation toxicokinetics of 1,2,4-
trimethylbenzene in volunteers: comparison between exposure to white spirit and 1,2,4-
trimethylbenzene alone. Sci Total Environ. 20;199(l-2):65-71.
Jarnberg J. and G. Johanson, A. Lof. 1996. Toxicokinetics of inhaled trimethylbenzenes in man. Toxicol
Appl Pharmacol. 140(2):281-8.
Jarnberg, J; and G. Johanson. 1999. Physiologically based modeling of 1,2,4-trimethylbenzene
inhalation toxicokinetics. Toxicol Appl Pharmacol 155: 203-214.
http://dx.doi.org/10.10Q6/taap.1998.8596
Juarez-Perez CA, A. Torres-Valenzuela, L.C. Haro-Garcia, V.H. Borja-Aburto,, and G. Aguilar-
Madrid. 2014. Ototoxicity effects of low exposure to solvent mixture among paint manufacturing
workers. Int JAudiol. 53(6):370-6.
Koch Industries (Koch Industries, Incorporated). 1995. 90-day oral gavage toxicity study of 1,3,5-
trimethylbenzene in rats with a recovery group. (44618). Wichita, KS: Koch Industries, Inc.
Korsak, Z. and K. Rydzynski. 1996. Neurotoxic effects of acute and subchronic inhalation exposure to
trimethylbenzene isomers (pseudocumene, mesitylene, hemimellitene) in rats. Int J Occup Med
Environ Health 9: 341-349.
Kostrzewski P., A. Wiaderna-Brycht, and B. Czerski. 1997. Biological monitoring of experimental
human exposure to trimethylbenzene. Sci Total Environ. 20;199(l-2):73-81.
Lee, CR; K.S. Jeong, Y. Kim, C.I. Yoo, J.H. Lee,; and Y.H. Choi, . 2005. Neurobehavioral changes of
shipyard painters exposed to mixed organic solvents. IndHealth 43: 320-326.
Lehotsky, J., V. Mezesova, P. Kaplan, and V. Mezes, 1985. The structure, function and development of
the membranes of sarcoplasmic reticulum and isolation of its subfractions. Fol. Med. Martiniana
11: 33-69.
Maltoni, C; A. Ciliberti, C. Pinto, M. Soffritti,;F. Belpoggi, and L. Menarini. 1997. Results of long-term
experimental carcinogenicity studies of the effects of gasoline, correlated fuels, and major
gasoline aromatics on rats. Ann N Y Acad Sci 837: 15-52. http://dx.doi.org/10.1111/j. 1749-
6632.1997.tb56863.
Meulenberg, C. and H. Vijverberg. 2000. Empirical relations predicting human and rat tissue: Air
partition coefficients of volatile organic compounds. Toxicol Appl Pharmacol 165: 206-216.
http://dx.doi.org/10.1006/taap.200Q.8929
Nong A., D.G. McCarver, R.N. Hines and K. Krishnan. 2006. Modeling interchild differences in
pharmacokinetics on the basis of subject-specific data on physiology and hepatic CYP2E1 levels:
A case study with toluene. Toxicology and Applied Pharmacology 214: 78-87.
NRC (National Research Council). 2011. Review of the Environmental Protection Agency's draft IRIS
assessment of formaldehyde. Washington, DC: National Academies Press.
http://vvvvvv.nap. edu/catalog/13 142. html
NRC (National Research Council). 2013. Acute Exposure Guideline Levels for Selected Airborne
Chemicals, Vol 13 (2013), Chap. 8, Trimethylbenzenes. National Academy Press, Washington
DC ISBN 978-0-309-29025-8 http://www.nap.edu/catalog.php7record id= 15852 Accessed
August 20, 2014
39
-------
Pelekis, M., L.A. Gephart, and S.E. Lerman. 2001. Physiological-model-based derivation of the adult
and child pharmacokinetic intraspecies uncertainty factors for volatile organic compounds. Regul
Toxicol Pharmacol. 33: 12-20.
Pratt, H., N. Karim., N. Bleich, and N. Mittelman. 2000. Short latency visual evoked potentials in
occupational exposure to organic solvents. Neurophysiol Clin. 30(5):306-12.
Qian, H.I., N. Fiedler, D.F. Moore, and C.P. Weisel. 2010. Occupational exposure to organic solvents
during bridge painting. Ann Occup Hyg. 54(4):417-26.
Quevedo, L. da S., T. Tochetto, M.A. Siqueira, and M.S. Machado. 2012. Auditory brainstem response
in gas station attendants. Braz JOtorhinolaryngol. 78(6):63-8.
Richie, G.D., K.R. Still, W.K. Alexander, A.F. Nordholm, C.L. Wilson, J.R. Ill, and D.R. Mattie. 2001.
A review of the neurotoxicity risk of selected hydrocarbon fuels. J Toxicol Environ Health Part
B Critical Reviews. 4(3):223-312).
Ruitjen, M W., J. Hooisma, J.T. Brons, C.E. Habets, H.H. Emmen, and H.I. Muijser. 1994.
Neurobehavioral Effects of Long-term Exposure to Xylenes and Mixed Organic Solvents in
Shipyard Spray Painters. Neurotoxicology 15(3): 613-620.
Saillenfait, AM; F. Gallissot; J.P. Sabate, and G. Morel. 2005. Developmental toxicity of two
trimethylbenzene isomers, mesitylene and pseudocumene, in rats following inhalation exposure.
Food Chem Toxicol 43: 1055-1063. http://dx.doi.oru/10.1016/i.fct.2005.02.008
Silva CL, M. Passos, and J.S. Camara . 2011. Investigation of urinary volatile organic metabolites as
potential cancer biomarkers by solid-phase microextraction in combination with gas
chromatography-mass spectrometry. Br J Cancer 105(12): 1894-904. Epub 2011 Nov 15.
