Human Health Risk Assessment
Research Strategy
EXTERNAL REVIEW DRAFT
FEBRUARY 1998
Research and
Data Collection
Research
Needs
/ Risk
Assessment
Risk
Character-
Ization
Risk
Management
Decisions
OSS?
U.S. Environmental Protection Agency
Office of Research and Development
J3TIC QUALITY INSPECTED X
February 1998
Draft—Do Not Quote or Cite
-------
/*
INTERNET DOCUMENT INFORMATION FORM
A . Report Title: Human Health Risk Assessment Research Strategy
B. DATE Report Downloaded From the Internet 6/8/98
Report's Point of Contact: (Name, Organization, Address, Office
Symbol, & Ph #): U.S. Environmental Protection Agency
Office of Research Development
D. Currently Applicable Classification Level: Unclassified
E. Distribution Statement A: Approved for Public Release
F. The foregoing information was compiled and provided by:
DTIC-OCA, Initials: V\ HA Preparation Date: Id
~O
The foregoing information should exactly correspond to the Title, Report Number, and the Date on
the accompanying report document If there are mismatches, or other questions, contact the
above OCA Representative for resolution.
19980617 155
-------
Disclaimer
This document is a draft for external review purposes only and does not constitute
U.S. Environmental Protection Agency policy. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
February 1998
ii
Draft—Do Not Quote or Cite
-------
Table of Contents
Page
List of Tables v
List of Figures vi
Authors and Contributors vii
Abbreviations and Acronyms viii
Executive Summary E-l
1. Introduction 1-1
1.1 Purpose: Achieving a Focused Research Agenda 1-1
1.2 Scope of the Research Problem 1-3
1.2.1 Reducing Uncertainties in Human Exposure Measurements and
Models 1-6
1.2.2 Developing and Applying Mechanistic Models and Data To Reduce
Uncertainties in Hazard Characterization and Dose-Response Assessment ... 1-7
1.2.3 Characterizing Variability in Human Exposure and Susceptibility
to Disease 1-7
1.3 Coordination with the Broader Environmental Research Community 1-8
1.4 Structure of This Document 1-9
2. Human Exposure Research 2-1
2.1 Background .2-1
2.2 Strategic Directions for Research To Reduce Uncertainties in Exposure-Dose
Measurements and Models 2-6
2.2.1 Problem Statement 2-6
2.2.2 Scientific Questions 2-8
2.2.3 Research Approaches, Products, and Uses 2-8
3. Dose and Effects Research 3-1
3.1 Background 3-1
3.1.1 Current Office of Research and Development Research 3-3
3.2 Dose Estimation Research 3-8
3.2.1 Uncertainties in Mechanistic Data for Hazard Characterization
and Dose-Response Assessment 3-8
3.2.1.1 Problem Statement 3-8
3.2.1.2 Scientific Questions 3-8
3.2.1.3 Research Approach, Products, and Uses 3-11
3.3 Effects Research 3-12
3.3.1 Problem Statement 3-12
3.3.2 Scientific Questions 3-12
3.3.3 Research Approach, Products, and Uses 3-12
3.3.3,1 Development of More Selective and Valid Tests for
Mechanistically Based Hazard Characterization 3-13
February 1998 iii Draft—Do Not Quote or Cite
-------
Table of Contents
(cont'd)
Page
3.3.3.2 Enhance Empirical Approaches for Dose-Response Assessment ... 3-14
3.3.3.3 Focus on Receptor-Mediated Mechanisms. 3-14
3.3.3.4 Focus on Health Effects Associated with Less-Than-Lifetime
Exposures 3-15
3.4 Characterizing and Assessing Variation in Human Susceptibility to Disease 3-16
3.4.1 Problem Statement 3-16
3.4.2 Scientific Questions 3-16
3.4.3 Research Approach, Products, and Uses 3-16
4. Risk Assessment and Characterization Research 4-1
4.1 Background 4-1
4.2 Strategic Directions 4-2
4.2.1 Problem Statement 4-2
4.2.2 Risk Assessment Questions 4-4
4.2.3 Risk Assessment Approach, Products, and Uses 4-5
4.2.3.1 Biological Measures of Exposure and Their Relationship
to Human Activity Patterns, Media, and Pathways 4-5
4.2.3.2 Use of Biological Information in Risk Assessment 4-5
4.2.3.3 Variation in Human Susceptibility 4-7
5. Science Directions for Human Health Risk Assessment Research 5-1
6. References 6-1
Appendix A: Recommendations for Strengthening Human Health Risk Assessment
in the U.S. Environmental Protection Agency A-l
Appendix B: Office of Research and Development Research Strategies, Priorities,
and Plans B-l
Appendix C: The Impact of Legislation and Regulation on Human Health Risk
Assessment Research C-l
February 1998
iv
Draft—Do Not Quote or Cite
-------
List of Tables
Number Page
2-1 Overview of Current Human Exposure Research Sponsored by the
Office of Research and Development 2-4
2-2 Scientific Constraints and Uncertainties on Exposure and Risk Assessment
in the U.S. Environmental Protection Agency 2-7
2-3 Future Approaches and Products for Human Exposure Research Sponsored
by the Office of Research and Development 2-9
3-1 Overview of the Current Dose Estimation and Health Effects Research
Program 3-4
3-2 Future Directions in Dose Estimation and Human Health Effects Research
To Improve Human Health Risk Assessment 3-9
4-1 Overview of the Office of Research and Development's Current Health Risk
Assessment Research Program 4-3
4-2 Future Research Directions To Improve Human Health Risk Assessment ........ 4-6
5-1 Summary of Priorities in Human Health Research 5-2
B-l Office of Research and Development Research Plans and Strategies B-9
February 1998
v
Draft—Do Not Quote or Cite
-------
List of Figures
Number Page
1-1 Relationship between core and problem-driven components of the Office
of Research and Development's human health research program 1-2
1-2 The elements of human health risk assessment 1-4
2-1 Scientific elements involved in human exposure research and exposure
assessment 2-2
3-1 Scientific elements in dose estimation research 3-2
4-1 Scientific elements in risk assessment and characterization research 4-2
B-l Implementing the Office of Research and Development's strategic plan B-3
B-2 Criteria for setting research priorities B-6
February 1998
vi
Draft—Do Not Quote or Cite
-------
Authors and Contributors
Principal Authors
Vicki Dellarco,* Senior Health Scientist
Office of Science and Technology, Office of Water
Washington, DC
Herman Gibb, Assistant Director for Multimedia Research
National Center for Environmental Assessment
Washington, DC
Suzanne McMaster, Assistant Director for Pesticides and Toxics Research
National Health and Environmental Effects Research Laboratory
Research Triangle Park, NC
Jennifer Orme Zavaleta, Assistant Director for Multimedia Research
National Health and Environmental Effects Research Laboratory
Research Triangle Park, NC
Dale Pahl, Assistant Director for Multimedia Research
National Exposure Research Laboratory
Research Triangle Park, NC
John Vandenberg, Assistant Director for Air Research
National Health and Environmental Effects Research Laboratory
Research Triangle Park, NC
Michael Waters, Assistant Director for Waste and International Programs
National Health and Environmental Effects Research Laboratory
Research Triangle Park, NC
Contributing Authors
Jeannctte Wiltse*
Health and Ecological Criteria Division
Office of Science and Technology, Office of Water
Washington, DC
Harold Zenick, Associate Director for Health
National Health and Environmental Effects Research Laboratory
Research Triangle Park, NC
~Formerly with ORD's National Center for Environmental Assessment, Washington, DC.
February 1998
vii
Draft—Do Not Quote or Cite
-------
List of Abbreviations and Acronyms
ATSDR
Agency for Toxic Substances and Disease Registry
CDC
Centers for Disease Control and Prevention
COPD
Chronic obstructive pulmonary disease
DBP
Disinfection by-product
DOE
Department of Energy
EPA
U.S. Environmental Protection Agency
FDA
Food and Drag Administation
FIFRA
Federal Insecticide, Fungicide, and Rodenticide Act
MCL
Maximum contaminant level
NAAQS
National Ambient Air Quality Standards
NAFTA
North American Free Trade Agreement
NCI
National Cancer Institute
NCTR
National Center for Toxicological Research
NHANES
National Health and Nutrition Assessment Surveys
NHAPS
National Human Activity Pattern Survey
NHEXAS
National Human Exposure Assessment Survey
NffiHS
National Institute of Environmental Health Sciences
NIOSH
National Institute for Occupational Safety and Health
NIST
National Institute for Standards and Technology
NOAA
National Oceanic and Atmospheric Administration
NRC
National Research Council
ORD
Office of Research and Development
PAH
Polycyclic aromatic hydrocarbon
PBPK
Physiologically based pharmacokinetic
PCB
Polychlorinated biphenyl
PM
Particulate matter
RCT
Research Coordination Team
SAB
Science Advisory Board
SAR
Structure-activity relationship
February 1998 viii Draft—Do Not Quote or Cite
-------
List of Abbreviations and Acronyms
(cont'd)
STAR Science to Achieve Results
TEF Toxic equivalency factor
THERdbASE Total Human Exposure Research Database and Advanced Simulation
Environment
TSCA Toxic Substances Control Act
VOC Volatile organic compound
February 1998
ix
Draft—Do Not Quote or Cite
-------
Executive Summary
Background
This document describes the direction in which the Office of Research and Development
(ORD) human health risk assessment research program is expected to evolve over the next
several years. The ORD research planning process involves a series of steps designed to identify,
verify, and document research priorities. This research strategy represents a step in this process;
it is both an elaboration of the description of ORD's human health risk assessment research
program contained in the ORD Strategic Plan and an outline for development of the more
specific laboratory/center implementation plans.
This document describes ORD's human health research program that addresses key
uncertainties in human health risk assessment. This research strategy is an attempt to build
consensus for a focused, integrated research agenda that will strengthen the scientific foundation
for future risk assessments.
Strategic Research Directions
Based on an evaluation of the needs of the U.S. Environmental Protection Agency's
(EPA's) regulatory and regional programs and consideration of recommendations made by
external advisory groups, three key strategic objectives have been identified for core human
health risk assessment research. These objectives, as listed below, will provide direction and
focus for ORD human health risk assessment research for the next 5 to 10 years:
(1) Reducing uncertainties in exposure measurements and measurement-derived models,
(2) Applying mechanistic information (to reduce uncertainties) in hazard characterization and
dose-response assessment, and
(3) Characterizing and assessing variation in human exposure and susceptibility to disease
Research directions for each of these objectives are provided, along with explanation of the
process used to prioritize the objectives. Discussion is presented in the context of ORD's
organization along the lines of the risk assessment paradigm.
Anticipated Results
Focusing human health risk assessment research on the strategic objectives identified in
this document will lead to the development of specific research products identified for each
research objective. The potential applications of these results are discussed within the document
in terms of products and anticipated uses. In addition, the impacts that the overall research
program and its individual components are expected to have on the quality of human health risk
assessments are identified and discussed.
February 1998
E-1
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
X
Introduction
1.1 Purpose: Achieving a Focused Research Agenda
The purpose of this research strategy is to present current and future directions for ORD's
core research program in human health risk assessment. This research strategy represents the
second step of a three-step research planning process. In the first step, ORD established, and
published in the 1997 Update to ORD's Strategic Plan (U.S. Environmental Protection Agency,
1997a), strategic research planning principles, ranking criteria, and six high-priority research
areas that will receive special, expanded attention within the broad program of research it
supports.
This document represents the second step. Essentially, this document expands the
description of the core program (see box below) in human health risk assessment beyond the
brief summary provided in ORD's Strategic Plan. During this second step, ORD will solicit and
incorporate inputs from the broad EPA community (both scientists and policy makers) and the
external scientific community on the most appropriate long-term research directions that will
improve the scientific foundation for the conduct and interpretation of health-related problem-
directed research (See research plans/strategies for these problem-directed areas in particulate
matter, microbes/disinfection by product, endocrine disrupters, arsenic). In the final step of the
research planning process, this document will be used by ORD's laboratories and centers to
prepare detailed research project plans.
Thus, this document is both an elaboration of the core research program in human health
risk assessment described in the ORD strategic plan (U.S. Environmental Protection Agency,
1997a) and a goal-oriented outline for the development of a more detailed laboratory/center
implementation plans. The critical question that this document addresses is What are the
appropriate strategic directions for this core research program that will develop the
fundamental methods, databases, and measurements to strengthen the scientific foundation for
health risk assessments across EPA ? The relationship between the core and problem-driven
components of ORD's human health research program is illustrated in Figure 1-1.
February 1998 1 -1 Draft—Do Not Quote or Cite
-------
In focusing this document on strategic directions for a core research program in health risk
assessment, ORD is adopting a recommendation of the National Research Council's Committee on
Research Opportunities and Priorities for EPA. "Core research should seek better understanding of
fundamental phenomena and generate broadly applicable research tools and information. These
goals will not vary much over time, and thus core research priorities will stay relatively constant."
Core research should include three basic objectives: "(1) Acquisition of systematic understanding
about underlying environmental processes..(2) Development of broadly applicable research
tools, including better techniques for measuring physical, chemical, biological, social, and economic
variables of interest; more accurate models of complex systems and their interactions; and new
methods for analyzing, displaying, and using environmental information for science-based decision
making; (3) Design, implementation, and maintenance of appropriate environmental monitoring
programs, with evaluation, analysis, synthesis, and dissemination of the data and results to improve
understanding of the status of and changes in environmental resources over time and to confirm that
environmental policies are having the desired effect" (National Research Council, 1997).
PM
Endocrine
Disrupters
Arsenic
Air
Toxics
Human Health
Risk
Assessment
Research
Microbes/
i"""^ tr—i P*"*
DBPs
Ozone
Pesticides/'
Toxics
O Core Research
O Problem-Driven
Research
Figure 1-1. Relationship between core and problem-driven components of ORD's human
health research program.
February 1998
1-2
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
This strategy is not intended to be a technical document. Rather, it is targeted to an
audience of senior scientific advisors, environmental policy and decision makers, and anyone
with a strong interest in establishing research priorities and directions to strengthen the scientific
foundation for EPA decision making.
1.2 Scope of the Research Problem
Human health risk assessment is a process that characterizes the potential adverse health
effects resulting from exposure to environmental hazards (National Research Council, 1983).
In 1983, the National Research Council described four primary steps of risk assessment that are
qualitative or quantitative in nature. They are: (1) hazard identification, (2) dose-response
assessment, (3) exposure assessment, and (4) risk characterization (Figure 1-2). Hazard
identification describes the likelihood that an environmental agent can produce an adverse effect
in humans under certain exposure conditions. Dose-response assessments quantitatively estimate
the relationship between exposure and the health effect. Elements of exposure assessment
include the identification and quantification of the population exposed, important routes of
exposure, and estimations of magnitude, duration, and frequency of contact between an
environmental agent and humans. The last step, risk characterization, integrates this information
into a qualitative or quantitative estimate of the likelihood that a hazard posed by exposure to the
agent would pose a human health risk (National Research Council, 1994). A risk
characterization describes the assumptions and uncertainties associated with the risk estimate.
Assumptions and uncertainties exist because of a lack of knowledge about the biological,
chemical, and physical processes within and between exposure and effect. It may not be possible
or practical to study the causal relationship for all the different health outcomes resulting from
numerous exposure scenarios. Thus, use of assumptions and defaults becomes necessary in
characterizing risk. Research that targets key assumptions can improve the scientific
underpinning of the resulting risk assessment by reducing the inherent uncertainties.
In recent years, advances in the state of environmental science have illustrated that new risk
assessment methods are needed to investigate complex environmental and human health issues
that were not contemplated in early environmental legislation. These advances illustrate the
importance of new risk management options for EPA, replacing, where appropriate, the "one-
size-fits-all" approach to risk management with a more population-specific approach where
February 1998 1 -3 Draft—Do Not Quote or Cite
-------
Risk
Characterization
Hazard
Identification
Dose-Response
Assessment
Internal
Dose
Risk
Characterization
Effect(s)
Emission
Source(s)
Environmental
Concentrations
Is the
environmental
agent capable of
causing an
adverse effect In
humans?
What is the
relationship
between dose to
the target tissue
and adverse
effects in
humans?
What
environmental
exposures occur
or are expected to
occur for human
populations, and
what is the
resulting dose to
the target tissue?
What is the
estimated
human
health risk
from
anticipated
exposures?
Figure 1-2. Hie elements of human health risk assessment.
1 risk management options are developed for infants and children, susceptible subpopulations, or
2 the general population (see text box below).
Emerging Emphases in Human Health Risk Assessment
and Management
Historic Approach
Emerging Emphases
General population
Sensitive subpopulation
Single source
Multiple sources
Single pollutant
Multiple pollutants
Single pathway
Multiple pathways
Single endpoint
Multiple endpoints ,
Central decision making
Community decision making
Command and control
Flexibility in achieving goals
Single stressor risk reduction
Holistic risk reduction
February 1998
1-4
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
In recognition of these changes, EPA-wide guidance recently was provided "to take into
account cumulative risk issues in scoping and planning major risk assessments and to consider a
broader scope that integrates multiple sources, effects, pathways, stressors, and populations for
cumulative risk analyses ..." (U.S. Environmental Protection Agency, 1997b),
The need for additional research in human health risk assessment is both urgent and
compelling (see Appendix A and the text box below). During the past 10 years, a number of
national scientific advisory groups have identified significant deficiencies in EPA-wide risk
assessment practices. These advisory groups also have developed specific recommendations to
assist EPA in identifying critical scientific issues that must be remedied to strengthen human
health risk assessment across EPA. However, the scope and number of scientific uncertainties
that need to be addressed with research is substantial and disproportionately large in comparison
to current EPA resources. In the words of the National Research Council, "Because EPA's task
of protecting the environment and human health is so vast and difficult, and because resources to
undertake the necessary research are very limited, choices will have to be made among many
worthwhile projects" (National Research Council, 1997).
