EPA/600/R-06/013F
August 2007
Concepts, Methods and Data Sources for
Cumulative Health Risk Assessment of Multiple
Chemicals, Exposures and Effects: A Resource
Document
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
in collaboration with
U.S. Department of Energy Argonne National Laboratory
Environmental Assessment Division
Argonne, IL 60439
-------
NOTICE
The U.S. Environmental Protection Agency through its Office of Research and
Development conducted, and funded research described here under two collaborative
Interagency Agreements, Numbers DW89921662 and DW89939210, with the
Department of Energy from 2001 to 2006. It has been subjected to the Agency's peer
and administrative review and has been approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
ABSTRACT
Public interest in the health impacts of environmental chemical exposures and
their interactions with other stressors continues to grow with increased information
about exposures to multiple chemicals in air, water and soil from different sources.
However, population vulnerability factors, such as diet, behaviors, genetic traits,
economic status and social characteristics are often not considered. Cumulative risk
assessment may be thought of as a population-based analysis, characterization and
possible quantification of the combined risks to health or the environment from multiple
route exposures to multiple agents or stressors. This current report serves as a
resource document for identifying specific elements of and approaches for implementing
cumulative risk assessments. This report is not a regulatory document and is not
guidance but rather a presentation of concepts, methods and data sources. It is
designed to assist EPA's development of specific approaches and cumulative risk
guidance for use by its Program Offices and Regions. It is intended as a resource for
EPA scientists and others in the broader risk assessment community with an interest in
locating data and approaches relevant to cumulative risk assessment. This report
focuses on two areas: initiating factors for a cumulative risk assessment with
procedures for data collection and organization; and technical approaches for assessing
and characterizing human health risks associated with a subset of cumulative risk
issues (i.e., multiple chemicals, exposures and effects). Schematics are shown for
evaluating data, profiling the population of concern, grouping chemicals into integrated
exposure and toxicity groups, performing toxicity assessments and conducting
cumulative risk characterizations. Issues discussed include toxicological interactions,
pharmacokinetics, multiple toxic effects, epidemiologic methods, biomonitoring data, the
temporal nature of exposures and environmental chemical transformations. Articulation
of variability and uncertainty is stressed as part of the final Risk Characterization.
Preferred citation:
U.S. EPA. 2007. Concepts, Methods and Data Sources for Cumulative Health Risk Assessment of
Multiple Chemicals, Exposures and Effects: A Resource Document. U.S. Environmental Protection
Agency, National Center for Environmental Assessment, Cincinnati, OH. EPA/600/R-06/013F.
-------
TABLE OF CONTENTS
LIST OF TABLES viii
LIST OF FIGURES ix
LIST OF TEXT BOXES xii
LIST OF ABBREVIATIONS xiv
PREFACE xvii
AUTHORS, CONTRIBUTORS AND REVIEWERS xviii
EXECUTIVE SUMMARY xxi
1. CUMULATIVE RISK ASSESSMENT INTRODUCTION 1-1
1.1. BACKGROUND 1-1
1.1.1. The Integrated Process for Cumulative Risk Assessment 1-1
1.1.2. Terminology 1-4
1.2. ABOUT THIS REPORT 1-6
1.2.1. Innovations Included in this Report 1-8
1.3. EXISTING EPA PUBLICATIONS RELATED TO CUMULATIVE
RISK 1-10
1.3.1. EPA Guidance Documents 1-12
1.4. THIS REPORT'S APPROACH TO CUMULATIVE RISK
ASSESSMENT 1-17
1.4.1. Technical Approaches for Multiple Chemicals,
Exposures and Effects 1 -22
1.4.2. Identify the Initiating Factor for the Cumulative Risk
Assessment 1 -22
1.4.3. Characterize the Community and Population Based on
the Initiating Factor 1 -23
1.4.4. Generate Initial List of Relevant Chemicals 1-24
1.4.5. Identify Links between Chemicals and Subpopulations 1-25
1.4.6. Quantify Human Exposures for Initial Exposure Grouping 1-27
1.4.7. Quantify Dose-Response for Initial Toxicity Grouping 1 -28
1.4.8. Integrate Exposure and Dose-Response Information 1-28
1.4.9. Conduct Risk Characterization 1 -28
1.8. SUMMARY 1-29
-------
TABLE OF CONTENTS cont.
2. IDENTIFICATION OF INITIATING FACTORS, POPULATION
CHARACTERISTICS, DATA COLLECTION AND ORGANIZATION 2-1
2.1. INITIATING FACTORS FOR CUMULATIVE RISK ASSESSMENT 2-1
2.1.1. Health Endpoint as the Initiating Factor 2-1
2.1.2. Chemical Concentrations as the Initiating Factor 2-3
2.1.3. Multiple Sources or Release Events as the Initiating Factor 2-4
2.2. INITIAL DESCRIPTION OF THE POPULATION 2-5
2.2.1. Preliminary Characterization of the Population Based
on the Initiating Factor 2-6
2.2.2. Refining the Population Profile Based on Vulnerable
Subpopulations 2-7
2.3. INITIAL ASSESSMENT OF EXPOSURE DATA 2-9
2.3.1. Initiating the Exposure Assessment when Health
Endpoint is the Initiating Factor 2-8
2.3.2. Initiating the Exposure Assessment when Elevated
Chemical Concentrations are the Initiating Factor 2-14
2.3.3. Initiating the Exposure Assessment when One or More
Sources is the Initiating Factor 2-15
2.3.4. Summary 2-16
2.4. INTEGRATION OF PUBLIC HEALTH INFORMATION 2-17
2.5. EPIDEMIOLOGIC INVESTIGATIONS IN CUMULATIVE RISK
ASSESSMENT 2-18
2.6. LINKING THE LIST OF RELEVANT CHEMICALS TO THE
POPULATION PROFILE THROUGH A CONCEPTUAL MODEL 2-20
3. EXPOSURE ASSESSMENT OF MULTIPLE CHEMICALS,
EXPOSURES AND EFFECTS 3-1
3.1. DEFINING EXPOSURE ASSESSMENT FOR CUMULATIVE
RISK ASSESSMENTS THAT EVALUATE MULTIPLE
CHEMICAL EXPOSURES 3-1
3.2. U.S. EPA EXPOSURE ASSESSMENT GUIDANCE 3-4
3.3. CUMULATIVE EXPOSURE ASSESSMENT: ANALYSIS PHASE 3-7
IV
-------
TABLE OF CONTENTS cont.
3.3.1. Exposure Setting 3-9
3.3.2. Exposure Pathways and Routes 3-14
3.3.3. Exposure Quantification 3-48
3.4. ILLUSTRATION OF CUMULATIVE CONCEPTS FOR THE AIR
PATHWAY AT A CONTAMINATED SITE 3-69
3.4.1. Em ission Inventories 3-71
3.4.2. Dispersion Modeling 3-74
3.5. RETROSPECTIVE STUDIES 3-80
3.6. SUMMARY COMPARISON AND SCREENING SUGGESTIONS 3-81
4. TOXICITY ASSESSMENT OF MULTIPLE CHEMICALS, EXPOSURES
AND EFFECTS 4-1
4.1. DEFINING CUMULATIVE TOXICITY ASSESSMENT 4-2
4.2. TOXICITY ASSESSMENT GUIDANCE AND METHODS 4-2
4.2.1. Practices for Evaluation of Toxicity for Various Durations 4-3
4.2.2. Practices for Evaluating Chemical Mixtures 4-4
4.2.3. Old, New and Enhanced Approaches for Cumulative
Toxicity Assessment 4-10
4.3. TOXICOLOGY OF INTERNAL CO-OCCURRENCE 4-13
4.3.1. Use of Internal Doses in the Hazard Index 4-18
4.4. CHEMICAL MIXTURES GROUPING AND TOXICITY
ASSESSMENT SCHEME 4-18
4.4.1. Chemical Groupings by Common Effects 4-23
4.4.2. Refinement of Toxicity Groups 4-26
4.4.3. Cumulative Toxicity Assessment Scheme 4-28
4.4.4. Evaluating Subpopulations 4-31
4.5. EVALUATING MULTIPLE EFFECTS 4-31
4.5.1. A Quantitative Method for Evaluating Multiple Effects 4-34
4.5.2. Interpretation 4-40
-------
TABLE OF CONTENTS cont.
4.6. EVALUATING INTERACTION EFFECTS 4-42
4.6.1. Toxicology of Interactions 4-42
4.6.2. A Quantitative Method for Evaluating Interaction Effects 4-44
4.7. EVALUATING MULTIPLE ROUTE EXPOSURES 4-47
4.7.1. Quantitative Approaches to Evaluating Multiple Route
Exposures to Mixtures 4-48
4.7.2. Internal Dose Estimates 4-52
4.8. SUMMARY RECOMMENDATIONS 4-53
5. CUMULATIVE RISK CHARACTERIZATION 5-1
5.1. CHARACTERIZATION OF RISK IN A CUMULATIVE RISK
ASSESSMENT CONTEXT: AN OVERVIEW 5-2
5.2. SPECIAL CONCERNS WITH CUMULATIVE RISK
CHARACTERIZATION 5-3
5.3. A RISK CHARACTERIZATION PROCESS FOR CUMULATIVE
RISK ASSESSMENT 5-9
5.3.1. Populations 5-9
5.3.2. Temporal Analysis 5-11
5.3.3. Integrative Cumulative Risk Assessment 5-12
5.3.4. Evaluation of Extrapolations, Simplifications and Omissions 5-13
5.3.5. Develop Quantitative Risk Estimates 5-13
5.3.6. Sensitivity and Uncertainty Analysis 5-13
5.3.7. Cumulative Risk Characterization 5-14
5.4. VARIABILITY AND UNCERTAINTY IN EXPOSURE AND
DOSE-RESPONSE 5-16
5.4.1. Usefulness of Variability and Uncertainty Analyses in
Cumulative Risk Assessments 5-16
5.4.2. Exposure Assessment Uncertainty and Variability 5-17
5.4.3. Uncertainty and Variability in Dose-Response Assessment
for Cumulative Risk Assessment 5-20
5.4.4. Variability and Uncertainty Summary 5-26
VI
-------
TABLE OF CONTENTS cont.
5.5. EXAMPLE EVALUATIONS OF QUANTITATIVE APPROACHES
TO CUMULATIVE RISK CHARACTERIZATION 5-26
5.5.1. Example Cumulative Risk Characterization: Cumulative
Hazard Index 5-27
5.5.2. Categorical Regression Calculations for Multiple Effects
and Pathways 5-31
5.5.3. Assumptions with Multi-route Formulas for Multiple Effects 5-33
5.5.4. Combination of Exposures of Different Time Frames 5-33
5.6. OUTCOMES FROM CUMULATIVE RISK CHARACTERIZATION 5-34
5.6.1. Interpretation of Results in the Context of Interaction Factors... 5-35
5.6.2. Interpretation of Results in Context of Problem Formulation 5-37
5.6.3. Interpretation of Results in Context of the Initiating Factor 5-37
5.7. SUMMARY 5-38
6. REFERENCES 6-1
7. GLOSSARY 7-1
APPENDIX A: CUMULATIVE RISK TOOLBOX A-1
APPENDIX B: TOXICITY INFORMATION TO SUPPORT GROUPINGS B-1
APPENDIX C: SEVERITY OF TOXIC EFFECT C-1
VII
-------
LIST OF TABLES
No. Title
2-1 Examples of Illnesses Linked to Multiple Environmental Factors 2-11
3-1 Properties of Selected Organic Chemicals and Degradation Products to
Demonstrate Availability of Information 3-17
3-2 Grouping Chemicals by Common Migration Behavior 3-30
3-3 Grouping Chemicals by Environmental Fate Measures 3-31
3-4 General Grouping Categories for Key Fate Parameters 3-36
3-5 Specific Parameter Values for Example Chemicals 3-37
3-6 Summary Comparison and Screening Suggestions 3-39
3-7 Example Groupings Based on Exposure Considerations (Media
and Timing) 3-41
3-8 Example of Cumulative Exposures for Current Land Use 3-51
3-9 Example of Cumulative Exposures for Future Land Use 3-54
4-1 Selected Reference Values for Different Exposure Durations 4-5
4-2 Severity Assignments for Cholinesterase Inhibition Data 4-36
4-3 Frequency of Categories of Effect Associated with Aldicarb Exposure
in Humans 4-37
4-4 Modeled Probabilities of an Adverse or Frank Effect 4-39
4-5 Joint Toxicity: Non-additive Effects of Metal Pairs on Systems/Organs
Using Oral Exposure 4-41
4-6 Default Weighting Factors for the Modified Weight of Evidence 4-46
5-1 Joint Toxicity: Summary of Pairwise Toxic Interactions by Organ/System 5-36
VIM
-------
LIST OF FIGURES
No. Title
1 -1 Integrated Process for Cumulative Risk Assessment 1-3
1-2 Assessing Integrated Multiples in Cumulative Risk Assessment: Focus
on Human Health 1-8
1-3 Key EPA Resources for this Report: Precedent U.S. EPA Guidance and
Reports Containing Specific Approaches for Assessing Major Parts of
Cumulative Health Risks 1-11
1-4 Highlights of Recent Cumulative Risk-Related Program Guidance and
Research Reports 1-13
1-5 Common Initiating Factors and Elements of Cumulative Assessments 1-18
1 -6 Variables Considered in Cumulative Risk Assessment and their
Relationship to Initiating Factors 1-19
1-7 Key Steps in a Cumulative Risk Assessment 1-21
1-8 Schematic of Cumulative Risk Characterization Approach in this Report 1-30
2-1 Example Initiating Factors and Data Elements for Cumulative Risk Analyses... 2-2
2-2 Conceptual Models for All Initiating Factors 2-22
2-3 Example Second-Tier Detail of the Analytical Approach for Health Risk 2-23
3-1 Conceptual Model for Hypothetical Cumulative Exposure Assessments
Illustrating Pathways Considered and Complete Pathways 3-8
3-2 Illustration of Global Background from Atmospheric Fallout of Tritium 3-26
3-3 Approach for Estimating Exposure in Cumulative Risk Assessments 3-29
3-4 Assessing Relative Mobility in Soil to Support Chemical Groupings 3-34
3-5 Example Changes in Exposure Profile from Degradation and Partitioning 3-44
3-6 Illustration of Changing Media Concentrations Affecting Potential
Exposures 3-45
IX
-------
LIST OF FIGURES cont.
No. Title
3-7 Ten Steps in Perform ing Aggregate Exposure and Risk Assessment 3-56
3-8 Pathway-Specific and Combined Exposure to a Single Hypothetical
Chemical 3-57
3-9 Dose Metrics for Environmental Contaminants 3-62
3-10 Linking Exposure Assessment Modeling with a PBPK Model for DBPs 3-63
3-11 Levels of Dose Specificity that can be Estimated in a Cumulative
Exposure Assessment 3-64
3-12 Multipathway Potential Doses and Target Organ Doses 3-66
3-13 Human Residence Time for Selected Contaminants 3-67
3-14 Conceptual Illustration Showing the Persistence of a Biological Effect
Exceeds the Duration of the Exposure 3-68
4-1 Approach for Assessing Mixtures Based on Available Data 4-6
4-2a Flow Chart Showing Approaches for Evaluating Whole Mixtures 4-11
4-2b Flow Chart Showing the Component Based Approaches for
Evaluating Multiple Chemicals, Exposure Routes, Effects and
Toxicological Interactions 4-12
4-3 Level of Specificity for Dose-Response Relationships 4-14
4-4 Human Residence Time for Selected Contaminants 4-16
4-5 Conceptual Illustration of Persistence of Mixture Components 4-17
4-6 Conceptual Illustration of Effects of Metabolism on Toxicity 4-17
4-7a Chemical Grouping by Co-occurrence in Media and Time 4-20
4-7b Chemical Groupings by Common Target Organs and Effects 4-20
4-7c Grouping Chemicals for Cumulative Risk Assessment 4-21
4-7d Grouping Chemicals for Cumulative Risk Assessment (cont.) 4-22
-------
LIST OF FIGURES cont.
No. Title
4-8 Information on Primary and Secondary Effects Linked with Hypothetical
Exposure Sources to Show Example Chemical Groups 4-24
4-9 Hypothetical Example of Chemical Groupings by Co-occurrence in Media
and Time, Similar Toxicity 4-25
4-10 Examples of Toxicity Group Refinements 4-29
4-11 Complex Mixture Reference Dose 4-32
4-12 Schematic for Relative Potency Factor Approach 4-49
4-13 Combining Grouped RPF Estimates Across Exposure Routes 4-51
5-1 Schematic of Cumulative Risk Characterization Approach in this Report 5-10
XI
-------
LIST OF TEXT BOXES
No. Title
1-1 Summary of Traditional Risk Assessment Paradigm 1-2
1 -2 Key Terms for Cumulative Health Risks 1-4
1-3 Challenges to Conducting Cumulative Risk Assessments 1-17
1-4 Chemical and Stressor Involvement in Cumulative Risk Assessment 1-24
2-1 Example of Pesticides and Farmer Characteristics 2-9
2-2 Example of Illness Initiating Factor from Pesticide Incident 2-10
3-1 Cumulative Exposure Assessment Questions 3-1
3-2 Selected Information Guides 3-4
3-3 Exposure Assessment: Analysis Steps 3-9
3-4 Example Data Sources and Uses 3-10
3-5 Information for Susceptibility Assessment 3-12
3-6 Exposure Pathway Elements 3-14
3-7 Illustration of Groupings Based on Properties and Fate 3-35
3-8 Example of Possible Release Sources 3-40
3-9 Weathering Example: Toxaphene 3-46
3-10 Chemical Groupings by Coexistence in Media/Time 3-47
3-11 Examples of Chemical Pairs Influenced by Exposure Timing 3-48
3-12 Examples of Chemical Groupings by Coexistence in Media/Time 3-59
3-13 Basic Steps for Cumulative Air Analysis 3-70
3-14 Benefits of Dispersion Models 3-71
3-15 Multiple Emissions During Cleanup 3-71
XII
-------
LIST OF TEXT BOXES cont.
No. Title
3-16 Emission Factors for Multiple Sources 3-72
3-17 Mobile Sources and Multiple Chemicals 3-72
3-18 Comparison of PM Properties 3-73
3-19 Example Particulate Factors 3-74
3-20 Air Dispersion Model Inputs 3-75
3-21 Example Model Input Considerations 3-76
3-22 Meteorologic and Receptor Data 3-76
3-23 Model Capabilities for Cumulative Air Analyses 3-78
3-24 Comparison of Exposure Assessment Processes 3-81
4-1 Selected Information Guides for Toxicity Assessment 4-2
4-2 Target Organ Toxicity Doses 4-23
4-3 Procedure for Estimating Whole Mixture Toxicity Values 4-31
4-4 EPA Uses of Route to Route Extrapolations U.S. EPA Workshop
Report on Inhalation Risk Assessment 4-47
4-5 RPF Formulas for Risk Estimation of a Two Chemical Mixture 4-50
5-1 Elements of Risk Characterization 5-1
5-2 ExampleSite Closure vs. Public Access 5-4
5-3 ExampleSite Safety 5-30
XIII
-------
LIST OF ABBREVIATIONS
The main acronyms and abbreviations used in this document are identified below.
Where use is essentially limited to tables or equations, the term is specified with those
tables and equations. Where use is primarily in an appendix, the term is specified in
that appendix.
atm atmosphere
ATSDR Agency for Toxic Substances and Disease Registry
BMD benchmark dose
BMDL lower confidence limit on BMD
BP boiling point
°C degrees Celsius or centigrade
Cd Cadmium
CDC Centers for Disease Control and Prevention
CEP Cumulative Exposure Project
CERCLA Comprehensive Environmental Response, Compensation, and Liability
Act
CHI cumulative hazard index
C\2 free chlorine
DBP disinfection byproduct
DCA 1,1-dichloroethane
DDT dichlorodiphenyltrichloroethane
DMA deoxyribonucleic acid
DNAPL dense nonaqueous phase liquid
DOE U.S. Department of Energy
EFH Exposure Factors Handbook
EPA U.S. Environmental Protection Agency
ETS environmental tobacco smoke
foe fraction of organic carbon
FQPA Food Quality Protection Act
GEP good engineering practice
CIS geographic information system
Hg mercury
HQ hazard quotient
ICED Index Chemical Equivalent Dose
IRIS Integrated Risk Information System (EPA database)
IUR inhalation unit risk
XIV
-------
LIST OF ABBREVIATIONS cont.
Kd soil-water partition coefficient
KH Henry's constant
Kow octanol-water partition coefficient
Ksp solubility product
L liter
LOAEL lowest-observed-adverse-effect level
m3 cubic meter
mg milligram
mg/kg milligram per kilogram
mg/kg-day milligram per kilogram body weight per day
mm millimeters
MOA mode of action
mol moles
MP melting point
MAS National Academy of Sciences
NCEA National Center for Environmental Assessment, EPA
NHEXAS National Human Exposure Assessment Survey
NOAEL no observed adverse effect level
NRC National Research Council (NAS)
OPP Office of Pesticide Programs
ORD Office of Research and Development (EPA)
OSWER Office of Solid Waste and Emergency Response
PAHs polycyclic aromatic hydrocarbons
Pb lead
PBPK physiologically-based pharmacokinetic (model)
PBTK physiologically-based toxicokinetic
PCBs polychlorinated biphenyls
PM2.5 particulate matter with a diameter of 2.5 |j,m or less
PM10 particulate matter with a diameter of 10 |j,m or less
ppb parts per billion
ppm parts per million
RAGS Risk Assessment Guidance for Superfund (EPA)
RAPIDS Regional Air Pollutant Inventory Development System
RfC reference concentration
RfD reference dose
RPF relative potency factor
RFV reference value
xv
-------
LIST OF ABBREVIATIONS cont.
Sw solubility in water
TCDD tetrachlorodibenzo(p)dioxin
TCE trichloroethylene
TD toxicodynamics
TEQ toxicity equivalents
TK toxicokinetics
TPA tris(2-ethylhexyl) phosphate
TSP total suspended particulates
TTD target organ toxicity doses
UF uncertainty factor
|j,g microgram
|j,m micrometer
VOC volatile organic compounds
VP vapor pressure
WOE weight of evidence
XVI
-------
PREFACE
This report was developed as a collaborative effort between the U.S.