Silva CL, M. Passos, and J.S. Camara. 2012. Solid phase microextraction, mass spectrometry and
metabolomic approaches for detection of potential urinary cancer biomarkersa powerful
strategy for breast cancer diagnosis. Talanta 89:360-8. doi: 10.1016/j.talanta.2011.12.041. Epub
2011 Dec 22.
Swiercz, R, D. Wiaderna, W. Wasowicz, and K. Rydzynski, K. 2003. Pseudocumene in brain, liver, lung
and blood of rats after single and repeated inhalation exposure. Int J Occup Med Environ Health
16: 61-66.
Tang, C.Y, D.M. Carpenter, E.L. Eaves, J. Ng, N. Ganeshalingam, C.P. Weisel, H. Qian, G. Lange, and
N.L. Fiedler. 2011. Occupational Solvent Exposure and Brain Function: An fMRI Study.
Environmental Health Perspectives 119 (7): 908-913.
U.S. EPA (U.S. Environmental Protection Agency). 2002. A Review of the Reference Dose and
Reference Concentration Process, EPA/630/P-02/002F
http://www.epa.gov/raf/publications/pdfs/rfd-fmal.pdf (accessed October 1, 2014)
U.S. EPA (U.S. Environmental Protection Agency). 2005. Guidelines for Carcinogen Risk Assessment.
EPA/630/P-03/001F. March 2005 Risk Assessment Forum Washington, DC
http://www.epa.gov/raf/publications/pdfs/CANCER GUIDELINES FINAL 3-25-05.PDF
(accessed September 10, 2014)
U.S. EPA (U.S. Environmental Protection Agency). 2012. Benchmark Dose Technical Guidance, Risk
Assessment Forum. EPA/100/R-12/001.June 2012. Washington, DC
http://www.epa.gov/raf/publications/pdfs/benchmark dose guidance.pdf (accessed August 1,
2014)
40
-------
U.S. EPA (U.S. Environmental Protection Agency). 2013. External Review Draft Toxicological Review
of Trimethylbenzenes (CASRN 25551-13-7, 95-63-6, 526-73-8, and 108-67-8) In Support of
Summary Information on the Integrated Risk Information System (IRIS) EPA/635/R-13/17la.
http://vosemite.epa.gov/sab/sabproduct.nsf/02ad90bl36fc21ef85256eba00436459/EElE28QE77
586DE985257B65005D37E7/$File/TMBs Draft TR IRIS HeroPublic Aug 16 2013.pdf
(accessed August 20, 2014)
Versar Inc. 2013. Peer review report: External peer review of the 1995 Koch Industries study report: 90-
day oral gavage toxicity study of 1,3,5-trimethylbenzene in rats with a recovery group. (EP-C-
12-045). Springfield, VA: Versar, Inc.
Win-Shwe TT, N. Kunugita,Y. Yoshida, D. Nakajima, S. Tsukahara, and H. Fujimaki. 2012.
Differential mRNA expression of neuroimmune markers in the hippocampus of infant mice
following toluene exposure during brain developmental period. JAppl Toxicol. 32: 126-34.
Win-Shwe TT, Y. Yoshida,N. Kunugita, S. Tsukahara, and H. Fujimaki. 2010. Does early life toluene
exposure alter the expression of NMD A receptor subunits and signal transduction pathway in
infant mouse hippocampus? Neurotoxicology 31: 647-53
Zahlsen K., I. Eide ,A.M. Nilsen, and O.G. Nilsen. 1992. Inhalation kinetics of C6 to C10 aliphatic,
aromatic and naphthenic hydrocarbons in rat after repeated exposures. Pharmacol Toxicol.
71(2): 144-9.
41
-------
APPENDIX A: CHARGE TO THE SAB
NCEA Charge to the Science Advisory Board for the
IRIS Toxicological Review of Trimethylbenzenes
August 2013 (Updated May 2014)2
Introduction
The U.S. Environmental Protection Agency (EPA) is seeking an external peer review of the scientific
basis supporting the draft Toxicological Review of Trimethylbenzenes (1,2,3-trimethylbenzene [1,2,3-
TMB], 1,2,4-trimethylbenzene [1,2,4-TMB], and 1,3,5-trimethylbenzene [1,3,5-TMB]) that will appear
on the Agency's online database, the Integrated Risk Information System (IRIS). IRIS is prepared and
maintained by the EPA's National Center for Environmental Assessment (NCEA) within the Office of
Research and Development (ORD). This is a new assessment; there is currently no entry on the IRIS
database for any isomer of trimethylbenzene.
IRIS is a human health assessment program that evaluates scientific information on effects that may
result from exposure to specific chemical substances in the environment. Through IRIS, EPA provides
high quality science-based human health assessments to support the Agency's regulatory activities and
decisions to protect public health. IRIS assessments contain information for chemical substances that
can be used to support the first two steps (hazard identification and dose-response assessment) of the
human health risk assessment process. When supported by available data, IRIS provides health effects
information and toxicity values for chronic health effects (including cancer and effects other than
cancer). Government and others combine IRIS toxicity values with exposure information to characterize
public health risks of chemical substances; this information is then used to support risk management
decisions designed to protect public health.
The external review draft Toxicological Review of Trimethylbenzenes is based on a comprehensive
review of the available scientific literature on the noncancer and cancer health effects in humans and
experimental animals exposed to 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB. This draft IRIS assessment
includes:
a Preamble to describe the methods used to develop IRIS assessments;
an Executive Summary to concisely summarize the major conclusions of the assessment;
a Literature Search Strategy and Study Selection section to describe the process for identifying and
evaluating the evidence for consideration in developing the assessment;
a Hazard Identification section to systematically synthesize and integrate the available evidence of
organ/system-specific hazards; and
a Dose-Response Analysis section to describe the selection of studies for consideration in calculating
toxicity values and to provide details of the analysis and methodology in deriving and selecting
toxicity values.
Additionally, appendices for chemical and physical properties, toxicokinetic information, summaries of
toxicity studies, and other supporting materials are provided as Supplemental Information (See
Appendix A to C) to the draft Toxicological Review. The draft assessment was developed according to
guidelines and technical reports published by EPA (see Preamble) and contains a qualitative
characterization of the hazards for TMBs, including a cancer descriptor of a chemical's human
carcinogenic potential, and noncancer toxicity values, including a chronic oral reference dose (RfD) and
2 The charge for TMBs was updated to include general charge question #4 requesting comment from the external peer review
panel on the adequacy of EPA's assessment revisions and response to the public comments. The CAAC Augmented for the
TMB Panel discussed and revised this charge question on the May 22, 2014 teleconference.