"In the absence of reliable risk assessment, enormous sums of money that might be better spent
elsewhere may be allocated to dealing with perceived risks. While it is essential to ensure public health
and environmental integrity, limited resources reinforce the need to assess risks as accurately as
possible Estimates have indicated that the cost of environmental regulations in the United States
will total between $171 and $185 billion by the year 2000 (Carlin et al., 1991). Compliance with air
pollution control regulations will cost an estimated $94 billion per year by the year 2000 (Carlin et al., 1991).
Russell et al.(1991) estimated that cleaning up all the major hazardous waste sites would cost between
$500 billion and $1 trillion over the next 30 years. The sums are enormous, and a convincing analysis must
be provided to demonstrate that these expenditures are justified as the most cost-effective way to reduce
risks to human health and to the environment" (National Research Council, 1997).
After considering recommendations from extramural advisory groups, as well as from
senior scientists from across ORD and EPA's program and regional offices, ORD has identified
three strategic directions for its core human health risk assessment research during the next
several years. When adopted, these strategic directions will focus future ORD research in three
areas that would have the broadest applicability for improving the scientific foundation for EPA
risk assessments (see Appendix B):
(1) reducing uncertainties in human exposure measurements and models;
February 1998 1-5 Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
(2) applying mechanistic models and data to reduce uncertainty in hazard characterization and
dose-response assessment; and
(3) characterizing variability in human exposure and susceptibility to disease.
The implications of these research problems for EPA health risk assessments are described
briefly in the following sections and explored in more detail in Chapters 2 through 5 of this
document.
1.2.1 Reducing Uncertainties in Human Exposure Measurements and Models
Risk assessors rarely have actual exposure information to assess environmental risks and
usually are dependent on a variety of models and assumptions. In the rare case where actual
exposure measurements have been made, there may remain a considerable lack of knowledge
about the internal dose to humans. Frequently, human exposure is multichemical and
multipathway in nature, but historic approaches to regulation have tended to focus on a single
chemical and a single exposure pathway. Examples include evaluation of dietary exposure to a
specific pesticide or outdoor inhalation exposure to VOC's.
There are many gaps in the knowledge of human exposure to environmental pollutants.
Currently, because of lack of data, risk assessment default assumptions are made that there are no
significant differences in time-activity patterns as a function of age, gender, socioeconomic
status, or ethnic origin; and there are no significant differences in time-activity patterns of the
population in relation to regional variability or rural, urban, or suburban place of residence.
In reality, the amount of time spent in different microenvironments can vary significantly over a
lifetime and can have a large impact on both the actual exposure and the risk assessment.
The pattern and frequency of exposure also affect the type of health effects produced.
Short-term exposures of intense magnitude result in a different pattern of target tissue insult than
does the same total dose delivered over a longer time period. Also, short-term exposures can
occur at critical times during growth and development with far greater effect than if the
exposures were to occur at other times.
February 1998
1-6
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
1.2.2 Developing and Applying Mechanistic Models and Data To Reduce Uncertainties
in Hazard Characterization and Dose-Response Assessment
Risk assessment often involves the extrapolation from observations obtained at exposures
orders of magnitude greater than the environmental exposure for which estimates of risk are
being made, as well as from test animals to humans. The uncertainties in such extrapolations are
considerable and represent major problems facing the risk assessor.
Extrapolation from animal data to estimate human risks involves a variety of assumptions
about interspecies differences between animals and humans.
Extrapolation from high to low dose from either animal or human data requires
assumptions about the potential high-to-low dose difference in the shape of the dose-response
curve. For carcinogens, EPA has taken the default approach that, in the absence of biological
information to the contrary, a linear low-dose approach to risk estimation is to be used, despite
recognition that the actual risk could be between the estimated risk and zero. For noncancer
risks, EPA uses uncertainty factors to establish a dose below which adverse effects are not
expected to occur. These estimates are generally conservative and are made with little if any
knowledge of whether biological effects actually occur at such low doses. Research to
investigate factors that affect the shape of the response curve at low doses will greatly improve
both hazard characterization and dose-response assessment.
1.2.3 Characterizing Variability in Human Exposure and Susceptibility to Disease
The significance of variation in human susceptibility to disease has been recognized for
many years. Similar variation is known to exist in response to environmental toxicants and may
be related to factors such as age, preexisting disease, lifestyle, genetic background, gender and
ethnicity (or some combination of these). For example, the developing nervous system of a child
is especially sensitive to lead exposure and young children have behaviors (e.g., eating paint
chips, hand-to-mouth activities) that increase their exposures to lead. Thus, adequately
protecting children from the risks of lead (or other susceptible subpopulations from other
chemicals) requires a fuller understanding of these factors. Such variation must be addressed to
develop improved exposure, health, and risk assessments. However, currently available
approaches are often crude (e.g., assuming a 10-fold uncertainty factor for susceptibility in
February 1998
1-7
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
noncancer health assessments) or rarely used because of intense data requirements (e.g., exposure
assessments of a specific vulnerable subpopulation such as children exposed to ozone).
As is obvious from this discussion, there is an immense set of possible combinations of risk
factors and chemicals, effectively preventing a direct measurement of each set. The only effect
approach is to carefully prioritize potential scenarios of high concern and conduct research to
understand the fundamental principles. Such information can serve as the basis of models
between measured and unmeasured scenarios or the basis of the design of problem-directed
research. For example, a more complete understanding of activity patterns of children would
allow estimation of factors that result in increased chemical contact and dose. Such historical
knowledge led to concerns about children exercising outdoors when ozone levels are high and
drove ozone-specific research to enable a quantitative assessment. As another example, a core
goal is to identify the mechanisms of sensitivity of children to pesticides and to quantify the
activity patterns of children. This information enables the design of separate problem-driven
research on what specific pesticides children are most susceptible to and what activity patters
increase their exposure to those specific pesticides. Even with the design and conduct of more
studies on this issue, risk assessment models will still need to make assumptions. Hence, this
core research on susceptibility must provide principles that can be translated to improved risk
assessment models. This need was also recognized in the Food Quality Protection Act which
required a protective factor for children.
1.3 Coordination with the Broader Environmental Research Community
The ORD has been a federal leader in human health risk assessment research for the past
15 years and sustains an in-house scientific capability in all the elements of human health risk
assessment research. ORD scientists have fostered research coordination and collaboration in
health risk assessment with their peers in other federal and state agencies (e.g., National Institute
of Environmental Health Sciences [NIEHS], Centers for Disease Control and Prevention [CDC],
Food and Drag Administration [FDA], National Oceanic and Atmospheric Administration
[NOAA], National Cancer Institute [NCI], Agency for Toxic Substances and Disease Registry
[ATSDR], NCTR, National Institute for Occupational Safety and Health [NIOSH], and
Department of Energy [DOE] laboratories, and states, including California, Texas, and New
Jersey), as well as in academic and private research organizations. In addition to peer
February 1998 1 -8 Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
collaboration, a major portion of ORD's human health risk assessment research program has
been sustained through cooperative agreements, grants, and interagency agreement with these
organizations. Moreover, ORD has established a number of formal agreements with several of
these agencies to sustain and improve current research coordination.
It is essential that future ORD research in human health risk assessment continue and
expand on current interagency research collaboration and formal research agreements to ensure
the broadest possible leverage of expertise to this complex research area. This is particularly
important for the resource-intensive elements of risk assessment research (e.g, human exposure
field studies) where current staffing levels are very limited. Currently, ORD's interagency
coordination and collaboration in these areas is quite strong (see, for example, the text box about
the National Human Exposure Assessment Survey [NHEXAS] at the end of Chapter 2).
1.4 Structure of This Document
The initial portion of this document includes an executive summary and introduction. The
main body of the document includes three chapters that explain the strategic directions for future
health risk assessment research and the research approaches and scientific contributions that
ORD expects will result from these strategic directions. The sixth chapter discusses the
improvements in the science of human health risk assessment that will result from these strategic
directions. The final chapter contains the references cited in preceding chapters, followed by
Appendixes A through D.
Within the main body of the document, the information presented in Chapters 2 through
5 begins with a background section, which describes the scientific elements of each component
of the risk assessment paradigm and examples of current research supported by ORD.
Subsequent sections of each chapter discuss the strategic directions for future research for each
research area. This is accompanied by a discussion of the principal scientific problems or areas
of uncertainty; the scientific questions that must be addressed to resolve the problems; and the
research approach and scientific contributions (or products) that will respond to the questions as
well as the contributions that this research will make to strengthen the scientific foundation for
risk assessment.
February 1998
1-9
Draft—Do Not Quote or Cite
-------
Summary of Document Structure
Executive Summary
Chapter 1: Introduction and Identification of Broad Strategic Directions for Core Research
Chapters 2 through 4:
Research Area
• Background information
- The scientific elements of the research area
- Examples of current research in the area supported by ORD
* Strategic directions for future research
- Principal scientific problems
- Scientific questions or areas of uncertainty
- Research approaches and scientific contributions or products
Chapter 5; Implications for reducing uncertainty in risk assessment
Chapters: References
Appendixes
February 1998
1-10
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
_2_
Human Exposure Research
2.1 Background
Figure 2-1 presents a conceptual diagram of the scientific elements involved in human
exposure research. This figure illustrates the relationships among sources of environmental
contamination, transport and transformation, environmental characterization, and human
exposure and dose. Source characterization and source-attribution research involve
quantifying, in time and space, emission source characteristics in such a fashion that source-
receptor relationships can be developed for single or multiple environmental contaminants.
Transport and transformation research involves quantifying physical transport processes (from
source to receptor), physical and chemical transformations, and biological processes.
Environmental characterization research focuses on the physical structure of an environment
and on determining ambient levels of chemical or biological contaminants in that environment.
In the human exposure context, environments of concern include settings where short- or long-
term exposures may be of concern (e.g, occupational, residential, and commuting environments).
Time-activity pattern research develops temporal profiles of those environments in which
humans are exposed to environmental contaminants during their daily activities, the duration of
those exposures, and the human activities or behaviors that may affect the exposure.
Conceptually, human exposure research investigates the magnitude, duration, and
frequency of contact between an environmental contaminant (or biological agent) and the human
body (National Research Council, 1991; Duan and Ott, 1989).1 Total human exposure
'A quantitative definition of exposure is more complex than this qualitative description implies. For
example, an air pollution scientist may characterize human exposure as the magnitude and duration of the
atmospheric contaminants at the interface with the human breathing zone. From the perspective of a health scientist,
the concept of human exposure to atmospheric contaminants may refer to an aerosol within the lung at the interface
between airway and alveoli that, because of interactions within the body, may possess a different chemical
composition from that of the aerosol before it was inhaled. A different type of complexity is introduced when
considering human exposure from multiple environmental pathways. For example, when considering an infant's
exposure to lead inhaled from motor vehicle exhaust and ingested through dermal-oral or pica activities, calculating
the resulting exposure requires that the pathway-specific exposures be expressed in comparable terms. In summary,
a mathematical definition of human exposure depends critically on where the human-environmental boundary is
located and on whether single-pathway or mutipathway exposures are being investigated.
February 1998
2-1
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
«Chemical
«Microbe
* Radiation
Transport, Transformation,
and Source-Attribution
Models
«Concentration
• Pathway
«Phase Distribution
* Morphology
Human Exposure
Models
Magnitude •
Duration •
Frequency •
Route •
. 'Frequency
- • Individual
J * Community
, • Population
• Sirbpopulation
Exposure
Distribution
• Ventilation Rata
• Ingestion Rate
Exposure-Dose Models
Internal
Dose
Environmental
ConeentraSons
Risk
Characterization
Human
Exposure
Risk
Management
Emission Source
Environmental
Characterization
Transport
Dose
Figure 2-1, Scientific elements involved in human exposure research and exposure
assessment.
integrates all relevant routes of exposure to a specific coruaminant(s). For example, people are
exposed to lead via inhalation, food, water, and hand-to-mouth behavior and evaluation of one
route only would result in erroneous exposure assessment and ineffective risk management. This
example also illustrates the importance of time-activity pattern research (e.g., what is the
relationship between the hand-to-mouth activity of a young child and lead exposure). Even when
total human exposure is known, dose must be understood to put the influence of different
pathways into perspective. For example, suppose food concentrations of a chemical are high, but
little is eaten, and little is absorbed compared to low concentrations of the same chemical in air
with a high rate of absorption into the body from the lungs.
The scope of ORD's current human exposure research includes projects that seek to
measure, evaluate, and model exposure-dose relationships illustrated in Figure 2-1 and to begin
to link this knowledge to source and fate research (described elsewhere) with the ultimate goal of
source-to-dose modeling.
February 1998
2-2
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
ORD human exposure research in this area has been developed over the past 15 years
through collaboration between ORD scientists conducting human exposure, environmental
health, and risk assessment research with their peers in other federal agencies (e.g., NIEHS,
CDC, FDA, NCI, N1CHD, ATSDR, NIOSH, DOE) and in academic and other research
organizations. Although ORD scientists participate in establishing strategic directions for EPA
research, they are not responsible solely for conducting the research to accomplish the strategic
goals. A major portion of ORD's human exposure research has been supported through
cooperative agreement and grant assistance mechanisms. The focus of ORD's current human
exposure research responds to the following four scientific questions.
(1) What methods are needed to measure multipathway human exposure and to develop
estimates of total exposure?
(2) What are the statistically representative time-activity patterns that affect
microenvironmental exposure at different scales (e.g., population, subpopulaiion, national,
regional)?
(3) What protocols are needed to develop measurement-based population distributions of
multipathway human exposure and to communicate the results of these studies?
(4) What models and systems are needed to mathematically represent microenvironmental and
population distributions of human exposure?
Current human exposure research sponsored by ORD (in cooperation with grantees from
academic and private research institutions, partnerships with other federal and state agencies, and
scientists in its laboratories and centers) is summarized in Table 2-1. Current human exposure
measurement research includes projects to develop and evaluate; (1) statistical and analytical
chemistry measurement methods, (2) microenvironmental (including residential) exposure
measurement databases, (3) pilot studies to develop population-scale multimedia exposure
protocols, and (4) time-activity pattern databases. Current human exposure modeling research
focuses on developing microenvironmental models and a framework for total human exposure
modeling. Microenvironmental models are designed to predict single- or multipathway human
exposure to contaminants in specific (e.g., residential, commuting, occupational) environments.
Total human exposure models are designed to predict multipathway human exposure and the
frequency distribution of exposures for a population or subpopulation, either from a probabilistic
sample of human exposure and activity pattern measurements or from the integration of
February 1998
2-3
-------
CD
CT
—*
C
w
—*
•<
CO
CD
00
ro
¦
a
a
T
D
o
z
o
»¦ ~
D
c
0
1 ~
CD
O
O
CD
Table 2-1. Overview of Current Human Exposure Research Sponsored by OF
ID
Scientific Questions
and Research Focus
Research Approach
Research Products
Future Emphasis
What protocols are needed
to develop measurement-
derived exposure databases
at different scales?
Develop, demonstrate, and evaluate protocols for single and
multipathway exposure measurement studies. ORD research
focuses on development of protocols for community-scale and
regional-scale population distributions. Examples include
NHEXAS, the Pesticide Residential Exposure Research
Guidelines, and the investigation of pesticide exposures in
children.
Reduced uncertainty in quantifying
population distributions of human
exposure; guidelines for environmental
health and human exposure investigations
by EPA offices, states, and industry
Guidelines for human exposure and
environmental health investigations for
EPA offices and states
Future emphasis will be on
multipathway protocols for
exposure surveillance.
Future research activities are
anticipated to be sustained at
current levels.
Develop, demonstrate, and evaluate protocols for single and
multimedia residential exposure studies that incorporate source-
pathway-exposure investigations.
Reduced uncertainty in characterizing
residential exposures and the relationships
between indoor and outdoor sources
Future research activity is
anticipated to increase.
What methods are needed
to measure multipathway
human exposure and total
exposure?
Develop, demonstrate, and evaluate methods for measuring
dermal, oral, and dietary exposure. Improving the accuracy
of exposure estimates for infants and children is one focus
of ORD research in this area.
Reduced uncertainty in estimating
multipathway and total human exposure for
infants and children
In general, future research
activity is expected to decrease
for methods development.
Develop, demonstrate, and evaluate single- and multimedia
methods for measuring mixtures and phase-distributed
compounds. ORD research focuses on aerosols, metals, VOCs,
semivolatile organic compounds, and microbiological
contaminants.
Reduced uncertainty in measuring mixtures
and interpreting multipathway exposures
for these compounds
In general, future research
activity is expected to decrease
for methods development.
Develop, demonstrate, and evaluate low-cost indicators of
multimedia exposure. ORD research focuses on immunoassay,
biosensor, and blood-breath techniques.
Low-cost measurement methods and near-
real-time sensor technology
Future research in development
of low-cost methods will be
sustained at current levels.
What are the statistically
representative time-activity
patterns that affect exposure
at different scales?
Develop, demonstrate, and evaluate statistical instruments
for identifying time-activity patterns for populations and
population subgroups. ORD research focuses on children
and farmworkers.