Environmental Protection Agency's (EPA) Office of Research and Development (ORD),
National Center for Environmental AssessmentCincinnati Office (NCEA-Cin) and the
Department of Energy's Argonne National Laboratory. It offers information that can be
used to implement basic cumulative risk assessment concepts within the framework set
forth by EPA. This current report serves as a resource document for identifying specific
elements of and approaches for implementing cumulative risk assessments. This report
is not a regulatory document and is not guidance but rather a presentation of concepts,
methods and data sources. It is designed to assist EPA's development of specific
approaches and cumulative risk guidance for use by its Program Offices and Regions.
It is intended as a resource for EPA scientists and others in the broader risk
assessment community with an interest in locating data and approaches relevant to
cumulative risk assessment. The aim is to illustrate approaches and resources that can
be used to more explicitly assess human health cumulative risks from multiple route
exposures to multiple chemicals found at sites or within communities. This scope can
involve evaluating many different sources and contaminants, several media (soil, water,
air and structures) and associated exposure pathways, various representative
individuals or population subgroups which could be exposed over time, multiple health
effects and toxicological interactions among chemicals. The overall goal of using
cumulative risk assessment approaches is to produce more accurate and effective
assessments of these sites and situations, leading to more informed and ultimately
better decisions for managing potential cumulative health risks. External peer review
included two categories of comments that were collected between March and July 2006:
(1) comments from an independent peer review panel, organized and implemented by
Eastern Research Group (ERG) under EPA Contract No 68-C-02-060, in a meeting
open to the public on May 25-26, 2006, in Cincinnati, Ohio and (2) public comments
using an E-docket during a 45 day public comment period from March 31-May 15, 2006.
The public comments received by EPA were issued to the Peer Review panel members
prior to the May 2006 review meeting for their consideration in making comments and
recommendations to EPA. Information concerning the peer review meeting results can
be found online at: http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=149983.
XVII
-------
AUTHORS, CONTRIBUTORS AND REVIEWERS
This research was sponsored by the U.S. Environmental Protection Agency
(EPA), Office of Research and Development, National Center for Environmental
AssessmentCincinnati Division (NCEA). Through an interagency agreement, NCEA
researchers collaborated with scientists from the Department of Energy's Argonne
National Laboratory to conduct this research and to author this report. These
individuals are listed below.
AUTHORS
National Center for Environmental Assessment, U.S. EPA, Cincinnati, OH
Richard C. Hertzberg (Project Lead, Retired)
Linda K. Teuschler
Glenn E. Rice
John C. Lipscomb
J. Michael Wright
Jason C. Lambert
Anthony Fristachi
Argonne National Laboratory, U.S. Department of Energy, Argonne, IL
Margaret MacDonell (Project Lead)
James Butler
Young-Soo Chang
Heidi Hartmann
John Peterson
Kurt Picel
Tetra Tech EM, Inc., Dallas, TX
Shanna Collie
Shannon Garcia
Alan Johns
Camarie Perry
ENVIRON Corporation, Emeryville, CA
Lynne Haroun
XVIII
-------
AUTHORS, CONTRIBUTORS AND REVIEWERS cont.
CONTRIBUTORS AND REVIEWERS
Gary Bangs
U.S. Environmental Protection Agency
National Center for Environmental
Assessment
Washington, DC
Edward Bender (retired)
U.S. Environmental Protection Agency
Office of Assistant Administrator
Office of Science Advisor
Washington, DC
Michael Callahan
U.S. Environmental Protection Agency
Region 6
Dallas, TX
James Carlisle (Expert Panel)
California EPA
Office of Environmental
Health Hazard Assessment
Meadow Vista, CA
David Carpenter (Expert Panel)
Institute for Health and the Environment
University of Albany
Rensselaer, NY
Paul Chrostowski (Expert Panel)
CPF Associates, Inc.
Takoma Park, MD
David Cooper
U.S. Environmental Protection Agency
Office of Solid Waste and Emergency
Response
Washington, DC
Audrey Galizia
U.S. Environmental Protection Agency
National Center for Environmental
Assessment
Cincinnati, OH
Ihor Hlohowskyj
U.S. Department of Energy
Argonne National Laboratory Team
Argonne, IL
Pat Jennings
U.S. Environmental Protection Agency
Office of Water
Washington, DC
Jeremy Johnson
U.S. Environmental Protection Agency
Region 7
Kansas City, KS
Kannan Krishnan (Chair, Expert Panel)
Human Toxicology Research Group
University of Montreal
Canada
Sarah Levinson
U.S. Environmental Protection Agency
Region 1
Boston, MA
Margaret McDonough
U.S. Environmental Protection Agency
Region 1
Boston, MA
Chuck Nace
U.S. Environmental Protection Agency
Region 2
New York, NY
Michael Posson
ENVIRON
Emeryville, CA
Kaitlin Prieur
TetraTech EM, Inc.
Dallas, TX
XIX
-------
AUTHORS, CONTRIBUTORS AND REVIEWERS cont.
CONTRIBUTORS AND REVIEWERS cont.
Stig Regli
U.S. Environmental Protection Agency
Office of Water
Washington, DC
Jon Reid
U.S. Environmental Protection Agency
National Center for Environmental
Assessment
Cincinnati, OH
Libby Stull
U.S. Department of Energy
Argonne National Laboratory Team
Argonne, IL
Robert Sullivan
U.S. Department of Energy
Argonne National Laboratory Team
Argonne, IL
David Tomasko
U.S. Department of Energy
Argonne National Laboratory Team
Argonne, IL
Nga Tran (Expert Panel)
ExPonent, Inc.
Washington, DC
xx
-------
EXECUTIVE SUMMARY
ES.1. BACKGROUND
Public interest in and awareness of the health impacts of environmental chemical
exposures and their interactions with other stressors continues to grow as more
information is assembled about exposures to multiple chemicals in air, water and soil
from different sources. Environmental Protection Agency (EPA) has responded to
increasing requests for ways to understand and evaluate the combined impacts of these
conditions by preparing a set of reports on various aspects of cumulative risk
assessment. The EPA's Framework for Cumulative Risk Assessment (herein referred
to as the Framework) defines the general concepts and considerations for these
assessments (U.S. EPA, 2003a), and earlier reports laid a broad foundation for the
initial Planning and Scoping phase needed to conduct a cumulative risk assessment
(U.S. EPA, 1997a, 2002a). This report is linked to, and relies upon these documents,
as well on several key EPA guidance documents, as illustrated by the examples in
Figure ES-1. This current report serves as a resource document for identifying specific
elements of and approaches for implementing cumulative risk assessments. This report
is not a regulatory document and is not guidance but rather a presentation of concepts,
methods and data sources. It is designed to assist EPA's development of specific
approaches and cumulative risk guidance for use by its Program Offices and Regions.
It is intended as a resource for EPA scientists and others in the broader risk
assessment community with an interest in locating data and approaches relevant to
cumulative risk assessment.
The Framework defines cumulative risk as the combined risks from aggregate
exposures (i.e., multiple route exposures) to multiple agents or stressors, where agents
or stressors may include chemicals, as well as biological or physical agents (e.g., noise,
nutritional status), or the absence of a necessity such as habitat (U.S. EPA, 2003a).
Cumulative risk assessment, then, is an analysis, characterization and possible
quantification of the combined risks to health or the environment from multiple agents or
stressors. Other important aspects of cumulative risk assessment include a population
focus, emphasis on stakeholder involvement, consideration of population vulnerabilities,
and a focus on both human health and ecology. Areas of vulnerability articulated in the
Framework for human and biological ecosystems, communities and populations include
susceptibility or sensitivity, differential exposure (e.g., caused by cultural practices or by
living in close proximity to pollutant sources), differential preparedness (e.g., lack of
disease immunizations) and differential ability to recover. Note that the conduct of a
XXI
-------
Risk Analysis Phase
Planning and
Scoping Phase
Risk Assessment Guidance
for Superfund
(1989a)
Planning and Scoping for
Cumulative Risk Assessment
(1997a)
Methodology for
Multipathway Combustor Emissions
(1998a)
Guidance for Assessing
Health Risks of Chemical Mixtures
(2000a)
Planning and Scoping
Lessons Learned
(2002f)
Guidance on
Cumulative Risk of Pesticides
(2002c)
Framework for
Cumulative Risk Assessment
(2003a)
FIGURE ES-1
Key EPA Resources for this Report: Precedent U.S. EPA Guidance and Reports
Containing Specific Approaches for Assessing Major Parts of Cumulative Health Risks
XXII
-------
cumulative risk assessment will not be appropriate to every investigation; it is most
useful when addressing the risks from multiple stressors acting together (U.S. EPA,
2003a).
The Framework incorporates the risk assessment paradigm (NRC, 1983) within
the three phases of a cumulative risk assessment that it identifies (see Figure ES-2): (1)
Problem Formulation, (2) Risk Analysis and (3) Risk Characterization. Planning and
Scoping, an iterative dialogue between the scientists, risk managers and stakeholders,
takes place mostly during the Problem Formulation phase but may be revisited as
needed during the Risk Analysis and Risk Characterization phases. The output from
Risk Characterization is then used to support environmental Decision-Making. Other
factors, such as economic, social and policy considerations, may enter into both the
Planning and Scoping and the Decision-Making stages of the cumulative risk process.
These may influence the design of the analysis or the final risk management decisions.
ES.2. SCOPE OF THIS REPORT
This report focuses on two areas: (1) concepts concerning the initiating factors
for a cumulative risk assessment with procedures for data collection and organization
(Chapters 1 and 2) and (2) technical approaches for assessing and characterizing
health risks associated with a subset of cumulative risk issues (i.e., multiple chemicals,
exposures and effects), with examples pertaining to contaminated sites, drinking water
and ambient air (Chapters 3, 4 and 5). Some of the innovations proposed in this
document include
developing a description of initiating factors for a Cumulative Risk
Assessment and procedures for population characterization, data collection
and organization based on the initiating factors (Chapters 1 and 2);
implementing chemical grouping, a potentially helpful way to scope analyses
into manageable pieces to be assessed as chemical mixtures with co-
occurring exposures (Chapters 3 and 4);
approaches and data sources for evaluating the timing of exposures,
including discussions of kinetics and dynamics (Chapters 3 and 4);
integrating internal dose measurements to account for multiple route
exposures (Chapters 3 and 4);
further developing the quantitative method for the interaction-based hazard
index, first introduced in the 2000 mixtures guidance document (U.S. EPA,
2000a) (Chapter 4);
XXIII
-------
Includes the four
analytic elements of the
NRC (1983) Risk
Assessment Paradigm
Updated management needs
Planning
and
scoping
(Technical,
stakeholder
& manager
dialogue)
Risk assessment
Problem
formulatio
Decision
Making
Analysis
Risk
characterization
"/'Economic, political-science
*^ social & other analyses
FIGURE ES-2
Integrated Process for Cumulative Risk Assessment
(Source: adapted from U.S. EPA, 2002f)
XXIV
-------
extending the Relative Potency Factors (RFP) method to cumulate across
exposure routes, an approach first presented in an earlier EPA report on
drinking water disinfection by-product (DBP) mixtures (U.S. EPA, 2000e)
(Chapter 4);
integrating output from multiple effects modeling (illustrated using a
categorical regression model) with the Hazard Index (HI) and response
addition models to express risks for multiple health effects (Chapter 4);
providing added detail on the cumulative HI approach used by the Superfund
program (U.S. EPA, 1989a), including discussion of the impacts for risk
characterization (Chapters 4 and 5);
presentation of a method for cumulative risk characterization that considers
factors unique to conduct of a Cumulative Risk Assessment, including the
recognition of uncertainties in cumulative dose-response and exposure
assessment (Chapter 5); and
a general emphasis on integrating exposure and dose-response analysis
(Chapters 3, 4 and 5).
This report covers only some of the many aspects of cumulative risk for human
health assessment. It does not address risk management decisions and risk
communication. This report also does not consider interactions with non-chemical
stressors, such as noise, nor other kinds of risks, such as microbial or ecological risks.
In addition, social, political and economic issues are not discussed and only some
aspects of vulnerability are highlighted.
ES.3. THIS REPORT'S APPROACH TO CUMULATIVE RISK ASSESSMENT
Many situations do not have a population focus or do not involve multiple
chemicals and so would not need a cumulative risk assessment. However, there are
certain initiating factors that would naturally lead to conducting a cumulative risk
assessment. Figure ES-3 shows these three identified initiating factors along with the
data elements that may be used to conduct a cumulative risk assessment. These
initiating factors are (1) multiple pollutant sources or releases, (2) elevated
concentrations from environmental monitoring or biomonitoring of chemicals and (3)
increased population illness in a community. Figure ES-4 illustrates the types of
information that may be considered for data collection and population characterization
and shows the relationship of this information to the initiating factors. It is noteworthy
that traditional source-based assessments are usually initiated when chemicals are
found or released into the environment from known sources. When this occurs,
population vulnerability factors, such as diet, behaviors, genetic traits, economic status
and social characteristics are often not included in the assessment. These traits are
xxv
-------
multiple industrial
facilities
and disposal areas,
accidental chemical
releases
cluster of
leukemia cases,
elevated cancer rates
organics in air
or soil,
transported to
water and
accumulated in
fish
incidence of infant
mortality, hospital
admission rates
public health
data
Population
illness
Sources,
releases
multiple-
chemical
fate
population
subgroup
sensitivities
Integrated
characterization
population
vulnerabilities
multi-route
exposures
mixtures
toxicity
. 9 Chemical
concentrations
genetic
susceptibility,
children,
elderly
inhalation,
ingestion, dermal
exposures from
air, water, soil,
fish, produce
high blood lead
levels in children,
high levels of
chemicals found in
soil or indoor dust
homes close
to pollutant
sources, poor
health care,
subsistence
fishers
aroclor: reproductive effects,
diesel exhaust: lung cancer,
drinking water disinfectant
byproducts: bladder cancer
FIGURE ES-3
Example Initiating Factors and Data Elements for Cumulative Risk Analyses
XXVI
-------
Initiating Factor:
Sources and
Releases
Initiating Factor: Elevated Environ-
mental Chemical Concentrations or
Biomonitoring Levels
Site
Sources:
stack
emissions,
surface
runoff,
leaching
Non-Site
Sources: food,
household
products,
indoor/outdoor
air pollution,
drinking water,
pesticides
Drug and
alcohol
abuse,
smoking,
cultural
practices
i
/
Poor
nutrition,
obesity,
physical
and
mental
health
Polymor-
phisms,
gender,
age,
race
/
Environmental
Contaminants
Poverty,
education,
minority
status,
unemploy-
ment,
income,
residential
proximity to
sources,
family
dysfunction,
health care
access
Diet and
Behavior
Biological and
Genetic Factors
Socio-Economic
Stressors
Initiating Factor: Population Illnesses
or Perceived Population Illnesses
FIGURE ES-4
Variables Considered in Cumulative Risk Assessment and their Relationship to Initiating Factors
XXVII
-------
more likely to be assessed when population illness or the potential for illness are the
initiating factor. Note that there may be challenges related to finding the needed
expertise and collaborative partners to carry out a cumulative risk assessment when
these non-traditional stressors are incorporated into an assessment. The EPA does
address a few of these factors (e.g., sensitive subgroups, children, elderly), however, it
may be useful to conduct additional research on analyzing health risks for vulnerable
populations and to collaborate with other organizations that may have access to
relevant data.
Figure ES-5 shows the key steps in a cumulative risk assessment, with a primary
focus of addressing multiple chemicals, pathways, timeframes and effects in a
population-based setting. These steps define the population of concern and its study
area, generate a list of environmental contaminants relevant to the initiating factor and
identify links between environmental chemical exposures and vulnerabilities within the
population. These steps form the initial collection and organization of information to
focus on the cumulative aspects of the risk assessment. These steps may not be
sequential and may involve a number of iterations as the analyst examines factors
related to population vulnerabilities, public health information, toxicological and
epidemiologic data, completed exposure pathways, differential exposures and contact
with environmental media and pollutant sources. Outputs include a population profile, a
list of relevant chemicals, chemical groups for use in risk analysis and characterization
and a conceptual model. Outputs may include additional epidemiologic evaluations that
assess the health of the community or that examine associations between health
impacts and pollutant exposures.
ES.4. EXPOSURE ASSESSMENT OF MULTIPLE CHEMICALS, EXPOSURES AND
EFFECTS
In cumulative risk assessments that examine risks posed by multiple chemicals,
exposure assessments evaluate a population's chemical exposures through multiple
routes of exposure over time. Such assessments may encompass multiple exposure
timeframes in which the timing and intensity of exposures to different chemicals are
examined relative to each other. It is also important to determine whether the
exposures to multiple chemicals can lead to toxicokinetic interactions1 or toxicodynamic
1 Toxicokinetic interactions refer to alterations in the absorption, distribution, metabolism or elimination of
a toxic chemical. For example, these interactions can be mediated by the induction or inhibition of
enzymes involved in xenobiotic activation or detoxification. See Appendix C U.S. EPA (2000a) for
complete discussion.
XXVIII
-------
STEPS
OUTPUTS
1) Identify Initiating Factor
2) Characterize Population
based on Initiating Factor
.-?
/
/
Population Profile
3) Generate Chemical List
I
.
"V
ii
j -
''
4) Identify Links between
Chemicals & Subpopulations
5) Quantify Exposure
for General Population
& Subpopulations, Form
Initial Exposure Groups
6) Quantify Dose-
Response for Initial
Toxi city-based
Chemical Groups
I
8) Conduct Risk Characterization
List of Relevant Chemicals
Conceptual Model
Epidemiologic Evaluations
Chemical Groups
by Media & Time
Chemical Groups
by Toxicity
7) Integrate Exposure & Dose-
Response. Refine Exposure and
Toxicity Assessments
Integrated
Chemical Groups
Final Cumulative RA
FIGURE ES-5
Key Steps in a Cumulative Risk Assessment. The interdependence of exposure and
toxicity assessments is indicated by blue arrows.
XXIX
-------
interactions2. In addition to providing information about multiple chemical exposures in
the general population, these exposure assessments identify potentially susceptible or
vulnerable subpopulations3 in the study area and potentially unique pathways of
exposure in those subpopulations.
Cumulative exposure assessments will likely rely on environmental monitoring
data and environmental fate models. The community's boundary may define the
geographic region of study for a cumulative exposure assessment, unlike chemical-
focused assessments or single source-focused assessments. If the timing of different
chemical exposures is important, the analyst can use fate models to estimate changes
in the concentrations in environmental media over time. The pollutants may occur in
these media as a consequence of releases from multiple and different sources that
could be either close to or distant from the population of concern. The environmental
fate information for such an assessment could be site dependent.
While approaches to exposure assessment modeling are stressed in this
chapter, the use of biomonitoring data (e.g., biomarkers of exposure) holds a great deal
of promise for future cumulative risk assessments. The use of biomarkers in cumulative
risk assessments currently is limited. They can provide key quantitative exposure
estimates in cumulative risk assessments (e.g., biomarker data are used to estimate
current chemical exposure levels in an affected population or the general population).
Such data also can be used to verify selected exposure model results (e.g., show that
specific chemical exposures and absorption are occurring in the population or, if the
data are collected in a different location or under different conditions, provide evidence
showing that human absorption of the chemical from environmental exposures are
possible). For example, some studies have used existing blood chemical or urine
chemical concentration data, such as data published in NHANES (NCHS, 2002).
Exclusive use of biomarker data in cumulative exposure assessment efforts is
currently not practicable when considering a large number of diverse chemicals due to
analytical and resource limitations. Analytical limitations include considerations such as
whether sensitive biomarkers for many types of environmental chemicals have been
2 Toxicodynamic interactions encompass all interactions that do not directly affect absorption, distribution,
metabolism or elimination of a toxic chemical. Toxicodynamic interactions affect a tissue's response or
susceptibility to chemically mediated toxic injury. Modes of toxicodynamic interactions include, among
others, depletion or induction of protective factors, alterations in tissue repair, changes in hemodynamics,
and immunomodulation. See Appendix C U.S. EPA (2000a) for complete discussion.
3 Vulnerable or susceptible populations in the study area can be identified during either the exposure or
dose-response assessment phases of a cumulative risk assessment. This identification is based on
properties of the chemicals being evaluated as well as social, cultural or genetic factors that influence
vulnerability or susceptibility.
xxx
-------
developed and whether the chemical's biological half-life after absorption is sufficient to
estimate exposure over a relevant exposure period. Collection of human biomarker
data can be invasive and costly, resource limitations may constrain the ability of
researchers to collect such data.
If collected, the interpretation of biomonitoring data and application to risk
assessment can be challenging. While biomonitoring identifies individuals who are
exposed and have measured internal doses reflecting absorption of a chemical, to
estimate the individuals' actual exposures, the biomonitoring data would need to be
integrated with additional information (e.g., exposure modeling information) to identify
the pathways, timing and routes of exposure. Additional exposure and environmental
modeling would be needed to identify sources of chemicals in the contaminated media.
Although the use of biomonitoring data holds great promise for cumulative risk
assessments, few methods exist at this time for such applications (U.S. EPA, 2003a).
Exposure models may be divided two general categories: screening and refined.
Screening models involve relatively simple estimation techniques and generally use
preset, worst-case conditions to produce conservative estimates of the environmental
quality impact of a specific source or source category. Analysts use these instead of
more detailed (and more expensive) models to assess sources that clearly will not
cause or contribute to ambient concentrations above established standards for public
health. If results of conservative screening analyses indicate that multiple chemical
concentrations from one source or a combination of sources might not meet ambient
standards and health criteria, then the analysts would apply refined models for a more
representative assessment.
An example of a refined approach for detailed consideration of exposure timing in
dose/response assessment is the EPA's Office of Pesticide Policy approach, identified
as the calendar approach, in General Principles for Performing Aggregate Exposure
and Risk Assessments (U.S. EPA, 2001 a). The calendar approach estimates
sequential, daily chemical exposures by linking episodic exposures (e.g., seasonal
exposures to pesticides through surface water contact following residential lawn
applications of pesticides in the spring and summer) with routine exposures (e.g.,
contaminants in the food supply).
Mixtures occurring in a community may originate from different sources. Thus,
information about sources of chemical pollutants, chemical properties and fate can be
organized to guide chemical groupings that reflect the coexistence in media to which
people can be exposed within contaminated communities. The grouping of the
chemicals could be based on the potential for their co-occurrence in each
XXXI
-------
compartment/medium, potential for interactions affecting transformation and potential
for co-occurrence and interaction along each transport pathway between media. Figure
ES-6 provides an overview of how this information might be organized according to
media and the processes of fate and transport.