A-l
-------
a chronic inhalation reference concentration (RfC) for all three trimethylbenzene isomers. A quantitative
cancer assessment for trimethylbenzenes was not conducted due to inadequate data.
Charge Questions
In April 2011, the National Research Council (NRC) released its Review of the Environmental
Protection Agency's Draft IRIS Assessment of Formaldehyde (NRC 2011). In addition to offering
comments specifically about EPA's draft formaldehyde assessment, the NRC included comments and
recommendations for improving the development of IRIS assessments. The IRIS Program's
implementation of the NRC recommendations is following a phased approach. Phase 1 of
implementation has focused on a subset of the short-term recommendations, such as editing and
streamlining documents, increasing transparency and clarity, and using more tables, figures, and
appendices to present information and data in assessments. Phase 1 also focused on assessments that had
been near the end of the development process and close to final posting. The IRIS Program is now in
Phase 2 of implementation which addresses all of the short-term NRC recommendations. The Program
is implementing all of these recommendations but recognizes that achieving full and robust
implementation of certain recommendations will be an evolving process with input and feedback from
the public, stakeholders, and external peer review committees. This phased approach is consistent with
the NRC's Roadmap for Revision as described in Chapter 7 of the formaldehyde review report. The
NRC stated that "the committee recognizes that the changes suggested would involve a multi-year
process and extensive effort by the staff at the National Center for Environmental Assessment and input
and review by the EPA Science Advisory Board and others."
Below is a set of charge questions that address scientific issues in the draft IRIS Toxicological Review
of Trimethylbenzenes. The charge questions also seek feedback on whether the document is clear and
concise, a central concern expressed in the NRC report. Please provide detailed explanations for
responses to the charge questions. EPA will also consider the Science Advisory Board review panel's
comments on other major scientific issues specific to the hazard identification and dose-response
assessment of trimethylbenzenes. Please consider the accuracy, objectivity, and transparency of EPA's
analyses and conclusions in your review.
General Charge Questions:
1. NRC (2011) indicated that the introductory section of IRIS assessments needed to be expanded
to describe more fully the methods of the assessment. NRC stated that they were "not
recommending the addition of long descriptions of EPA guidelines to the introduction, but rather
clear, concise statements of criteria used to exclude, include, and advance studies for derivation
of [toxicity values]." Please comment on whether the new Preamble provides a clear and concise
description of the guidance and methods that EPA uses in developing IRIS assessments.
2. NRC (2011) provided comments on ways to improve the presentation of steps used to generate
IRIS assessments and indicated key outcomes at each step, including systematic review of
evidence, hazard identification, and dose-response assessment. Please comment on the new IRIS
document structure and whether it will increase the ability for assessment to be more clear,
concise and easy to follow.
3. NRC (2011) state that "all critical studies need to be thoroughly evaluated with standardized
approaches that are clearly formulated" and that "strengthened, more integrative, and more
transparent discussions of weight of evidence are needed." NRC also indicated that the changes
suggested would involve a multiyear process. Please comment on EPA's success thus far in
implementing these recommendations.
4. EPA solicited public comments on the draft IRIS assessment of trimethylbenzenes and has
A-2
-------
revised the assessment to respond to the scientific issues raised in the comments. A summary of
the public comments and EPA's responses are provided in Appendix F of the Supplemental
Information to the Toxicological Review of Trimethylbenzenes. Are there scientific issues that
were raised by the public as described in Appendix F that may not have been adequately
addressed by EPA?
Chemical-Specific Charge Questions
A. Executive Summary
1. The major conclusions of the assessment pertaining to the hazard identification and dose-
response analysis have been summarized in the Executive Summary. Please comment on the
whether the conclusions have been clearly and sufficiently described for purposes of condensing
the Toxicological Review information into a concise summary.
B. Literature Search Strategy/Study Selection
1. The process for identifying and selecting pertinent studies for consideration in developing the
assessment is detailed in the Literature Search Strategy/Study Selection section. Please comment on
the whether the literature search approach, screening, evaluation, and selection of studies for
inclusion in the assessment are clearly described and supported. Please identify any additional peer-
reviewed studies from the primary literature that should be considered in the assessment of
noncancer and cancer health effects of 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB.
C. Hazard Identification
Synthesis of Evidence
1. A synthesis of the evidence for trimethylbenzene toxicity is provided in Chapter 1, Hazard
Identification. Please comment on whether the available data have been clearly and appropriately
synthesized for each toxicological effect. Please comment on whether the weight of evidence for
hazard identification has been clearly described and scientifically supported.
Summary and Evaluation
1. Does EPA's hazard assessment of noncancer human health effects of trimethylbenzenes clearly
integrate the available scientific evidence (i.e., human, experimental animal, and mechanistic
evidence) to support the conclusions that trimethylbenzenes pose potential hazards to the nervous
system, respiratory system, the developing fetus, and the circulatory system (i.e., blood)?
2. Does EPA's hazard assessment of the carcinogenicity of trimethylbenzenes clearly integrate the
available scientific evidence to support the conclusions that under EPA's Guidelines for Carcinogen
Risk Assessment (U.S. EPA. 20051 there is "inadequate information to assess the carcinogenic
potential" of trimethylbenzenes?
D. Toxicokinetics and Pharmacokinetic Modeling
Data characterizing the toxicokinetics of 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB following inhalation
and oral exposures in humans and experimental animals supports the use of physiologically-based
pharmacokinetic (PBPK) models for 1,2,4-TMB. For the purposes of this assessment, the Hissink et al.
(2007) model, originally describing 1,2,4-TMB toxicokinetics following exposure to white spirit (a
complex mixture of volatile organic compounds), was modified by EPA to calculate internal dose
A-3
-------
metrics following exposure to 1,2,4-TMB alone for the derivation of an inhalation RfC for 1,2,4-TMB.