Reduced uncertainty in human exposure
models that incorporate time-activity
pattern data
Future research activity is
anticipated to be sustained
at current levels.
What are the statistically
representative time-activity
patterns that affect exposure
at different scales?
Research to develop, demonstrate, and evaluate time-activity
pattern data and a database system to incorporate both exposure
measurement and time-activity pattern data. ORD research
focuses on NHAPS and on THERdbASE.
Reduced uncertainty in time-activity
pattern profiles; databases for construction
of time-activity pattern profiles, simulation
of exposure distributions, and evaluation of
exposure mitigation and risk management
options
Future research activity in
this area is expected to be
sustained at current levels.
-------
Table 2-1 (cont'd). Overview of Current Human Exposure Research Sponsored by ORD
Scientific Questions
and Research Focus
Research Approach
Research Outputs
Future Emphasis
What models and
systems are needed to
mathematically represent
microenvironmental and
population distributions
of human exposure?
Research to develop, demonstrate, and evaluate measurement-
based models that represent personal and microenvironmental
exposures, exposure-source relationships, and the physical
and chemical factors that affect exposure magnitude,
duration, frequency, and variability. ORD research focuses
on developing models that can reduce uncertainty in risk
assessment. Exposure measurement data from NHEXAS and
pesticide exposure studies will be used in this research.
Reduced uncertainty in both
microenvironmental models and models
based on population distributions of
exposure
Development of prospective and
retrospective exposure models that are
evaluated with measurement data
Reduced uncertainty in risk assessment
models
Future research activity in this
area is expected to increase and to
focus on both single and
multipathway models that
represent exposure in different
microen vironments.
Research to develop, demonstrate, and evaluate measurement-
based models that represent exposure-biomarker-dose
relationships and the physical and chemical factors that affect
potential and absorbed dose.
Reduced uncertainty in exposure-PBPK
models
Future research activity in this
area is expected to increase.
Research to develop, demonstrate, and evaluate measurement
databases for baseline comparisons to interpret exposure data
and exposure mitigation options with study participants, their
communities, and governments.
Baseline data for interpreting future
exposure studies at community to regional
scales
New approaches for design and
interpretation of exposure and biomarker
databases
Future research activity in this
area is expected to increase.
What are the important
biomarkers of exposure
and effect?
Research to develop and evaluate biomarkers of exposure and
effect to priority pollutants and for multiple endpoints,
including cancer, respiratory toxicity, neurotoxicity,
immunotoxicity, and developmental and reproductive
toxicity. ORD research is focused on DNA adducts of
products of incomplete combustion, PAHs, and drinking
water disinfection by-products (DBPs); biochemical markers
for model neurotoxicants; and cellular markers for
reproductive toxicants, dioxins, and PCBs.
Improved exposure and dose-response
assessment: qualitative and quantitative
characterization of target tissue exposure
Research activities to address this
objective are anticipated to
increase.
o
Z
o
r+
D
c
o
r+
CD
O
o
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
microenvironmental models and time-activity pattern data to predict daily exposure profiles or
population exposure distributions.
2.2 Strategic Directions for Research To Reduce Uncertainties in
Exposure-Dose Measurements and Models
2,2.1 Problem Statement
In 1995, EPA's Science Advisory Board (SAB) completed a report that reviewed the state
of exposure assessment science, identified constraints on exposure and risk assessment within
EPA, and formulated recommendations for strengthening the scientific foundation for exposure
and risk assessment through future research (U.S. Environmental Protection Agency, 1995a).
Significant concerns about the lack of exposure measurements, databases, and models across
EPA are prominent among the findings and recommendations in this report. The implications for
exposure and risk assessment posed by these deficiencies are summarized in Table 2-2. The
SAB report acknowledged the capability and relevance of ORD's current research for addressing
these agency-wide problems. However, it also concluded that a substantial and long-term future
research effort to improve exposure measurements and to develop exposure databases and
models would be required to remedy these scientific deficiencies. Relevant findings from this
and other national advisory panels are summarized in Appendix A.
In addition to this SAB study, the National Research Council (NRC; 1994) completed a
report on science in risk assessment that made wide-ranging recommendations to improve EPA's
risk assessment procedures. The report identifies the need for research into variability in human
exposure and the extent to which this contributes to variability in susceptibility to disease
prominently among its recommendations because of the substantial scientific uncertainty in this
area. Variability and susceptibility are related also to age, lifestyle, genetic background, gender,
and ethnicity (see also Chapter 3)—at individual-to-population-scales. The NRC panel
concluded that the amount of variation could have a significant effect on current estimates of
individual exposure and risk and, depending on the homogeneity of the population from which
exposure and risk are determined, on the estimation of population risk as well.
This chapter integrates discussions about exposure research for the general population and
susceptible subpopulations because susceptibility (from the exposure perspective) is investigated
February 1998
2-6
Draft—Do Not Quote or Cite
-------
Table 2-2. Scientific Constraints and Uncertainties on Exposure and Risk Assessment
in EPA (U.S. Environmental Protection Agency, 1995a)
Scientific Components of
Exposure Assessment
Examples of Constraints and Scientific Uncertainties
Environmental and
1.
There are virtually no measurement studies or protocols that characterize multipathway exposures
Exposure Measurements
either at microenvironmental or population scales.
2.
EPA typically measures pollutant emissions without determining actual human exposures or biological
markers of exposure and effect.
3.
Although EPA supports costly ambient monitoring networks to implement regulations that protect the
public or environmental health, these networks do not measure exposure or biological markers of
exposure and effect.
4.
When EPA conducts exposure and risk assessments, sources of emissions and dispersion models
typically are used in place of actual exposure data. Despite evidence that determining less-than-
lifetime exposures is essential to defining relationships between acute exposure, dose, and response,
EPA's assumptions about emission sources and their associated ambient concentrations fix them as
constants during a 70-year human lifetime.
5.
When EPA conducts environmental measurement studies for screening or exposure assessment, the
studies rarely investigate the multiple environmental pathways that are essential for a scientifically
valid estimate of total human exposure.
6.
Typically, EPA's exposure and risk assessments are conducted on a pollutant-by-pollutant basis
without regard to the nature of pollutants during actual exposures.
7.
Despite evidence that people spend 50 to 80% of their time in residential environments, EPA exposure
and risk assessments typically assume that residential exposures are equivalent to outdoor ambient
concentrations of pollutants and are not affected by either the building or indoor sources of pollution.
8,
Methods of adequate sensitivity and accuracy that are inexpensive enough for broad use in multimedia
exposure measurements are often not available.
Exposure Modeling,
1.
There are virtually no databases of human time-activity pattern data at regional, population, or
Databases, Time-Activity
subpopulation scales.
Patterns, and Susceptible
2.
EPA exposure and risk assessments typically assume that an individual's time-activity patterns are
Subpopuiations
invariant over a lifetime.
3.
EPA exposure and risk assessments typically assume no difference in time-activity patterns across a
population as a function of region, residential location (urban versus rural), gender, age.
socioeconomic status, or ethnic origin.
4.
EPA exposure and risk assessments typically da not identify characteristics of susceptible
subpopuiations (including time-activity pattern behavior or acute exposure information) related to
elevated exposures or effects.
5.
EPA exposure and risk assessments typically ignore residential time-activity pattern data related to
indoor and residential exposures.
6,
There are virtually no measurement-derived databases of multipathway human exposure.
7.
EPA exposure and risk assessment models rarely, if ever, are validated with actual human exposure
measurements.
8,
EPA exposure and risk assessments assume statistical distributions of population exposures that are
not validated and do not include information about highly exposed individuals or susceptible
subpopuiations.
9.
There are no protocols for communicating exposure, risk, and mitigation information to residents in
communities or regions.
10.
There are virtually no multipathway human exposure models that represent relationships between
exposure and dose.
11.
There are virtually no multipathway human exposure models that represent prospective or
retrospective relationships between pollutant sources, pathways, environmental concentrations,
exposures, and dose.
12.
EPA rarely achieves the integration of models and measurements required by the scientific method for
investigation of actual human exposures,
Pollutant- or Media-Specific
1,
The distribution of exposure to common pollutants such as particulate matter (PM), microbes, DBFs
Issues
pesticides, and other toxics in susceptible subpopuiations is not known.
2.
Whether populations are exposed to sufficient concentrations of endocrine disrupters to cause adverse
effects cannot be estimated.
3.
In some significant instances (e.g., microbes, pollutants in drinking water, pesticides, PM), adequate
analytic methods do not exist,
February 1998
2-7
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
through studies of time-activity pattern-exposure-dose. Thus, one typieal research project serves
both.
2.2.2 Scientific Questions
The scientific uncertainties posed by these limitations in exposure and risk assessment can be
represented within a framework for future research that is composed of the following three
fundamental and related scientific questions.
(1) What are the pathway-specific measures of human exposure for contaminants of concern?
(2) What are the behavioral and time-activity determinants of human exposure for populations
and susceptible subpopulations?
(3) What are the mathematical relationships among contaminant sources, environmental fate
processes, pathway-specific environmental concentrations of contaminants, total human
exposure, and dose for average and susceptible subpopulations?
This framework for future research acknowledges the importance of direct measures of
exposure, activity pattern data, and biological indicators of exposure and of the integration of
measurements and modeling. Creating these measurements requires the development,
evaluation, and application of appropriate methods. In addition, research to develop and apply
statistical techniques and time-activity questionnaires that represent the distribution of exposures
across subpopulations (e.g., infants, children) is essential for the development of scientifically
valid models of exposure and dose.
2.2.3 Research Approaches, Products, and Uses
The three scientific questions highlighted in the previous section provide the strategic
framework to define future research approaches and products required to improve the scientific
foundation for exposure and risk assessment. These future research approaches, products, and
outcomes are summarized in Table 2-3 and described briefly in the discussion in the rest of this
section. As Table 2-3 indicates, ORD will direct its human exposure research program to
respond to the most critical deficiencies and constraints in EPA-wide exposure and risk
assessment practices. This will be accomplished by increasing ORD's research emphasis on
(1) developing, demonstrating, and evaluating protocols for measurements of actual human
exposure; (2) developing human time-activity pattern data and on interpreting and extending this
February 1998
2-8
Draft—Do Not Quote or Cite
-------
Tl
CD
cr
Table 2-3. Future Approaches and Products for
Human Exposure Research Sponsored by ORD
Scientific Uncertainties/Future Focus
Research Approach
Research Products
Risk Assessment
What are the pathway-specific measures
of human exposure for contaminants of
concern?
Develop and demonstrate and evaluate
methods that reduce uncertainty or costs
in measuring multipathway exposures.
Develop, demonstrate, and evaluate
protocols and databases to characterize
microenvironmental exposure.
Develop and demonstrate protocols and
databases for measuring, and
communicating the results of population
distributions for multipathway and total
exposure.
Validated methods for measuring
pathway-specific and multipathway
exposure.
Validated "next generation", low-cost
methods for measuring human exposure.
Validated residential and
microenvironmental measurement
protocols and databases.
Validated protocols for determining
and communicating the result of
multipathway population exposure
distributions at community-to-regional
scales.
New exposure methods to reduce
uncertainty in determining multimedia
and total exposures.
Enhanced ability to design and conduct
exposure measurement studies.
New exposure measurement databases to
evaluate existing exposure models, to
determine the reasonableness of
exposure assessments, and to develop
more accurate models.
Enhanced ability to communicate results
of exposure studies.
What are the behavioral and time-activity
determinants of human exposure for
populations and susceptible
subpopulations?
Develop, demonstrate, and evaluate a
national human activity pattern database.
Determine relationships between time-
activity patterns and exposures at various
scales.
Investigate relationships between
exposure and factors that may affect
susceptibility.
Apply time-activity pattern data to
investigate exposures for population
subgroups that may have increased
susceptibility.
A national human activity pattern
database.
Linked activity pattern/exposure
modeling databases that permit statistical
analysis of relationships between time-
activity patterns and exposure.
An evaluation of time-activity pattern
and exposure data to identify susceptible
subpopulations.
Enhanced ability to apply and interpret
statistically representative time-activity
pattern profiles.
Reduced uncertainty in models,
exposure assessments, and risk
assessments that rely on time-activity
data.
Enhanced ability to characterize
subpopulations and to identify
differential exposures for
subpopulations and regions where
variability plays a significant role for
exposure and risk assessment.
What are the mathematical relationships
among contaminant sources,
environmental fate, pathway-specific
environmental concentrations of
contaminants, and total human exposure?
Develop, demonstrate, and evaluate
microenvironmental exposure models.
Develop, demonstrate, and evaluate
multipathway and total human exposure
models.
Develop, demonstrate, and evaluate
models that represent prospective and
retrospective relationships among
sources, pathway-specific environmental
concentrations, total exposure, and dose.
Validated models that represent
multipathway residential and
microenvironmental exposure.
Validated models that represent
population distributions of total human
exposure at community-to-regional
scales.
Validated prospective and retrospective
exposure models.
New measurement-derived exposure
models to reduce uncertainty in
exposure and risk assessments.
Enhanced ability to apply exposure
models to investigate designs for future
exposure measurement studies for
microenvironments, and populations at
community-to-regional scales.
Reduced uncertainty in characterizing
exposure-dose relationships and in risk
assessments.
0
1 ¦
CD
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
data to increase the understanding of exposure; and (3) developing exposure models and on
evaluating these models with measurement-derived databases.
Future exposure research shall emphasize developing and evaluating protocols and
databases of exposure measurements for the general population and for susceptible
subpopulations. Despite the importance of direct measurements of exposure, current exposure
assessments in single- and muliple-media continue to be hampered by a significant lack of
exposure measurement databases, A survey of exposure-related databases in the United States
(Sexton et al., 1992) has identified only a relatively small number that report actual measures of
exposure or dose and virtually none that collect measures of exposure across all relevant
environmental pathways.
Although there is a large body of environmental and occupational measurement data for
airborne pollutants (especially for the criteria air pollutants, volatile organic compounds (VOCs),
and some inorganic constituents of aerosols such as acidity and sulfates), few exposure databases
exist to characterize airborne or multimedia human exposure in residential environments (where
humans spend the majority of their time) or the relative residential/ambient outdoor contributions
to these exposures. This is because research has shown that the air pathway alone may not be the
most important route of exposure for some aerosol constituents (such as polycyclic aromatic
hydrocarbons [PAHs]) and that the personal aerosol cloud in the human breathing zone contains
contaminants that did not originate from conventional stationary air pollutant sources. Future
multimedia exposure measurement studies are needed also to characterize exposure to other
semivolatile organic compounds, particularly pesticides.
In some cases, protocols for conducting future human exposure measurement studies of
subpopulations will evolve from current research sponsored by ORD in partnership with other
federal agencies and internationally recognized academic leaders. For example, ORD pilot
studies to evaluate protocols for residential exposure measurements, population-scale exposure
measurements, and exposure communication and mitigation procedures currently are being
developed or evaluated.
Research has clearly demonstrated that the persons who are most at risk are members of
susceptible subpopulations (e.g., the elderly, the infirm, the poor, the very young, those who
engage in frequent strenuous physical activity, those who are highly exposed occupationally).
In the context of residential exposure, infants and children may represent one of the largest and
February 1998
2-10
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
most susceptible subpopulations, both in terms of their potential for exposure to environmental
contaminants and the likelihood of adverse responses to these exposures. Their behavioral
patterns may result in greater exposures to contaminants in the environments where they live and
play; their small body size may increase their dose, and their developing organ systems may put
them at greater risk from these exposures relative to adults. For example, infants and children
may consume greater amounts of some foods and may ingest greater amounts of some
contaminants (from dermal-oral mouthing of lead or pesticide residue in household dust or of
lead from soil) than do older children and adults. Thus, measurements (particularly for metals
and persistent organic pollutants) shall be made in food, water, and other beverages; indoor,
outdoor, and other microenvironmental air; and interior and exterior dust and soil. Activity
pattern determinations, dermal-oral patterns of activity and ingestion, and biomarker
measurements shall be made to allow calculation of potential contaminant dose and actual dose.
Studies now are being completed that will furnish survey, sampling, analysis, and interpretation
methods for childrens' total exposure to several organic pollutants, including PAH, pesticides
such as DDT, chlordane, chlorpyrifos, and 2,4-D; phthalate esters; phenols, especially bisphenol-
A (a potential endocrine disrupter); and polychlorinated biphenols (PCBs). Children are also
likely to be at increased risk from outdoor exposure because they typically exercise more
outdoors, thereby increasing their dose of air pollutants.
In addition to these methodological shortcomings for characterizing susceptible
subpopulations, research will develop methods to characterize microenvironmental and
population exposures. However, future exposure methods research shall be justified within the
context of human health risk assessment, for example, when current methods for high-priority
contaminants do not include adequate detection limits, accuracy, or precision, or when current
methods are so costly as to effectively preclude their use.