While chemicals can be easily grouped based on common sources and releases
(e.g., chemicals in diesel exhaust), the usefulness of groupings for various chemical
classes can be improved based on typical primary release mechanisms that would be
expected to control initial contamination and migration behavior in the environment.
Released chemicals can disperse quickly over a fairly wide area by convection (such as
via wind or surface water flow), and they can also migrate following waste placement.
The dominant processes at a given location determine what will be the "receiving
medium" into which a particular class of chemicals is introduced and from which they
can migrate.
Groups of chemicals may be expected to be distributed to various environmental
compartments (or media). An implicit assumption is that sufficient time has passed for
transport and system equilibration to occur. In some cases, such as deposition in
aquatic sediments or transport through the food chain, this process can take from
months to years following an initial release of contaminants. By the same token, after
an extended time, chemicals from a variety of different sources would be expected to
ultimately reach similar environmental sinks. In some cumulative risk assessments, it
may be important to examine when these chemical movements would occur.
This concept is illustrated for an example release scenario (industrial spill) in
Table ES-1. This concept applies to any environmental release, so other scenarios can
also be considered, such as combustor emissions related to routine operations or
temporary releases (e.g., due to excursions from a continuous-operation facility or
discrete releases from a mobile facility). The result is an initial set of chemical groups
that can be further refined in the toxicity assessment and then used for Risk
Characterization and uncertainty analysis.
Text Box ES-1 summarizes a general comparison of the processes involved in
conducting a basic versus a cumulative exposure assessment. As this summary
shows, the basic topics and outcomes are the same. The cumulative column simply
highlights additional attention that would be paid to certain features in explicitly
considering cumulative risk issues. Cumulative risk assessments evaluate aggregate
exposures by multiple pathways, media and routes over time, plus combined exposures
to multiple contaminants from multiple sources.
XXXII
-------
Pollution Source
1
aTransfoimation
Transformation
transformation refers to a group of
processes that can act to change the
composition of a mixture.
blntracompartment transport refers to the
processes that move a mixture through an
individual compartment (e.g., turbulence and
wind will move a mixture through the
atmosphere) and intercompartment transport
refers to processes that move a chemical
mixture from one medium to another.
Pollution Source-
Pollution Source,
Receiving Media
"*\ lntracompartmentb
Transport
Intercompartment Transport13
Other Media
Concentrations at
Points of Exposure in
Multiple Media
Human Activity Patterns
Exposed
Subpopulations
Toxi co kinetics
Target Tissue Doses
FIGURE ES-6
Approach for Estimating Exposure in Cumulative Risk Assessments
XXXIII
-------
TABLE ES-1
Example Groupings Based on Exposure Considerations (Media and Timing)*
Exposure
Duration
Environmental Medium -
Transport/ Removal Process
Soil upper horizon - volatilization and leaching
from surface, biodegradation
Air-
volatilization from soil
Surface water (river) -
overland flow and particle transport
from surface soil
Aquatic sediments - precipitation from water,
adsorption on particles, deposition
Soil lower horizons - leaching from surface soil,
adsorption and biodegradation
Groundwater-
leaching from soil
Release Scenario
Industrial Spill on Soil near a River
(VOCs, SVOCs and Metals)
Acute to Short-Term
-------
* Projected intervals reflect physical-chemical properties and fate data, including half-lives; other factors also affect partitioning and timing,
including local conditions such as temperature (for volatilization); organic content (for soil and sediment sorption), which for this example is
assumed to be relatively low; and depth to aquifer (for leaching to groundwater), which is assumed to be moderate to deep.
As = arsenic; CCI4 = carbon tetrachloride; Cd = cadmium; Cr= chromium; DCA = 1,1-dichloroethane; DCE = 1,1-dichloroethylene; Hg = mercury;
Ni = nickel; PCBs = polychlorinated biphenyls; SVOCs = semivolatile organic compounds; TCE = trichloroethylene; VC = vinyl chloride;
VOCs = volatile organic compounds.
xxxv
-------
Comparison of Exposure Assessment Processes (Text Box ES-1)
Basic Assessment Cumulative Assessment
What general question is being addressed?
How could people be exposed to chemicals, Similar, but emphasizing combined source
what would the amount of exposure be? contaminants and cumulative exposures
What is evaluated?
individua, Sources/reieases of ohemioais **~' ^^ "
sets of chemicals
Concentrations of chemicals at points of Emphasis on sets of chemicals that coexist initially and
human contact those that move together
r, i u « , , .... Representative receptors as for the basic case, paying
Peope who represent current conditions ,, K ,. , ... K . . . K y M
.,.. . f . , . attention to sensitive subgroups and unique exposure
and ike y future and use ,. ... , u \ t- \
y activities (e.g., per cultural practices)
Routes by which people could be exposed to Emphasis on combined chemicals and routes over
each chemical time, considering sequencing
Amoun, o, each onemica, taKen in over time ^£%%«SS£ ' *"""
How are results used?
Estimated intakes are considered in groups to guide
of ioint toxictty to assess
ES.5. TOXICITY ASSESSMENT OF MULTIPLE CHEMICALS, EXPOSURES AND
EFFECTS
Assessments of adverse health effects from exposures to multiple chemicals
through multiple routes of exposure over time may account for multiple health effects
and for joint toxic action resulting from exposure to a chemical mixture. Risk
assessments may include evaluation of the timing and intensity of exposures to different
chemicals, including the examination of internal co-occurrence of multiple chemicals
and toxicological interactions in the target tissue(s). Cumulative risk assessments add
layers of complexity to the evaluation of chemical mixtures. Methods for cumulative risk
assessment may be developed by expanding on the theory and methods presented in
the EPA's Supplementary Guidance for Conducting Health Risk Assessment of
Chemical Mixtures (U.S. EPA, 2000a) to evaluate various aspects of cumulative risk.
Figures ES-7a and ES-7b present both established methods along with new or
enhanced methods for cumulative risk assessment. For example, Figure ES-7a shows
the development of toxicity values (i.e., Reference Doses [RfDs], Reference
Concentrations [RfCs] and slope factors) as presented in the 2000 Supplementary
Mixtures Guidance for whole mixtures and sufficiently similar mixtures, but Figure ES-7a
also includes additional epidemiologic evaluations that may be useful when illnesses in
xxxvi
-------
Whole Mixture
Data Available
/ \
/ Whole \
Mixture
\pf Concern/
' Sufficiently^
Similar
k Mixture /
V 1 '
1
Toxicological
Evaluations
1
Epidemiologic
Evaluations
Derive RfDs/RfCs;
Slope Factors
I
Exposure Assessment of Whole Mixture
of Concern; Assessment of Similarity
I
I
Hazard Quotient;
Risk Estimate
Epidemiologic
Risk Measures
FIGURE ES-7a
Flow Chart Showing Approaches for Evaluating Whole Mixtures
XXXVII
-------
r Component Data Available J
Multiple
Toxicological
Effects for Each
Component
Multivariate
Statistical
Models, e.g.,
Categorical
Regression
/ lexicologically
Similar
N. Components
Dose Addition
Mix of
lexicologically
SimilarS Independent
Components
Available
Interactions
Data
I
Relative
Potency
Factors
/ lexicologically \
Independent
k Components )
Response
Addition
PBPK*
Modeling
Integrated
Additivity
Methods
Component Exposure Assessment
i 1 1
Risk
Estimates;
Hazard
Index
Binary
Weight of
Evidence;
Interaction
Profiles
Interaction-
Based
Hazard
Index
Hazard
Index;
Cumulative
Hazard
Index
Summing
of Route-
Specific
Index
Chemical-
Based
Risk
Estimates
Internal
Dose
Hazard
Index;
Multiple
Route
Internal
Doses
Risk
Estimate
Index
Chemical-
Based Risk
Estimate;
Hazard
Quotient
*PBPK = physiologically-based pharmacokinetic
FIGURE ES-7b
Flow Chart Showing the Component Based Approaches for Evaluating Multiple Chemicals, Exposure Routes, Effects and
Toxicological Interactions
XXXVIII
-------
the population initiates a cumulative risk assessment. Figure ES-7b presents
established component-based chemical mixtures methods (i.e., RPFs, HI, Response
Addition and the Interaction-Based HI), but several other approaches for use in
cumulative risk assessment are also reflected in this figure. Further, Figure ES-7b
handles not only toxicologically similar and dissimilar mixtures, but also addresses
mixes of these, as well as addressing the case of multiple toxicological effects. Finally,
additional methods are shown that include the use of PBPK models to estimate internal
doses of chemicals and examine the potential for toxicological interactions.
Grouping chemicals by the potential for co-occurrence and joint toxic action is a
key simplifying concept in this report. Chemical components of mixtures can be
screened for inclusion in a cumulative risk assessment using the elements of
component-based methods. Figures ES-8a, ES-8b and ES-8c outline a process for
classifying chemicals into groups suitable for analysis and the application of the
methods shown in Figures ES-7a and ES-7b. This process includes the following steps:
1) Figure ES-8aClassify all chemicals of concern into initial groups by their
potential to occur in the same or different media and at the same or different
time.
2) Figure ES-8aDivide these exposure/time groups further into subgroups in
which chemicals are thought to cause toxicity by the same mode of action or
affect the same target organ. Include all target organs or effects for which
positive evidence exists of adverse health effects. An initial step is to collect
toxicological and pharmacokinetic data on each of the individual chemicals to
be considered in the risk assessment. Factors to consider in forming these
toxicity groups include pharmacokinetic parameters, persistence of the
chemicals in the body and the formation of metabolites.
3) Figure ES-8b and ES-8cAssess the toxic potential of the chemicals and
whole mixtures of concern using methods shown in Figures ES-7a and
ES-7b. Figure ES-8b shows a flow chart that first evaluates the whole
mixtures and single chemicals for toxicity potential, ensuring that those with
the greatest potential to cause toxicity are maintained in the cumulative risk
assessment. Then, the chemical groups formed in Figure ES-8a are
evaluated for joint toxicity, addressing multiple effects, interactions and
exposure routes; these groups are then screened into or out of the cumulative
risk assessment. Figure ES-8c provides additional detail on the processes
shown in Figure ES-8b, indicating the methods and outputs from this data
analysis.
The methods developed for cumulative toxicity assessment may be used in
several different ways depending on data availability and on the goals of the
assessment. They may be applied as screening tools (e.g., to decide whether or not
toxicological interactions are of importance for a certain group of chemicals) or as tools
xxxix
-------
Exposure
Group.*
Exposure
Scenarios:
Exposure Groups
Same Media;
Same Time
Air: Daily
Exposure to
Municipal Waste
Combustor
Emissions
Air: Daily
Inhalation
Exposure to
Disinfection By-
products via
Showering
Same Media;
Different Time
Drinking Water:
Acute Accidental
Exposure to
Source Water
Contaminants
Drinking Water:
Exposure to
Uranium
Contaminated
Ground Water,
Years Later
Different Media;
Same Time
Drinking Water:
Daily Exposure to
Disinfection By-
products via
Ingestion and
Showering
Fish: Daily
Exposures via
Local Fish
Consumption
Different Media;
Different Time
Air: Short Term
Exposure to
Emissions from
Temporary
Combustor
Drinking Water:
Acute Accidental
Exposure to Source
Water
Contaminants,
Months Later
Chemicals in Exposure Groups (Above) Further Grouped Based on Similar Toxicity
Kidney
Brain
Fetus
Heart
Lung
Hg, Cd, BDCM
Hg, DCA
Hg, BDCM, DCA
Hg, Cd
Hg
Ni, TCE, U, Cr
TCE, As, Ni, CCI4
TCE, Ni, Cr
TCE, Ni, As, Cr
Ni, Cr
Hg, BDCM
Hg, DCA, PCB
Hg, BDCM, DCA,
PCB
Hg
Hg
Cd, TCE, Ni, Cr
TCE, As, Ni, CCI4
TCE, Ni, Cr
Cd, TCE, Ni, As, Cr
Ni, Cr
FIGURE ES-8a
Hypothetical Example of Chemical Groupings by Co-occurrence in Media and Time, Similar Toxicity
Terms: As = Arsenic (inorganic); BDCM = Bromodichloromethane; Cd = Cadmium; CCI4 = Carbon tetrachloride; Cr
Chromium; DCA = Dichloroacetic Acid; U = Uranium (soluble salts); Hg = Mercury (based on mercuric chloride); Ni =
Nickel (soluble salts); PCB = Polychlorinated Biphenyls (Arochlor 1016); TCE = Trichloroethylene
xl
-------
Component or
Whole Mixtures
Data for Toxicity
Group(s)?
Apply Whole
Mixture Risk
Assessment
Methods
to Toxicity
Group(s)
Component
Calculate Single
Chemical
HQ's, Cancer Risks,
Public Health Data
Apply Component Mixture Risk Assessment
Methods to Toxicity Group(s)
Any HQ>1 or
cancer risk >10-6?
Or Exceeding of
Public Health
Levels?
Continue with
Group(s
Optional Evaluation of Multiple Effects
Optional Evaluation of Multiple Route Exposures
Optional Evaluation of Interaction Effects
AnyHM,
Health Risks >1Q-6,
Odds Ratios >1?
Screen out Group(s)
from Cumulative
Risk Assessment
Conduct Cumulative Risk
Assessment for Single
Chemicals and Group(s)
FIGURE ES-8b
Grouping Chemicals for Cumulative Risk Assessment. The mixture risk methods are applied to each group, with
"concern" judged by the appropriate screening value (e.g., mixture RfD for whole mixture oral exposure). Groups can be
screened out only if both whole mixture and component methods indicate no concern.
xli
-------
Apply Component Mixture Risk Assessment
Methods to Toxicity Group(s)
Optional Evaluation of Multiple Effects
Methods: Multivariate Modeling (e.g., Categorical
Regression, Multivariate Normal Linear Regression)
Outputs: Use Results in Hazard Index or Response Addition
Apply Whole Mixture Risk Assessment Methods to
Toxicity Group(s)
Methods: Calculate Mixture RfD/C or Estimate
Risks, Conduct Epidemiologic Study
Outputs: HQ, Risk Estimate, Odds Ratios or Other
Epidemiologic Relative Risks Measures
Optional Evaluation of Multiple Route Exposures
Methods: Cumulative Hazard Index (CHI),
Sum of Risks from Route Specific RPF's,
Cumulative Relative Potency Factors (CRPF),
PBPK Model Estimates of Internal Doses;
Outputs: Use Results to Indicate Risk Potential (CHI)
or Estimate Health Risks (CRPF, PBPK, RPFs)
Optional Evaluation of Interaction Effects
Methods: Locate Interactions Data (e.g., ATSDR
Interaction Profiles, ARCOS & MIXTOX Databases,
Journal Articles on Toxicological Interaction Studies);
Outputs: Use Results to Qualitatively Assess Potential
Interactions or Calculate the Interaction-Based Hazard Index
Outputs:
Risk Indicator
of Concern?
e.g., CHI, HI or
Qualitative
Judgment of
Interaction
Potential
v y
No
Outputs:
Risk Estimate
of Concern?
e.g., Health
Risks>10-6,
Elevated
Odds Ratios
Screen out
Group(s)
from Cumulative
Risk Assessment
Yes
Conduct
Cumulative
Risk Assessment
for Single
Chemicals
and Group(s)
Conduct Thorough
Risk Characterization,
Uncertainty Analysis
FIGURE ES-8c
Grouping Chemicals for Cumulative Risk Assessment (cont). Specific mixture risk methods are applied depending on
which multiples are being evaluated, with "concern" judged by the appropriate screening value as determined during the
Problem Formulation stage of CRA.
xlii
-------
for estimating quantitative risk numbers (e.g., estimating the risk of an adverse level of
cholinesterase inhibition by applying a RPF approach to a group of pesticides). In some
cases all of the methods shown might be applied, and in other cases, only a select few
methods would be useful depending on the exposure scenario.
ES.6. RISK CHARACTERIZATION
A Risk Characterization is usually described as having two parts: an integrative
analysis, which contains the risk estimates and can be highly technical, and a risk
characterization summary, which focuses on recommendations and uncertainties.
Figure ES-9 provides an overview of the final Risk Characterization process for a
cumulative risk assessment. It is an expansion of the final Risk Characterization step
shown in Figure ES-5, beginning with outputs from the steps shown in Figure ES-5,
such as, the population profile and the integrated chemical groups. The cumulative
Risk Characterization may differ from a traditional Risk Characterization in several ways
that are often caused by missing data or a lack of understanding of the various multiples
and their interactions. Some of the more important differences are listed below:
Recommendations could be multivariate (i.e., it may be difficult to identify a
single chemical, pathway or critical effect that drives the risk)
Recommendations might be based on groupings of chemicals, pathways and
effects, but such groupings can be based on subjective judgments
Recommendations might be based on epidemiological findings relevant to a
population illness, for which it is useful to articulate confounding factors and
exposure uncertainties
Uncertainty analysis might be predominantly qualitative because of the use of
numerous defaults (e.g., for addressing interactions and multiple effects).
In summary, in the Risk Characterization phase of cumulative risk assessment, it
may be useful to consider issues in the context of evaluating multiple chemicals,
exposures and effects, including interaction effects, with respect to the population
characteristics. Issues regarding uncertainty, variability and sensitivity analysis are
important to present. An integrative technical analysis of the predicted risks is useful,
as well as a summary of the results and uncertainties of the Risk Analysis. Risk
Characterization results may be used by risk managers in the final Decision-Making
stage of a cumulative risk assessment; thus the Planning and Scoping process, data
sources, analytical techniques, logic used to make various technical decisions and
uncertainty analysis are more useful if they are scientifically sound and presented in a
transparent manner.
xliii
-------
Populations ConsideredGeneral Population or Subpopulations
Exposure ** Toxicity
Groups ^ > Groups
X If
Integrated Groups for Cumulative
Risk Assessment
Temporal Analysis
Do relevant multi-chemical
exposures or persistent
effects overlap?
Extrapolations,
Simplifications
and Omissions
Acceptable?
Conduct Qualitative
Assessment Only or
Revise Analytic Scope
Develop Quantitative Risk Estimates
For General Population and Relevant
Subpopulations
For each Relevant Chemical Group
For each Relevant Pathway and Exposure Route
For each Relevant Timeframe
Conduct Single Chemical Assessment Only,
Discuss All Cumulative Risk Elements Considered,
Describe Data Gaps for Future Research
Document Cumulative Risk Characterization
Including Risk Estimates and
Uncertainty Discussion
Describe Data Gaps for Future Research
Sensitivity and Uncertainty Analysis
Identify Sources of Uncertainty
Develop Integrated Sensitivity Analysis, if
Possible
-Identify Sources of Model Uncertainty
-Identify Sources of Parameter Variability and
Uncertainty
Identify Critical Research Needs
FIGURE ES-9
Schematic of Cumulative Risk Characterization Approach in this Report
xliv
-------
1. CUMULATIVE RISK ASSESSMENT INTRODUCTION
1.1. BACKGROUND
Public interest in and awareness of the health impacts of environmental chemical
exposures and their interactions with other stressors continues to grow as more
information is assembled about exposures to multiple chemicals in air, water and soil
from different sources. In the United States, organizations such as the United
States Environmental Protection Agency (EPA) and Agency for Toxic Substances and
Disease Registry (ATSDR) have developed documents that support the development of
cumulative risk assessment (see ATSDR, 2002c; U.S. EPA, 1997a, 2000a, 2002a,
2003a). Internationally, organizations such as the World Health Organization's
International Programme on Chemical Safety, the European Food Safety Authority and
NoMiracle (Novel Methods for Integrated Risk Assessment of Cumulative Stressors in
Europe) are sponsoring workshops and authoring publications to help increase
knowledge on the transfer of pollutants between different environmental compartments,
on food safety, and on the impact of cumulative stressors, including chemical mixtures
(EFSA, 2006; IPCS, 2006; NoMiracle, 2006).
EPA has responded to increasing requests for ways to understand and evaluate
the combined impacts of these conditions by preparing a set of reports on various
aspects of cumulative risk assessment. Those documents provide information that
organizes and helps explain the scope of cumulative risk assessment. The EPA's
Framework for Cumulative Risk Assessment (herein referred to as the Framework in
this report) defines the general concepts and considerations for these assessments
(U.S. EPA, 2003a), and earlier reports laid a broad foundation for the initial Planning
and Scoping phase needed to conduct a cumulative risk assessment (U.S. EPA, 1997a,
2002a). This current report serves as a resource document for identifying specific
elements of and approaches for implementing cumulative risk assessments. This report
is not a regulatory document and is not guidance but rather a presentation of concepts,
methods and data sources. It is designed to assist EPA's development of specific
approaches and cumulative risk guidance for use by its Program Offices and Regions.
It is intended as a resource for EPA scientists and others in the broader risk
assessment community with an interest in locating data and approaches relevant to
cumulative risk assessment.
1.1.1. The Integrated Process for Cumulative Risk Assessment. The Framework
defines cumulative risk as the combined risks from aggregate exposures (i.e., multiple
1-1
-------
route exposures) to multiple agents or stressors, where agents or stressors may include
chemicals, as well as biological or physical agents (e.g., noise, nutritional status), or the
absence of a necessity such as habitat. Cumulative risk assessment, then, is an
analysis, characterization and possible quantification of the combined risks to health or
the environment from multiple agents or stressors. Other important aspects of
cumulative risk assessment include a population focus, emphasis on stakeholder
involvement, consideration of population vulnerabilities, and a focus on both human
health and ecology. Areas of vulnerability articulated in the Framework for human and
biological ecosystems, communities, and populations include susceptibility or sensitivity,
differential exposure (e.g., caused by cultural practices or by living in close proximity to
pollutant sources), differential preparedness (e.g., lack of disease immunizations), and
differential ability to recover. Note that the conduct of a cumulative risk assessment will
not be appropriate to every investigation; it is most useful when addressing the risks
from multiple stressors acting together (U.S. EPA, 2003a).