Additionally, the model was further modified by the addition of an oral route of exposure for use in a
route-to-route extrapolation for the derivation of an oral RfD for 1,2,4-TMB.
1. Please comment on whether the selected PBPK model (Hissink et al.. 2007) with EPA's
modifications adequately describe the toxicokinetics of 1,2,4-TMB (Appendix B). Was the PBPK
modeling appropriately utilized and clearly described? Are the model assumptions and parameters
scientifically supported and clearly described? Are the uncertainties in the model structure
adequately characterized and discussed?
2. The internal dose metric selected for use in the derivation of the RfC and RfD for 1,2,4-TMB was
the steady-state weekly average venous blood concentration (mg/L) of 1,2,4-TMB for rats exposed
for 6 h/day, 5 days/week. Please comment on whether the selection of this dose metric is
scientifically supported and clearly described. If a different dose metric is recommended for deriving
the RfC, please identify this metric and provide scientific support for this choice. Are the
uncertainties in the selected dose metric adequately characterized and discussed?
E. Inhalation Reference Concentration (RfC) for 1,2,4-TMB
1. A 90-day inhalation toxicity study of 1,2,4-TMB in male rats (Korsak and Rvdzynski. 1996) was
selected as the basis for the derivation of the RfC. Please comment on whether the selection of this
study is scientifically supported and clearly described. If a different study is recommended as the
basis for the RfC, please identify this study and provide scientific support for this choice.
2. Decreased pain sensitivity (measured as an increased latency to pawlick response after a hotplate
test) in male Wistar rats was concluded by EPA to be an adverse effect on the nervous system and
was selected as the critical effect for the derivation of the RfC. Please comment on whether the
selection and characterization of this critical effect is scientifically supported and clearly described.
If a different endpoint(s) is recommended as the critical effect(s) for deriving the RfC, please
identify this effect and provide scientific support for this choice.
3. In order to characterize the observed dose-response relationship comprehensively, benchmark dose
(BMD) modeling was used in conjunction with dosimetric adjustments for calculating the human
equivalent concentration (HEC) from a rat and human PBPK model (Hissink et al.. 2007) to identify
the point of departure (POD) for derivation of the RfC. Please comment on whether this approach is
scientifically supported for the available data, and clearly described.
a. Has the modeling been appropriately conducted and clearly described, based on EPA's
Benchmark Dose Technical Guidance (U.S. EPA. 2012)?
b. Has the choice of the benchmark response (BMR) for use in deriving the POD (i.e., a
BMR equal to 1 standard deviation change in the control mean for the latency to pawlick
response) been supported and clearly described?
4. Please comment on the rationale for the selection of the uncertainty factors (UFs) applied to the POD
for the derivation of the RfC for 1,2,4-TMB. Are the UFs appropriate based on the recommendations
described in Section 4.4.5 of A Review of the Reference Dose and Reference Concentration
Processes (U.S. EPA. 2002). and clearly described? If changes to the selected UFs are proposed,
please identify and provide scientific support for the proposed changes.
F. Inhalation Reference Concentration (RfC) for 1,2,3-TMB
1. A 90-day inhalation toxicity study of 1,2,3-TMB in male rats (Korsak and Rvdzynski. 1996) was
selected as the basis for the derivation of the RfC. Please comment on whether the selection of this
study is scientifically supported and clearly described. If a different study is recommended as the
A-4
-------
basis for the RfC, please identify this study and provide scientific support for this choice.
2. Decreased pain sensitivity (measured as an increased latency to pawlick response after a hotplate
test) in male Wistar rats was concluded by EPA to be an adverse effect on the nervous system and
was selected as the critical effect for the derivation of the RfC. Please comment on whether the
selection and characterization of this critical effect is scientifically supported and clearly described.
If a different endpoint(s) is recommended as the critical effect(s) for deriving the RfC, please
identify this effect and provide scientific support for this choice.
3. In order to characterize the observed dose-response relationship comprehensively, benchmark dose
(BMD) modeling was used in conjunction with default dosimetric adjustments (U.S. EPA. 1994b)
for calculating the human equivalent concentration (HEC) to identify the point of departure (POD)
for derivation of the RfC. Please comment on whether this approach is scientifically supported for
the available data, and clearly described.
a. Has the modeling been appropriately conducted and clearly described, based on EPA's
Benchmark Dose Technical Guidance (U.S. EPA. 2012)?
b. Has the choice of the benchmark response (BMR) for use in deriving the POD (i.e., a
BMR equal to a 1 standard deviation change in the control mean for the latency to
pawlick response) been supported and clearly described?
4. Please comment on the rationale for the selection of the uncertainty factors (UFs) applied to the POD
for the derivation of the RfC for 1,2,3-TMB. Are the UFs appropriate based on the recommendations
described in Section 4.4.5 of A Review of the Reference Dose and Reference Concentration
Processes (U.S. EPA. 2002). and clearly described? If changes to the selected UFs are proposed,
please identify and provide scientific support for the proposed changes.
G. Inhalation Reference Concentration (RfC) for 1,3,5-TMB
One developmental toxicity study (Saillenfait et al.. 2005) following inhalation exposure to 1,3,5-TMB
was identified in the literature and was considered as a potential principal study for the derivation of the
RfC for 1,3,5-TMB. However, the candidate RfC derived for 1,3,5-TMB based on this study (and the
critical effect of decreased maternal weight gain) was 20-fold higher than the RfC derived for 1,2,4-
TMB (based on decreased pain sensitivity). Given the available toxicological database for 1,2,4-TMB
and 1,3,5-TMB, there are several important similarities in the two isomers' neurotoxicity that support an
RfC for 1,3,5-TMB that is not substantially different than the RfC derived for 1,2,4-TMB. Additionally,
the available toxicokinetic database for the two chemicals indicates that internal dose metrics would be
comparable. Thus, EPA concluded that deriving such disparate RfCs for these two isomers was not
scientifically supported. Rather, EPA concluded that given the similarities in toxicokinetics and toxicity
between the two isomers, there was sufficient evidence to support adopting the RfC for 1,2,4-TMB as
the RfC for 1,3,5-TMB.