Future exposure methods shall be developed to measure mulri pathway exposures
(particularly for biological fluids and in dermal and dietary routes of exposure) to semivolatile
compounds such as PAHs and pesticides and their metabolites. Methods are needed also to
measure human exposure to microbial pathogens in drinking water (see microbe/DBF research
plan). Potential urinary biomarkers of exposure have been identified for several PAHs,
pesticides (e.g., chlorpyrifos, pentachlorophenol, DDT), and other organic pollutants that are
persistent in the environment and may be bioaccumulated. However, they will be validated and
February 1998
2-11
Draft—Do Not Quote or Cits
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
biomarkers for other high-priority persistent pollutants will be identified and their measurement
methods developed or improved. Potential biomarkers will be examined and validated in other
easily obtained biological excreta such as breath sweat, saliva, or sebum. Screening methods that
have low limits of detection and high sensitivity will be necessary to estimate exposures from
sampling such media. These methods are likely to include enzyme-linked immunosorbent, assays
(ELISA). Improved low-cost sampling methods, such as dermal wipes, will be tested for
application to persistent organic contaminants. Rapid, low-cost screening techniques shall be
developed, evaluated, and used to determine whether simple screening methods (e.g., immuno-
chemical methods, such as immunoassay-based tests) can identify those situations where high
exposures are likely and warrant further investigation. Rapid, generic extraction methods such as
supercritical fluid extraction shall be improved for use as screening tools.
Future exposure research shall emphasize developing and evaluating databases for
behavioral and time-activity determinants of human exposure for susceptible subpopulations.
Significant uncertainties exist about how variations in time-activity patterns and behaviors
contribute to variations in human exposure and susceptibility to disease. Two major types of
variability that contribute to this uncertainty are (1) exposure profiles (magnitude, duration, and
frequency) and (2) sensitivity to toxic insults (i.e., responsiveness to a given dose, such as that of
a person with asthma being more responsive to some air pollutants than is a person with healthy
lungs. Exposure and sensitivity are related also to age, lifestyle, genetic background, gender,
ethnicity, socioeconomic status, and preexisting disease. Until recently, the time-activity pattern
information required to investigate these issues had limited spatial, geographic, and demographic
coverage. However, with the recent completion of the National Human Activity Pattern Survey
(NHAPS) supported by ORD, national time-activity pattern data is being compiled by categories
such as gender, age, spatial location, occupation, socioeconomic status, race, day of the week,
and years of education (Nelson et al., 1994).
This database will be evaluated in the future to identify relationships between time-activity
patterns and high-end exposure for the general population, as well as for population subgroups
and regions. These investigations will enable future research to develop more accurate exposure
models and to identify and characterize population subgroups (e.g., infants, children, the elderly,
ethnic groups) more accurately and will contribute to exposure models. Such time-activity
pattern data will also be used to improve survey methodology. For example, a standard
February 1998
2-12
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
residential exposure questionnaire will be developed to obtain more detailed time-activity data
for all age groups and for underrepresented subpopulations such as those who are not fluent in
English. Automated passive collection devices that record events and microenvironmental
locations on a real-time basis will be refined and field tested.
Future exposure research shall emphasize developing, demonstrating, and evaluating
mathematical models that represent relationships between environmental contaminants and
multipathway human exposure and dose. Currently, the science of total human exposure
modeling is in its infancy. Although mathematical formulations for total exposure models have
been developed (Georgopoulos et al., 1997), no total exposure model has been demonstrated and
evaluated using field exposure measurements.2 Research support must be provided to achieve
this objective and to link total exposure models with dose models (i.e., physiologically based
pharmacokinetic (PBPK) models), as well as with models that predict source-environmental
concentration relationships (i.e., prospective and retrospective total human exposure models).
In developing the pathway-specific components of total exposure models, dietary and dermal
exposure pathways require particular emphasis because of their current higher degree of
uncertainty. Future dietary exposure models under development will be able to utilize food
consumption data, dietary behavior characteristics (e.g., characteristics related to regional and
ethnic influences), chemical residue data, and microbial contamination data. Future dermal
exposure models will be able to incorporate dermal contact and transfer data, data on skin
permeability to adsorption or absorption for various contaminants, and dermal-oral transfer and
ingestion data. In addition to this research, computational research will focus on developing a
modular multipathway modeling system that can incorporate measurement databases; time-
activity pattern data; demographic data; and contaminant emission, transport, and transformation
processes.
In directing ORD resources to accomplish these future research objectives, some elements
of ORD's current human exposure research program will be sustained at current levels, some
will be increased compared to current levels, and some will be decreased. ORD expects to
decrease support for basic source characterization and source attribution research, transport and
2With the possible exception of Ott et al. (198E).
February 1998 2-13
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
transformation research, and environmental characterization research. It is anticipated that future
research in these areas will be supported by other components of ORD's research program,
ORD also expects to decrease significantly its current support for exposure measurement
methods research after ongoing projects to develop multimedia methods for dietary exposure,
dermal exposure, biological markers, and semivolatile organic compounds are complete.
Methods will focus on resolving measurement-related uncertainties for contaminant exposures
where health risks are considered to be highly uncertain but of substantial concern and on
developing the next generation of low-cost and rapid-response methods (e.g., biological markers
of exposure, biosensors).
ORD will continue to support research to investigate relationships between human
exposure and time-activity patterns, and will increase future efforts to investigate subpopulations.
Future research will focus initially on analysis and dissemination of the survey results from the
NHAPS.
ORD will change the focus of the exposure research it conducts and sponsors to emphasize
the integration of measurement and modeling disciplines that have, in many instances, developed
historically as independent scientific functions. This exposure section addresses two components
of this strategy concurrently, namely exposures and susceptibility because the research must be
concurrent. For example, in evaluating population distributions of exposure, several
subpopulations must be considered and it is expected that they would have a range in
susceptibility. Also, for exposure, susceptibility is often defined by the extent of exposure, one
group versus another, again requiring concurrent comparisons. A close research relationship
between all parts of the risk assessment process in required for success. However, a significant
level of coordination is required to understand the exposure-dose-response relationship. In this
chapter, dose is primarily considered in close relationship to exposure, as in development of
models that predict the dose to the target with certain multipathway exposures; such research is
dependent significantly on the pharmocokinetic research described in Chapter 3, which is
conducted in close relationship to effects. Biomarkers research is also a continuum. This chapter
focuses on exposure biomarkers (e.g., blood or breath levels of a chemical); Chapter 3 focuses on
effects biomarkers (e.g., DNA adducts, endpoint markers). Of course, often biomarkers are
indicators of exposure, effects, and/or susceptibility, leading to the need for close coordination of
such research.
February 1998
2-14
Draft—Do Not Quote or Cite
-------
An Example of Strategic Partnerships for Human Exposure Research and Exposure and
Risk Assessment Development in ORD:
The National Human Exposure Assessment Survey
The National Human Exposure Assessment Survey (NHEXAS) is perhaps the most ambitious
study ever undertaken to examine a wide range of environmental pollutants and chemicals that humans are
exposed to in daily life. Whereas previous studies have focused on exposure to one chemical through one
environmental pathway, the goal of this study is to better understand the complete picture of human
exposure to toxic chemicals, by looking at humans' many exposures to all types of toxic chemicals through
all routes of exposure. Based on their experience with previous single- and multipathway exposure studies
in the United States and with the World Health Organization, ORD research scientists developed the initial
concept and design for this survey and coordinated this major research effort with colleagues in the FDA,
CDC, and the National Institute for Standards and Technology (NIST), NHEXAS studies are being
conducted in three different regions of the United States:
{1) a study in Arizona is being conducted by the University of Arizona, Battelle Memorial Institute, and
the Illinois Institute of Technology;
(2) a study in Illinois, Indiana, Michigan, Minnesota, Ohio, and Wisconsin is being conducted by the
Research Triangle Institute and the Environmental Occupational Health Sciences Institute of Rutgers
University; and
(3) a study in Maryland is being conducted by Harvard University, Emory University, Johns Hopkins
University, and WESTAT,
Scientists from ORD, FDA, CDC, and NIST are collaborating members of the research teams in each of
these studies.
During the course of these studies, researchers work with participants to measure the level of
chemicals in the air they breathe, in the foods and beverages they consume (including drinking water),
and in the soil and dust around their homes. Chemicals being analyzed include VOCs in air and water,
metals such as lead and mercury, and pesticides in food and soils. Researchers also are measuring
chemicals in participants' blood and urine samples. Participants complete questionnaires to help identify
possible sources of chemical exposure. At the conclusion of the study, each participant will receive a report
on the results of exposure and biological measurements, with an explanation of the findings' significance.
Confidentiality of participants is strictly protected, although they are free to inform others if they choose.
Data collected during these studies are expected to enable the human heath risk assessment
research community to accomplish the following scientific goals.
• Improve estimates of total human exposure to chemicals and identify population subgroups that are
highly exposed to environmental chemicals.
* Provide a baseline of the normal range of human exposure to chemicals in the general population, to
allow comparisons with specific studies on particular exposure routes.
~ Relate Identifiable pollution sources to the actual exposures that people experience and compare
short-term exposures to longer term exposures.
• Ultimately, enable researchers to improve the accuracy of human exposure assessment models and of
human health risk assessments.
Sample collection for the studies began in mid-1995 and is expected to be completed by late
1997. Sample analyses are expected to be completed by early 1998. After statistical analysis and
summary, significant findings from each of the studies are expected to be peer reviewed and published
in 1998, with databases becoming publicly available in 1999.
1 ORD expects that current and future resources for human health risk assessment research
2 will not be adequate to support large-scale population exposure studies (i.e., a regional- or
3 national-scale population exposure or surveillance study such as that contemplated as the
4 long-term goal for the NHEXAS program). However, recognizing the importance of population-
February 1998
2-15
Draft—Do Not Quote or Cite
-------
1 scale exposure research to reducing uncertainty in risk assessment and to the development of
2 total human exposure models, ORD will continue efforts to build a broad partnership and support
3 to achieve this objective. This partnership will include other federal agencies with intramural
4 and extramural research programs directly related to human health risk assessment (e.g., NIEHS,
5 CDC, FDA) and scientific experts from academic and private research institutions.
6
February 1998
2-16
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
_3_
Dose and Effects Research
3.1 Background
Figure 3-1 illustrates the scientific components of dose estimation and health effects
research that lay the framework for this section of the proposed strategy. Figure 3-1 illustrates
the scientific elements involved in dose estimation research. Dose estimation serves as the link
between exposure and effects. That portion of the environmental contaminant that is transferred
into the body surfaces (i.e, by inhalation, dermal contact, or ingestion) is known as the applied
dose. The applied dose ultimately is absorbed, leading to a dose at a target organ that causes the
effect of concern at that site. Investigation of dose and biological markers of exposure and
effects (e.g., DNA adducts, cholinesterase inhibition) represents the point of transition between
exposure assessment and effects assessment. Exposure biomarkers demonstrate that exposure to
a given agent has occurred, whereas effects biomarkers identify an effect of a particular type that
has occurred. Also at this transition point, quantitative relationships between exposure,
absorption rate, distribution, metabolism, and elimination rate are represented mathematically by
PBPK models. There is clearly a continuum between exposure-dose-response and between
biomarkers of exposure, effects and susceptibility. Those aspects closely aligned with exposure
are contained within Chapter 2 (Exposure). Those more related to effects are given here.
In practice, there is collaboration between the ORD researchers in these areas.
The assessment of effects includes both hazard characterization and dose-response
evaluations. ORD's hazard characterization research involves the development of methods that
demonstrate a qualitative relationship between exposure and effect, Dose-response research then
characterizes this relationship to link exposure-dose with incidence and severity of effect,
considering mechanisms and factors that may affect dose- response relationships. This
information is then used to develop quantitative models for estimating risk. In this chapter, the
term dose-response is used because it is the NAS terminology and the ultimate goal. In most
cases, the exposure-response is the object of study and assessments.
February 1998
3-1
Draft—Do Not Quote or Cite
-------
Emission
Source(s)
i
r
Environmental
Fate
f
Environmental
Concentrations
'
'
Human
Exposure
*
r
Internal
Dose
r
Effect(s)
Risk
Characterization
*
Risk
Management
PBPK Models-
Human
Exposure
r
Applied
Dose
f
Absorbed
Dose
1
>
Target
Dose
1
r
Hazard
Characterization
'
Dose-Response
Evaluation
• Time-Activity Studies
> Suseptible Populations
Exposure Rate
- Inhalation
- Ingestion
- Dermal
Inhalation
Ingestion
Dermal
' Metabolism
• Distribution
• Clearance
¦ Epidemiology
«Human Clinical Studies
¦ Animal Studies
¦ Cancer
¦ Noncancer
- Pulmonary
- Neurologic
- Liver
- Kidney
- Immunologic
- Developmental
- Reproductive
- Other
Figure 3-1. Scientific elements in dose estimation research.
1 Traditionally, EPA has taken different quantitative approaches of risk assessment for cancer
2 and noncancer effects. In cancer dose-response assessment, the default assumption, in the
3 absence of relevant biological evidence on mechanism of action, has been that increased risk
4 varies linearly with dose, even at low doses. Thus, exposure to any dose would result in some
5 increase in cancer risk. Under the proposed new cancer guidelines (U.S. EPA, 1996) a similar
6 default assumption is retained; however, data on the mode-of-action of a chemical is emphasized,
7 and such data will guide the process of risk estimation. Mode-of-action refers to the interaction
8 of the chemical with specific targets or pathways. It is important to recognize that a given
9 chemical may display more than one mode of action. For example, one of the targets of
February 1998
3-2
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
carcinogens is the genes (DNA) that control cell growth; other targets are the biochemical
processes that are involved in cell growth, cell growth regulation, cell signaling, and cell-to-cell
communication. Still other targets of chemical carcinogens may include processes involved in
cell toxicity and death, alterations in hormone levels, effects on receptors involved in cell growth,
effects on enzymes that metabolize carcinogens, effects on the immune system, and effects on the
cellular repair systems that allow cells to repair damage caused by carcinogens. Concomitant
with the recognition of these facts has been the realization that the currently used, statistically
based cancer risk assessment models (e.g., the linearized multistage model) are probably not
appropriate for all types of chemical carcinogens. Despite the recognition of various targets and
events in carcinogenesis, mechanistic information remains largely incomplete, and, for most
direct DNA reactive carcinogens, the assumption of linearity will still apply.
For many noncancer toxicities, it is assumed that dose thresholds exist, that below a certain
dose, no overt toxicity will be expressed. This assumption is based on the known capacity of the
organism to detoxify the chemical or repair a certain amount of damage at the molecular,
cellular, tissue, or organ level. In addition, multiple insults at the molecular or cellular level may
be required, given that a population of cells often must be affected to produce an effect on the
whole organism. Newer research, such as that on dioxins, has shown that the cancer-noncancer
dichotomy, as reflected in the preceding discussion, may have limited relevance in a unified
concept of health risk assessment.
3.1.1 Current ORD Research
ORD's current program concerning dose estimation research focuses on three scientific
questions (also see Table 3-1).
(1) How can estimations of deposition (for chemicals having portal-of-entry effects) or
absorption (for chemicals having systemic effects) following inhalation, oral, or dermal
exposures be improved?
(2) What are the critical factors affecting estimation of target tissue dose?
(3) What are the biomarkers of effects?
February 1998 3-3 Draft—Do Not Quote or Cite
-------
Table 3-1. Overview of the Current Dose Estimation and Health Effects Research Program
Scientific Questions and
Research Focus/Objective
Research Approach
Research Outputs/Products
Future Emphasis
How can models linking
exposure scenario and target
tissue dose be improved?
Develop and evaluate PBPK models to predict
target dose and to elucidate factors affecting dose
estimation. ORD research is focused on model
compounds such as trichlorethylene, chloroform,
carbon tetrachloride, dibromochloromethane,
arsenic, acrylamide, PAHs, DBPs, and dioxins.
Improved dose-response assessment
and research targeting: improved
extrapolation and interpolation of data
(cross-species, cross-route, cross-dose
scenario, etc.) using PBPK models
Research activities to address this objective
are anticipated to increase.
How can estimations of
absorption for inhalation, oral,
and dermal exposures be
improved?
Research to identify and characterize factors
affecting absorption, including physicochemical
characteristics; exposure conditions, including
dosing pattern and vehicle; and portal of entry
factors, including contact location. ORD research
is focused on VOCs, such as trichloroethylene
and carbon tetrachloride; on respirable PM; on
metals such as arsenic; on respiratory irritants
such as ozone; and on dioxins and PCBs.
Reduced uncertainty in dose-response
assessment: complete and consistent
reference values for use in dose
estimation and identification and
characterization of the factors
producing greatest uncertainty in
estimation of absorption
Research activities to address this objective
are anticipated to decrease.
What are the critical factors
affecting estimation of target
tissue dose?
Research on factors affecting distribution and
tissue dose, including metabolism and clearance;
tissue binding; blood flow and tissue volumes;
tissue/blood partitioning; and co-pollutant
exposures. ORD research is focused on human
and rat physiological parameters, dioxin
sequestration, arsenic metabolism, and pulmonary
particle clearance.
Improved dose-response assessment:
improved interspecies estimates of
target dose and evaluation of the
effects of dosing scenario on target
dose
Research activities to address this objective
are anticipated to be sustained at current
levels.
How can models linking
exposure scenario and target
tissue dose be improved?
Develop and evaluate PBPK models to predict
target dose and to elucidate factors affecting dose
estimation. ORD research is focused on model
compounds such as trichlorethylene, chloroform,
carbon tetrachloride, dibromochloromethane,
arsenic, acrylamide, PAHs, DBPs, and dioxins.
Improved dose-response assessment
and research targeting: improved
extrapolation and interpolation of data
(cross-species, cross-route, cross-dose
scenario, etc.) using PBPK models
Research activities to address this objective
are anticipated to increase.
o
r+
D
c
o
p+
CD
O
n
o
-------
CD
cr
—*
c
0}
•3
CO
CO
00
Table 3-1 (cont'd). Overview of the Current Dose Estimation and Health Effects Research Program
Scientific Questions and
Research Focus/Objective
Research Approach
Research Outputs/Products
Future Emphasis
How can estimations of
deposition and absorption for
inhalation, oral, and dermal
exposures be improved?