The National Research Council (NRC) issued Risk Assessment in the Federal
Government: Managing the Process (NRC, 1983), commonly called the Red Book, over
20 years ago. This document identified four basic steps for risk assessment, called the
Risk Assessment Paradigm: hazard identification, dose-response assessment,
exposure assessment and risk characterization, as explained in Text Box 1-1. These
general steps provide the original foundation for risk-based programs across many
federal agencies and are an integral part of cumulative risk assessment. The
Framework incorporates the
risk assessment paradigm
within the three phases of a
cumulative risk assessment
Summary of Traditional Risk Assessment Paradigm
(Text Box 1-1)
Hazard identification/ Identify contaminant hazards and determine
data evaluation their levels in various media (soil, water, air)
Evaluate who could be exposed, how much,
Exposure assessment ^ow frequency
Dose-response Quantify dose-response relations and
assessment define toxicity values from scientific studies
_. . . , . ,. Describe cancer risks, noncancer effects
Risk characterization . ... . . '
and re ated uncertainties
that it identifies (see
Figure 1-1): (1) Problem
Formulation, (2) Risk
Analysis and (3) Risk Characterization. Planning and Scoping, an iterative dialogue
between the scientists, risk managers and stakeholders, takes place mostly during the
Problem Formulation phase but may be revisited as needed during the Risk Analysis
and Risk Characterization phases. The output from Risk Characterization is then used
to support environmental Decision-Making. Other factors, such as economic, social and
policy considerations, may enter into both the Planning and Scoping and the Decision-
Making stages of the cumulative risk process. These may influence the design of the
analysis or the final risk management decisions.
1-2
-------
Includes the four
analytic elements of the
NRC (1983) Risk
Assessment Paradigm
Updated management needs
Planning
and
scoping
(Technical,
stakeholder
and manager
dialogue)
Risk assessment
Problem
formulatio
Risk
characterization
-^Economic, political-science,
social and other analyses
FIGURE 1-1
Integrated Process for Cumulative Risk Assessment
(Source: adapted from U.S. EPA, 2002f)
1-3
-------
During the Problem Formulation phase, the risk analysts, risk managers and
other stakeholders jointly establish the goals, breadth, depth and focus of the
assessment, producing a conceptual model and an analysis plan. The conceptual
model identifies the stressors to be evaluated and the health or environmental effects to
be evaluated; it also describes the possible relationships among various stressors and
potential effects. The analysis plan lays out the data needed, the approach to be taken
and the types of results expected during the Analysis phase.
The Analysis phase of the Framework includes the determination of the analytical
and quantitative methods to be used for exposure assessment, dose-response
assessment and risk estimation. The exposure and dose-response processes for
cumulative risk are expected to occur iteratively to ensure information compatibility.
This phase also includes the initial estimates of joint health risk from the multiple
stressors to which the study population and sensitive population subgroups are exposed
(U.S. EPA, 2003a, p. xviii).
The final phase of a cumulative risk assessment, Risk Characterization, involves
further analysis so that the risk
estimates are explained in
terms of their significance and
uncertainties. This is also
where the risk assessment
process is evaluated to
determine whether the
objectives and goals of the first
phase (Planning and Scoping
and Problem Formulation) have
been met.
1.1.2. Terminology.
Terminology often used for
cumulative risk assessment
overlaps with terms used in
environmental science,
chemical mixtures risk
assessment and public health.
Some common terms are
defined in Text Box 1 -2. The
Cumulative risk
Effect
Exposure pathway
Key Terms for Cumulative Health Risks (Text Box 1-2)
Aggregate exposure Combined exposure to one chemical;
can be from multiple sources or
pathways
Combined risk from exposures to
multiple chemicals or stressors;
exposures may be aggregate
Health end point estimated from toxicity
studies (first-observed is critical effect;
secondary effect seen at higher doses)
A complete pathway includes (1) source
and mechanism of release,
2) contaminant fate & transport (through
environmental media), (3) point of
receptor contact with the source or
affected medium and (4) exposure route
How a contaminant gets inside a person
(e.g., via inhalation, ingestion, or dermal
absorption)
One chemical acting on another to
influence fate or transport
Toxic action exerted by two or more
chemicals acting together
Joint toxicity that is greater or less than
expected under additivity (note: forms of
additivity include summing of doses,
risks or biological measurements across
chemical components of a mixture)
Group actually or potentially exposed
Origin of contaminant (e.g., a landfill)
Exposure route
Environmental
interaction
Joint toxicity
Toxicological
interaction
Receptor population
Source
1-4
-------
glossary (Chapter 7) provides detailed definitions for these and other terms in this
report.
For EPA, cumulative risk assessment involves combined risks from multiple
exposures to multiple stressors from all contributing sources. This assessment
addresses a given receptor population, whether this is an actual community or a
hypothetical population of future inhabitants of a geographic region. This integrated
approach then extends beyond assessments that produce separate estimates for each
contributing source (such as releases from a waste pit, emissions from an incinerator or
effluent from a wastewater treatment facility) by estimating risk from the joint exposure
via all identified sources.
A cumulative risk assessment can involve multiple exposure pathways and
exposure routes that reflect different ways contaminants can enter the body from
different media (e.g., breathing air and drinking water). An exposure pathway describes
how chemicals are transported from a source to a person or subpopulation (e.g.,
through the air or water). An exposure route identifies the way the contaminant actually
enters the body.
These assessments also consider multiple effects within two main categories:
cancer and noncarcinogenic systemic effects. For the latter, in a cumulative risk
assessment involving multiple chemicals, it is important to include an evaluation of both
critical and secondary effects. The critical effect is the first effect observed as the
chemical's dose is increased above a no-effect range in the relevant toxicity study, and
it serves as the basis for the Reference Dose (RfD, see definition in Chapter 7) or other
noncancer toxicity value; secondary effects are typically those seen at higher doses in
the same target organ or tissue and/or different physiological compartment(s) and are
rarely incorporated into a single chemical risk assessment beyond uncertainty analysis
of the entire relevant toxicity database. In the assessment of chemical mixtures, an
important difference from single chemical assessments is that the health effects
observed as a result of combined chemical exposures may differ in phenotype and/or
magnitude from the critical effects caused by the individual chemical exposures. Thus,
it is important to evaluate secondary effects for those chemicals to which humans may
be exposed in combination. In these cases, the doses of the chemicals in the mixture
may act in an additive manner to cause one of these secondary or higher level effects,
or the responses (effects or risks) themselves may be additive. In addition, co-exposure
to multiple chemicals may result in toxicological interactions (e.g., synergism or
antagonism) that lead to secondary or higher level effects. Thus, in a cumulative risk
1-5
-------
assessment involving more than one chemical as a stressor, it is important to consider
evaluating critical effects as well as secondary and higher level effects.
Multiple stressors are central to cumulative risk analyses. Multiple stressors
include vulnerability factors and chemical, physical and biological exposures. The EPA
defines as an aggregate exposure assessment as an assessment that seeks to
characterize a single-chemical exposure that could involve multiple exposure pathways
(be present in many sources or media) and potentially taken in by multiple routes (oral,
dermal, etc.). Because an aggregate assessment only addresses a single chemical, it
is not formally considered a cumulative assessment. However, if a set of aggregate
exposures is combined, addressing two or more chemicals and their joint effects, then
that would constitute a cumulative assessment.
Interactions that consider location and timing are a main emphasis of this report.
In the environment, interactions can alter the fate and transport of chemicals, e.g., by
facilitating mobility in soil or sorption onto air particulates. Once taken into the body, a
key emphasis of this evaluation \sjointtoxicity, which is defined as the collective toxicity
of two or more chemicals. This can be additive (the default assumption), less than
additive (antagonism), or more than additive (synergism). The EPA has defined the
specific term, toxicological interactions, to represent interactions that are other than
additive (U.S. EPA, 2000a). The EPA has developed an interaction formula based on
departures from dose addition (see Chapter 4). Toxicological interactions are then
commonly defined by EPA as those that result in effects that are either lower or higher
than expected from the individual chemicals acting under an assumption of dose
additivity, such as the reported synergistic effect of cadmium (Cd) and lead (Pb) on the
neurological system or the reported antagonistic effect of Cd and Pb on the kidney (see
Chapters 4 and 5). Such interactions are a common concern at contaminated sites.
1.2. ABOUT THIS REPORT
As discussed above, cumulative risk assessment covers a breadth of topics
which may include combination toxicology, chemical mixtures, multiple exposure
pathways and exposure durations, and can extend from identifying how the assessment
was initiated to determining how the analysis will be conducted and how results will be
presented. Building on the concepts that have been identified in earlier reports and
offering examples to illustrate how those concepts can be applied, this report addresses
only human health assessment (as shown in Figure 1-2), and focuses on two areas: (1)
concepts concerning the initiating factors for a cumulative risk assessment with
procedures for data collection and organization (Chapters 1 and 2) and (2) technical
1-6
-------
approaches for assessing and characterizing health risks associated with a subset of
cumulative risk issues (i.e., multiple chemicals, exposures and effects), with examples
pertaining to contaminated sites, drinking water and ambient air (Chapters 3, 4 and 5).
The report's organization is as follows:
Chapters 1 and 2 present information on cumulative risk assessment initiating
factors, data collection and organization. Both chapters describe elements of the
Problem Formulation phase, i.e., the preliminary characterization of the population
assessed, the initial identification of the chemicals, exposures and effects of concern
and an evaluation of the potential relationship between population illness and
chemical exposures. Procedural steps for a cumulative risk assessment are
described, and the differences between population-based cumulative risk
assessments and traditional source-based or chemical-based risk assessments are
highlighted.
Chapter 1 discusses the development of cumulative risk assessment theory
and procedures, provides background information and describes the
organization and content of the current report. Chapter 1 also presents an
overview of cumulative risk processes and a summary of the approach
proposed in this report to address cumulative risk, emphasizing the factors that
could initiate the decision to undertake a cumulative assessment.
Chapter 2 discusses the initial characterization of the population and
identification of relevant chemicals as influenced by the initiating factor
that initiated the cumulative risk assessment, It discusses data collection
and organization, the use of public health information and epidemiologic
approaches, and it ends with a discussion of conceptual models for
identifying links between environmental exposures and target
populations.
Chapters 3, 4 and 5 present information on technical approaches to the Analysis and
Risk Characterization of multiple chemicals, exposures and effects. (Figure 1-2
illustrates this narrow focus on a subset of cumulative risk issues.) These include
evaluations of exposures and risks using chemical mixtures methods; approaches
for grouping chemicals for Risk Analysis and Risk Characterization; evaluating
assumptions and uncertainties; and deciding whether to conduct a qualitative or a
quantitative assessment.
Chapter 3 offers exposure assessment concepts, resources and approaches
for a cumulative risk assessment that can help characterize the setting, quantify
exposures and group the chemicals and pathways based on joint and
interactive processes. The influence of toxicity information on the exposure
assessment is discussed.
Chapter 4 explains and illustrates key toxicity concepts and chemical
mixture risk assessment methods that may be used to evaluate multiple
effects, exposures and toxicological interactions. The chemical groups
first established using exposure information are further defined based on
1-7
-------
common toxicological action. The influence of exposure information on
the toxicity assessment is discussed.
Chapter 5 provides information for the Risk Characterization phase,
including a discussion of issues to be addressed, methods for evaluating
some of the uncertainties inherent in cumulative risk assessments and
the need for comparison of results with the goals from the Planning and
Scoping phase.
Supporting Information
Chapter 6 identifies reference information for the documents and articles cited
in this report.
Chapter 7 defines basic terms used in cumulative risk assessments.
Appendix A presents a toolbox of selected resources that can be useful
in conducting cumulative risk assessments.
Appendix B illustrates how primary toxicity information can be organized
to support grouping for cumulative risk assessments.
Appendix C presents a discussion on the history and use of
toxicological severity concepts in risk assessments.
As shown in Figure 1-2, this report covers only some of the many aspects of
cumulative risk for human health assessment (not ecological assessment), so it is
important to note the areas that it does not consider. For example, while multiple
chemicals and exposures and both cancer and noncancer health endpoints are
addressed, approaches for interactions with non-chemical stressors, such as noise, or
for other kinds of risks, such as microbial or ecological risks, are not included. The
important issues related to stakeholder involvement in Planning and Scoping and risk
communication are also not included as they are described in previous documents (U.S.
EPA, 1997a, 2002a). In addition, social, political and economic issues are not
discussed and only some aspects of vulnerability are highlighted. Finally, this report
does not address the final risk management decision or the communication of such a
report to interested audiences.
1.2.1. Innovations Included in this Report. Within its targeted scope, this report
addresses certain aspects of the Problem Formulation, Risk Analysis and Risk
Characterization phases involved in implementing a cumulative risk assessment. In
actual applications, some of the approaches shown in this report may be extended more
broadly to assess other types of stressors complex exposures and vulnerability issues.
Many of the techniques have roots in previous EPA documents, such as the 2000
Supplementary Guidance for Conducting Health Risk Assessment of Chemical Mixtures
1-8
-------
Human
Health
Ecology
Multiple
sources
%/ sfressors
path ways/routes
%/ population groups
%/ effects
time frames
Sociocultural
Resources
Economics
FIGURE 1-2
Assessing Integrated Multiples in Cumulative Risk Assessment: Focus on Human Health
1-9
-------
(U.S. EPA, 2000a) (herein referred to as the 2000 Supplementary Mixtures Guidance in
this report) but new material is also presented with detail on how existing methods can
be extended to address areas the 2000 Supplementary Mixtures Guidance does not
cover (e.g., multiple route exposures to multiple chemicals and effects). This report also
brings information together from various sources to show how existing methods and
data can be used (see the toolbox in Appendix A). Innovations in methodology include
the following:
developing a description of initiating factors for a Cumulative Risk
Assessment and procedures for population characterization, data collection
and organization based on the initiating factors (Chapters 1 and 2);
implementing chemical grouping, a potentially helpful way to scope analyses
into manageable pieces to be assessed as chemical mixtures with co-
occurring exposures (Chapters 3 and 4);
approaches and data sources for evaluating the timing of exposures,
including discussions of kinetics and dynamics (Chapters 3 and 4);
integrating internal dose measurements to account for multiple route
exposures (Chapters 3 and 4);
further developing the quantitative method for the interaction-based hazard
index (HI), first introduced in the 2000 mixtures guidance document (U.S.
EPA, 2000a) (Chapter 4);
extending the Relative Potency Factors (RFP) method to cumulate across
exposure routes, an approach expanded an earlier EPA report on drinking
water DBP mixtures (U.S. EPA, 2000e) (Chapter 4);
integrating output from multiple effects modeling (illustrated using a
categorical regression model) with the HI and response addition models to
express risks for multiple health effects (Chapter 4);providing added detail on
the cumulative HI approach used by the Superfund program (U.S. EPA,
1989a), including discussion of the impacts for risk characterization (Chapters
4 and 5);
presentation of a method for cumulative risk characterization considers
factors unique to conduct of a Cumulative Risk Assessment, including the
recognition of uncertainties in cumulative dose-response and exposure
assessment (Chapter 5); and
a general emphasis on integrating exposure and dose-response analysis
(Chapters 3, 4 and 5).
1.3. EXISTING EPA PUBLICATIONS RELATED TO CUMULATIVE RISK
This report is linked to, and relies upon, several key guidance documents across
EPA, as illustrated by the examples in Figure 1-3. EPA's Office of Research and
1-10
-------
Development (ORD) has prepared and coordinated a number of major reports that
cover the topics shown in the next paragraph, and other EPA Program Offices have
developed issue papers and guidance documents on some of the key factors in
cumulative risk assessment. The general scope and timeline of these documents are
highlighted in Figure 1-4. (There are several other EPA guidance documents and
reports that address issues related to risk assessment, such as Data Quality Objectives,
but do not explicitly address the issues related to cumulative risk; they are discussed in
Appendix A.) Dates shown on that figure are for selected major reports within the
program areas; additional publications are described in the balance of this report (e.g.
see U.S. EPA, 2001 a, 2002a,b, 2003b). Other sections of this report describe
publications developed by ATSDR and other organizations that support cumulative risk
analyses. The publications shown in Figure 1-4 focus on distinct parts of cumulative
risk assessment rather than on all aspects described in the Framework. This is
because those documents were prepared to address specific issues as defined by (1) a
regulatory requirement, e.g., for air toxics, pesticides and drinking water, (2) a public
demand, e.g., for community-based studies or (3) a new assessment approach or
policy, e.g., for chemical mixtures or Planning and Scoping. Other reports will continue
to be developed to address the various steps and issues in the Framework.
To illustrate how certain cumulative risk topics are not covered when the scope is
limited to a targeted issue, consider three reports highlighted in Figure 1-4, each of
which focuses on human health risks (but addresses only one type of risk). The 2001
national air toxics assessment of more than 30 priority urban air toxics does not address
toxic interactions; however, default chemical mixture methods based on additivity
concepts are applied. The 2002 pesticide assessment only focuses on a limited set of
organic compounds, which act by the same toxic mode of action to exert the same
general effect. The 2000 Supplementary Mixtures Guidance does not address
aggregate exposures, only multiple chemicals by the same exposure route.
1.3.1. EPA Guidance Documents. The four steps of the risk assessment paradigm
(NRC, 1983), hazard identification, exposure assessment, dose-response assessment,
and risk characterization, provide the original foundation for risk-based programs across
many federal agencies (see Text Box 1-1). They are reflected in most EPA guidance
for assessing risks, such as the Risk Assessment Guidance for Superfund (RAGs)
(U.S. EPA, 1989a), which has served for many years as the common basis for
contaminated site cleanups and federal and state waste management programs. Other
programmatic risk assessment guidance documents, such as those addressing national
1-11
-------
Risk Analysis Phase
Planning and
Scoping Phase
Risk Assessment Guidance
for Superfund
(1989a)
Planning and Scoping for
Cumulative Risk Assessment
(1997 a)
Methodology for
Multipathway Combustor Emissions
(1998a)
Guidance for Assessing
Health Risks of Chemical Mixtures
(2000a)
Planning and Scoping
Lessons Learned
(-20029
Guidance on
Cumulative Risk of Pesticides
(2002cJ
Framework for
Cumulative Risk Assessment
(2003a)
FIGURE 1-3
Key EPA Resources for this Report: Precedent U.S. EPA Guidance and Reports
Containing Specific Approaches for Assessing Major Parts of Cumulative Health Risks
1-12
-------
Chemical mixtures
What: health risks for whole mixtures, for combinations of similar,
independent, & interacting chemicals
Why: update 1986 guidelines for multiple chemicals to enhance methods
Who: National Center for Environmental Assessment
When: 2000a (guidance)
Pesticides
What: health risks for common mode of action, multiple exposure routes
Why: address Food Quality Protection Act "no harm" requirements
Who: Office of Pesticide Programs
When: 2002a,c (organophosphates assessment and guidance)
Planning and scoping for cumulative risk assessment
What: description of concepts for up-front thinking to lay out process
Why: guide the first step, emphasizing broad scope & integrated dialogue
Who: Office of Science Policy
When: 1997a (guidance)
Planning and scoping lessons learned
What: summary of experience from studies since the 1997 guidance
Why: encourage formal planning & scoping of environmental assessments
Who: Office of Science Policy
When: 2002f (report with case studies)
Community-based pilot studies
What: range of multiple urban chemicals/sources, exposures, health effects
Why: address public concerns about combined risks in urban communities
Who: Regional Offices, with local organizations and citizen groups
When: Iate1990s - 2004 (individual studies)
Research needs for cumulative risk assessment
What: user-based evaluation of current programs, approaches, and needs
Why: focus and prioritize Agency research, leverage interagency efforts
Who: Office of Science Policy, with Regional Offices
When: 2002b (workshop summary)
National air toxics assessment
What: inhalation health risks of outdoor air toxics from multiple sources
Why: define baseline & driving chemicals/sources, prioritize data collection
Who: Office of Air Quality Planning and Standards
When: 2001 (national-scale report for 1996 data, updates coming)
Framework for cumulative risk assessment
What: description of umbrella issues, concepts, and general approaches
Why: guide overall integrated organization for many types of assessments
Who: Risk Assessment Forum
When: 2003a (framework report)
Disinfection byproducts in water
What: health risks of multi-route exposures to water treatment residuals
Why: address Safe Drinking Water Act "complex mixtures" requirements
Who: National Center for Environmental Assessment
When: 2000e, 2003b (initial risk report, other reports coming)
Multi-pathways Exposures forCombustor Emissions
What: health risks of multi-pathway exposures to combustor emissions
Why: address Clean Air Act requirements
Who: National Center for Environmental Assessment
When: 1998a (methods document)
Case studies for cumulative risk assessment
What: summary of examples, including community-based pilot studies
Why: provide insights to help others conduct cumulative risk assessments
Who: Risk Assessment Forum
When: 2006 (effort underway, no report yet)
Developing health risk assessment approaches for addressing
multiple chemicals, exposures and effects
What: combined health risks for multiple chemicals, pathways, effects
Why: provide simplifying methods and show feasibility
Who: National Center for Environmental Assessment
When: 2007 (this report)
FIGURE 1-4
Highlights of Cumulative Risk-Related Program Guidance and Research Reports
1-13
-------
air standards, drinking water standards and regulation of pesticides, also are structured
roughly along the four steps of the risk assessment paradigm.
In risk-based standard setting (e.g., setting a national safe exposure level for a
chemical), contaminants have historically been evaluated one at a time. Consider,
however, the example of the assessment of contaminated sites, where more complex
exposures are included; in RAGs, chemical exposures are summed across
environmental media and exposure pathways to estimate total exposures, cancer risks
and the combined potential for noncancer effects (U.S. EPA, 1989a). Although RAGs
calls for considering multiple chemicals, exposure routes and effects (thus cumulative
risks), few specific suggestions are provided that would enable an analyst to extend
analysis beyond the basic additive approach in the original EPA mixture guidelines
(U.S. EPA, 1986a), primarily because of limitations in current understanding of
environmental and toxicological interactions.
As knowledge of the environmental fate and toxicology of chemicals has
increased through ongoing research, the risk assessment process has kept pace. The
National Research Council has recommended moving away from the single-chemical
assessment focus (NRC, 1994), and the emphasis has continued to shift toward a
receptor- (population-) based focus. As noted in the 2000 Supplementary Mixtures
Guidance, the four originally distinct steps of the risk assessment paradigm are now
closely linked; in particular, it is useful to jointly conduct the exposure and toxicity
evaluations so that the exposure assessment can be refined based on toxicity
information and vice versa. During the past several years the EPA has published
several cumulative risk documents (as illustrated in Figure 1-4) that capture this shift
and extend assessment concepts beyond the original basic approach.
For example, the EPA Planning and Scoping documents identify iterative
Problem Formulation as a key element of the cumulative risk assessment process
(U.S. EPA, 1997a, 2002a). This broadens the process beyond the four original data-
driven analytic steps by bringing in the key scoping (or deliberative) component. The
Framework document defines a flexible structure that includes Planning and Scoping,
and Problem Formulation as well as specific assessment and characterization issues
(U.S. EPA, 2003a). That document describes main concepts and the underlying
technical factors across a range of risk types and applications. Together, this set of
EPA publications provides a general view of how risk analyses can better reflect real-
world conditions. These include complex exposure and effect processes as well as
"human interactions" that involve stakeholders and regulators discussing a given risk
issue to better understand and address cumulative risks.