1. Please comment on EPA's conclusion to not base the RfC derivation for 1,3,5-TMB on isomer-
specific data. Is the scientific justification for not deriving an RfC based on the available data for
1,3,5-TMB supported and has it been clearly described?
2. Please comment on whether EPA's approach to developing the RfC for 1,3,5-TMB is scientifically
supported for the available data and clearly described.
A-5
-------
H. Oral Reference Dose (RfD) for 1,2,4-TMB
The oral database for 1,2,4-TMB was considered inadequate for derivation of an RfD. However,
available evidence demonstrates similar qualitative profiles of metabolism and patterns of parent
compound distribution across exposure routes (i.e., oral and inhalation). Furthermore, there is no
evidence that would suggest the toxicity profiles would differ to a substantial degree between oral and
inhalation exposures. Therefore, route-to-route extrapolation, from inhalation to oral, using the modified
Hissink et al. (2007) PBPK model was used to derive a chronic oral RfD for 1,2,4-TMB. In order to
perform the route-to-route extrapolation, an oral component was added to the model, assuming a
constant infusion rate into the liver. Specifically, in the absence of isomer-specific information, an
assumption was made that 100% of the ingested 1,2,4-TMB would be absorbed by constant infusion of
the oral dose into the liver compartment. The contribution of first-pass metabolism was also evaluated.
I. Please comment on whether EPA's conclusion that the oral database for 1,2,4-TMB is inadequate for
derivation of an RfD is scientifically supported and clearly described. Please comment on whether
oral data are available to support the derivation of an RfD for 1,2,4-TMB. If so, please identify these
data.
2. A route-to-route extrapolation from inhalation to oral exposure using the modified Hissink et al.
(2007) PBPK model has been used to derive an oral RfD for 1,2,4-TMB. Please comment on
whether the PBPK modeling been appropriately utilized and clearly described. Are the model
assumptions and parameters scientifically supported and clearly described? Are the uncertainties in
the model structure adequately characterized and discussed? Please comment on whether this
approach is scientifically supported and clearly described in the document.
3. Please comment on the rationale for the selection of the uncertainty factors (UFs) applied to the POD
for the derivation of the RfD for 1,2,4-TMB. Are the UFs appropriate based on the recommendations
described in Section 4.4.5 of A Review of the Reference Dose and Reference Concentration
Processes (U.S. EPA. 2002). and clearly described? If changes to the selected UFs are proposed,
please identify and provide scientific support for the proposed changes.
I. Oral Reference Dose (RfD) for 1,2,3-TMB
The oral database for 1,2,3-TMB was considered to be inadequate for derivation of an RfD. Based on
the similarities in chemical properties, toxicokinetics, and toxicity profiles between the 1,2,4-TMB and
1,2,3-TMB isomers, EPA concluded that there was sufficient evidence to support adopting the 1,2,4-
TMB RfD as the RfD for 1,2,3-TMB.
1. Please comment on whether EPA's conclusion that the oral database for 1,2,3-TMB is inadequate for
derivation of an RfD is scientifically supported and clearly described. Please comment on whether
oral data are available to support the derivation of an RfD for 1,2,3-TMB. If so, please identify these
data.
2. Please comment on whether EPA's approach to developing the RfD for 1,2,3-TMB is scientifically
supported and clearly described.
J. Oral Reference Dose (RfD) for 1,3,5-TMB
The oral database for 1,3,5-TMB was considered to be inadequate for derivation of an RfD. EPA
concluded that given the similarities in the chemical properties, toxicokinetics, and toxicity profiles
between the two isomers, there was sufficient evidence to support adopting the RfD for 1,2,4-TMB as
the RfD for 1,3,5-TMB.
1. Please comment on whether EPA's conclusion that the oral database for 1,3,5-TMB is inadequate for
derivation of an RfD is scientifically supported and clearly described. Please comment on whether
A-6
-------
oral data are available to support the derivation of an RfD for 1,3,5-TMB. If so, please identify these
data.
2. Please comment on whether EPA's approach to developing the RfD for 1,3,5-TMB is scientifically
supported and clearly described.
K. Carcinogenicity of 1,2,4-TMB, 1,2,3-TMB, and 1,3,5-TMB
1. The draft Toxicological Review of Trimethylbenzenes did not conduct a quantitative cancer
assessment for any isomer due to the lack of available studies. Please comment on whether data are
available to support the derivation of a quantitative cancer risk estimate.
A-7
-------
APPENDIX B: RESULTS OF REVIEW OF TRIMETHYLBENZENE PBPK MODEL
INTERNAL METRICS
Model Overview
A PBPK model for whi te spirit constituents was developed and published by TNO Quali ty of Life, The
Netherlands (Hissink et al.. 2007). This model was reviewed along with other trimethyl benzene (1,2,4-
TMB) models by the U.S. EPA (the Agency) and chosen to use for internal dose metric estimation (U.S.
EPA, 2013). In this process, a detailed computer code analysis was conducted, and generally found to be
acceptable, but some corrections were necessary.
The changes to the model code (.csl file) consisted of addressing a coding error in the supplied file (not
discussed in the manuscript) that resulted in metabolic rate changing over the course of exposure (I A /AX
= KVMAX*(ABlS(I'-T]J
1
~ ,
Figure 1. Model predicted biood concentrations for the study described in Hissink et al., 2007. Compare this figure
to B-10(b) of U.S. EPA (2013). This figure represents the fit to the final model parameters and thus replicates Figure
B-10(b).
B-1
-------
(a)
_ U
I
s
Hi) -.yr i ctirmir hi Cflft f- IV. jrl f Writ*.!) I J .* 1HI Hwii rl i . .T*Y"|
! ¦ V"; - - -
8
I »1
< t -J
L J
11 SD 1 a 3D 43 "0 6 3 7D ID 4* I'3 3L E 33 D 130 Ml 15 D I tJl
TIhmi |hr)
- ; i
L I
-¦
1/^ ¦ f
[ |
*
¦¦___
i
f-
ZD Jfl 4 0 ID ID TD IV ID W» 11D N D 130 1« D 15 D 1D0
Note: Rots eiposed to 1^,4-TMB in write spirit (WS] -{Hissirvfc- ct al.. 2007] {a) before and (d| after numerical oot Tiizatori. See
Legend, figures B-7 and B-S.