Research to identify and characterize factors
affecting absorption, including
physicochemical characteristics; exposure
conditions, including dosing pattern and
vehicle; and portal of entry factors, including
contact location. ORD research is focused on
VOCs, such as trichloroethylene and carbon
tetrachloride; on metals such as arsenic; on
respiratory irritants such as ozone; and on
dioxins and PCBs.
Reduced uncertainty in dose-
response assessment: complete and
consistent reference values for use in
dose estimation and identification
and characterization of the factors
producing greatest uncertainty in
estimation of absorption
Research activities to address this
objective are anticipated to decrease.
What are the critical factors
affecting estimation of target
tissue dose?
Research on factors affecting distribution and
tissue dose, including metabolism and
clearance; tissue binding; blood flow and tissue
volumes; tissue/blood partitioning; and
co-pollutant exposures. ORD research is
focused on human and rat physiological
parameters, dioxin sequestration, arsenic
metabolism, and pulmonary particle clearance.
Improved dose-response assessment:
improved interspecies estimates of
target dose and evaluation of the
effects of dosing scenario on target
dose
Research activities to address this
objective are anticipated to be sustained
at current levels.
Improve hazard
characterization
Develop and evaluate new toxicological test
methods to identify and characterize the
hazards posed by environmental exposure to
chemicals (i.e., pesticides, industrial
chemicals), coupled with research to improve
the interpretation of toxicological data.
Includes research to elucidate underlying
mechanisms of pollutant toxicity and the repair
or adaptation of damaged tissues using animal
models and human studies, as well as
computational chemistry/SAR. ORD research
emphasizes neurotoxicity, developmental and
reproductive toxicity, pulmonary toxicity,
immunotoxicity, and cancer.
New and refined test methods.
Impaired methods of data
interpretation.
Research activities to address this
objective are anticipated to be sustained.
CO
I
CJl
D
-*
O
o
D
c
0
1 I1
CD
O
o
« I
CD
-------
Table 3-1 (cont'd). Overview of the Current Dose Estimation and Health Effects Research Program
Scientific Questions and
Research Focus/Objective
Research Approach
Research Outputs/Products
Future Emphasis
Improve biological basis for
dose-response assessment
Research to elucidate underlying mechanisms of
pollutant toxicity and the repair or adaptation of
damaged tissues using animal models and human
studies, as well as computational chemistry/SAR.
ORD research focuses on reproductive and
developmental toxicity, cancer, neurotoxicity,
pulmonary toxicity, immunotoxicity, and hepatic
and renal toxicity. Model compounds (e.g.,
chemicals for which existing data provide a
strong basis for study) are employed in
hypothesis-driven research. Solvents, metals, PM
and other air pollutants, dioxins, mercury, and
drinking water DBP are being studied.
Improved near-term understanding of
the level of confidence to associate
with existing methods for dose-
response assessment
In the longer term, evaluation of the
potential for biological models to
estimate dose response based on
mechanistic understanding
Research activities to address this objective
are anticipated to increase.
Improve empirical methods for
dose-response assessment
Research to support development of empirical
methods, such as the benchmark dose model and
the categorical regression approach. ORD
research focuses on evaluation of these methods
for developmental, pulmonary, immunological,
and neurotoxicological endpoints.
Improved near-term methods for risk
assessors to estimate dose response
from existing data
Research activities to address this objective
are anticipated to decrease.
"FIFRA = Federal Insecticide, Fungicide, and Rodenticide Act.
'TSCA = Toxic Substances Control Act.
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
To address these questions, ORD conducts research that identifies and characterizes various
factors that may affect deposition and absorption, such as the physicochemical characteristics of
the pollutant and the exposure conditions, including exposure patterns and portal of entry.
Research also has been conducted on the factors that affect the distribution, metabolism,
clearance, and other dynamics, such as tissue binding, that help to estimate target-tissue dose.
Both the research on absorption characteristics and target tissue dose estimation have helped to
reduce the uncertainty in dose-response assessment through the development of reference values
and through improved methods for characterizing interspecies extrapolation of dose and effect.
Specific areas of emphasis include research on trichloroethylene and other volatiles; PM; and
inorganics, such as arsenic, ozone, dioxins, and PCBs. The intent is to improve the qualitative
and quantitative characterization of target tissue exposure for endpoints such as respiratory,
developmental, neuro-, immuno-, and reproductive toxicity, as well as cancer. Biomarkers
research has been initiated for products of incomplete combustion, PAHs, DBFs, dioxins,
arsenic, PCBs and pesticides.
Current ORD health effects research includes investigations on improved methods of
hazard characterization and on biologically based and empirical dose-response models
(Table 3-1). This research area seeks to develop an improved scientific basis for risk assessment
by developing new toxicological test methods to identify and characterize hazards, and to define
underlying mechanisms of toxicity and carcinogenicity to facilitate methods and model
development and validation. The goals are to elucidate the critical physiologic and mechanistic
factors that contribute to health effects in laboratory animals and humans; to determine the
effects of varying route, dose, dose-rate, duration, and cumulative dose on health outcomes; and
to develop data for and to evaluate biologically based dose-response models for application in
human health risk assessments. The overall scientific approach is to conduct and link laboratory
studies and model development activities to understand and describe the mechanisms of toxicity
and methods to estimate toxic response to target tissue concentrations. The continued
development of biologically and physiologically based dose-response models will support dose
extrapolation to humans and will refine risk assessments through consideration of mechanism-of-
action. Results of this research will support the development of empirical methods such as the
benchmark dose and categorical regression approaches to health assessment.
February 1998
3-7
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
The following sections identify research needed to reduce significant scientific
uncertainties in dose and effects research. The following areas will be the focus of this research.
3.2 Dose Estimation Research
3.2.1 Uncertainties in Mechanistic Data for Hazard Characterization and Dose-Response
Assessment
3.2.1.1 Problem Statement
For quantitative noncancer health assessments, EPA typically estimates the daily exposure
from a particular route of exposure that is not anticipated to cause significant adverse effects over
a lifetime. In many cases, data are available only from studies in laboratory animals and may not
be available for the route or pattern of exposure of interest. Under these conditions, risk
assessors must determine whether available data can be extrapolated to the route, species, and
exposure conditions being assessed. In general, the lack of data, difficulty in data interpretation,
and underutilization of existing data because of insufficient models and statistical reliability
reduce the validity of extrapolations used to estimate target dose. Although the current research
program has focused on characterizing absorption, distribution, and clearance, there remains a
need to continue research to improve the knowledge of absorption characteristics, the potential
for portal-of-entry effects, the potential for first-pass metabolic effects to modulate target dose,
and the influence of exposure pattern on target tissue dose and response for acute, intermittent,
and longer-term exposures.
3.2.1.2 Scientific Questions
The scientific question that provides the strategic direction to define the research products
and their use in risk assessment (i.e., to improve the application of mechanistic data for hazard
characterization), exposure-dose-response research, and risk assessment is (see Table 3-2);
How can dose estimation across species and exposure scenarios be improved?
(A related question concerning variability in response to toxicity is addressed partially in
Chapters 2 and 4.)
The concentration of a pollutant to which a human is exposed is often not the same as the
dose (i.e., the amount of pollutant delivered to the target organ). A number of mechanisms,
February 1998
3-8
Draft—Do Not Quote or Cite
-------
Table 3-2. Future Directions in Dose Estimation and Human Health Effects Research
To Improve Human Health Risk Assessment
Scientific Questions
and Research Focus
Research Approach/Tasks
Research Products
Use in Risk Assessment
How can we improve the accuracy of
dose estimation across species, exposure
routes, and scenarios?
Develop, validate, and apply new
biological markers of exposure and
effect to reflect exposure-dose-
response relationships. Develop
better PBPK models for dose
estimation.
PBPK models for classes of
compounds to estimate blood and
tissue concentration and time
course.
Validated biomarkers of exposure
and effect for use in dose
response estimation.
Reach reliance on assumptions and
route-to-route extrapolations.
How can mechanistic information be
used to improve the ability to detect
hazards?
Develop screening methods to set
testing priorities
Develop cost-effective methods for
toxicity data collection
Validated screening protocols
using, for example, in vivo, in
vitro, and SAR methods
New and revised standard toxicity
testing protocols
To identify and rank existing
pesticides and industrial chemicals
in terms of potential toxicity
To screen new chemicals as they
enter the regulatory system and to
assess relative toxicity
To develop EPA test guidelines
To support regulatory activities (e.g.,
TSCA test rules and consent
agreements, FTFRA data call-ins)
How can toxicity data be better
interpreted to predict and define
hazards?
Develop improved methods for
data interpretation; for example,
identify biomarkers of exposure
and effect and validate the use of
biomarkers in human populations;
focus on hazards resulting from
less-than-lifetime exposures
Guidance document on
interpretation of toxicity data
For incorporation into risk
assessment guidelines
-------
Table 3-2 (cont'd). Future Directions in Dose Estimation and Human Health Effects Research
To Improve Human Health Risk Assessment
How can uncertainty in extrapolations
(e.g., from high doses in animals to
environmental exposures in humans) be
reduced?
Develop quantitative models for
predicting tissue and organism
response to target tissue dose (i.e.,
biologically based dose-response
models)
Develop improved empirical dose-
response models (i.e., benchmark
dose models)
Models for predicting toxicity from
chemical exposures that can be
modified and applied in chemical-
specific risk assessments
Validated benchmark dose models
and guidelines for applications
To provide critical examples of
development and use of mechanistic
models and to evaluate the potential of
these models for replacing default
approaches for cancer and noncancer
risk assessment
To provide a state-of-the-science basis
for replacing default, primarily
empirical risk assessment approaches
To improve reference dose/reference
concentration procedures and thereby
improve the basis for risk
management decisions
What are the factors influencing human
susceptibility to disease? How do they
influence human health risk assessment?
Determine biomarkers of
susceptibility within the human
population
Determine the magnitude of
contribution of susceptibility factors
to human health risk assessment
Methods for detection of
susceptible individuals
Models for predicting the
distribution of susceptible
individuals within the human
population
Application of biomarkers in risk
assessment
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
many poorly understood, affect the transport of the pollutant through the portal of entry (e.g., the
lung for air pollutants, the digestive tract for pollutants in drinking water and food) to the target
organs. Also, the physical and chemical status of the pollutant within the body is affected by a
number of mechanisms, physicochemical and metabolic, that can alter the disposition of the
pollutant and, ultimately, the dose of the active agent to the target organs.
In risk assessment, however, the exposure concentration of a pollutant often is used as a
surrogate for the dose because data on the dose of the active agent to the target organs are not
available. Research to improve target dose data, including methods and models, is needed to
reduce uncertainties associated with extrapolation from one route of exposure to another; from
high to low exposure; from one species to another; and among exposure scenarios of varying
magnitude, duration and frequency. One important aspect of this research is the development of
biological markers for exposure and the quantitative linkage of these markers with markers of
effect. Improved quantitative PBPK models to relate actual exposures to target tissue dose in
humans under a variety of exposure conditions are needed to provide more accurate "dose" input
for dose-response assessment.
3.2.1.3 Research Approach, Products, and Uses
Table 3-2 identifies the research tasks and products that respond to the scientific question
mentioned above, and it also describes the products of this research that are related to
improvements in the scientific foundation for risk assessment. Given the substantial criticism
associated with current EPA practices for estimating cancer risks (e.g., linear extrapolation) and
noncancer risks (application of uncertainty factors to a NOAEL/LOAEL), an area of increased
emphasis will be obtaining fundamental pharmacokinetic and mechanistic data and tools for
application in deriving more biologically defensible risk assessments. The pharmacokinetic data
and models will serve as the linchpin for linking exposure and effects. Pharmacokinetic research
will address issues related to route-to-route and cross-species extrapolation and identification of
markers of actual target tissue dose. The mechanistic data will allow for clarification of the
relevance of animal models (cross-species extrapolation) and the validation of biomarkers of
toxic effects that may serve as early indicators of effects and be used as the basis for low-dose
extrapolation.
February 1998
3-11
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
3,3 Effects Research
3.3.1 Problem Statement
Characterization of hazard potential and extrapolation of dose-response data from animals
to humans is fraught with uncertainty. The interpretation of animal toxicological data with
regard to interspecies hazard and selection of a dose-response model that fit experimental data
well may result in estimates that can span several orders of magnitude at environmental exposure
levels. The uncertainties stem from fundamental gaps in knowledge regarding interspecies and
intraspecies extrapolation, variability in susceptibility and response, and the shape of the dose-
response curve at environmentally relevant doses. Default assumptions are used in the face of
these uncertainties and knowledge gaps. Research is needed to define and reduce the
uncertainties, minimizing the need for default assumptions and providing a stronger mechanistic
basis for human health risk assessment.
3.3.2 Scientific Questions
Three scientific questions provide the strategic focus to define the research required to
improve the accuracy of hazard characterization, exposure-dose-response research, and risk
assessment.
(1) How can mechanistic information be used to improve the ability to detect hazards?
(2) How cm the methods to interpret human health effects data be impaired?
(3) How can uncertainty in extrapolations (e.g., from high doses in animals to environmental
exposures in humans) be reduced?
3.3.3 Research Approach, Products, and Uses
The above questions provide the strategic framework to define the human health effects
research approaches and associated research products required to improve the scientific
foundation for risk assessment. These approaches, products, and their anticipated benefits to
improving risk assessment are summarized in Table 3-2. Research emphasis should include
development of more selective and valid tests for mechanistically based hazard identification and
characterization; enhancement of empirical approaches for dose-response assessment, using
mechanistic information to move beyond benchmark and no-observed-adverse-effect-level
approaches for cancer and noncancer risk assessments; performing research to improve the
February 1998
3-12
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
understanding of receptor-mediated mechanisms; and focusing on health effects associated with
less-than-lifetime exposures. The following sections are intended to amplify the specific
research directions.
3,3.3.1 Development of More Selective and Valid Tests for Mechanistically Based Hazard
Characterization
Improved tests for hazard characterization will be developed to assess the potential effects
of chemicals on various health endpoints. For example, use of "transgenic mice" allows for the
investigation of the influence of selective gene expression/nonexpression on the effect of a
chemical Similarly, more relevant in vitro models will be built by expressing human receptors
in cell reporter assays. Such test systems may be validated using conventional test methods for
which a greater degree of mechanistic understanding or database is available. Development of
biomarkers will also enhance the identification of hazards.
Other new approaches, such as computational chemistry and structure-activity relationships
(SAR), will improve the ability to conduct hazard identification on a large number of compounds
for which there is little or no health effects information. These new approaches also will make
the use of bioassays more cost-effective by improving the capacity to choose the most relevant
bioassays to be performed. Computational chemistry and SAR approaches will complement
ongoing experimental studies, involving hazard identification and mechanisms-of-action for
important pollutant classes. These efforts will yield insights into underlying reaction
mechanisms associated with chemical toxicity (e.g., computed energies to evaluate and compare
plausible reaction pathways for metabolic activation), thus aiding in the design of research issues
and approaches. SAR modeling also will be used to guide experimental studies into productive
new areas, directing the application of assays to fill data gaps for SAR analysis and, in some
cases, to providing a basis for extrapolation to untested chemicals. ORD will use this research
information to support the process of guideline development, especially for emerging areas of
health risk assessment (e.g., health risks associated with short-term exposures, such as
pulmonary, neuro-, and immunotoxicity and complex mixtures).
February 1998 3-13 Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3.3.3.2 Enhance Empirical Approaches for Dose-Response Assessment
Although the benchmark dose and other empirical approaches are seen as improvements
over traditional (i.e., reference dose) noncancer risk assessment approaches through the use of
more dose-response data, these approaches do not fully incorporate mechanism-of-action data.
The continued development of biologically and physiologically based dose-response models will
support animal-to-human extrapolation to humans and to refine risk assessments based on
mechanism-of-aetion. The results of research on biological mechanisms and toxicokinetics will
improve the quantitative estimation of human risk posed by environmental chemicals (including
multiple chemical sensitivity) that have been described only empirically. Evaluation of the
applications and limitations of these methods and the characterization of their strengths and
weaknesses for risk assessment are essential.
3.3.3.3 Focus on Receptor-Mediated Mechanisms
The reassessment of health risks posed by dioxin found that compounds acting through the
same receptor were additive in causing effects, whereas nonadditive interactions occurred when
multiple mechanisms were involved. Thus, ORD will conduct additional research to assess the
role of receptor-mediated mechanisms in lexicological effects produced by other compounds.
An important focus in this research concerns how toxicants can interfere with critical cellular
pathways (e.g., signal transduction pathways and receptors involved in cell growth).
Computational ehemistry/SAR studies will be used in conjunction with laboratory studies to
identify key features of such receptors, to study receptor-mediated mechanisms of action, and to
model the interaction of environmental chemicals with receptors.
More work also is needed to understand the effects of receptor mediation on the dose
response of toxic chemicals and mixtures. A stracture-activity-based toxic equivalency factor
(TEF) approach has been applied to mixtures of dioxin-like compounds. Future research will
examine the utility of the TEF methodology to predict biochemical and toxicological responses
of environmentally relevant mixtures of dioxin-like chemicals in animal models.