1-14
-------
These EPA publications respond to the public's desire to bring together individual
pieces of the environmental risk picture (many of which are regulated under separate
federal programs) so risks that encompass all sources, stressors, exposures, affected
population groups and effects can be better understood and ultimately better assessed.
Thus, while the four-step NRC paradigm from two decades ago provided an essential
foundation, the approach for assessing health risks from exposures to chemicals in the
environment has evolved considerably since then.
One major difference from the historical approach is that today's analyses, in
terms of the scope of this report, are more closely integrated with careful attention paid
to potential interactions among them. Emerging science is offering new ways to
evaluate how one chemical could affect the behavior of another in the environment; how
one could affect how another is absorbed in, metabolized by and distributed in or
eliminated from the body; and whether their combined toxicity could differ from that
estimated from the single chemical toxicities. This report illustrates how this new
information can be applied to better address cumulative health risks. Sections
1.5.1 -1.5.3 provide detail on three existing EPA guidance documents that form a
foundation for addressing the multiplicity issues with the exposure and toxicity
assessment steps of cumulative risk, along with brief discussion of cumulative risk
areas not addressed in those documents.
1.3.1.1. Mixtures Risk AssessmentA common application of mixture risk
assessment methods is to Superfund waste sites. The applicable legislation passed in
1980, the Comprehensive Environmental Response, Compensation and Liability Act
(CERCLA), specifically calls for the evaluation of risks from mixtures. In the original
EPA mixtures guidelines (U.S. EPA, 1986a), the recommended approach was dose or
response additivity based on evaluations of individual chemicals. While interactions
were discussed and addressing them was recommended (if data were available), no
specific approaches were described because toxicological understanding and
quantitative data on interactions were limited. To help address this issue, the EPA
released the 2000 Supplementary Mixtures Guidance, which updates the earlier
guidelines by providing further methodologic detail that reflects evolving toxicological
knowledge. By describing a process for quantitatively evaluating toxic interactions of
multiple chemicals, that guidance offers a clear step forward from past practice.
Specific approaches address complex mixture risk values, environmental
transformations of complex mixtures, toxicological similarity based on varying evidence
(from similar toxic mechanisms to similar target organs) and toxicological interactions.
1-15
-------
A main issue not addressed in the 2000 Supplementary Mixtures Guidance is
approaches for assessing multipathway exposures to chemical mixtures as well as
approaches for multiple effects from chemical mixture exposures.
1.3.1.2. Superfund Site AssessmentRAGs, the standard guidance for
assessing health risks at Superfund sites, (U.S. EPA, 1989a) and subsequent
companion documents require the consideration of multiple chemicals, sources,
exposure routes, receptors and effects. Thus, a basic cumulative assessment is
already being conducted at Superfund sites. As mentioned previously, RAGs does not
explain how to assess toxic interactions because quantitative methods were limited
when it was published. Instead, a default approach was defined under which chemicals
are evaluated individually, and doses and toxic responses were assumed to be additive,
providing the first EPA Program Office approaches to component-based mixture risk
assessment. For independent toxic endpoints, such as different types of cancer,
component risks are added. For toxicologically similar endpoints, component doses are
scaled and added to form the familiar HI (see Chapter 4). RAGs also developed the
quantitative evaluation of multiple pathway exposures with the total Hazard Quotient
(HQ) concept (see Chapters 4 and 5). Because the HI and risk addition formulas of
these exposures used by the Superfund program relied on single chemical risk values
readily available from EPA's Integrated Risk Information System (IRIS) database (U.S.
EPA, 2007), the mixture assessment was feasible and continues to be widely
implemented. While RAGs represents a significant step in the development of
cumulative risk assessment methods, it does not discuss toxic interactions, the
screening approaches for multiple pathways are minimally described and key details on
how and when to use the total HQ concept are not presented.
1.3.1.3. Pesticide Group Cumulative Risk AssessmentFollowing the
passage of the Food Quality Protection Act (FQPA) in 1996, EPA programmatic
guidance was developed to address a much more focused risk than that of previous site
assessments. FQPA called for the estimation of health risk from combinations of
pesticides with a common toxicological mode of action, regardless of source. The
resulting cumulative risk guidance includes a modified HI formula for the mixture aspect
and an aggregate risk formula that is functionally similar to the total HQ formula in the
Superfund guidance (U.S. EPA, 2002c). Sophisticated guidance was developed for
evaluating toxicity data to decide which pesticides qualify for the common mode of
action group and for estimating the likely intakes from aggregate exposure from dietary
1-16
-------
and other sources based on multiple types of national or regional information (U.S.
EPA, 2001 a, 2002c). The cumulative risk guidance was then demonstrated by an
extensive risk assessment of the organophosphate pesticide group and its common
mode of action, cholinesterase inhibition (U.S. EPA, 2002a). An issue not addressed in
the pesticide guidance is that only the toxic effect for the common mode of action is
assessed, chemicals not sharing the common mode of action are not included and toxic
interactions are not addressed.
1.4. THIS REPORT'S APPROACH TO CUMULATIVE RISK ASSESSMENT
Many situations do not have a population focus or do not involve multiple
chemicals and so would not need a cumulative risk assessment. However, there are
certain scenarios which would naturally lead to conducting a cumulative risk
assessment, denoted here as initiating factors. Figure 1-5 shows these three identified
initiating factors along with the data elements that may be included in a cumulative risk
assessment. These initiating factors are (1) multiple pollutant sources or releases, (2)
elevated concentrations from environmental monitoring or biomonitoring of chemicals
and (3) increased population illness in a community. Figure 1-6 illustrates the types of
information that may be considered for data collection and population characterization
and shows the relationship of this information to the initiating factors. It is noteworthy
that traditional source-
based assessments are
usually initiated when
chemicals are found or
released into the
environment from known
sources. When this occurs,
population vulnerability
factors, such as diet,
behaviors, genetic traits,
economic status and social
characteristics are often not
included in the assessment.
(See Text Box 1-3 for a
discussion of the
challenges related to
expertise and
Challenges to Conducting Cumulative Risk Assessments
(Text Box 1-3)
A challenge to conducting cumulative risk assessments that include
non-traditional stressors is identifying expertise in the risk
assessment community for evaluating risks posed by such stressors
and developing organizational support for such efforts. In the U.S.
Federal Government, different Agencies have purview for related
exposures; no individual Agency has as its mission to evaluate all
chemicals and stressors together, so collaboration may become
important. For example, the Food and Drug Administration is
responsible for dietary stressors and Pharmaceuticals, so
collaborations with EPA would be useful regarding if and how these
stressors would impact a population also exposed to environmental
chemical pollutants. Furthermore, the development of a collaborative
network with the medical community and industry (e.g.,
pharmaceutical companies) also would help to integrate
environmental risk assessments with public health information,
exposure data and dose-response study results on toxic chemicals
and Pharmaceuticals. Establishing cross-organizational workgroups
and within-organizational structures would be an initial step towards
conducting and completing cumulative risk assessments. Similarly,
within the EPA, collaboration may become important among
established organizational units, e.g., among the program offices for
water, air, solid waste and pesticides.
1-17
-------
Initiating **'
Factors
Data
Elements
public health
data
Population
illness
Sources,
releases
multiple-
chemical
fate
population
subgroup
sensitivities
Combined
characterization
population
vulnerabilities
multi-route
exposures
mixtures
toxicity
Chemical
concentrations
FIGURE 1-5
Common Initiating Factors and Elements of Cumulative Assessments
1-18
-------
Initiating Factor:
Sources and
Releases
Initiating Factor: Elevated Environ-
mental Chemical Concentrations or
Biomonitoring Levels
Site
Sources:
stack
emissions,
surface
runoff,
leaching
Non-Site
Sources: food,
household
products,
indoor/outdoor
air pollution,
drinking water,
pesticides
Drug and
alcohol
abuse,
smoking,
cultural
practices
Poor
nutrition,
obesity,
physical
and
mental
hoalth
Polymor-
phisms,
gender,
age,
race
/
Environmental
Contaminants
Poverty,
education,
minority
status,
unemploy-
ment,
income,
residential
proximity to
sources,
family
dysfunction,
health care
access
Diet and
Behavior
Biological and
Genetic Factors
Socio-Economic
Stressors
Initiating Factor: Population Illnesses
or Perceived Population Illnesses
FIGURE 1-6
Variables Considered in Cumulative Risk Assessment and their Relationship to Initiating Factors
1-19
-------
organizational structure when these non-traditional stressors are incorporated into an
assessment.) These traits are more likely to be assessed when population illness or the
potential for illness are the initiating factor.
Figure 1-7 shows the key steps in a cumulative risk assessment, with a primary
focus of addressing multiple chemicals, pathways, timeframes and effects in a
population-based setting. These steps define the population of concern and its study
area, generate a list of environmental contaminants relevant to the initiating factor and
identify links between environmental chemical exposures and vulnerabilities within the
population. These steps form the initial collection and organization of information to
focus on the cumulative aspects of the risk assessment. These steps may not be
sequential and may involve a number of iterations as the analyst examines factors
related to population vulnerabilities, public health information, toxicological and
epidemiologic data, completed exposure pathways, differential exposures and contact
with environmental media and pollutant sources. Outputs include a population profile, a
list of relevant chemicals, chemical groups for use in risk analysis and characterization
and a conceptual model. Outputs may include additional epidemiologic evaluations that
assess the health of the community or that examine associations between health
impacts and pollutant exposures. The activities in a cumulative risk assessment that
are summarized in this chapter include:
Characterize the population or community of concern and the study area based
on the initiating factor
Develop a list of relevant chemicals
Compile information on exposure conditions and toxicity
Identify population subgroups who are sensitive to the relevant chemicals or
vulnerable to differential exposure
Iterate those steps to improve the relevance of the exposure and population
factors to the health risks of greatest concern
Conduct a risk characterization, including uncertainty and sensitivity analyses.
One important goal of the risk assessment process is to evaluate the strength of
any links between the chemical exposures to the receptor population and the
information or event that initiated the cumulative risk assessment. For example,
consider the case where awareness of multiple pollutant sources raises concerns of
cumulative health risk. The data from EPA's Toxic Release Inventory (TRI) might
include more than 20 chemicals, but it does not provide exposure levels or evidence
that all 20 chemicals reach anyone in the population of concern. Establishing those
links (e.g., between the TRI data and actual exposure) is a key part to many of the initial
1-20
-------
STEPS
OUTPUTS
1) Identify Initiating Factor
Population Profile
2) Characterize Population
based on Initiating Factor
I
List of Relevant Chemicals
3) Generate Chemical List
Conceptual Model
4) Identify Links between
Chemicals & Subpopulations
Epidemiologic Evaluations
5) Quantify Exposure
for General Population
& Subpopulations, Form
Initial Exposure Groups
6) Quantify Dose-
Response for Initial
Toxicity-based
Chemical Groups
7) Integrate Exposure & Dose-
Response. Refine Exposure and
Toxicity Assessments
8) Conduct Risk Characterization \-
Chemical Groups
by Media & Time
Chemical Groups
by Toxicity
Integrated
Chemical Groups
Final Cumulative RA
FIGURE 1-7
Key Steps in a Cumulative Risk Assessment. The interdependence of exposure and
toxicity assessments is indicated by blue arrows.
1-21
-------
steps of cumulative risk assessment. In this chapter, the steps are briefly described in
order to show their contributions to the cumulative risk assessment and their
interconnections (Figure 1-7). Subsequent chapters have more complete discussions
on exposure assessment (Chapter 3), toxicity assessment (Chapter 4), which are both
part of the Risk Analysis phase and Risk Characterization (Chapter 5).
1.4.1. Technical Approaches for Multiple Chemicals, Exposures and Effects.
Chapters 3, 4 and 5 deviate from the broad definition of cumulative risk assessment by
providing technical methods for evaluating only multiple chemicals, exposures and
effects. The approaches do not account for additional population variables, such as
those that are associated with vulnerabilities (i.e., those factors in Figure 1-6 related to
diet, behavior, genetics and biology, and social and economic factors). EPA does
address a few of these factors (e.g., sensitive subgroups, children, elderly), however, it
may be useful to conduct additional research on analyzing health risks for vulnerable
populations and to collaborate with other organizations that may have access to
relevant data (see Text Box 1 -4). The intent of these three chapters and Appendices A,
B and C is to provide a library of approaches and resources to more explicitly assess
the multifactor aspects of cumulative risks for specific scenarios and sites. Because of
the variability in these scenarios, such an assessment can involve evaluating many
different sources and contaminants, several media (soil, water and air) and associated
exposure pathways, various representative individuals or population groups who could
be exposed over different time frames and multiple health effects. The overall goal of
using cumulative risk approaches is to produce more accurate and informative
assessments of these sites and situations, leading to better decisions for managing
potential cumulative risks.
Additionally, Chapters 3, 4 and 5 of this report provide a structured collection of
approaches for addressing the chemical interactions and joint toxicity issues in
cumulative risk assessment. Chemical and toxicological interactions are a primary
focus because these are areas where methodological advances allow the traditional
process (evaluating chemicals individually) to be enhanced. Approaches for chemical
grouping are presented in order to simplify complexities and combine components for
joint analysis, so attention can be focused on the factor combinations that could
contribute most to causing adverse cumulative risks.
1.4.2. Identify the Initiating Factor for the Cumulative Risk Assessment. The initial
stage of a cumulative risk assessment (Planning and Scoping) forms a systematic,
1-22
-------
iterative process that defines the risk problem to be assessed and the technical
elements to be emphasized (U.S. EPA, 2003a). The main backdrop for Problem
Formulation and initial data review is provided by the regulatory context and the
particular information or technical factors that led to the decision to consider undertaking
a cumulative risk assessment. Three typical initiating factors include multiple pollutant
sources within the community, increases in illnesses in the population and elevated
chemical concentrations due either to monitoring of environmental levels or
biomonitoring of chemicals in humans (e.g., in blood or urine samples). These initiating
factors could occur in any community, but environmental justice considerations may
cause a cumulative risk assessment to be undertaken more readily because of the
proximity of a community to pollutant sources or cultural practices of a population that
may cause it to be differentially exposed. Figure 1-5 shows these initiating factors and
the common data elements that link the initiating factors with the population. These
initiating factors can be displayed within the preliminary conceptual model that is
developed during the Problem Formulation phase. The identification and discussion of
initiating factors in the planning stages may improve the understanding of any links
between the population risk estimate, which is the result of the cumulative risk
assessment, and the initiating factor, which initiated the assessment.
After the initiating factor has been characterized, the next steps involve defining
the population of concern and its study area, generating a list of environmental
contaminants relevant to the initiating factor and identifying links between environmental
chemical exposures and vulnerabilities within the population. Then, data are collected
and organized with a focus on the cumulative aspects of the risk assessment. These
steps, 2-4 in Figure 1-7, may not be sequential but may involve a number of iterations
as the risk analyst examines factors related to population vulnerabilities, public health
information, toxicological and epidemiologic data, completed exposure pathways,
differential exposures and contact with environmental media and pollutant sources.
Outputs from steps 2-4 include a population profile, a list of relevant chemicals and a
conceptual model. They may also include additional epidemiologic evaluations that
assess the health of the community or examine associations between health impacts
and pollutant exposures.
1.4.3. Characterize the Community and Population Based on the Initiating Factor.
The population characterization usually would include a physical description of the study
area and a demographic description of the population in that study area. The study
area could be a political unit, such as that defined by a county or city boundary or could
1-23
-------
be delimited by geographic features, such as a lake and surrounding watershed. The
population may be a neighborhood or the community in an entire city, perhaps an Indian
reservation or the public using a resource, such as a lake. The population description
would also include sensitive or susceptible subgroups based on increased exposure,
genetic or physical traits or vulnerability. This description may include cultural practices
or housing locations that differentially affect a population's exposure. If population
illness is the initiating factor for the cumulative risk assessment, then it would be
consistent with good epidemiologic practices to determine the extent of the morbidity or
mortality and the uniqueness of the disease or the disease rates in comparison to
baseline levels in other communities. Often the definition of the population and study
area could be influenced by the initiating factor. Because a cumulative risk assessment
is population-focused so that all relevant exposures and effects are considered, as the
potential exposures and health effects are further investigated, the population
characterization will be refined.
1.4.4. Generate Initial List of Relevant Chemicals. The Framework distinguishes
cumulative risk assessment from traditional risk assessments by its population focus.
Consequently, once the initial population description is complete, including population
demographics and the boundaries of the study area, information on chemical releases,
biomonitoring data, public health information and environmental concentrations are
evaluated in light of the identified population to develop the initial list of relevant
chemicals (see discussion in Text Box 1-4 where it is recognized that other confounding
factors may also be responsible for
health effects in the population).
Existing EPA approaches for
exposure assessment are likely to
be sufficient for this step. Partly
because of stakeholder
involvement in the cumulative risk
assessment, this initial list of
relevant chemicals is likely to be
closely tied to the initiating factor.
The influence of the initiating factor
is discussed in more detail in the
exposure assessment chapter
(Chapter 3).
Chemical and Stressor Involvement in Cumulative Risk
Assessment (Text Box 1-4)
When a cumulative risk assessment is initiated by health
effect(s) in the population, and there is reason to believe an
environmental exposure may be the cause, the initial goal
of the investigation is to determine if environmental
chemicals present in the affected community can be linked
to those health endpoints. It may be noted that other
stressors within the population may be responsible for either
causing health effects or for contributing to their expression
in conjunction with chemical exposures. Examples of such
confounding factors include contributions to various cancers
from smoking, hearing loss from co-exposure to noise and
chemicals and associations between high blood pressure
and stress. As such, investigators conducting a cumulative
risk assessment may find it useful to make note of such
stressors that may contribute to occurrence of a health
endpoint in addition to developing a list of relevant
chemicals. This information can then be taken into account
during uncertainty analysis and risk characterization.
1-24
-------
1.4.4.1. Use Program and Regional Office Procedures Determination of
relevant chemicals is covered in several guidance documents from the EPA Program
Offices (e.g., Office of Water, Office of Pesticide Programs, Office of Solid Waste and
Emergency Response, Office of Air Quality, Planning and Standards) and Regional
Offices (see Appendix A). For exposures by multiple media, the chemicals may be
identified using approaches from several programs or guidance from EPA's ORD (e.g.,
U.S. EPA, 1996a). The initial chemical list may be overly inclusive to allow for the
examination of potential interactions from joint exposures so that joint toxicity can be
evaluated in later steps of the assessment. For example, a chemical could be
evaluated based on the ratio of its exposure level to its safe level, i.e., its HQ. A
chemical that might be screened out in a single chemical assessment because its HQ is
less than 1 might be retained in a cumulative assessment (e.g., unless it's HQ is less
than 0.1) in order to allow for potential dose additivity or interactions.
1.4.4.2. Identify Chemicals Related to the Initiating factor The three types
of initiating factors in this report have only subtle differences in their influence on the
chemical list. When health endpoints are the initiating factor, the preliminary list of
chemicals could include any that have been shown in human or animal studies to cause
or contribute to those health effects. When environmental concentrations, biomonitoring
data or pollutant sources are the initiating factor, the preliminary chemical list could at
first be restricted to those measured or likely to be found in environmental emissions.
Those that lack toxicity information or are initially deemed unlikely to pose significant
health risks based on human or animal data may be placed on a watch list pending
further analysis during the iteration of the exposure and toxicity assessment steps.
Chemicals known to be similar to or toxicologically interactive with those on the
preliminary chemical list might then be added if their exposure to the identified
population is considered plausible, such as similar chemicals in food. It is consistent
with chemical mixtures risk assessment practices to consider multiple endpoints for
each chemical, not just the critical effect used to define the EPA's IRIS risk values (U.S.
EPA, 2007) to allow for determination of potentially interactive chemicals. In any case,
the resulting list of chemicals is preliminary and perhaps most useful in refining the
population description by identifying subgroups that could be sensitive to chemicals on
this list.
1.4.5. Identify Links between Chemicals and Subpopulations. Once the general
receptor population has been identified and characterized and the preliminary chemical
1-25
-------
list exists, the next step is to examine the potential for exposures and differential
exposures to those chemicals among population groups, including sensitive or
vulnerable population subgroups in the defined population of concern. Certain
population groups may be particularly sensitive to toxic chemicals because of higher
exposure or increased vulnerability. Higher exposures can often be estimated by
considering lifestyle information (e.g., U.S. EPA, 1997a) and occupational data (e.g., as
conducted by the National Institute for Occupational Safety and Health
http://www.cdc.gov/niosh/homepage.html). One difference for cumulative risk
assessment is that elevated exposures can include the combined exposure to multiple
toxicologically-similar chemicals, such as chemicals in the workplace or lifestyle
exposures (e.g., food sources) that are not on the preliminary chemical list. Because of
the population focus and stakeholder involvement, cultural or other lifestyle factors
might be identified by stakeholders that could suggest additional sources of chemicals
or exposure levels of significance that could lead to additional sensitive population
subgroups (Figure 1-6). Vulnerability can be more complex, ranging from existing
disease (e.g., hospital patients, individuals receiving outpatient treatment) to genetic
predisposition (e.g., for some lung cancers) to socioeconomic factors (e.g., access to
health care). Vulnerability is discussed in some detail in the next chapter but many
issues are poorly understood and are the foci of current research.
The chemical list may then be combined with the description of likely sensitive
population subgroups. This information could be arranged in several ways. For
example, a table could list the chemicals ranked by the strength of their link to the
initiating factor. Such a table might be arranged as follows:
Tier 1 Chemicals are linked directly to population subgroups through biomonitoring
and are identified in emissions from one or more sources
Tier 2 Chemicals are linked indirectly to population subgroups by association with
elevated disease in the population and are identified in emissions from one or
more sources
Tier 3 Chemicals are linked to sensitive subgroups of the population of concern
based on human data and are identified in emissions from one or more
sources
Tier 4 Chemicals are linked to sensitive subgroups of the population based on
extrapolations from experimental animal studies and are identified in
emissions from one or more sources
Tier 5 Chemicals are identified in emissions from one or more sources and are
identified by their potential for joint exposure (e.g., by multiple routes) or joint
toxicity with other chemicals on the list
1-26
-------
Tier 6 Chemicals are identified in emissions from one or more sources but lack
toxicity information or are initially deemed unlikely to pose significant health
risks based on human or animal data; these chemicals are placed on a watch
list pending further analysis during the iteration of the exposure and toxicity
assessment steps
1.4.6. Quantify Human Exposures for Initial Exposure Grouping. Up to this point,
there has been no actual exposure assessment, only a listing of chemicals. Extensive
EPA documents provide guidance for conducting assessments for the three major
routes of exposure: dermal, oral and inhalation (see Chapter 3 for details and citations).