Figure B-10. Comparisons of model predictions to measured blood concentrations in
rats exposed to 1,2,4-TMB in VVS.
0.01
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600
Time (hr)
Venous 1,2,4-TMB in rats repeatedly exposed to 25, 100, and 250 ppm 1,2,4-TMB (Swiercz et al., 2003)
B-2
-------
Venous 1,2,4-TMB in rats repeatedly exposed to 25, 100, and 250 ppm 1,2,4-TMB (Swiercz et al., 2003)
100
10
1
.01
606
Time (hr)
Figure 2. Model predicted blood concentrations for the study described in Swiercz et al., 2003. Rats were exposed
to TMB 6 hr/day, 5 days/wk for 4 weeks. Blood was collected from the tail vein after the last exposure. Top) whole
timecourse, Bottom) last 6 hr. Compare this figure to B-12 of U.S. EPA (2013).
tup-HMuij iotc* jStipm i ibrttiu*! m , jwc-ji
k
¦
S-
* 1
1
f
i
^ ¦
1 ' "= ¦ 1
I-
r
I
r i
HI m » «» 4ia t-ii
(b| 1" 'T'l
Siwena et al. 110031 in rats repeated^ exposed ta l.Z.-fl-TMB: {a| before ana (b| after numerical opt niartj'on. See Legend in
Figures S-7 and B-S.
Figure B-12, Comparisons of mode] predictions to measured venous blood
concentrations by Swiercz et al. (2003] in rats repeatedly exposed to 1,2,4-TMB.
Internal blood 1,2,4-TMB metrics predicted by the model were compared to a few other studies and
consistently over-predicted the data, as reported in U.S. EPA 2013 (Tables 4 and 5).
Human Internal Dose Metrics
In the report (U.S. EPA 2013), the human exposure data of Hissink et al., 2007 was shown with the Vmax
and Km optimized to fit the rat data from the same study, and was not shown using the Vmax and Km
optimized from the Swiercz et al. (2003) rat data which was used in the final model (Table 2). Figure 3
shows the fit of that data using the VmaxC/Km used for internal dose metric determinations.
In agreement with figures B-14 and B-15, the model also under-predicts the data from Jarnberg et al.
(1998, 1997a; 1996) (Figure 4) and Kostrzewki et al, 1997 (Figure 5).
B-3
-------
White spirit exposure of humans at 45 mg/L (100 ppm) (Hissink et al., 2007)
0.16
0.15
0.14
0.13
0.12
0.11
0.1
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Figure 3. Comparisons of model predictions to measured human venous blood in human volunteers exposed to
100 ppm WS with 7.8% 1,2,4-TMB (39 mg/m3 1,2,4-TMB). The red line shows the fit when the metabolic parameters
optimized to fit rat data from the same study (Hissink et al., 2007) are used, the blue line shows the fit when the
VmaxC and Km optimized from the study of Swiercz et al., 2003 is used.
Blood concentrations of 1,2,4-TMB in volunteers exposed to 2 or 25 ppm 1,2,4-TMB (Jarnberg and coworkers)
1
0.1
0.01
0.001
0
Time (hr)
Figure 4. Comparisons of model predictions to measured human venous blood concentrations of Jarnberg et al.
(1998, 1997a; 1996) in volunteers exposed to 2 or 25 ppm (~10 or 123 mg/m3) 1,2,4-TMB for 2 hours while riding
a bicycle (50 W).
B-4
-------
Blood 1,2,4-TMB in human volunteers exposed to 154 mg/m3 1,2,4-TMB (Kostrzewski et al., 1997)
1
.1
.01
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96
Figure 5. Comparisons of model predictions to measured human venous blood concentrations in Kostrzewki et al.
(1997) in human volunteers exposed to 154 mg 1,2,4-TMB/m3 for 8 hours.
Conclusions
Agency changes are consistent with state of the art PBPK modeling and well-implemented. The Agency
version of the model consistently underpredicts compared to the Hissink parameterization (Data not
shown). The model still consistently over-predicts rat data. According to U.S. EPA 2013:
The measured Wistar rat arterial blood and tissue concentrations were consistently overpredicted by
the model, suggesting collection delays in the studies. The model also consistently overpredicted the
measured Sprague-Daw ley rat tissue and blood concentrations, including the "recovery " (12 hr post-
exposure) samples, which should not be subject to collection delays. Many of the "validation"
comparisons were made at exposure concentrations (250ppm [1,230 mg/msjor greater) for which the
optimized model did not provide accurate venous blood concentrations. It cannot be determined with
the available data whether the 2-3-fold differences between the model and Spr ague-Daw ley rat blood
concentrations at lower concentrations (75 and 150 ppm [369 and 738 mg/nuf) are due to
methodological differences (e.g., in sample collections and analysis) or true strain differences.
Overall, we conclude that the optimized model produces acceptable simulations of venous blood 1,2,4-
TMB for chronic exposure to < 100 ppm (492 mg/nn) for rats or < 30 ppm (147.6 mg/rm) for humans
1,2,4-TMB by inhalation
Because the overprediction is consistent between rodent strains and across studies, the model optimization
choices should maybe be reconsidered. An attempt was made to evaluate the model optimizations, but the
data files used to conduct those optimizations (e.g. swiercz-2003-ven-low.csv) were not found and thus
the optimizations would not run.
Conversely, the human model may be underpredicting blood concentrations. A comparison of Figure B-
16 (U.S. EPA, 2013) to the output produced in this assessment indicates that the fit to the human data of
Hissink et al 2007 matches for the elimination phase, but -25% lower peak blood concentrations are
predicted (Figure 3). Because fat content in these volunteers was measured, the study-specific fat
percentage was used, resulting in a slight additional decrease in the peak. Although holding the Km
constant and optimizing the Vmax did not result in a significant improvement to the fit to the data (U.S.