Finally, receptor-mediated toxicity will be studied in humans as a function of genetic
background and age. ORD will incorporate information on receptor-mediated mechanisms and
toxicokinetics, as well as information obtained from human studies (e.g., receptor
February 1998
3-14
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
polymorphisms, isoforms, levels, cross talk) into dose-response models that are relevant to
specific segments of the human population.
3.3.3.4 Focus on Health Effects Associated with Less-Than-Lifetlme Exposures
Noncancer-related toxic endpoints such as developmental, pulmonary, neuro-, and
immunotoxicity may result from less-than-lifetime exposure scenarios. In neuro toxicology,
animal-to-human extrapolation research focuses on developing animal models of neurotoxic
effects that can be more precisely extrapolated to humans. Two effects that are particularly
difficult to extrapolate from animals to humans are cognitive dysfunction and sensory alterations.
Thus, research in neurotoxicology, will focus on developing and validating animal models of
these endpoints that also can be measured in humans.
Research will also seek to improve key default assumption. For example, results of
neurological and pulmonary toxicity research concerning the relationship between duration and
concentration of exposure have suggested that dose rate is more critical for estimating effects
than is cumulative exposure for some short-term and intermittent exposures (unless chemicals
are persistent and bioaccumulative). Thus, research will characterize the relationships between
dose rate (or dose metric) and toxicity and repair/compensation from short-term intermittent
exposures to environmental chemicals.
ORD also seeks to improve the following quantitative models to further characterize and
predict effects in humans: animal-to-human extrapolation models, models to evaluate the
variability of exposure scenarios and the impact on time when predicting effects of pollutants on
humans, and pharmacokinetic models in which physiological parameters (e.g., CO and C02
levels, other blood gases) can be taken into account when assessing effects of pollutants on
humans. Another default assumption concerns cancer. It is presumed that cancer results from
lifetime exposure to an agent. Research is needed to determine the time course for the
development of cancer. In summary, ORD will pursue research on the effects of short-duration
exposure and on relationships between exposure level and exposure duration. This research will
be used to develop assessment methods, dose-response models, and guidance for assessing
effects from less-than-lifetime exposures.
February 1998
3-15
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3.4 Characterizing and Assessing Variation in Human Susceptibility to
Disease
3.4.1 Problem Statement
Uncertainties regarding human variation in susceptibility and response to environmental
pollutants support the need for increased research on multidisciplinary endpoints to identify the
factors that affect human susceptibility, the magnitude and distribution of these factors in the
human population (see also Chapter 2), and the quantitative relationship between these factors
and increased risk among specific subpopulations. Epidemiology, human clinical studies, animal
toxicology studies, and in vitro assays are important methods to identify and assess factors that
may contribute to observed variability in susceptibility. These factors, including age, lifestyle,
genetic background, gender, and ethnicity, will be studied to determine how they contribute to
human health risk.
3.4.2 Scientific Questions
Two scientific questions provide the strategic focus for future research needed to
characterize variation in human susceptibility for exposure-dose-response research and risk
assessment (see also Table 3-2):
(1) How can hazards be better defined/predicted, dose-response extrapolation be improved, and
variation related to human susceptibility be further characterized?
(2) How can risk assessments from varying exposure scenarios be improved?
3.4.3 Research Approach, Products, and Uses
ORD's research will improve understanding how differences in susceptibility contribute to
dose-response models representing various human subpopulations (e.g., infants and children,
women, the elderly, individuals with preexisting diseases, and different races and ethnic groups).
Particular emphasis will be placed on embryos/fetuses, infants, and children as a vulnerable
population. This emphasis is consistent with the EPA Administrator's directive to consider risks
that environmental pollutants pose to infants and children and with the national commitment to
ensure a healthy future for children. The research proposed here, compliments the research
directions outlined in the draft Research Strategy, being developed for childrens health
February 1998
3-16
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
protection. Research will also be conducted to determine the conditions under which there are
age-dependent quantitative and qualitative differences in responsiveness to pesticides.
Investigations on the toxicokinetics of susceptibility factors, on underlying mechanisms of
increased sensitivity, and on disease-related physiological parameters will help to identify the
critical genetic and biological biomarkers of susceptibility. As an example, research will be
performed on chronic low-dose effects at the molecular and cellular levels that may result in the
induction of genetic polymorphisms in the human germ line.
In adult volunteers, clinical investigations using carefully controlled exposures and dietary
interventions can provide a wealth of data on the potential influence of specific polymorphisms
on the likelihood of an adverse response to an environmental agent. In a clinical setting,
exposure and dose-response relationships can be characterized for individuals with, for example,
chronic pulmonary disease (e.g., asthma, bronchitis, chronic obstructive pulmonary disease
[COPD]), cardiovascular disease, acute respiratory disease (e.g., upper respiratory infections), or
multiple chemical sensitivity. In addition, potential susceptibilities associated with gender, age,
or race can be studied.
Research will focus also on the underlying biological mechanisms responsible for
individual susceptibility to pollutants. In particular, the mechanism by which different pollutants
cause injury to cells in the respiratory tract will be studied. Cells and fluids from the upper and
lower respiratory tract will be analyzed for biochemical and molecular responses of induced in
vivo or in vitro (e.g., signal transduction systems and transcription factors involved in the
responses of the cells to pollutant exposure). Again, new transgenic and knockout mouse models
offer the possibility to directly examine the genetic regulatory mechanisms that influence these
toxicological responses, thus directing the researcher to new hypotheses regarding possible
mechanisms of action of environmental chemicals. Among the pulmonary toxicology models
being studied in experimental animals are COPD, pulmonary and systemic hypertension, asthma
and reactive airway diseases, degenerative heart disease, and pulmonary fibrosis. Appropriate
animal strains and species assessments need to be determined for comparison to responses
observed in humans.
Using a combined mechanistic approach of clinical and toxicological investigation, it will
be possible to identify, select, and apply critical human biomarkers for the characterization of
susceptible subpopulations in conjunction with epidemiologic field studies. These field studies
February 1998
Draft—Do Not Quote or Cite
-------
1 will provide the validation in the field of health effects that is seen in the laboratory and clinic.
2 With sufficiently sensitive biomarkers, early changes can be detected, thereby improving EPA's
3 ability to prevent effects. Public health programs of newborn screening could have major
4 benefits in identifying susceptible individuals so that exposures to agents to which these infants
5 are highly susceptible can be reduced or avoided.
6
February 1998
3-18
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
A
Risk Assessment and Characterization Research
4.1 Background
Figure 4-1 highlights the primary elements of ORD's current program in risk assessment:
development of risk assessment methodology, risk assessments of chemicals that demonstrate
new approaches, and guidance and training. EPA's risk assessment research utilizes not only the
results of research conducted by ORD, but health and exposure research conducted outside EPA
(e.g., National Institutes of Health, universities, etc.) as well. The current ORD risk assessment
program is summarized in Table 4-1. Current research in risk assessment include the following.
• Methodology
- Methodologies for quantitative assessment (e.g., benchmark dose approach for noncancer
endpoints, biological models for cancer dose-response assessment)
• Prototype Assessments
- Assessments of contaminants and sites of national significance that demonstrate new
approaches to risk assessment and that respond to contentious or sensitive issues
• Guidance and Training
- Health and exposure risk assessment guidelines that incorporate the most recent and relevant
scientific information (see text box)
- Training and consultation in risk assessment (e.g., training on the various guidelines,
consultation to EPA regions, and programs on various risk assessment problems)
- Guidance on selected topics of interest, such as the relevance of rat kidney tumors to humans,
the relevance of thyroid tumors produced at high chemical exposures in animals to the human
situation, and Monte Carlo approaches to the use of information on exposure distribution
- Risk information databases (e.g., the Exposure Factors Handbook currently available on CD
with search capabilities, which provides information on exposure parameter distributions of
interest to the risk assessor, such as fish consumption rates, respiratory rates, daily volume of
drinking water consumed, etc.)
February 1998
4-1
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
Emission
Source(s)
Environmental
Concentrations
<
f
Human
Exposure
internal
Dose
r
Effect(s)
*
Risk
Characterization
'
Risk
Management
Risk
Assessment
Risk
Characterization
• Risk Assessment Guidlines
> Integrate Hazard, Dose-
Response, and Exposure
¦ Qualitative and
quantitative
characterization
of hazard and
exposure as It
relates to human
health risk
Figure 4-1. Scientific elements in risk assessment and characterization research.
- Risk information expert systems (e.g., Risk Assistant, which is an interactive software
program guiding the risk assessor through various choices on a risk assessment problem)
4.2 Strategic Directions
4.2.1 Problem Statement
Inherent in all risk assessment guidance and methodology are uncertainties and gaps in
scientific knowledge. Many of these gaps and uncertainties likely will continue for years to
come, but health risk assessment, because of public health considerations, cannot wait for
complete information. The challenge to the risk assessor is to develop approaches and default
options (i.e., policy judgments to accommodate uncertainties and gaps in scientific knowledge)
that make maximum use of existing information.
February 1998
4-2
Draft—Do Not Quote or Cite
-------
Table 4-1. Overview of ORD's Current Health Risk Assessment Research Program
Research Objectives
Research Approach
Research Products
Future Emphasis
Develop risk assessment
methodology
Quantitative models for
dose-response assessment
Dermal exposure
methodology
Uncertainty analysis for
reference concentrations
Improved methodology
for multipathway and
multichemical exposure
assessment
Provide new methods to
address risk assessment
questions
Research activities in this
area are anticipated to
increase.
Conduct prototype risk
assessments
Selection for risk
assessment of chemicals
of high visibility to EPA
or for which new data has
become available that
allows a demonstration of
new risk assessment
approaches
Provide assessment of the
chemical under study and
provide advanced
methods of assessment
that may have
applicability to other
chemicals
Research activities in this
area are expected to
continue at a level similar
to that of the past.
Develop risk assessment
guidance and databases
and provide risk
assessment consultation
and training
Risk assessment
guidelines
Guidance documents on
topics of interest, such as
rat kidney tumors, Monte
Carlo approaches, etc.
Risk assessment
databases (e.g., Exposure
Factors Handbook,
MIXTOX Data Base,
Integrated Risk
Information System, etc.)
Risk assessment training
Consultative advice to
regions and programs
Expert system software
(e.g., Risk Assistant)
Provide an improved
framework for systematic
risk analysis and guidance
on difficult risk
assessment issues
Provide information on
parameters of interest to
the risk assessor
Improved knowledge and
capability for risk
assessors in the regions
and programs
Research activities on
risk assessment databases
are expected to increase;
other areas are expected
to continue at a level
similar to that of the past,
with the exception of
expert systems, which is
expected to decrease.
February 1998
4-3
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
EPA has developed the following health risk assessment guidelines:
• Carcinogen Risk Assessment (1996) (Proposed)
• Reproductive Toxicity (1996)
• Neurotoxicity (1995) (Proposed)
• Exposure Assessment (1992)
• Developmental Toxicity (1991)
• Health Assessment of Chemical Mixtures (1986) (Currently being revised for proposal in FY 1998)
• Mutagenicity (1986)
These and other guidelines are revised as new information and understanding becomes available.
The Carcinogen Risk Assessment Guidelines are a good example of the changes and developments that
have occurred in risk assessment thinking over the years. The first Carcinogen Risk Assessment
Guidelines in 1978 introduced some basic ideas: risk assessment versus risk management and hazard
identification versus dose-response assessment. The second iteration of the guidelines in 1986
incorporated the thinking of NRC's 1983 Risk Assessment in the Federal Government: Managing the
Process (the Red Book) and the Office of Science and Technology Policy's 1985 Scientific Principles of
Carcinogenesis. The 1986 guidelines provided guidance on classification for hazard identification,
approaches to dose-response assessment, and an outline of what should be covered in risk
characterization. The currently proposed guidance on carcinogen risk assessment places an emphasis on
using all the relevant biological information in the assessment. It eliminates a matrix approach to hazard
identification, expands and simplifies the discussion on dose-response assessment, and expands the
guidance on risk characterization.
4.2,2 Risk Assessment Questions
Although there are many gaps and uncertainties that exist in human health risk assessment,
the areas of primary concern chosen by ORD for this strategy are the three areas articulated in
Chapter 1. The risk assessment questions that arise as a result are presented below.
• What are the distributions of chemical exposure for children, adults, and selected vulnerable
populations, the exposure pathways and activity patterns associated with these distributions,
and the relationships and trends associated with such data?
• What and how should biological information, including information on short-term exposures,
be incorporated into qualitative and quantitative risk assessments?
• What are the factors that affect variation in exposure and variation in human susceptibility to
disease from environmental pollutants, how are such factors distributed in the population, and
how can they be incorporated into human health risk assessments?
February 1998
4-4
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
4.2.3 Risk Assessment Approach, Products, and Uses
The approaches that will be taken in response to the questions identified above, the
products of the research, and the anticipated uses are summarized in Table 4-2. Additional detail
is provided below,
4.2.3.1 Biological Measures of Exposure and Their Relationship to Human Activity Patterns,
Media, and Pathways
NHEXAS will provide information on biological assays (e.g., urine, blood, hair, nails, and
other biomarkers) for chemical exposures, human activity patterns, exposure by different media
(e.g., air, food, water, soil), and pathways (e.g., ingestion, inhalation, dermal absorption).
Analysis of these data in the coming years is expected to provide the basis for exposure
assessment guidance major pathways. Guidance also expected to be developed as a result of the
NHEXAS analyses are recommendations on more accurate and cost-effective methods of
measuring exposures (e.g., utility of cross-sectional survey data such as 24-h dietary recall, 4-day
duplicate diet, and 24-h personal air sample; types of dust sampling methods such as wipe,
vacuum, or deposition).
The primary database for developing such guidance at least in the near term, is expected to
be NHEXAS. Data from currently available and future National Health and Nutrition
Assessment Surveys (NHANES), administered by the U.S. Department of Health and Human
Services, and other surveys are expected to figure more prominently in the assessment work in
this area in the next 5 to 10 years.
4.2.3.2 Use of Biological Information in Risk Assessment
There has been a rapid increase in the understanding of the underlying biological basis of
toxicological reactions to compounds, and emerging techniques promise to fuel continued
progress. Thus, an important direction in health risk assessment is to incorporate the results of
research on biological mechanisms and toxicokinetics into the quantitative description of human
risk posed by environmental chemicals and to reduce reliance on toxicological endpoints.
Research on mechanisms is particularly important given that EPA's revised guidelines for
carcinogen risk assessment pay considerable attention to the use of mechanistic models and data
February 1998
4-5
Draft—Do Not Quote or Cite
-------
Table 4-2. Future Research Directions To Improve Human Health Risk Assessment
Risk Assessment
Questions
Approach in Response
to Question
Products
Uses
What is the baseline
multichemical exposure
distribution for children,
adults, and selected
vulnerable populations; the
exposure pathways and
activity patterns associated
with these distributions; and
the relationships and trends
associated with such data?
Analysis of large
databases on exposure
(e.g., NHEXAS,
NHANES, Department
of Agriculture
Marketbasket, FDA
Total Diet Study, North
American Free Trade
Agreement (NAFTA),
total exposure
assessment monitoring
studies pesticides and
particulate exposure,
NHAPS, etc.)
Report on population
exposure to chemicals
and the factors affecting
the distribution
Improvements in the
Exposure Factor
Handbook
Update Exposure
Assessment Guidelines
and Health Assessment
of Chemical Mixtures
Guidelines
Improved probabilistic
exposure assessment
methods derived from
field measurements of
exposure
What and how should
biological information,
including information on
short term exposures, be
incorporated into qualitative
and quantitative risk
assessments?
Analysis of scientific
literature
Revisions to risk
assessment guidelines
Prototype assessments
for chemicals for which
biological information
can improve the
assessment
Report on current
knowledge concerning
acute-to-chronic
extrapolations
Report on the use of
mechanistic information
in low-dose risk
assessments for cancer
and noncancer endpoints
Improved use of all
relevant biological
information in risk
assessment
What are the factors in
human susceptibility to
disease from environmental
pollutants? How are such
factors distributed in the
population, and how can
they be incorporated into
carcinogen and
noncarcinogen risk
assessments?
Analysis of scientific
literature, census data,
large databases on
distribution of potential
human susceptibility
factors (e.g., NHANES,
Harvard Nurses Study,
etc.)
Report on the extent of
exposures to susceptible
populations to identify
for follow-up study those
groups at increased risk
Methods and guidance
on how variation in
susceptibility and
exposure should be
factored into risk
assessments
Assessments that
evaluate the risk to
susceptible
subpopulations, as well
as to the general
population
February 1998
4-6
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
in both hazard characterization and dose-response assessment (U.S. Environmental Protection
Agency, 1996),
The pollutant exposure scenarios that EPA must assess reflect a continuum from acute
(e.g., accidental releases and spills) or intermittent bursts of exposures (e.g., during pesticide
application) to longer durations of exposure (e.g., via drinking water) that are still less than
lifetime exposures. Adverse health effects can be elicited in some cases after only a few periods
of exposure; in others, longer term exposure is required. Understanding the biological processes
involved is critical to understanding the dose-rate phenomenon. ORD will use research data on
the effects of short-term exposure and on relationships between exposure level and exposure
duration to develop guidance for assessing risk from less-than-lifetime exposures. For example,
ORD currently is developing a standard method to assess risk from short-term exposures with
regard to inhaled substances (i.e., Acute Risk Assessment Methodology for Inhaled Chemicals).