For multiple sources and pathways, detailed exposure guidance exists for combustor
emissions (U.S. EPA, 1998a, 2005b) along with programmatic guidance on Superfund
sites and multiple pesticide exposures (U.S. EPA, 1999a,b). In general, the
assessment might rely on guidance across several Programs or from ORD. For
example, general exposure guidance and information on exposure factors are available
from the National Center for Environmental Assessment (NCEA; U.S. EPA, 1992a,
1997c, 1998a, 2002i), guidance on aggregate exposures to pesticides is available from
the Office of Pesticide Programs (U.S. EPA, 1999e, 2001 a), guidance on exposure from
hazardous waste combustion facilities is published by the Office of Solid Waste and
Emergency Response (OSWER) (U.S. EPA, 2005b) and dermal exposure to soil is
covered by the supplemental OSWER guidance (U.S. EPA, 2004a).
Quantification of exposure for cumulative risk assessment begins with a clear
definition of the population and study area so that the analyst can identify all existing
and future completed pathways. Monitoring data for chemical concentrations and
information from epidemiologic studies or public health databases may be used as
starting points for any exposure modeling that is done. The assessment may also
identify the relevant exposure factors, with particular attention to unique factors for the
sensitive subpopulations; such factors, e.g., cultural practices, may be used to adjust
the exposure assessment based on differential exposures. Once the exposure is
characterized for the population of concern and its sensitive and vulnerable
subpopulations, the next step is to attempt to simplify the combinations of chemicals,
pathways and timing (including duration and intermittency of exposure) by grouping the
chemicals according to timing and either medium or pathway (see Chapter 3 for details).
Any issues that cannot be quantified may be described qualitatively regarding
their relative importance to the population exposure and for possible future
quantification, should information become available. Information from the dose-
1-27
-------
response assessment would be useful in this evaluation of those unquantified issues,
particularly in terms of exposures of sensitive or vulnerable subpopulations.
1.4.7. Quantify Dose-Response for Initial Toxicity Grouping. The focus of toxicity
assessment regarding cumulative risk revolves around timing issues of exposure and
toxicity. The chemical grouping resulting from the exposure quantification (by timing,
media and pathway) is further evaluated in terms of toxicological timing factors:
toxicological overlap of internal dose, kinetics interactions, toxicodynamic interactions
and persistence of effects (see Chapter 4 for details and additional references).
Simultaneous exposures are the ones most often evaluated for potential joint toxicity,
but sequential exposures can also result in joint effects. Initiators and promoters of
cancer and delayed or persistent toxicity are examples where potential joint toxicity
could occur from exposures at different times.
During this step, chemicals previously put on the watch list may be re-evaluated
by considering the potential or expected toxicities at the estimated exposure levels.
Toxicological interactions could be further considered for the watch list chemicals,
structure-toxicity relationships or other similarity procedures, as could interactions
involving characteristics of the sensitive subpopulations. An example of the latter
interaction is nutritional deficiencies enhancing toxicity of some metals (U.S. EPA,
2004b). Any dose-response or other toxicity issues that cannot be quantified may be
described qualitatively, especially regarding importance to potential health effects in the
sensitive subpopulations.
1.4.8. Integrate Exposure and Dose-Response Information. In this final analysis
stage, the exposure assessment is interfaced with the dose-response assessment in
order to refine the information on joint exposures of main toxicological significance and
to identify timing issues of most concern regarding increased toxicity. Any matches of
toxicity overlaps (toxic interactions or persistent effects) with exposure overlaps are
highlighted for consideration of improvements in the exposure information. Ideally, this
step would occur throughout the assessment process. The refined exposure and
toxicity characterizations and the resulting initial risk estimates, the products of this step,
are the main inputs to the Risk Characterization.
1.4.9. Conduct Risk Characterization. A Risk Characterization is usually described in
EPA guidance as having two parts: an integrative analysis, which contains the risk
estimates and can be highly technical, and a risk characterization summary, which
1-28
-------
focuses on recommendations and uncertainties. Figure 1-8 provides an overview of the
final Risk Characterization process. It is an expansion of the final Rick Characterization
step shown in Figure 1-7, beginning with outputs from the steps shown in Figure 1-7,
such as, the population profile and the integrated chemical groups. Figure 1-8 is
presented again in Chapter 5 (as Figure 5-1) and explained in greater detail. The
cumulative Risk Characterization may differ from a traditional Risk Characterization in
several ways (detailed in Chapter 5) that are often caused by missing data or a lack of
understanding of the various multiples and their interactions. Some of the more
important differences are listed below:
Recommendations could be multivariate, i.e., the analyst might not be able to
identify a single chemical, pathway or critical effect that drives the risk
Recommendations might be based on groupings of chemicals, pathways and
effects, but such groupings can be based on subjective judgments
Recommendations might be based on epidemiological findings relevant to a
population illness, for which it is useful to articulate confounding factors and
exposure uncertainties
Uncertainty analysis might be predominantly qualitative because of the use of
numerous defaults, e.g., for addressing interactions and multiple effects
Time dependence of exposure and mixture composition might be addressed by
surrogates (e.g., annual averages) or simplified factors (e.g., index chemical
concentration) resulting in complex analyses and unknown information gaps
1.8. SUMMARY
Many site and situation health risk assessments can be adequately addressed
using single chemical and single pathway evaluations. At other sites and in other
situations risk analysts may choose to evaluate population vulnerabilities, multiple
chemicals and complex exposures; in these cases cumulative risk assessments will be
undertaken. Many basic cumulative risk conceptsincluding consideration of multiple
sources, chemicals and exposuresare in the standard guidance from the last
15 years. This report builds on those standard EPA guidance approaches along with
new approaches so that together they provide the conceptual and procedural
methodology that in many cases will be feasible and sufficient for addressing the
multiple factor issues with cumulative risk assessment.
1-29
-------
Populations ConsideredGeneral Population or Subpopulations
Exposure
Groups
Toxicity
> Groups
If
Temporal Analysis
Do relevant multi-chemical
exposures or persistent
effects overlap?
Integrated Groups for Cumulative
Risk Assessment
Extrapolations,
Simplifications
and Omissions
Acceptable?
Conduct Qualitative
Assessment Only or
Revise Analytic Scope
Develop Quantitative Risk Estimates
For General Population and Relevant
Subpopulations
For each Relevant Chemical Group
For each Relevant Pathway and Exposure Route
For each Relevant Timeframe
Conduct Single Chemical Assessment Only,
Discuss All Cumulative Risk Elements Considered,
Describe Data Gaps for Future Research
Document Cumulative Risk Characterization
Including Risk Estimates and
Uncertainty Discussion
Describe Data Gaps for Future Research
Sensitivity and Uncertainty Analysis
Identify Sources of Uncertainty
Develop Integrated Sensitivity Analysis, if
Possible
-Identify Sources of Model Uncertainty
-Identify Sources of Parameter Variability and
Uncertainty
Identify Critical Research Needs
FIGURE 1-8
Schematic of Cumulative Risk Characterization Approach in this Report
1-30
-------
2. IDENTIFICATION OF INITIATING FACTORS, POPULATION CHARACTERISTICS,
DATA COLLECTION AND ORGANIZATION
This report has identified three initiating factors of a cumulative risk assessment.
These include multiple pollutant sources within the community, increases in illnesses in
the population, and elevated chemical concentrations, due to either monitoring of
environmental levels or biomonitoring of chemicals in humans. Figure 2-1 shows
examples of these initiating factors and of the data elements that may be considered in
a cumulative risk assessment. This chapter discusses these initiating factors and data
elements and shows their interconnections. Section 2.1 describes the initiating factors.
Sections 2.2 and 2.3 describe the preliminary evaluations of population and exposure
information including the influence of the initiating factors on those evaluations. Section
2.4 discusses the importance of incorporating public health data on a community into
the cumulative risk assessment. Section 2.5 describes epidemiologic approaches to
addressing community concerns when a cumulative risk assessment is initiated.
Section 2.5 describes the linking of population and exposure information to identify any
subgroups within the population that would be sensitive to effects from those exposures
and the use of conceptual models to help organize the information and analysis.
Chapters 3-5 provide a more detailed evaluation and quantification of exposure, dose-
response, the interface between these, and then the cumulative Risk Characterization.
2.1. INITIATING FACTORS FOR CUMULATIVE RISK ASSESSMENT
2.1.1. Health Endpoint as the Initiating Factor. When there is a perceived or
documented increased incidence of one or more health effects in a community with no
clear cause, there can be a demand for an investigation by the public. Initial
investigations should focus on examining whether the health endpoints are, in fact,
elevated. If an increased incidence of disease is not found, or if the elevated rates are
considered a statistical artifact, then further investigations may not be warranted. If
additional investigations are needed, then exposure assessments may be conducted
either separately or as part of an epidemiologic study. Many health endpoints have
been associated with several possible chemical causes, so these investigations may
initiate a cumulative risk assessment. For example, in the 1970s a cluster of leukemia
cases in Woburn, Massachusetts initiated exposure assessment studies (Durant et al.,
1995; Parker and Rosen, 1981) and epidemiologic investigations (Lagakos et al., 1986;
Cutler etal., 1986; Public Health Service, 1981; Telles, 1981) in the area. Although the
eventual focus was on trichloroethylene exposures, an initial investigation focused on
2-1
-------
multiple industrial
facilities
and disposal areas,
accidental chemical
releases -
cluster of
leukemia cases,
efeyated cancer rates
organics in air
or soil,
transported to
water and
accumulated in
fish
inhalation,
ingestion, dermal
exposures from
air, water, soil,
, fish, produce
high blood lead
levels in children,
high levels of
chemicals found in
soil or indoor dust
genetic
susceptibility,
children,
elderly
incidence of infant
mortality, hospital
admission rates
public health
data
Population
illness
multiple-
chemical
fate
population
subgroup
sensitivities
Integrated
characterization
population
vulnerabilities
m u It i-route
exposures
mixtures
toxicity
Chemical
concentrations
homes close
to pollutant
sources, poor
health care,
subsistence
fishers
aroclor: reproductive effects,
diesel exhaust: lung cancer,
drinking water disinfectant
byproducts: bladder cancer
FIGURE 2-1
Example Initiating Factors and Data Elements for Cumulative Risk Analyses
2-2
-------
several organic compounds while a later study investigated metal exposures (U.S. EPA,
2005a). In many cases, existence of higher than expected health effects (based on
disease prevalence or incidence measures) is not easily connected to a cause, so the
initial investigation might begin with a critical examination of the available health effects
information. Variation in the quality of such information can be high, ranging from
anecdotal articles in the press to peer-reviewed data published in scientific journals. An
examination of the details and the quality of initiating factor information by stakeholders
and investigators is a primary objective of the Planning and Scoping stage.
2.1.2. Chemical Concentrations as the Initiating Factor. The detection of toxic
chemicals in the environment, or in the human body, at elevated concentrations may
initiate a cumulative risk assessment. For example, elevated levels of urban smog due
to ground level particulates and ozone can frequently lead to public health intervention
(e.g., advisories for young children and elderly to stay indoors). When community
members become aware that such exposures may be combined with elevated chemical
concentrations in soil and groundwater (e.g., heavy metals), a cumulative risk
assessment may be conducted, such as the Cumulative Risk Initiative for Cook County,
IL and Lake County, IN (see U.S. EPA, 2003a, p. 32).
What is considered "elevated" may be situation specific and could be determined
through various interactions among environmental engineers, regulators, exposure
analysts and toxicity experts. For example, exposures to environmental contaminants
may be higher than engineering goals, exceed regulatory levels issued by various
Agencies (see Appendix B for examples), or be of concern based on positive
toxicological data either on single chemicals or from studies where several of the
chemicals have been studied as a mixture. As when health effects are the initiating
factor, it is important to document the quality and variability of the concentration data
and whether such measurements indicate possibly complete exposure pathways for use
in the Risk Analysis and Risk Characterization phases. Available concentration data
have even greater influence in initiating a cumulative risk assessment when there are
elevated levels of additional chemicals elsewhere, such as in food, that also impacts the
same population. For example, in the 1990s, elevated levels of 2,3,7,8-chlorine
substituted polychlorinated dibenzo-p-dioxins (PCDDs) were found in fryer chickens as
the result of a contaminated mineral additive (i.e., ball clay) in their feed (Ferrario et al.,
2000); such an exposure in the population's food supply would be important to consider
in a cumulative risk assessment with additional exposures to PCDDs from other
sources.
2-3
-------
Biomonitoring data may also serve as a cause for health concerns for a given
community. The Center for Disease Control (CDC) has published its third National
Report on Human Exposure to Environmental Chemicals, which provides an
assessment of the exposure of the U.S. population to environmental chemicals using
biomonitoring (CDC, 2006). Through the National Health and Nutrition Examination
Survey (NHANES), which is a series of surveys designed to collect data on the health
and nutritional status of the U.S. population, the CDC measures levels of chemicals or
their metabolites in blood and urine samples from randomly selected participants. This
report includes exposure data for the U.S. population for 148 environmental chemicals
over the period 2001-2002 and can be used to compare these national distributions with
levels measured in a given community. If elevated levels of toxic chemicals are
observed in biomonitoring data (e.g., elevated blood lead levels) in a specific
community, then information may be gathered to examine increases in morbidity or
mortality in the community or potential sources of exposure to such pollutants. Although
biomonitoring for toxic chemicals may not be as routine as sampling for chemical
concentrations, collection of such data is becoming more frequent and may be useful in
identifying community concerns and potential health risks.
2.1.3. Multiple Sources or Release Events as the Initiating Factor. Multiple sources
of chemical contamination can be an initiating factor for a cumulative risk assessment,
often when they are the pending consequence of a proposed change, such as an
upcoming siting decision for a new manufacturing plant. Observations of multiple,
uncharacterized releases can also elevate concerns. For example, repeated
discharges from multiple outfalls into streams have led to actions by Georgia
Riverkeeper conservation groups, ranging from lawsuits to scientific sampling of the
water and biota (Richardson, 2004).
For a cumulative risk assessment initiated by a multiple sources or release
events, one of the first activities is to identify all relevant sources of potential exposure
to the population of concern, particularly sources releasing chemicals similar to those in
the sources. For example, an investigation into possible pesticide drift to a residential
neighborhood from nearby farms may warrant a concurrent evaluation of exposures
from household use of similar pesticides by rural residents in the geographic area of
concern.
2-4
-------
2.2. INITIAL DESCRIPTION OF THE POPULATION
In contrast to the source-based approach that begins with releases and
addresses all populations impacted by those releases, a receptor-oriented assessment
begins by defining the population group of interest and addressing all sources impacting
that population. The population group could be determined by geographic,
demographic or other criteria. This population group can be identified from the findings
of a recent exposure study, or the population may be chosen simply to reflect locations
of concern to U.S. EPA or to stakeholders. These locations can range from school
yards or parks to homes and Native American lands. During this process, vulnerabilities
of the population may be identified, including sensitivities and susceptibilities, general
health and nutritional status, and factors that may cause differential exposure (e.g.,
lifestyle factors, cultural practices, dietary factors such as subsistence fishing, activity
patterns and proximity of homes, playgrounds or farms/gardens to a pollutant source,
etc.). Under this approach, vulnerable subpopulations can be identified and exposures
traced back to evaluate all pathways by which a given subpopulation could be exposed
to a variety of chemicals. As described in the EPA's Framework document, this
approach is often applied to community-based cumulative risk assessments (U.S. EPA,
2003a). It can also play a role in other applications that are typically source-based. For
example, the assessments for contaminated sites could use a population-based
approach to address a specific group for which unique exposure or
vulnerability/susceptibility issues are of concern (see Chapters 3 and 4). The analysis
plan for a cumulative risk assessment could then reflect a combination of source- and
receptor-based approaches.
When considering health effects in the population, the cumulative risk
assessment addresses both existing health effects found in actual populations and also
the potential for effects that may occur in later years (e.g., cancers that are expressed
only after a long latency period). This is consistent with current Agency practices, for
example, in Superfund site assessments where risk assessments evaluate health risks
based on both current and future land uses and possible exposure pathways in the
present and in the future (U.S. EPA, 1989a). When a cumulative risk assessment
initiating factor is tied to a specific population (e.g., actual or perceived elevations in
adverse health effects or the presence of chemicals found through biomonitoring), then
the population may be specifically characterized by different vulnerability factors such
as age distributions and other socio-demographic data. However, an equally important
case occurs when a cumulative risk assessment initiating factor does not necessarily
point to a given population or community, e.g., multiple pollutant sources are in a
2-5
-------
general area or levels of monitored environmental chemicals are increased. This
population at risk may not be easily identified initially, but may still be at potential risk of
expressing health effects at some time in the future. Thus, although it is preferable to
characterize existing populations with known exposures to environmental chemicals and
observable health conditions, there is a need to define future populations with expected
or anticipated exposures to environmental chemicals which may have uncertain impacts
on human health. In either case, a cumulative risk assessment may be appropriate
given community interests and perceptions and if the weight of evidence suggests that
exposures to multiple chemicals may lead to significant health effects in the population
of interest.
2.2.1. Preliminary Characterization of the Population Based on the Initiating
Factor. The initial population characterization usually includes a description of the
study area and the relevant population. The initiating factor could influence whether the
study boundary or the population is defined first. Consequently, the initial population of
concern could be the community in an entire city or county, especially any identified
sensitive or susceptible population subgroups. Alternatively, the initial population could
be those in frequent contact with a geographic area, such as a park or lake. However,
sometimes the stakeholders and analysts agree after further evaluation that the
initiating factor is of lesser significance, and that another initiating factor will be the key
motivation for continuing the cumulative risk assessment. The initial description of the
study area and population of concern are considered to be preliminary and are subject
to change during the course of the risk assessment.
2.2.1.1. Population Defined by the Health Endpoint If a population group is
associated with the health effect initiating factor, then this group would automatically be
included in the initial population of concern. For example, if the initiating factor is an
increased absence from school for children 12 years and younger because of
respiratory problems, then that group of children forms the initial population of concern
and certain sensitive subgroups could be further examined (e.g., asthmatic
schoolchildren). Because cumulative risk assessment can include multiple endpoints,
the population could be initially defined in broad and somewhat vague terms, with
refinement following the later steps when links are determined between the initiating
factor health endpoints (as well as other endpoints) and chemical exposures.
2-6
-------
2.2.1.2. Population Defined by Chemical Concentrations When elevated
environmental chemical concentrations are detected, the monitoring locations can act
as initial bounds of the study area. If transport is plausible for those chemicals, then the
study area and population can be much larger than the initial release sites or monitoring
locations. When elevated biomonitoring data are detected, then the homes or business
locations of those people being tested can act as initial bounds of the study area. If
feasible, determining the source(s) of the chemicals found in biological samples (e.g., in
blood or urine) may be a priority. Identification of the source(s) would then provide
information to further refine the study area and population at risk for the study.
Chemical concentrations limited to specific resources or geographic features can
define a study population according to those with likely access to that resource or
location. Contamination of a recreational lake might lead to the population being
defined as those known and potential users of the lake; this might include recreational
anglers and their families and friends who might consume fish caught in the lake. At
this stage, the identification of sensitive population subgroups might be based only on
known sensitive groups in the defined population.
2.2.1.3. Population Defined by Multiple Sources When multiple sources
are the initiating factor, exposures have typically not yet been estimated. The initial
boundaries of the population of concern then might be roughly defined by possible
dispersion or deposition characteristics of existing and possible future emissions as well
as populations with possible future exposures (U.S. EPA, 1998a). The initiating factor
sources could initially be considered in isolation. As the assessment proceeds, the
refinements would consider all sources so that each pollutant source would be
evaluated both for its individual incremental population risk as well as in the context of
combined risk with other sources.
2.2.2. Refining the Population Profile Based on Vulnerable Subpopulations. Once
the initial population characterization and study area have been defined, vulnerabilities
within those populations may be identified in a cumulative risk assessment. EPA's
Framework adopts "vulnerability" concepts that encompass the topic of receptor
characteristics. Four areas are articulated where "human and biological ecosystems,
communities, and populations may be vulnerable: susceptibility/sensitivity, differential
exposure, differential preparedness (e.g., disease immunizations), and differential ability
to recover." Given this context, receptor population characteristics may include diverse
factors such as genetic susceptibility, age, stress, disease state, economic status,
2-7
-------
ethnicity, health status, proximity to sources, activity patterns, etc. Once the potential
vulnerability factors are identified, risks may be calculated separately for populations
with specific receptor characteristics. Risk assessments stratified by subpopulations
can be conducted in a stepwise manner, beginning with single chemical assessments
for that subpopulation and expanding the analysis to examination of risk associated with
cumulative exposures.
Epidemiologic studies often involve examination of whether certain receptor
characteristics (i.e., vulnerability factors) contribute to the toxicity caused by chemical
mixture exposures. If certain factors may confound associations of interest they are
addressed by statistical adjustment or through specific design features. Effect measure
modification may also be considered if vulnerability factors modify the main association
between exposure and disease. For example, Perera et al. (2003) reported differential
effects of polycyclic aromatic hydrocarbons(PAH)-related exposures with lower mean
birth weight and smaller head circumference among African Americans versus
Dominican infants born in New York City. These data suggested that minorities may be
differentially exposed to environmental tobacco smoke (ETS), increasing their
susceptibility to environmental PAH levels. Additional research showed a multiplicative
effect between ETS exposure and a molecular marker of PAH exposure (benzo [a]
pyrene-DNA adduct), despite no PAH-related developmental effects in the absence of
ETS (Perera et al., 2004). For a cumulative risk assessment, a factor such as
differential exposure to ETS may be taken into account when evaluating the potential
health effects of an environmental mixture. In the assessment of rural communities, the
literature suggests that impacts from exposures to mixtures of pesticides may be
evaluated from a cumulative risk perspective (see Text Box 2-1).