B-5
-------
EPA, 2013), since human data is available, it might be advisable to determine human-specific metabolic
rates. Three different human exposure studies were identified and blood TMB concentrations are under-
predicted post-exposure in all of them (Figures 3-5 and U.S. EPA, 2013 figures B14 and B15).
Apart from the consistent over-prediction of rat data and under-prediction of human data, this model
simulates the data overall and parameterization and implementation seem correct, although a complete
model review was not conducted.
Suggested Conventions to Facilitate PBPK Model Review
The US EPA needs to implement a rigorous and consistent approach to having their PBPK models and
approach is peer-reviewed. This peer-review should be implemented in a consistent and thorough
manner and should be conducted by an external panel, either the CAAC or some other assembled peer-
review panel. This peer-review should yield a report detailing the findings of the peer-review. The
review can follow EPA's own method for reviewing PBPK models (McLanahan et al., 2012). As the
CAAC reviews assessments that utilize PBPK models, the Agency can facilitate the panels ability to
review and confirm the uses of the PBPK model. These include:
The inclusion of an "about these files" script is excellent and highly recommended. This file is
very important and should be checked carefully. The file should include information to:
o Describe generated figures (publication and figure #1)
o Dosing and parameters.
o other pertinent information.
Over-arching setup files should be included. Parameters set in individual .m files should be
discouraged to assure a unified parameterization is in place.
o Because files may not be run in order, each file must setup all parameters through the use
of standardized setup files and must either contain the data needed to produce figures or
must call a central data file.
Files should be put organized in a logical progression. Suggested order might be:
o Setup files for difference species/conditions
o Rodent studies via a route
o Rodent studies via alternate routes...
o Human studies
o Simulations
All files should be annotated
o Especially note changes or different from standardize approaches
o Should indicate which, if any figures they reproduce from EPA reports and/or manuscripts,
o Data source should be identified (Digitized from figure, supplied by author...)
Files should show the model mass-balance
B-6
-------
References
Brown, R.P., Delp, M.D., Lindstedt, S.L., Rhomberg, L.R., Beliles, R.P., 1997. Physiological parameter
values for physiologically based pharmacokinetic models. Toxicol. Ind. Health 13, 407-84.
Hissink, A.M., Kriise, J., Kulig, B.M., Verwei, M., Muijser, H., Salmon, F., Leenheers, L.H., Owen, D.E.,
Lammers, J.H.C.M., Freidig, A.P., McKee, R.H., 2007. Model studies for evaluating the
neurobehavioral effects of complex hydrocarbon solvents III. PBPK modeling of white spirit
constituents as a tool for integrating animal and human test data. Neurotoxicology 28, 751-60.
U.S. EPA (2013). Toxicological Review of Trimethylbenzenes (CASRN 25551-13-7, 95-63-6, 526-73-8,
andl08-67-8). In Support of Summary Information on the Integrated Risk Information System
(IRIS). Supplemental Information. EPA/635/R-13/171b Revised External Review Draft
www.epa.gov/iris
B-7
-------
Table 1. Comparison of Rat Model Input Parameters
Parameter
< I lissink el nl.
rransmilled lo
rransmilled lo
21107)
\\
Summit
Commenls
Partitioning
Saline:Air
3
QC by EPA, as reported in Hissink et al
Olive oil:Air
13200
QC by EPA, as reported in Hissink et al
Blood: Air - rat
148
QC by EPA, as reported in Hissink et al
Rapidly perfused:Blood
2.53
QC by EPA, as reported in Hissink et al
Slowly perfused:Blood
1.21
QC by EPA, as reported in Hissink et al
Fat:Blood
62.7
QC by EPA, as reported in Hissink et al
Brain:Blood
2.53
QC by EPA, as reported in Hissink et al
Liver:Blood
2.53
QC by EPA, as reported in Hissink et al
Anatomical and Physiological
Alveolar ventilation rate (L/hr/kg0 7)
20
14*
(Brown et al., 1997)
Total cardiac output (L/hr/kg0 7)
20
14*
(Brown et al., 1997)
Blood flow (% cardiac output)
Liver (total)
25
17.6
(Brown et al., 1997)
Fat
9
Brain
1.2
2
(Brown et al., 1997)
Rapidly perfused (total)
49.8
76*
57.4§
(Brown et al., 1997)
Slowly perfused (total)
15
NA
Calculated
Tissue volume (% body weight)
Liver
4
Fat
7
(Brown et al., 1997)
Brain
0.72
0.57
(Brown et al., 1997)
Rapidly perfused
4.28
NA
9§
(Brown et al., 1997)
Slowly perfused
75
NA
82§
(Brown et al., 1997)
Metabolism
VmaxC (mg/hr/kg0.7)
3.5
4.17
Hissink et al visibly optimized: US EPA
Km (mg/L)
0.25
0.322
used ACSL.x to numerically optimize. Also
used Swiercz et al. (2003) inhalation data to
optimize.
B-8
-------
* Within EPA version of model code, this is raised to the 0.74 power, not 0.7. QP = QPC*BW**0.74, QC = QCC*BW**0.74. Since this is
generally thought of as a "body surface area" correction, either is acceptable, the use of a different power is noted in footnote of table B-13. In
addition, the EPA version of the model uses two different QPC values to correct for alveolar volume (QC/QC2=0.7).
I I parameter is the same as reported in Hissink et al., 2007.
§In the final EPA version of the model, values for total rapid flow and volume (QRTOTC,VRTOTC) and for total slow volume (VSTOTC), are
used to calculate blood flow to rapidly perfused tissues (designated Rich within the .csl) and slow compartment volumes and flows. For example,
QR = QRTOTC*QC-QL-QBR. Where QC is total cardiac output, QL and QBR are liver and brain flows, respectively. The EPA did this to
correct mass-balance issues. Therefore, a direct comparison cannot be made to the values from Hissink et al.
₯ According to USEPA 2013, this should have been 9%
* The way in which total rapid compartment is presented in the updated version of the model, it is unclear what this value represents here. It may
be a calculation performed by the EPA to approximate the initial value.