This document will include dose-response models and dosimetric considerations that readily
address models for acute exposures. The sort of work being done for inhaled chemicals will be
extended to other routes of exposure.
4.2.3,3 Variation in Human Susceptibility
Intermdividual variation in susceptibility currently is not considered in EPA's cancer risk
assessments and is addressed only in default fashion in its noncancer assessments (the variation
across the population is assumed to be 10-fold). A factor of 10 may be inadequate to protect
certain subgroups and may be too conservative in other situations. An important strategic
direction for ORD is to develop assessments, guidance, and dose-response models that
incorporate data from different subgroups (e.g., young and old, women and men, healthy and
diseased individuals, different races, different ethnic groups, different genetic profiles) and from
the variability in exposure profiles.
More specifically, risk assessment guidelines will be improved, based on research that
• validates or improves the default assumption that, on average, the general population has the
same susceptibility to that of humans in the relevant epidemiologic studies, the most sensitive
rodents tested, or both;
• assesses the need for presenting age-specific risk estimates and integrated lifetime risk
estimates; and
February 1998
4-7
Draft—Do Not Quote or Cite
-------
1 • estimates the interindividual variability in the parameters of biologically based dose-response
2 models.
3
February 1998
4-8
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
5
Science Directions for Human Health Risk
Assessment Research
The preceding chapters each focused on separate components of the research needs and
activities affecting health risk assessment but did not provide integrated perspective on the total
ORD research strategy. The purpose of this chapter is to paint a comprehensive picture of how
ORD is focusing the human health risk assessment research program on the highest priority
needs and to describe the approaches and results anticipated as the research is conducted.
Table 5-1 links the three overarching priorities identified in Section 1 to selected questions that
focus the research program. Approaches to address the key questions, which were described in
the previous chapters, result in products such as those listed in this table. Finally, die research
products and scientific capabilities resulting from ORD's program are applied through improved
methods, models, and data used by clients; through improved risk assessment guidelines for
clients; and through improved training and consultation to clients. Through these applications,
improved risk assessments for more confident risk management are the end result, as is improved
targeting of research and collection and synthesis of exposure- and health-related data and
models. Clients are becoming more numerous as EPA seeks to empower the public with usable
information. Historically, clients were primarily EPA program and regional offices. While they
still are primary clients; states, local governments, tribes, and the public are more involved in
environmental decision making which is founded on scientifically sound risk assessments.
As discussed in Section 1.1, the methodological research and measurement data obtained
through ORD's core research in human health risk assessment research program is
complemented by research results obtained through more problem-specific research (see
Figure 1-1). Further, the application of the more generic methods, models, and data generated
through this research program results in improved problem-specific risk assessments and
improved targeting of future research efforts.
February 1998
5-1
Draft—Do Not Quote or Cite
-------
Table 5-1. Summary of Priorities in Human Health Research
Key Priorities
ORD Emphasis
Example Products
Applications
Reducing
uncertainties in
exposure
measurements and
measurement-
derived models
What are the pathway-
specific measures of total
human exposure in
microenvironments
(including residences) and
populations?
What are the mathematical
relationships between
sources of contaminants,
fate pathway-specific
environmental
concentrations, total human
exposures, and dose-
estimation?
Validated residential and other
microenvironmental exposure
measurement protocols
Validated protocols for
determining human exposures
at community-to-regional
scales
A national human activity
database
Multipathway exposure
models incorporating new
measurement and activity
patterns data
Validated source-dose models
incorporating multipathway
transport and transformation
processes
Report on population
exposure to contaminants and
the factors affecting the
distribution of exposures
New exposure methods to
reduce uncertainty in
determining total human
exposure
Enhanced ability to design
and conduct future exposure
measurement studies
Improvements in Exposure
Factors Handbook
Update to Exposure
Assessment Guidelines and
Health Assessment of
Chemical Mixtures
Guidelines
Training and consultation
support to risk assessors
New measurement-derived
exposure databases arid
models to reduce uncertainty
for future exposure and risk
assessments
Prototype risk assessments to
demonstrate application of
advanced data and methods
Applying
mechanistic models
and data in hazard
characterization and
dose-response
assessment
How can the accuracy of
dose estimation across
species and exposure routes
and scenarios be improved?
How can the ability to detect
hazards be improved?
How can toxicity data to
predict and define hazards
be improved?
How can uncertainties in
extrapolations (e.g., from
high doses in animals to
environmental exposures in
humans) be reduced?
PBPK models for classes of
compounds to estimate blood
and tissue concentration and
time course
Validated biomarkers for use
in dose-response estimation
Quantitative models for
predicting toxicity resulting
from chemical exposures,
which can be modified and
applied in chemical-specific
risk assessments
Validated benchmark dose
models and guidelines for
applications
New and refined test methods
Incorporation in updates to
endpoint-specific risk
assessment guidelines (e.g.,
cancer, developmental,
reproductive, neurological)
New and revised standard
toxicity testing protocols
Validated screening
protocols using, for example,
in vivo, in vitro, and SAR
methods
Guidance documents on
interpretation of toxicity data
Prototype risk assessments to
demonstrate application and
evaluation of mechanistic
data
February 1998
5-2
Draft—Do Not Quote or Cite
-------
Table 5-1 (cont'd). Summary of Priorities in Human Health Research
Key Priorities
ORD Emphasis
Example Products
Applications
Characterizing and
assessing variation in
human exposure
and human
susceptibility to disease
What are the behavioral and
time-activity determinants
of human exposure and for
exposures to susceptible
subpopulations (e.g.,
infants, children, different
socioeconomic status,
preexisting disease)?
How can hazards be better
defined/predicted, dose-
response extrapolation be
improved, and variation
related to human
susceptibility be
characterized?
Measurement data on
multipathway exposure
(including less-than
lifetime) and time-activity
patterns for highly exposed
and susceptible populations
Report on conditions under
which there are age-
dependent quantitative and
qualitative differences in
responsiveness to pesticides
Identification of critical
genetic and biological
markers of susceptibility
Reports on enhanced
susceptibility of individuals
with pre-existing disease
conditions (e.g., COPD,
asthma, CVD) to
environmental agents and
biological mechanisms
responsible for enhanced
responsiveness
Incorporation into endpoint-
specific guidelines for risk
assessment
More accurate identification
and characterization of
highly exposed
subpopulations, where
variability plays a significant
role for exposure and risk
assessment
Completion of prototype risk
assessments to demonstrate
incorporation of new data on
human variability in
exposure and response
February 1998
5-3
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
_6_
References
. 1997, The Presidential/Congressional Commission on Risk Assessment
and Risk Management,
Carlinetal. 1991.
Duan, N., and Ott, W. 1989. "Comprehensive Definitions of Exposure and Dose to
Environmental Pollution," in Proceedings of the EPA/AWMA Specialty Conference on Total
Exposure Assessment Methodology, November, Las Vegas, Nevada.
. 1997.
Georgopoulos et al. 1997. ES&T, 31(1).
National Research Council. 1983. Risk Assessment in the Federal Government.
National Research Council. 1991. Human Exposure Assessment for Airborne Pollutants:
Advances and Opportunities, page 19.
National Research Council. 1993.
National Research Council. 1994. Science and Judgement in Risk Assessment.
National Research Council. 1997. Building a Foundation for Sound Environmental Decisions,
page 64,
Nelson, W., Ott, W.. and Robinson, J, 1994. "The National Human Activity Pattern Survey:
Use of Nationwide Activity Data for Human Exposure Assessment," 94-WA75A.01,
Proceedings of the AWMA 87th Annual Meeting, Cincinnati, OH; see also EPA/600.
Office of Technology Assessment. 1993.
Office of Science and Technology. 1985.
Ott, W., Thomas, J., Mage, D., and Wallace, L. 1988. "Validation of the Simulation of Human
Activity Pattern and Pollutant Exposure (SHAPE) Model Using Paired Days from Denver, CO,
Carbon Monoxide Field Study," Atmospheric Environment, Vol. 22, No. 10, 2101-2113.
Russell et al. 1991.
February 1998
6-1
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Sexton, K„ Selevan, S., Wagener, D., and Lybarger, J. 1992. "Estimating Human Exposures to
Environmental Pollutants; Availability and Utility of Existing Data Bases." Arc. Emir. Health.
47(6): 398-407.
Slovic, Paul. Risk Perception.
U.S. Environmental Protection Agency. 1995a. Human Exposure Assessment: A Guide to Risk
Ranking, Risk Reduction, and Research Planning. EPA Science Advisory Board Committee on
Indoor Air Quality and Total Human Exposure. EPA-SAB-1AQC-95-005, March.
U.S. Environmental Protection Agency. 1995b. Memorandum from Robert J. Huggett, Ph.D.,
Assistant Administrator for Research and Development, to Fred Hansen, Deputy Administrator,
"April Planning Forum," page 3, April 3.
U.S. Environmental Protection Agency. 1995c.
U.S. Environmental Protection Agency. 1996. Proposed Revisions to Guidelines for Carcinogen
Risk Assessment,
U.S. Environmental Protection Agency. 1997a. 1997 Update to ORD's Strategic Plan,
EPA/600/R-97/015, April, page ix.
U.S. Environmental Protection Agency. 1997b. Cumulative Risk Assessment Guidance-Phase I
Planning and Scoping, July 3 memorandum signed by EPA Administrator Carol Browner and
Deputy Administrator Fred Hansen.
U.S. Global Change Research Program. 1997.
February 1998
6-2
Draft—Do Not Quote or Cite
-------
Appendix A:
Recommendations for Strengthening Human Health
Risk Assessment in EPA
February 1998
A-1
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
During the past 5 years, a number of scientific, advisory, and legislative groups have
evaluated challenges to and strategic directions for strengthening human health risk assessment
research within EPA, A synopsis of relevant recommendations is presented below.
In 1997, The Presidential/Congressional Commission on Risk Assessment and Risk
Management recommended a new framework for risk assessment and risk management.
It stressed that "Failure to account for multiple and cumulative exposures is one of the primary
flaws of current risk assessment and risk management. Whenever possible, measurements
should be obtained to support or validate any generic values in exposure assessment, to check
modeling results or to provide more realistic estimates of exposure than can be obtained with
models,"
During 1995, in Beyond the Horizon: Using Foresight to Protect the Environmental
Future (U.S. Environmental Protection Agency, 1995c), EPA's SAB recommended that "the
Agency should place equal emphasis (with the cancer endpoint) on noncancer human health
risks; EPA should broaden its human health research and regulatory focus to include respiratory,
cardiovascular, immunologic, neurologic, and reproductive endpoints ...; EPA should continue
broadening its approach to human health risk assessment by explicitly considering risks to
susceptible populations ...and that "new dose-response models (for the noncancer endpoints)
should be considered."
In 1995, in Human Exposure Assessment; A Guide to Risk Ranking, Risk Reduction, and
Research Planning (U.S. Environmental Protection Agency, 1995a), EPA's SAB concluded that
exposure and risk assessment are hampered by persistent and "severe limitations in the currently
available exposure measurement techniques, by severe limitations of the currently available
databases containing exposure and exposure-relevant data, by reliance on numerous assumptions
which have been proven incorrect or are not supported by common experience and/or direct
observations, and by the current fragmentation and lack of coherence of available models for
different media, pathways, chemicals . . . The report recommended that EPA undertake an
extensive exposure research program, as well as a more integrated exposure, effects, and risk
assessment research program, to ameliorate these deficiencies.
During 1994, in Science, Judgment, and Risk Assessment, (National Research Council,
1994) NRC made more than 70 recommendations regarding risk assessment and risk assessment
research, including recommendations for continued research to improve cancer guidelines, risk
February 1998
A-2
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
characterization and communication, noneancer risks, uncertainty analysis, and interindividual
susceptibility to chemicals.
In 1993, the U.S. Congress's Office of Technology Assessment report, Researching Health
Risks (Office of Technology Assessment, 1993), identified several areas that hold promise for
improving risk assessment: research into new methods for toxicity studies, biomedical and
molecular epidemiology, mechanistically based effects and dose-response extrapolation methods,
improved methods for measuring or estimating human exposures, mechanistic studies of the
actions of toxic substances, attention to methods evaluation and validation, and techniques for
characterizing and communicating risks and information management.
Also in 1993, NRC issued a report, Pesticides in the Diets of Infants and Children
(National Research Council, 1993), which recommended that EPA place increased emphasis on
understanding the relationship between health effects and dietary exposures and residues in food
eaten by children, multiple pollutants with common toxic effect, and total exposure estimates that
include dietary ingestion and also account for all nondietary intake (e.g., air, dirt, indoor surfaces,
lawns).
February 1998
A-3
Draft—Do Not Quote or Cite
-------
Appendix B:
ORD Research Strategies, Priorities, and Plans
February 1998
B-1
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Introduction
Figure B-l illustrates the procedures that ORD follows to determine its strategic directions
and research priorities, to translate these priorities into detailed research plans, and to implement
these plans through its extramural Science to Achieve Results (STAR) program and its
intramural program of laboratory research. As this figure indicates, there are three steps which
are essential to these procedures.
In the first step, ORD establishes its overarching strategic directions, together with its
strategic research planning principles and ranking criteria, and identifies a number of
high-priority research areas that will receive special, expanded attention within the broad
program of research it supports. This information is discussed in detail in the ORD Strategic
Plan and the 1997 Update to ORD's Strategic Plan (U.S. Environmental Protection Agency,
1997a). The high-priority research areas that ORD has identified include;
• core research in methods, models, and approaches to advance the science of risk assessment or
risk management (i.e., research to improve human health risk assessment, research to improve
ecological health assessment, and pollution prevention research); and
• research targeted at specific problems for which EPA has legislative or regulatory responsibility
(e.g., safe drinking water, high-priority air pollutants) and at emerging scientific problems (e.g.,
endocrine disrupters).
In the second step, ORD prepares more detailed descriptions of its core and problem-
directed research strategies. This is accomplished by considering the most important scientific
questions or issues that must be addressed, as well as the scientific projects and accomplishments
that will be needed to resolve the questions or issues. ORD then solicits the widest possible
scientific review (e.g., from the EPA scientific community including program and regional
offices, and the extramural community of national scientific experts) on the appropriateness of
these strategic directions.
After integrating the recommendations from this review, ORD completes the third step of
the research planning process by developing a detailed research plan to provide guidance on
implementing future research projects. Typically, these detailed research plans are prepared by
ORD's laboratories and centers (ORD's center responsible for the STAR program develops
research plans that result in requests for assistance). The plans discuss how research will be
February 1998
B-2
Draft—Do Not Quote or Cite
-------
Stratigic
Principles
Vision
Role of
Science
at EPA
Mission
Long-Term
Goals
Objectives
Priority
Setting
Activities to
Meet the
Objectives
Gritena
Strategic
Federal/State Research Partnerships
(Interagency Agreements)
•Strategic Plans
for Core
Research
and Problem-
Driven
Research
f
Research
Plans
f
Laboratory
and
Center
Research
Plans
Budget
Operating
Plans
Peer
Review
Collaborative
Research
t
I
. ^ ,
I
r
i
i
• i- •
i
i
i
Laboratory
Implementation
Plans
Intramural
Research
0 = EPA Program and Regional Office Involvement
A = External Scientific Community Involvement
Program
Review and
Evaluation
Grants
Research
Requests for
Applications
Peer Review
of Proposals
Award of
Grants
Figure B-1. Implementing ORD's strategic plan.
1 implemented, identify expected outcomes or scientific contributions, and explain provisions for
2 accountability.
February 1998
B-3
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Establishing a Partnership To Identify and Focus EPA's Diverse Needs for Science and
Research
Perhaps the most challenging aspect of this process is creating a consolidated research
agenda that meets the needs of ORD's diverse clients. The magnitude of this challenge was
illustrated in 1995 when ORD conducted a 4-month assessment to document research needs and
priorities identified by all parts of EPA. This assessment identified literally thousands of
needs—far more than could be accommodated through years of effort by ORD's entire staff and
research budget. The review also indicated that, while ORD's research priorities generally
respond well to the highest priority scientific problems identified by individual EPA client
offices, it was very difficult to fashion an agency-wide agreement on a single, consolidated
research agenda that would strengthen the scientific foundation for all of EPA's programs. This
difficulty stems in part from EPA's science and research requirements and also from the
substantial differences in legislative and regulatory mandates that EPA's program offices and
regions are responsible for implementing. Recognizing the importance of this challenge, ORD
has created an objective and inclusive research planning process which is described below.
• This process engages all parts of EPA in helping to identify and describe potential research
priorities. Members of the Research Coordination Teams (RCTs) (which consist of senior
representatives from ORD's national laboratories and centers as well as from EPA's program
and regional offices), the Research Coordination Council (which consists of the assistant
administrators, regional administrators, or their designated senior representatives), and the
Science Council (which consists of the associate directors from each ORD laboratory and
center) each identify important and relevant environmental research needs for consideration.
In addition, ORD solicits recommendations from EPA's extramural scientific advisors
(e.g., SAB, NRC) about strategic scientific directions and priorities for ORD's research.
• The process empowers the EPA-wide representatives on the RCTs to narrow the pool of
potential research needs by identifying those that are considered essential to strengthen EPA's
scientific foundation and to enable it to respond to legislative mandates and regulatory issues.
Subsequently, the RCTs define the components of these essential research needs by identifying
the scientific questions or issues that must be addressed to reduce uncertainty in each element
of the risk assessment or risk management paradigm. This step results in a series of research
activities that correspond to identified scientific questions and research needs.