2.3. INITIAL ASSESSMENT OF EXPOSURE DATA
Once a cumulative risk assessment is initiated and the population and study area
are defined, an initial exposure assessment is conducted. This section provides a
general description of the types of chemical information likely available and initially
needed in the early part of the exposure assessment process and its dependence on
the initiating factor. It also discusses specific population data to be collected initially in
order to conduct the exposure assessment. Chapter 3 discusses specific approaches
to cumulative exposure assessment.
2-8
-------
2.3.1. Initiating the Exposure Assessment when Health Endpoint is the Initiating
Factor. When an increased incidence of health endpoints initiates an assessment, and
exposures to environmental chemicals are suspected to be the cause, the initial goal of
the investigation is to determine if environmental chemicals present in a community are
linked in some way to those health endpoints. Specific populations may also be
evaluated for sensitivity to
the identified health
effects, the potential for
chemical exposures to
exacerbate an existing
condition in sensitive or
vulnerable individuals or
people who may have an
impaired ability to resist
these specific illnesses
due to social factors (e.g.,
poor nutrition or health
care access). These
types of analyses are
similar to epidemiologic
investigations, such as
those conducted to
determine //and why
there are elevated rates
Example of Pesticides and Farmer Characteristics
(Text Box 2-1)
A large, prospective epidemiologic study, The Agricultural Health
Study, is an ongoing effort to evaluate health effects in agricultural
cohorts in North Carolina and Iowa from pesticide exposures
(Alavanja et al., 1996). One component of this study examines the
impacts of lifestyle, cultural, ethnic and genetic factors (i.e.,
vulnerability factors) on the health of farmers in conjunction with
pesticides exposures, making it an important contribution to the
literature on cumulative risk assessment. Results from this study will
likely be published for years to come, but a few articles are already
available. Current results include
increased prostate cancer risk for study subjects with a
family history of prostate cancer (Alavanja et al., 2003);
increased prostate cancer risk for applicators over 50 years
in age who used chlorinated pesticides (Alavanja et al.,
2003);
identification of poor financial condition of the farm, limiting
the purchase of safety equipment, as a significant risk factor
for acute effects from high pesticide exposure events
(Alavanja etal., 2001);
higher pesticide exposures, resulting in more pesticide-
related health effects in white farmers than in black farmers.
The higher pesticide exposures may be explained by farm
characteristics or economics (Martin et al., 2002) and
association of specific pesticides (i.e., paraquat, parathion,
malathion, chlorpyrifos, thiocarbamate) with respiratory
symptoms of farmers (Hoppin et al., 2002).
of female breast cancer in a region (Aschengrau et al., 2003; Paulu et al., 2002). Text
Box 2-2 provides an example of an illness initiating factor which was initially attributed to
general organophosphate poisoning but later focused on exposure to a single pesticide.
In this case, a sensitive subpopulation, i.e., a group of children, who may have also
been differentially exposed due to their activity patterns, became ill due to an illegal
pesticide application.
Health registries can serve as important resources for evaluating the potential
health impacts of environmental exposures for cumulative risk assessments. For
example, most states maintain cancer registries, as do national organizations and some
federal agencies, e.g., the National Cancer Institute. Birth defect registries also exist in
over 30 states, but the quality of most data in these registries is considered inadequate
2-9
-------
for an effective tracking program (EHTPT, 2000), particularly regarding the implications
of linking such effects with environmental exposure to multiple chemicals.
For health
Example of Illness Initiating Factor from Pesticide Incident
(Text Box 2-2)
Information reported
to health officials
Setting observations
Investigative
discovery
Specific chemical
toxicity
Exposure
assessment
Risk management
action
Source
Seven siblings presenting with abdominal pains
and respiratory arrest, symptoms of
organophosphate poisoning. Two children died.
Adult resident recently sprayed an unknown
insecticide in the home.
Illegal pest-control application of methyl parathion
inside home at 3 times the concentration used in
agricultural spraying (this OP pesticide is only
intended for outdoor use).
Affects central nervous system: nausea, dizziness,
headache, vomiting. High levels can be fatal.
Samples from sprayer, food, water, air.
Biomonitoring (e.g., blood or urine samples) to
identify people exposed (multi-pathway).
Decontamination of house and increased
publication of dangers of inappropriate OP uses.
CDC (1984).
endpoints initiating
factors with suspected
environmental
etiologies, the initial
phase of the risk
assessment involves a
data collection effort that
focuses on identifying
chemicals (individually
or in groups) that are
known to cause the
effect in humans or
some animal species
(e.g., effect identified in rodent bioassays or in an occupational epidemiologic study).
Although Table 2-1 identifies a number of illnesses that are linked to environmental
contaminant exposures, chemical combinations and exposure conditions can be highly
situation-specific, so that identification of chemicals and chemical mixtures related to
specified health effects is typically initiated through a literature review of both
toxicological and epidemiologic data. Because multiple chemicals are involved, it is
consistent with best risk assessment practice to include both critical (primary) and other
secondary effects in the literature review. The critical effect is the first effect observed
as the chemical's dose is increased above a no-effect range in the relevant toxicity
study, while secondary effects are typically those seen at higher doses in the same
target organ or tissue and/or different physiological compartment(s). For example, the
acceptable level of a chemical to which humans may be safely exposed could be based
on hepatotoxicity, the most sensitive endpoint identified in a toxicological bioassay, but
the available literature indicate that the chemical is also a potential reproductive toxicant
at doses higher than those where hepatotoxicity was observed. This initial data
collection for the exposure analysis may be conducted in conjunction with the dose-
response and toxicity analysis, so that specific chemical mixtures of concern (given the
health endpoints) are identified, and chemicals with known toxic interactions can be
considered for additional exposure measurements and analysis. In summary, the goal
of this first step is to determine the pollutants of concern (either individually or in groups)
2-10
-------
TABLE 2-1
Examples of Illnesses Possibly Linked to Multiple Environmental Factors3
Illness/
Health Effect
Hypothesized Causes/
Epidemiologic Links
Associated Levels
Remarks
Reference
Acute mylogenous
leukemia (AML)
Benzene, ionizing
radiation, alkylating
agents and
topoisomerase
inhibitors.
Increased incidence of
leukemia observed in
lifetime occupational
studies at 10-50 ppm
benzene and higher.
These levels exceed the
U.S. occupational 8-hour
standard of 1 ppm for
benzene in air.
Benzene is present in gasoline,
automobile exhaust and cigarette
smoke. The latter also emits
radiation. AML is also a secondary
cancer after treatment for primary
cancers, and links between AML and
genetic (inherited) conditions and
viruses have also been established.
Hricko, 1994;
U.S. EPA,
1997b
Allergic contact
dermatitis
Nickel and chromium
The European Union (EU)
has prevented sale of
nickel-containing objects
that release over 0.5 ug
nickel/cm2 skin per week.
Delayed skin inflammation and rash
can occur; nickel is commonly used
in some jewelry. Note that the EU
nickel limit might not protect all
sensitized persons (no similar
U.S. limit has been placed on nickel
content in jewelry or other consumer
products).
Nickel Institute,
1999; Amduret
al., 1993
Asthma
Particulates, including
high molecular weight
(HMW) allergens
(polymers or proteins of
animal, plant, bacterial
or fungal origin in range
of 20-50 kilodaltons).
A 14% increase in
emergency room visits due
to asthma was associated
with very fine particulate
matter (PM2.5) averaging
12 ug/m3 (for 15 months).
Asthma is exacerbated by both
indoor and outdoor pollutants as well
as allergens. Correlations have
been observed between asthma and
sensitivity to cockroaches and to
HMW allergens.
Norris etal.,
1999; O'Connor
and Gold, 1999
2-11
-------
TABLE 2-1 cont.
Illness/
Health Effect
Hypothesized Causes/
Epidemiologic Links
Associated Levels
Remarks
Reference
Blackfoot disease
Arsenic
Observed in people
consuming well water with
170 ug/L arsenic and
higher. (This concentration
is much higher than the
U.S. drinking water
standard of 10 ug/L.)
Blackfoot disease, a severe form of
arteriosclerosis, is a vascular
complication of arsenic exposure.
Blackfoot incidence increases with
age.
U.S. EPA, 2007;
Amduret al.,
1993
Liver cancer
Many (MOO) chemicals
and risk factors,
including chlorinated
solvents, aflatoxin, and
animal products (meat,
eggs).
For aflatoxin (which can be
found in peanut butter),
Americans could consume
up to 0.15-0.50 ug/day.
Organic solvents are
ubiquitous at low levels in
urban air and hazardous
waste sites.
Causes of liver cancer are many and
varied; this organ is the most
common site for mutagens and non-
mutagens. To illustrate for aflatoxin,
effects can be confounded by
hepatitis B infection, which is
endemic in areas where high intake
is common.
NTP, 2002;
Goldetal.,
2001;CPDP,
2004; ATSDR,
2001
Lung cancer
Dozens of chemicals,
including those in
cigarette smoke and
radon.
Average U.S. radon levels
of 4.4-11 becquerels/m3.
Tobacco smoke is the leading cause
of lung cancer. Lung cancers
increase multiplicatively when radon
exposure occurs in addition to
cigarette smoking.
NTP, 2002
Neurological
damage/
reduced
intelligence
quotient (IQ)
Lead in lead-based
paint; mixtures of
polychlorinated
biphenyls (PCBs) and
dioxins; fetal irradiation;
methylmercury.
An increase in blood lead
levels from 10-30 ug/dL
resulted in an IQ reduction
of 4-5% (4.4-5.3 points) in
7-year-old children.
Increased maternal blood
mercury concentrations or
hair mercury
People can be exposed to lead via
many sources, e.g., paint, soil and
dust, drinking water, food,
occupational exposure, burning
candles with lead wicks and hobbies.
Baghurstetal.,
1992; NYSDOH,
2003; Birnbaum,
1995; Kjellstrom
etal., 1986,
1989;
Grandjean et al.
1997; Crumpet
2-12
-------
TABLE 2-1 cont.
Illness/
Health Effect
Hypothesized Causes/
Epidemiologic Links
Associated Levels
Remarks
Reference
concentrations resulted in
IQ reductions in 6-year old
children (regression
coefficient -0.5 IQ pts/1
ppm increase hair mercury)
al.,1998
Parkinson's
Disease and
Parkinsonism
(which can be
reversible)
Many pesticides,
including
organophosphates,
organochlorines,
carbamates, various
herbicides and
household fumigants;
manganese, carbon
monoxide and carbon
disulfide
Increased risk of
Parkinson's disease has
been observed in
connection with chronic
pesticide exposures.
Reversible Parkinsonism
has been seen following
acute pesticide exposures.
One occupational study
found 6% of workers
exposed to >5 mg/m3
manganese exhibited acute
Parkinson's symptoms.
Risk factors have been identified for
people using well water and living in
farming areas, especially those with
a history of pesticide exposure.
Higher levels of organochlorine
pesticides in brain tissue from
Parkinson's patients than the
general population. ldiopathicb
causes account for >85% of all
cases; suspected links exist to
MPTP,C organomercury,
encephalitis, major tranquilizing
drugs, carbon monoxide or disulfide
poisoning and frequent head
injuries.
Feldman, 1992;
Gorell etal.,
1999; Wright
and Keller-
Byrne, 2005;
Stephenson,
2000; Engel,
2001
a This table illustrates illnesses or health effects that have been linked with various environmental exposures (some lifestyle factors
are also shown) and that might initiate a cumulative risk assessment concern because of the number of possible chemical causative
agents and their likely joint toxicity.
b Idiopathic is defined as having an unknown cause.
= MPTP is the drug 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine.
2-13
-------
that have been linked to the initiating factor effect and similar health effects and to
identify the combinations of subpopulations and pollutants of concern that might require
more detailed exposure assessment because of higher exposure and/or enhanced
toxicity in those subpopulations.
2.3.2. Initiating the Exposure Assessment when Elevated Chemical
Concentrations are the Initiating Factor. When increased environmental chemical
concentrations or biomonitoring results initiate a cumulative risk assessment, the initial
goal of the investigation is to determine if those concentrations could result in exposures
or doses that could lead to potentially important health effects in the community,
including secondary health endpoints and the potential for effects due to toxicological
interactions among chemicals. In addition, the population profile for the community may
be examined to identify any increased incidence of morbidity or mortality measures that
may be considered during the exposure assessment. The initial phases of these types
of analyses are similar to the steps undertaken in traditional risk assessment analyses
such as those presented in the Risk Assessment Guidance for Superfund (U.S. EPA,
1989a). From an exposure perspective, following identification of the chemicals of
interest, such analyses will determine the spatial bounds of the assessment, examine
the fate of the identified pollutants, determine whether (and which) individuals in the
community are or could be exposed, and quantify such exposures. These are standard
components of an exposure assessment.
When increased chemical concentrations initiate an assessment, the initial phase
of the data gathering focuses on identifying the chemicals present in the community,
documenting the locations of these elevated concentrations (existing data on the
locations of these elevated concentrations could be supplemented with information
provided by stakeholders about the locations of previous polluting operations in the
community) and examining the health effects associated with these chemicals. In
conjunction with dose-response analyses, the primary and secondary health effects
associated with the individual chemicals or groups of chemicals are identified. Because
a cumulative risk assessment is being initiated, the investigation includes an evaluation
of the potential for other chemical exposures in the community that could increase the
toxicity of the chemicals known to be at high concentrations. This could involve an
examination of potential sources of pollution in the community (e.g., using the Toxic
Release Inventory reports on pollutants typically released from industrial sources)
followed by monitoring of related environmental media. Other EPA documents (e.g.,
Human Health Risk Assessment Protocol for Hazardous Waste Combustion Facilities
2-14
-------
[U.S. EPA, 2005b]) can further aid in the identification of the types of compounds
typically released from a source class. In summary, the goals of this first phase are
(1) to identify likely multi-chemical exposures among chemicals with high
environmental concentrations or elevated biomonitoring levels;
(2) to characterize the primary and secondary health effects potentially
associated with those chemicals and identify any related morbidity or mortality
in the population; and
(3) to determine if there are other pollutants (either individually or in groups) to be
monitored in other media (e.g., household pesticide use) because of their
influences on exposure or because they produce similar health effects.
2.3.3. Initiating the Exposure Assessment when One or More Sources is the
Initiating Factor. When one or more sources initiate a cumulative risk assessment, the
initial goals of the investigation are to determine if the chemicals released from those
sources could cause exposures high enough to cause health effects in the community
and to examine the community's population profile to identify any increased incidences
of morbidity or mortality that may be considered during the exposure assessment. With
multiple sources it is important to determine which chemicals from those sources will
reach the population(s) of concern. For example, releases of highly volatile chlorinated
solvents into ambient air are usually only considered significant for populations close to
the source as they disperse rapidly (ATSDR, 2001). Sources of chemical pollutants
include (1) point sources, such as industrial and commercial boilers, electric utility
boilers, turbine engines, wood and pulp processers, paper mills, industrial surface
coating facilities, refinery and chemical processing operations and petroleum storage
tanks and (2) area sources such as industrial wastewater treatment ponds, quarry
operations, tank farms and on-road and off-road vehicles. The initial phases of these
types of analyses are similar to the steps undertaken in traditional risk assessment
analyses that analyze single sources such as those presented in the Risk Assessment
Guidance for Superfund (U.S. EPA, 1989a) and those presented in the Methodology for
Assessing Health Risks Associated with Multiple Pathways of Exposure to Combustor
Emissions (U.S. EPA, 1998a) and the Human Health Risk Assessment Protocol for
Hazardous Waste Combustion Facilities (U.S. EPA, 2005b). Following identification of
the source(s) and chemicals of potential interest, aspects of which are discussed next,
such analyses will
characterize the source(s) by compiling basic facility information;
determine the spatial bounds of the assessment;
2-15
-------
examine the fate of the released pollutants;
determine whether (and which) individuals in the community could be exposed;
and
quantify such exposures.
These steps are standard components of an exposure assessment.
When one or more sources initiate a cumulative risk assessment, the initial
phase of the data gathering focuses on identifying the types of chemicals released from
those sources, including potential future releases, that could impact the community.
Different types of sources may be involved, so that exposure assessment guidance
from several U.S. EPA Program Offices might have to be consulted. Most of these
Program Offices have procedures for determining the important chemicals released
from different point sources of concern. For example, Chapter 2 of the draft Human
Health Risk Assessment Protocol for Hazardous Waste Combustion Facilities (U.S.
EPA, 2005b, Volume 1) presents an approach for identifying compounds of potential
concern that are emitted from hazardous waste combustors. In addition to the
chemicals released from the identified sources, the examination of other sources,
including nonpoint sources, of specific pollutant exposures to the community may also
be considered. In conjunction with dose-response analyses, the primary and secondary
health effects associated with the individual chemicals or groups of chemicals are
identified so that the exposure assessment can provide information to categorize the
identified chemicals from multiple sources into groups that jointly influence the same
health effects. In summary, the goal of this first phase is to determine those pollutants
(either individually or in groups) from the identified sources that are of concern for the
community because of likely co-exposures at concentrations of toxicological
significance.
2.3.4. Summary. During the initial assessment step, the focus of the cumulative risk
assessment is on determining what emissions sources, chemicals or population
locations to include and what chemicals to evaluate together. The population profile for
the community is also being examined to identify any increased incidence of morbidity
or mortality measures that may be considered during the exposure assessment. In
evaluating which chemicals are of concern for a community, it is useful to consider the
specific initiating factors of the cumulative risk assessment and any issues, sensitivities
or vulnerabilities that might be of special interest to the stakeholders.
Although more detailed approaches to exposure assessment are discussed in
Chapter 3, some insight on focusing the assessment can be gained from criteria
2-16
-------
commonly used for retaining or excluding chemicals. The chemical selection criteria
recommended by EPA Program Offices typically include
toxicity;
mass released or mass present in media;
the potential for physical or chemical interactive effects with other chemicals in
the area and with other media;
the tendency to persist, bioaccumulate and/or be transported between
environmental media and
the potential for relatively high exposures to sensitive or vulnerable populations.
In addition, for a population-focused cumulative risk assessment, the chemical selection
criteria also include consideration of
the possible contribution to induction of health effects that exist at relatively high
levels in the study population;
likelihood of exposure to the population of concern;
potential for overlapping exposures (times and routes) to toxicologically similar or
interacting chemicals;
specific genetic traits or other physical characteristics of the population that
would increase susceptibility to chemicals linked to the illnesses observed in the
population;
cultural practices that might cause the population to be differentially exposed to a
chemical or group of chemicals;
public health monitoring data and
chemicals that may be linked to illnesses or exposures in identifiable population
subgroups such as children or the elderly.
Depending on the community and the initiating factor, these criteria could be adapted or
augmented.
2.4. INTEGRATION OF PUBLIC HEALTH INFORMATION
Regardless of the initiating factor, it may be useful to collect and evaluate
available public health information relevant to the investigation and to the identified
population. Examples of such data include cancer and other disease rates, blood lead
levels, hospital admissions, and mortality records. Such information can be used for
comparison with initiating factor data, e.g., to verify suspected health effects given the
chemicals found in the environment or, conversely, to explain health effects in a
community that may not be caused by chemical exposures.
2-17
-------
Public health information was used in a community-level investigation in Chester,
PA (U.S. EPA, 2002b) after the EPA was approached by community representatives
regarding possible excessive chemical exposures (e.g., to diesel emissions and drinking
water disinfection by-products) and health effects (e.g., cancer). As part of the
cumulative risk assessment that was conducted, EPA used public health information,
examining blood lead levels and also comparing disease rates between Chester and the
state of Pennsylvania. In both males and females respiratory cancer rates in Chester
were found to be much higher than the state average. Incidence rates for leukemia,
prostate cancer, and all cancers combined were statistically significantly higher for
males compared to incidence rates for the state, the county and Philadelphia. EPA also
identified a serious public health problem in Chester by examining venous blood lead
test level results for 6783 children over a 5-year period. Results indicated that
approximately 50% of the children tested had blood levels in the range where lead
poisoning is a concern and approximately 67% had blood lead levels above the
accepted level of concern of 10 ug/dl (U.S. Consumer Product Safety Commission,
2005). These two investigations using public health data confirmed that adverse health
effects were being observed in Chester, PA. As a result, the cumulative risk
assessment included all identified carcinogens, lead and other environmental chemicals
as relevant chemicals to being evaluated.
2.5. EPIDEMIOLOGIC INVESTIGATIONS IN CUMULATIVE RISK ASSESSMENT
Different approaches to the cumulative risk assessment may be necessary
depending on the type of initiating factor that is identified. When population illness is
the initiating factor, an epidemiologic investigation may be warranted to ascertain, if
possible, the relationship of environmental or other exposures (or stressors) to the
occurrence of illness. When the initiating factor is a particular source, environmental
concentration or biomonitoring result, the investigation may apply epidemiologic
methods that include an examination of chemicals and their sources. In general, many
health effects can have multiple risk factors and the process of attributing risks to
individual stressors is often complex. An example that reflects this multi-factorial risk
perspective is an impoverished population whose principal diet is fish and foods with
high fat content, with a subpopulation characterized by lifestyle risk factors such as
alcohol abuse and smoking. The population may have elevated heart disease and
cancer rates that could be due to dietary or other behavioral risk factors and/or
environmental exposures, yet the community focuses their concerns on environmental
2-18
-------
exposures. The initiating factor (e.g., elevated cancer mortality) in this case may be
addressed by an epidemiologic investigation. When health endpoints are initiating
factors for cumulative risk assessment, the analytic methods may need to be in context
of this multi-factorial risk perspective.
Several key steps are involved if health endpoints are initiating community
concerns which may lead to a cumulative risk assessment. An initial step is a thorough
review of the expected etiology of the health endpoint(s) and identification of known risk
factors is necessary at an early stage. Epidemiologic studies of varying degrees of
complexity can be conducted depending on the nature of the identified initiating factor
data. A preliminary analysis of the health endpoint (e.g., identified by a disease cluster)
may include an assessment of the potential magnitude of the problem. This includes
identification of the population at risk and delineation of geographic and temporal
boundaries for disease ascertainment. Complete case ascertainment is another critical
step in determining the magnitude of a potential disease cluster. Although a preliminary
assessment of identified cases may include those identified by the community (e.g.,
through self-report), clinical confirmation of case diagnosis (e.g., medical chart review or
diagnostic confirmation by physicians) will be needed. Diagnostic or pathologic
verification will help distinguish whether the reported cases were in fact truly similar
etiologically and should be considered as part of a disease cluster. This preliminary
assessment also includes an examination of population characteristics of identified
cases in order to determine if the disease incidence varies among any susceptible
populations (e.g., children).