NA - Because the way in which total rapid and slow compartments are presented in the updated version of the model, these values would not be
used in the model and were not provided to Summit for review.
B-9
-------
Table 2. Comparison of Human Model Input Parameters
Parameter
(1 lissink el ill . 2<)')7)
Transmitted lo Summit
Comments
Partitioning
Saline:Air
3
QC by EPA, as reported in Hissink et al
Olive oil:Air
13200
QC by EPA, as reported in Hissink et al
Blood: Air - human
85
QC by EPA, as reported in Hissink et al
Rapidly perfused:Blood
2.53
QC by EPA, as reported in Hissink et al
Slowly perfused:Blood
2.11
QC by EPA, as reported in Hissink et al
Fat:Blood
62.7
QC by EPA, as reported in Hissink et al
Brain:Blood
2.53
QC by EPA, as reported in Hissink et al
Liver:Blood
2.53
QC by EPA, as reported in Hissink et al
Anatomical and Physiological
Alveolar ventilation rate (L/hr/kg0 7)
20
15*
(Brown et al., 1997)
Total cardiac output (L/hr/kg0 7)
20
16*
(Brown et al., 1997)
Blood flow (% cardiac output)
Liver (total)
26
17.5
(Brown et al., 1997)
Fat
5
8.5
(Brown et al., 1997)
Brain
14
11.4
(Brown et al., 1997)
Rapidly perfused (total)
30
66.6
Slowly perfused (total)
25
Calculated§
Tissue volume (% body weight)
Liver
2.6
Fat
14.6
21.4
Hissink et al., 2007, were describing the
specific population from their study -
average body fat (measured using calipers
was 14.6%.
Brain
2
Rapidly perfused
3
7.6
(Brown et al., 1997)
Slowly perfused
66.4
81§
(Brown et al., 1997)
VmaxC (mg/hr/kg0.7)
3.5
4.17
Scaled from rat Optimization
Km (mg/L)
0.25
0.322
Scaled from rat Optimization
B-10
-------
* Within EPA version of model code, this is raised to the 0.74 power, not 0.7. QP = QPC*BW**0.74, QC = QCC*BW**0.74. Since this is
generally thought of as a "body surface area" correction, either is acceptable, the use of a different power is noted in footnote of table B-13. In
addition, the EPA version of the model uses two different QPC values to correct for alveolar volume (QP/QP2=0.7).
I I Parameter is the same as reported in Hissink et al., 2007
§In all versions of the model, values for total rapid flow and volume (QRTOTC,VRTOTC) and for total slow volume (VSTOTC), are used to
calculate blood flow to rapidly perfused tissues (designated Rich within the .csl) and slow compartment volumes and flows. For example, QR =
QRTOTC*QC-QL-QBR. Where QC is total cardiac output, QL and QBR are liver and brain flows, respectively. The EPA added a mass-balance
equation (QSTOTC=l-QRTOTC) to correct mass-balance issues. Therefore, a direct comparison cannot be made to the values from Hissink et
al. for Flows to the slow compartment.
B-ll
-------
Table 3. Study-reported Cmax compared to predicted Cmax
Exposure
Concentration
(mg/1)
Data AVG*
Model Prediction
Model
Prediction/Data
Hissink et al. 2007 (8 hr)
0.047
0.16 ±0.010
0.27
1.7
0.19
0.81 ±NA
1.2
1.5
0.37
4.0 ±0.70
3.7
0.93
Swiercz et al. 2003
0.12
0.56
0.55
0.98
0.49
4.1
4.7
1.1
1.23
14
21.0
1.5
Comparison of model-predicted Blood 1,2,4-TMB to study-specific data. For Hissink et al. 2007, data is
at the end of the 8 hr exposure, for Swiercz et al., 2003 data is first collected on the last day of repeated
exposures. For .* ± SD when available.
Table 4. Model simulated and experimental measured concentrations of 1,2,4-TMB in male Sprague-
DAWLEY RATS EXPOSED TO 1,2,4-TMB AT THE END OF 12 HOUR EXPOSURE (ZAHLSEN, 1996).: TABLE B-11 FROM
U.S. EPA 2013
Exposure
Concentration
(mg/1)
Experiment
(mg/L)
Model Prediction
Model
Prediction/Data
0.37
1.7
4.2
2.5
0.74
6.9
18
2.6
1.5
14
48
3.5
B-12
-------
Table 4. Model simulated and experimental measured concentrations of 1,2,4-TMB in male Sprague-
DAWLEY RATS EXPOSED TO 1,000 PPM (4,920 MG/M3) 1,2,4-TMB (12 HR/DAY, FOR 14 DAYS) AT THE END OF
exposure: Table B-12 from U.S. EPA 2013
Day
Experiment
(mg/L)
Model Prediction
Model
Prediction/Data
1
63.5
181
2.8
3
43.1
293
6.8
7
33.4
372
11.1
10
34.0
395
11.6
14
35.2
399
11.3
Non-Cancer endpoint dose-response modeling for 1 ,2,4-TMB:Korsak et al., 2000
1
US EPA 2013
Average mg/1
2
Model Average
mg/1
3
Hissink Model
Average mg/1
4
Hissink/Model
Average mg/1
Low
0.1339
0.13
0.16
1.2
Mid
0.8671
0.87
1.9
2.2
High
5.248
5.4
12.2
2.3
Column 1 is the data taken from U.S. EPA, 2013 Table C-l (Korsak et al., 2000a). Column 2 are the
weekly average blood concentrations produced using average exposures and body weights from that study
in this assessment. Column 3 shows the same assessment using the rat parameters from Hissink et al, 2017
(Table 1). Column 4 shows the difference between the Hissink and U.S. EPA, 2013 parameterization.
Human internal metric comparison after continuous inhalation exposure: venous TMB
CONCENTRATION (SS)
Exposure Concentration
(mg/m3)
Model mg/1
Hissink Model
mg/1
Hissink/Model
mg/1
16
0.09
0.10
1.1
24.5
0.13
0.15
1.1
84
0.50
0.62
1.2
134
0.89
1.4
1.6
B-13
------- |