February 1998
B-4
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
U
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
• This planning process enlists the expertise of the RCTs and the Research Coordination Council
to recommend a consolidated research agenda. These two groups evaluate and rank research
activities through the application of a series of risk assessment, methods/models, and risk
management criteria. These criteria (Figure B-2) are designed to identify the most pressing
problems; assess the potential for each research area to support effective risk assessment, risk
management, or risk reduction; and ascertain research areas where ORD's scientific capability
can make a significant contribution.
Ail Example of the Priority-Setting Process: Establishing Research Priorities for Human
Health Risk Assessment
Before evaluating the significance and potential priority for research in human health risk
assessment, ORD's RCTs considered recommendations from several sources. For this particular
research area, the recommendations from scientists in program offices and regions, in ORD
laboratories, and on extramural advisory boards all underscored the fundamental need for more
scientifically defensible methods, measurement databases, models, and risk assessment protocols.
Based on this clear consensus, the research coordination teams used the methods and models
criteria (Figure B-2) to evaluate the significance of the research needs for human health risk
assessment. Based on this evaluation, ORD's RCTs ranked the need for future research in this
area as one of the highest. When these 31 research project areas were disaggregated into their
constituent future research activities, all human health risk assessment research activities ranked
in the highest priority tier of potential future research. The methods and models criteria
considered the potential applicability of health risk assessment research, the potential
contribution that future research would make to improving the science, and the size or extent of
the community that would use or benefit from the research. Application of these criteria
indicated the following.
(1) Core research in human health risk assessment would have broad applicability. One of the
reasons for the broad applicability of this research is that virtually all of EPA's major
legislative mandates (those which require EPA to promulgate regulations to protect the
public health from environmental contaminants) require EPA to develop human health risk
assessments. These include the Clean Air Act, the Safe Drinking Water Act, the
February 1998
B-5
Draft—Do Not Quote or Cite
-------
Setting Priorities for
Effects, Exposure, and
Assessment Research
More
Un
certainty
Reductio
Narrow Applicability
Broad Applicability
Less
Uncertainty
Re-
ductio
Setting Priorities for
Methods and Models
Research
Options
Too
Costly,
Inefficlen
Risk Problem
Well Characterized
Risk Problem
Poorly Characterized
Options
Already
Optimize
Setting Priorities for
Risk Management
Research
Figure B-2. Criteria for setting research priorities.
Source: Adaped from Paul Slovic, Risk Perception.
February 1998
B-6
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Clean Water Act, TSCA, FIFRA, the Resources Conservation and Recovery Act, the Superfund
Amendments Reauthorization Act, and the Food Quality Protection Act. In addition, in 1988,
Congress enacted legislation that mandated EPA to undertake research to improve health risk
assessment,
(2) Core research in human health risk assessment would reduce significantly the uncertainties
for EPA-wide risk assessment and risk management. These research outputs would reduce
significant uncertainties in the ability to quantify, model, and assess human exposures,
exposure-dose-response relationships, and risk from environmental contaminants at
community-to-regional geographic scales. Moreover, the research outputs would provide the
first measurement-derived models for multipathway risk assessment and risk management
decisions. In the absence of improved risk assessment and risk management methods,
models, and measurement databases, "... enormous sums of money that might be better
spent elsewhere may be allocated to dealing with perceived risks. While it is essential to
ensure public health and environmental integrity, limited resources reinforce the need to
assess risks as accurately as possible .... Estimates have indicated that the cost of
environmental regulations in the United States will total between $171 and $185 billion by
the year 2000 (Carlin et al., 1991). Compliance with air pollution control regulations will
cost an estimated $94 billion per year by the year 2000 (Carlin et al., 1991). Russell et al.
(1991) estimated that cleaning up all the major hazardous waste sites would cost between
$500 billion and $1 trillion over the next 30 years. The sums are enormous, and
a convincing analysis must be provided to demonstrate that these expenditures are justified
as the most cost-effective way to reduce risks to human health and to the environment"
(National Research Council, 1997).
(3) Core research in human health risk assessment would benefit and be used by a large and
diverse constituency. During the 1995 base review of all ORD research and client office
needs, research in human health risk assessment was identified as one of the most important
research needs across all EPA regional and program offices. When considered from the
broader national perspective, the 1993 study of health risk assessment conducted by the
Office of Technology Assessment (1993) demonstrated that the benefits of research in this
area would extend substantially beyond EPA's research, regulatory, and regional offices.
In addition to EPA and other federal agencies, the "user community" would include states,
February 1998
B-7
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
the private sector, academic research organizations, Congress, and international
environmental health organizations.
Developing a Focused Research Agenda
In summary, once ORD has identified its broad strategic directions, two subsequent levels
of planning are essential to the development of a focused research agenda. Typically, each step is
accompanied by the development, review, and publication of an ORD document.
Research strategies frame the scientific questions associated with these research areas, and
explain the direction, priorities, and outcomes of future research programs required to respond to
these questions. All parts of EPA are involved through ORD's research coordination teams in
helping define and describe research strategies. These strategies then undergo an external
scientific peer review. Thus, research strategy documents present research goals that have been
reviewed by the broad EPA community, by the extramural research community, and by scientists
in ORD laboratories.
Within ORD's national laboratories and centers, the strategy provides senior scientists and
research managers with a "blueprint" for designing and implementing research programs for a
5- to 10-year time frame. In addition, the research strategy enables ORD staff to relate their
individual research projects to ORD's strategic goals. For ORD's many stakeholders (e.g.,
EPA's program offices, regional offices, academia, other government agencies, the public),
a research strategy identifies the future directions, priorities, and scientific outcomes that can be
used to measure the focus and progress of environmental research.
This "blueprint" is used to develop more detailed and narrowly focused laboratory/center
implementation plans within ORD's national laboratories and centers that describe in detail the
research projects, outcomes, and outputs that will be produced to accomplish the strategic goals
or outcomes. Table B-l lists and briefly describes the research strategies and plans that ORD is
preparing during 1997 and 1998.
February 1998
Draft—Do Not Quote or Cite
-------
Table B-1. ORD Research Plans and Strategies
Title
Short Synopsis of Document's Focus
Microbes/Drinking
Water Disinfection
The continued occurrence of waterborne disease outbreaks demonstrates that drinking
water contaminated with bacteria, viruses, and parasites still poses a serious health risk
when treatment is inadequate. A large number of DBPs have been identified that result
from the disinfection of drinking water source waters. These DBPs have the potential to
cause adverse health effects in the exposed public. The key areas of research will focus on
assessing the health effects from exposure of waterborne pathogens and DBPs; the
assessment of the potential exposures of pathogens and DBPs in various U.S. populations,
especially in susceptible populations; assessing the risks from pathogen and DBP
exposures and comparing the trade-offs between risks; and determining cost-effective
technologies to treat source waters to achieve low-pathogen and DBP concentrations in
final consumer drinking water.
Particulate Matter
The overarching mission of EPA's PM research program is to provide an improved
scientific basis for future regulatory decisions concerning public health risks posed by
airborne particles (emphasizing fine particle [PM2 5] risks). The areas of PM health effects
research that need to be addressed to effect these decisions and implementation activities
are threefold: (1) development of a more complete interpretation of the PM epidemiologic
data; (2) an understanding of the biological mechanisms of PM to explain the observed
effects, the reported independence of effects from particle composition, and the lack of an
obvious threshold for effects (i.e., every exposure concentration may cause an effect in
some individuals in the population); and (3) an understanding of the composition, size,
physical properties, and sources of PM that may cause health effects.
Arsenic in
Drinking Water
The current arsenic drinking water maximum contaminant level (MCL) is 50 pg/L and was
set in 1942 by the Public Health Service. This MCL is not based on health risk
assessments as MCLs now are. The key areas of research will focus on the development
of cost-effective arsenic control technologies for small drinking water systems;
development and validation of analytical methods to speciate arsenic in water, soils, foods,
and biological tissues; assessment and risk characterization of human and animal studies
for arsenic exposures; and effects research on cancer and noncancer health effects,
mechanisms of action, and human susceptibility.
Endocrine
Disrupters
At present, the hypothesis that endocrine disrupting chemicals are causing adverse health
in wildlife and humans remains simply an intriguing hypothesis. Most of the knowledge
and concerns to date have risen from situations with relatively high-level exposure to
persistent organic pollutants or therapeutic use of pharmacological agents. For proper
regulatory action to occur, the understanding of the potential scope of endocrine disruption
in humans and wildlife must be increased to include defining the range of health effects,
critical life stages, sensitive species, and exposures relevant to alterations in endocrine
function, and developing risk management options to reduce or prevent additional adverse
effects in populations.
EMAP
This program develops the science of measuring ecosystem health and of monitoring the
condition and trends of natural resources at the regional scale. Using the Committee on
Environment and Natural Resources National Monitoring Framework and interagency
workgroups as guides, EMAP supports complementary intramural and extramural (STAR)
research programs to develop more cost-effective ecological indicators and to design
multiple-tier monitoring methods capable of detecting trends and associating ecological
impacts with likely stressors. The indicators and monitoring designs intended to support
state-, regional-, and national-level environmental report cards encompass multiple
stressors and many resource classes such as estuaries, streams, lakes, wetlands, forests,
and grasslands.
February 1998
B-9
Draft—Do Not Quote or Cite
-------
Table B-1 (cont'd). ORD Research Plans and Strategies
Title
Short Synopsis of Document's Focus
Human Health
Risk Assessment
Virtually all environmental legislation enacted by Congress requires EPA to conduct
human health risk assessments to ensure a strong scientific foundation for decisions about
the need for environmental regulations to protect human health and welfare. In recent
years, increasingly complex environmental and human health issues have challenged EPA
to develop more sophisticated regulations. At the same time, however, national scientific
advisory panels have voiced increasingly strong concerns about the scientific adequacy of
EPA's human health risk assessments. Responding to these concerns, ORD has developed
a human health risk assessment research program that integrates the expertise of scientists
in human exposure, dose-response, health effects, and risk assessment. This document
describes the strategic directions and priority research objectives for this ORD research
program during the next 10 years and explains how this strategy will respond to the key
recommendations from EPA's scientific advisory panels. Specific research priorities
discussed in the document include reducing uncertainties in exposure measurements and
measurement-derived models, applying mechanistic models and data to reduce uncertainty
in hazard identification and dose-response assessment, and characterizing and assessing
variation in human exposure and susceptibility to disease.
Ecosystems
Protection
In virtually every major environmental act, Congress has required EPA to protect human
health as well as the environment. This document provides the strategic direction and
priority research objectives for the ORD's Ecological Research Program. The goal of the
program is to provide the scientific understanding required to measure, model, maintain, or
restore, at multiple scales, the integrity and sustainability of ecosystems now and in the
future. Fundamental research areas include monitoring, modeling, assessment,
remediation, and restoration. Specific problems of importance discussed in the document
include ecological research on ozone, acid deposition, ecocriteria, wet weather flow,
pesticides, hazardous waste, global change, endocrine disrupters, ultraviolet-B radiation,
contaminated sediments, exotic species, habitat alteration and restoration, and regional risk
assessment.
Global Change
Based on the findings of the Intergovernmental Panel on Climate Change, guidance in
ORD's strategic plan, and the priorities specified in Our Changing Planet (U.S. Global
Change Research Program, 1997), ORD will strategically invest in global change research.
ORD's Global Change Research Program will focus on ecological vulnerabilities of
ecosystems to climate change, the implications for human health, and mitigation and
adaptation approaches. The research conducted will provide policy makers with
information on potential ecological and human health consequences of climate change and
technical data needed to evaluate alternative GHG emission reduction and adaptation
approaches.
Pollution
Prevention
For pollution prevention to be a success, all stakeholders (e.g., regulators, industry,
environmental groups) must have access to scientifically sound pollution prevention
technologies and approaches. They must also be able to measure and objectively evaluate
the viability and comparative environmental performance of these pollution prevention
technologies and approaches. There is a lack of user-friendly tools and methods to
compare pollution prevention solutions with each other and to end-of-the-pipe solutions,
and there is also a lack of proven pollution prevention technologies and approaches for
many pollutant sources in a number of economic sectors. Research is being undertaken in
pollution prevention to address fundamental knowledge gaps in both of these areas.
February 1998
B-10
Draft—Do Not Quote or Cite
-------
Table B-1 (cont'd). ORD Research Plans and Strategies
Title
Short Synopsis of Document's Focus
Waste
The goal of the ORD Waste Research Strategy is to set forth an effective research
program to understand and reduce human and ecological exposure to toxic materials
released during waste management and to assess and remediate contamination that has
occurred because of improper waste management. Focus is directed toward research on
groundwater, soils, and the vadose zone at contaminated sites; active waste management
facilities; and emissions from waste combustion facilities. Associated technical support
activities to assist EPA program offices and regions and other stakeholders also are
described.
February
1998
B-11
Draft—Do Not Quote or Cite
-------
Appendix C:
The Impact of Legislation and Regulation on Human
Health Risk Assessment Research
February 1998
C-1
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
The past 25 years have witnessed the enactment of a series of legislative mandates that
require EPA to protect the public health and welfare from environmental contaminants. In the
aggregate, this body of legislation mandates that EPA assume responsibility for conducting
research, developing human health risk assessments, and establishing regulations and standards
in all of the following areas.
• Clean Air. One section of the Clean Air Act mandates National Ambient Air Quality
Standards (NAAQS) to protect the public health and welfare from criteria pollutants. State and
federal air pollution programs are required to establish air pollution regulations that maintain
air quality levels at or below the NAAQS levels. Other sections establish national standards
for emissions of hazardous air pollutants from stationary and mobile sources that are hazardous
to human health and require evaluation of public health risk from exposure to urban air toxics.
The act authorizes EPA to conduct extensive research into the causes, effects, and extent of air
pollution.
• Drinking Water. One section of the Safe Drinking Water Act establishes standards for
drinking water quality known as maximum contaminant levels, which are based on human
health endpoints. Recent changes to this legislation call for investigation of human exposure
and health effects from drinking water contaminants such as disinfectant by-products,
microbes, and endocrine disrupting compounds.
• Clean Water. The Clean Water Act requires EPA to develop ambient pollutant limits for
surface waters and groundwater based, in part, on consideration of human health endpoints.
Regulations for disposal of sludge are based on an assessment of health risks. The act
authorizes EPA to conduct research on the harmful effects of water pollutants on human health
and welfare.
• Toxic Substances. The Toxic Substances Control Act (TSCA) requires industry to submit
exposure and health data that are used to determine whether to implement restrictions on the
manufacture, use, or disposal of toxic chemicals. The act authorizes EPA to conduct research
to develop techniques to screen and test for human health and ecological effects of chemical
substances and mixtures.
• Pesticides. The Federal Insecticide, Fungicide, and Rodenticide Act (FTFRA) requires the
collection, review, and evaluation of toxicity and other health-related data to assess the effects
February 1998
C-2
Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
of pesticide products. The act authorizes EPA to conduct research to ensure implementation of
its provisions.
• Hazardous Waste. The Resource Conservation and Recovery Act requires the evaluation of
toxicity and other health-related data to determine which wastes are considered to be
hazardous. Regulations for facilities that accept waste are designed to protect the health of
residents near disposal sites. The act authorizes EPA to conduct research on the adverse health
and welfare effects of solid and hazardous wastes.
. Superfund Waste Sites, the Superfund Amendments Reauthorization Act requires emergency
response and cleanup actions that are designed to protect the health of populations near waste
sites. The act authorizes EPA to conduct research to detect, assess, and evaluate the effects of
hazardous substances on human health.
• Food Quality. The recently enacted Food Quality Protection Act of 1996 requires
consideration of food consumption patterns, pesticide residues, human exposure and effects
data, data on susceptible subpopulations, analysis of cumulative risk, potential effects from
endocrine disrupters, and risk communication techniques, all with an emphasis on protecting
infants and children.
In promulgating regulations that implement the numerous provisions of this legislation,
EPA employs the human health risk assessment paradigm wherever it is scientifically relevant
and authorized by statute to do so. Collectively, however, the scientific burden imposed on the
human health risk assessment community by this legislation is significant, and it provides no
"lowest common denominator" for human health risk assessment or regulatory decision making
across EPA.
A related challenge for human health risk assessment is that, with limited exceptions (e.g.,
the Food Quality Protection Act of 1996), the body of legislation directs EPA to regulate one
pollutant at a time, often from a single source and from one environmental pathway. The result
inhibits EPA from considering human exposure to the same pollutant from different sources or
environmental pathways, even where the cost or effectiveness of alternative risk management
options may be significantly lower. This legislative focus on single sources, pollutants, and
media has inhibited research to develop multimedia and multistressor risk assessment methods
that are needed to investigate the complex environmental and human health issues present on
community, regional, national, and international scales today. Another result of this
February 1998 C-3 Draft—Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
"one-size-fits-all" approach is that EPA has had very limited ability to develop risk management
options that are flexible in their focus (e.g., on infants and children, cumulative risk, and specific
geographic regions).
Within the past year, EPA has initiated a new policy that recognizes the importance of
multimedia risk assessment, flexibility in developing risk management options, and community-
to-regional-scale issues for its stakeholders. This policy directs each EPA program to invest a
portion of its resources in community-to-intefnational-scale environmental issues and affords an
opportunity for ORD's human health research program to develop, demonstrate, and provide
protocols to strengthen multimedia risk assessment methods, as well as to sponsor environmental
health studies at community-to-international scales.
February 1998
C-4
Draft—Do Not Quote or Cite
------- |