Once baseline levels of disease occurrence are determined, statistical analyses
are conducted to determine if geographic or temporal excesses of disease are
occurring. This may include use of existing environmental monitoring or public health
surveillance data. Expected rates of disease can be compared to observed levels in a
community to determine if disease rates are in excess. Comparisons of disease
occurrence (or measures of mortality) can be made between the population residing in
the area of concern and populations in other geographic areas. Temporal trends in
disease incidence can also be examined in a community using time-series analyses if
longitudinal health data are available. If the initial assessment determines that elevated
risks are occurring in the population of concern, then additional epidemiologic studies
can be conducted to further determine whether or not specific exposures are linked to
the health endpoints of concern. Descriptive epidemiologic studies (e.g., an ecological
study) using population-level data could help examine whether disease occurrence
rates over time can be compared to existing exposure occurrence data. Analytical
2-19
-------
epidemiologic studies using individual-level data to test specific etiologic hypotheses
could include retrospective studies using existing exposure data or, alternatively,
prospective studies involving collection of additional environmental samples, confounder
data and biomarker data.
In keeping with best statistical practices, statistical analyses often take into
consideration the likelihood that type I and type II errors result in the presence or lack of
an association between disease and the exposures or stressors being examined.
Considerable caution is advisable in most instances of rare disease clusters, since
statistical power may be inadequate for cluster identification. The ability to infer
causality of elevated disease rates to specific exposures is often limited in
epidemiologic investigations of clusters. Although epidemiologic studies can be
conducted to address health illness and other initiating factors of cumulative risk
assessments, the utility of using epidemiologic data to draw causal inferences is beyond
the scope of this report. Determining causality for specific exposures generally includes
weight of evidence considerations across all existing epidemiologic and toxicological
studies. Effective risk communication is also critical throughout the investigation and
assessment processes in order to keep the public and stakeholders informed of the
status of ongoing investigations, including the study objectives, expectations and
limitations of the data and analyses. This is especially important if prospective studies
are later conducted as a result of initial assessments and may help bridge differences in
perception of risk between the public and risk assessors and other investigators.
2.6. LINKING THE LIST OF RELEVANT CHEMICALS TO THE POPULATION
PROFILE THROUGH A CONCEPTUAL MODEL
Following the development of population profiles and the initial data collection
activities, the relevant chemicals and endpoints of concern may be evaluated for
linkages to sensitive population subgroups in the community or the population being
assessed. In addition to identifying and examining chemical releases from local
sources, the cumulative risk assessment could include an examination of possible
regional and national sources of these potentially hazardous chemicals. The
assessment could also include an evaluation of any unique exposure sources or
pathways for the sensitive populations and an examination of the spatial relationships
between the identified sources and residences, sources of food, playgrounds, schools,
etc. to identify individuals or groups of people in the community who might be exposed.
Other community-based methods highlight the importance of community involvement in
the risk assessment planning process (U.S. EPA, 1997a, 1998a).
2-20
-------
One of the desired outputs from the Planning and Scoping phase of cumulative
risk assessment (U.S. EPA, 2003a) is a conceptual model. Conceptual models provide
both a written and visual representation of the structure and dynamics of the system
(e.g., the community or physical site) being assessed that can be subsequently
converted into an implemented approach (Suter, 1999; Suter et al., 2003). Conceptual
models typically identify the links between main system components (i.e., the sources,
chemicals, exposure pathways, exposure routes, subpopulations and health endpoints)
that will be analyzed. Conceptual models identify which sources, endpoints and
processes are included and which are excluded and what assumptions are being made.
Once the initial exposure and population descriptions are completed, the exposure and
dose-response analysts jointly develop a preliminary conceptual model to ensure that all
relevant exposures and endpoints are included. During the analysis phase of the
exposure assessment, the preliminary conceptual model is refined by incorporating
further information gained during the analysis steps (Chapter 3, Section 3.3).
Figure 2-2 illustrates some key elements of a conceptual model for evaluating
cumulative exposures and shows the complexity of the exposure scenario. From left to
right, Figure 2-2 begins with the initial focus on the health of the population and the
identification of vulnerabilities that influences the collection of appropriate exposure and
dose-response data. From bottom to top, the figure shows the factors that influence
cumulative risk assessments that are associated with vulnerability and multiple
chemicals, exposures and routes. From right to left, Figure 2-2, depicts the typical flow
of information for developing a risk assessment, depicting sources, processes,
receptors and flows between them.
Conceptual models for cumulative risk cannot present all the complexities that
are involved, especially those dealing with physical and toxicological interactions.
Consideration of all combinations and their potential interactions can be conceptually
difficult and impractical to present, so it is useful to first prioritize the potential
combinations of chemicals, routes, effects. That step is better represented by a
decision tree or influence diagram. A site-oriented, second-tier conceptual model may
also be useful, as depicted in Figure 2-3. In addition to the usual boxes describing the
scenario, processes, receptors, etc., there are also indications of places where
environmental, toxicokinetic and toxicodynamic interactions could be considered.
Those potential interactions can then be simplified by grouping (e.g., Section 3.3.2.2 for
exposure-based grouping) and prioritized using decision criteria. For example,
toxicological interactions could be screened based on toxicological significance, as
indicated by the relative importance of each chemical's environmental concentration
2-21
-------
Starting point
CRAis
Initiated
Observed,
concern or
potential for
human health
effects
Effects
influenced by
sensitivity or
susceptibility,
ability to
recover,
preparedness:
e.g., age,
genetics,
health status,
nutrition
Possibly linked to
Humans,
including
sensitive or
susceptible
subgroups,
vulnerable
populations
What has happened
or could happen?
Are there multiple
heath effects to
consider?
Exposures influenced
by differential
exposure factors:
e.g., cultural practices,
subsistence fishing,
activity patterns,
proximity of homes,
playgrounds or
farms/gardens
to pollutant source,
Exposures to
Concentrations in
Environmental Media
Air
Ground -*
water
N
-> Surface
water
Soil
Media concentrations
influenced by fate &
transport of pollutants
(surface runoff,
volatilization, leaching) and
environmental
transformations over varying
timeframes
What, where, when, how much?
Are exposure pathways complete from source to receptor?
Who, where.
when, how much?
Contaminated with
Chemicals Relevant
to Health Effects
Chemical hazards:
biomonitoring
data, media
monitoring data,
stack emissions,
industrial releases,
drinking
water, pesticide
use, etc.
Relevance to
health effects
based on
toxicology and
epidemiology
data, evidence of
joint toxic action
What, where, how much?
Are known health effects
of chemicals consistent
with observed health
outcomes in humans?
FIGURE 2-2
Conceptual Models for All Initiating Factors
2-22
-------
Geographic
area
Medium/
transport
pathway
Exposure
activity/use
Receptor
Effect
Environmental
1
land use
LUUdllUII/OUUlUe 1
setting Contaminants (Other stres
multiple chemicals
demography
i > n 1
landscape i famj|y^|ass
I form/matrix
! volatility,
1 environm
1
physicochemica
properties
1
J
amount
(volume, concentratio
nobility, persistence
sntal transformation
[
containment
status
n)
1
environmental interactions \ «
|
soil
1 1 1 1
^ | air surface || seeps | ground
"
,
' 1 1
i exposure point 1 route 1
! concentration ' ' s
1 1
1
1
1 1 sediment 1 1 plants || animals
1 ,
ntake/
mount
1
internal
dose
1
! general population 1 1 sensitive subgroups | | individual
<&
i
toxicokinetic
interactions
(part of joint toxicity
Note:
- potential for
environmental or
internal interaction.
carcinogenicity
1 joint toxicity
noncarcinogenicity
|_ toxicodynamic interactions \
genotoxicity
non-genotoxicity
mutagenicity
reproductive
deve
ter
opmental/
atogenic
musculoskeletal
multiple systemic effects:
neurological
immunological
integumentary
| | cardiovascular
| | gastrointestinal
1 1 respiratory
hepatic
renal/
urinary
whole body
FIGURE 2-3
Example Second-Tier Detail of the Analytical Approach for Health Risk
2-23
-------
using screening values such as the HQ. Schematic diagrams and decision flowcharts
for joint toxicity and toxicological interactions are given in Figures 4-6a, b, c and d.
Once that initial screening or grouping is completed, a revised conceptual model could
be created, followed by more detailed analysis of the toxicological interactions such as
is described in Chapter 4.
For cumulative risk assessments that encompass multiple exposure scenarios
(e.g., sources, chemicals, pathways, effects), such as at a contaminated site, it is
preferable to develop a hierarchy of conceptual models instead of trying to represent the
multiples in one model (Suter, 1999). As described in the EPA Risk Assessment
Guidance for Superfund (U.S. EPA, 1989a) and from a cumulative exposure
assessment perspective, unique exposures in populations living near a site might
require a detailed evaluation. In the next chapter, Figure 3-2 displays in more detail the
components of a cumulative exposure assessment along with primary exposure routes
for potential receptors, suggesting possible populations of elevated exposure, such as
individuals who consume large quantities of local fish. More detailed conceptual models
and diagrams for cumulative exposure are presented in Chapter 3 (e.g., Figures 3-5,
3-9, 3-11, 3-12) that suggest specific processes to be evaluated.
2-24
-------
3. EXPOSURE ASSESSMENT OF MULTIPLE CHEMICALS, EXPOSURES AND
EFFECTS
This chapter provides detailed information on the exposure assessment of
multiple chemicals, exposures and effects, a subset of cumulative risk issues that are
described in Chapters 1 and 2. Chapters 1 and 2 of this document address several
important cumulative exposure assessment concepts including a discussion of the
initiating factor, the identification the exposed population and the development of a
conceptual model for a cumulative risk assessment. Chapter 3 highlights existing data,
methods and approaches that can be used to address cumulative exposure assessment
issues that are posed as questions in Text Box 3-1. These methods can be used to
determine if individuals are co-exposed to multiple pollutants and over which time
periods these co-exposures occur. In
collaboration with the toxicity analyst,
toxicologically relevant time periods
and at high enough doses to be of
toxicological concern. Section 3.1
Cumulative Exposure Assessment Questions
.. . . . . (Text Box 3-1)
the exposure analyst can evaluate
How are people exposed to multiple chemicals?
whether the co-exposures occur over ,._..__,. ._ , _, ._
In which media, at what levels, where and when?
What are the intensity and duration of these
exposures?
Are there uniquely susceptible or vulnerable
subpopulations?
defines cumulative exposure
assessment as conducted in this chapter. Section 3.2 provides an overview of some
exposure assessment documents that describe current EPA practice. Section 3.3
discusses approaches for conducting of a population-focused cumulative exposure
assessment, giving a brief overview of the basic steps an analyst undertakes in an
exposure assessment and highlighting the issues that are not routinely evaluated in a
conventional (i.e., single chemical-focused or single source-focused) exposure
assessment. This includes grouping potential chemicals of concern by exposure
pathway and media with examples from different chemical groups (Section 3.3.2.2). In
Section 3.4, cumulative concepts for atmospheric pollutants are illustrated.
Retrospective studies are discussed in Section 3.5. Section 3.6 summarizes the
information in this chapter.
3.1. DEFINING EXPOSURE ASSESSMENT FOR CUMULATIVE RISK
ASSESSMENTS THAT EVALUATE MULTIPLE CHEMICAL EXPOSURES
In cumulative risk assessments that examine risks posed by multiple chemicals,
exposure assessments evaluate a population's chemical exposures through multiple
3-1
-------
routes of exposure over time. Such assessments may encompass multiple exposure
timeframes in which the timing and intensity of exposures to different chemicals are
examined relative to each other. The analysts seek to determine whether the
exposures to multiple chemicals can lead to toxicokinetic interactions1 or toxicodynamic
interactions2. In addition to providing information about multiple chemical exposures in
the general population, these exposure assessments identify potentially susceptible or
vulnerable subpopulations3 in the study area and potentially unique pathways of
exposure in those subpopulations.
Cumulative exposure assessments will likely rely on environmental monitoring
data and environmental fate models. The community's boundary may define the
geographic region of study for a cumulative exposure assessment, unlike chemical-
focused assessments or single source-focused assessments. If the timing of different
chemical exposures is important, the analyst can use fate models to estimate changes
in the concentrations in environmental media over time. The pollutants may occur in
these media as a consequence of releases from multiple and different sources that
could be either close to or distant from the population of concern. The environmental
fate information needed for a such an assessment could be site dependent; for
example, the data could include the degradation of chemicals or chemical mixtures in
the environment, interactions of pollutants in the environment that influence their fate
and interactions between chemicals and the environment (e.g., killing off or promoting
soil microbes that normally degrade some of the chemicals or altering the soil binding
so that chemical transport through soils is enhanced).
While approaches to exposure assessment modeling are stressed in this
chapter, the use of biomonitoring data (e.g., biomarkers of exposure) holds a great deal
of promise for future cumulative risk assessments. The use of biomarkers in cumulative
risk assessments currently is limited. They can provide key quantitative exposure
estimates in cumulative risk assessments (e.g., biomarker data are used to estimate
1 Toxicokinetic interactions refer to alterations in the absorption, distribution, metabolism or elimination of
a toxic chemical. For example, these interactions can be mediated by the induction or inhibition of
enzymes involved in xenobiotic activation or detoxification. See Appendix C U.S. EPA (2000a) for
complete discussion.
2 Toxicodynamic interactions encompass all interactions that do not directly affect absorption, distribution,
metabolism or elimination of a toxic chemical. Toxicodynamic interactions affect a tissue's response or
susceptibility to chemically-mediated toxic injury. Modes of toxicodynamic interactions include, among
others, depletion or induction of protective factors, alterations in tissue repair, changes in hemodynamics
and immunomodulation. See Appendix C U.S. EPA (2000a) for complete discussion.
3 Vulnerable or susceptible populations in the study area can be identified during either the exposure or
dose-response assessment phases of a cumulative risk assessment. This identification is based on
properties of the chemicals being evaluated as well as social, cultural or genetic factors that influence
vulnerability or susceptibility.
3-2
-------
current chemical exposure levels in an affected population or the general population).
Such data also can be used to verify selected exposure model results (e.g., show that
specific chemical exposures and absorption are occurring in the population or, if the
data are collected in a different location or under different conditions, provide evidence
showing that human absorption of the chemical from environmental exposures are
possible). For example, some studies have used existing blood chemical or urine
chemical concentration data, such as data published in NHANES (NCHS, 2002).
Exclusive use of biomarker data in cumulative exposure assessment efforts is
currently not practicable when considering a large number of diverse chemicals due to
analytical and resource limitations. Analytical limitations include considerations such as
whether sensitive biomarkers for many types of environmental chemicals have been
developed and whether the chemical's biological half-life after absorption is sufficient to
estimate exposure over a relevant exposure period. Collection of human biomarker
data can be invasive and costly, resource limitations may constrain the ability of
researchers to collect such data.
If collected, the interpretation of biomonitoring data and application to risk
assessment can be challenging. While biomonitoring identifies individuals who are
exposed and have measured internal doses reflecting absorption of a chemical, to
estimate the individuals' actual exposures, the biomonitoring data would need to be
integrated with additional information (e.g., exposure modeling information) to identify
the pathways, timing and routes of exposure. Additional exposure and environmental
modeling would be needed to identify sources of chemicals in the contaminated media.
Although the use of biomonitoring data holds great promise for cumulative risk
assessments, few methods exist at this time for such applications (U.S. EPA, 2003a).
When conducting cumulative risk assessments, the analyst may identify and, in
some situations, wish to quantify the uncertainties associated with exposure estimates.
Identifying the uncertainties in the exposure assessment is critical to a cumulative risk
assessment, because limitations and uncertainties in the exposure assessment need to
be highlighted in the risk characterization.
When possible, the analyst may consider developing a sensitivity analysis or
quantitative analysis of uncertainty for the exposure assessment. The sensitivity
analysis will be used to identify key input values to the exposure model (i.e., parameters
that significantly influence the exposure modeling results), highlighting important input
parameters to analyze in a quantitative uncertainty analysis.
In quantitative uncertainty analyses, the uncertainty of each input parameter can
be characterized through a probabilistic distribution for use in Monte Carlo simulations.
3-3
-------
Although detailed discussion of this topic is beyond the scope of this document, many
available resources provide guidance on performing probabilistic exposure and risk
assessments (e.g., Cullen and Frey, 1999); these include the following EPA sources:
Guiding Principles for Monte Carlo Analysis (U.S. EPA, 1997f)
Superfund's Process for Conducting Probabilistic Risk Assessment (U.S. EPA,
2001 e)
Framework for Cumulative Risk Assessment (U.S. EPA, 2003a)
Finally, we note that using the methods described in this chapter for a large-scale
cumulative exposure assessment at this time would be very resource intensive.
However, the application of approaches discussed in this chapter is considered feasible
for more focused cumulative analyses (e.g., for relatively small populations, small
geographic areas and limited numbers of chemicals, sources and pathways). In
addition, the cost and time needed to conduct a cumulative risk assessment are
expected to decrease as the data, approaches and tools to support these analyses
evolve, experience is gained and the analyses become more routine.
3.2. U.S. EPA EXPOSURE ASSESSMENT GUIDANCE
The general methods the EPA uses to evaluate human exposures are presented
in the Guidelines for Exposure Assessment (U.S. EPA, 1992a). EPA Program Offices
follow these guidelines and develop additional guidance documents that describe
exposure assessment methods relevant to the specific types of chemicals they
evaluate. For example, the basic process for assessing exposures at Superfund sites is
described in the Risk Assessment
Guidance for Superfund (U.S. EPA,
1989a) (see Text Box 3-2) and U.S. EPA
(1992a) provides overall guidance in this
area.
The assessment of exposures to
chemicals released during combustion is
described in Methodology for Assessing
Health Risks Associated with Multiple
Pathways of Exposure to Combustor
Emissions (U.S. EPA, 1998a) and in
Human Health Risk Assessment Protocol
for Hazardous Waste Combustion
Selected Information Guides (Text Box 3-2)
Guidelines for Exposure Assessment (U.S. EPA,
1992a)
Risk Assessment Guidance for Superfund
(U.S. EPA, 1989a)
Methodology for Assessing Health Risks
Associated with Multiple Pathways of Exposure to
Combustor Emissions (U.S. EPA, 1998a)
Human Health Risk Assessment Protocol for
Hazardous Waste Combustion Facilities (U.S. EPA,
2005b)
General Principles for Performing Aggregate
Exposure and Risk Assessments (U.S. EPA,
2001 a)
Guidance for Developing Ecological Soil Screening
Levels (U.S. EPA, 2003g)
3-4
-------
Facilities (U.S. EPA, 2005b). While these documents focus on conventional exposure
assessment approaches, they also present many cumulative exposure assessment
issues.
At times, Program Office guidance is developed specifically to address
cumulative exposure issues. For example, in response to the 1996 Food Quality
Protection Act, the Office of Pesticide Programs developed General Principles for
Performing Aggregate Exposure and Risk Assessments (U.S. EPA, 2001 a). Finally,
EPA documents that describe exposure approaches to chemical mixtures, such as the
Site-Specific Assessment Procedures volume in the review draft Exposure and Human
Health Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related
Compounds (U.S. EPA, 2003c), describe methods for examining cumulative exposure
issues for specific chemical classes that can be applied in other situations.
In summary, there are a number of EPA resources that describe methods and
approaches that can be used to address various aspects of exposure assessments that
can comprise cumulative risk assessments. Exposure models are commonly applied to
help integrate data, fill gaps and focus the scope of a more detailed phase of the risk
assessment process. Models can also contribute an analytic rigor to the analysis.
Various models and several tools are described in subsequent sections of this chapter
and several others are highlighted in Appendix A.
Prior to selecting any model, analysts may seek to understand the development
and evolution of a model. Analysts also would evaluate the strengths and limitations of
an exposure model to be used in a cumulative risk assessment and determine whether
the model's accuracy and the conditions under which the model was developed are
consistent with the goals defined during Planning and Scoping (i.e., examine the
conditions for which the model was developed and determine if the use of the model in
the exposure assessment will necessitate an extrapolation beyond the conditions for
which the model was developed) (NRC, 2007). Verification, validation and calibration
are three key elements of this model evaluation process.
Verification focuses on assuring that the model reflects the processes it aims to
characterize, by evaluating the breadth, accuracy and conformance or compliance of
the underlying concepts and model framework with established guidelines. This
process helps answer whether the model results are logical and whether they reflect the
current understanding of relationships among exposure, dose-response and risk
characterization. The EPA has drafted guidance to support the evaluation of models for
various applications (U.S. EPA, 2003k) as well as guidance to address issues of
verification and validation (U.S. EPA, 2002J). These terms are described in the
3-5
-------
following discussions and have also been defined to support specific program
applications (see U.S. EPA, 2006a).
Validation focuses on evaluating the analytic quality and soundness of the model,
documenting its scientific basis and verifying the code, comparing the output with that of
other models and conducting empirical comparisons of model predictions with field
study data. This process also involves ensuring that the goals identified for the model
during the planning or scoping phase are met, determining the causes of any failure to
meet requirements and documenting the results.
Calibration involves comparing model results with information of known accuracy
and making adjustments to the model until its results lie within reasonable bounds of the
accurate information. Following calibration, the model is tested with a different set of
data also of known accuracy. If the model results lie within reasonable bounds of this
new dataset, this indicates success and suggests that the model can provide valid
predictions when applied to other independent data sets.
Complex cumulative health exposure assessments may utilize many different
models. These models will always be limited since they rely on assumptions to address
specific knowledge gaps or for analytic simplification. While some of the models may
have undergone verification, validation and calibration, other models may not.
Depending on the availability of data for development and testing, for some of these
models, a model evaluation, that is an examination of whether the model provides
reasonable results (often assumed to represent a lower level of assessment for model
viability) may be all that is feasible given resource constraints placed on these
assessments. In some studies model evaluation also can involve benchmarking against
other models to assess differences among outputs compared with expectations based
on experimental data, to quantify uncertainties and sensitivities and to guide refinement
of model components and underlying assumptions. Note that a similar process of
verification, validation and calibration also applies to the data input to these models
(U.S. EPA, 2002J). Models that have been verified or validated may be given more
credibility in the assessment and may be more useful than models that have not
undergone validation. This likely will be the case if the conditions under which the
model will be applied are similar to those under which it was developed. If a model has
not undergone model evaluation, it may not be useful in a cumulative risk analysis
(NRC, 2007).
Given difficulties in obtaining sufficient data for full validation of a cumulative risk
model, the first step may simply be to determine whether results are reasonable based
on current scientific knowledge as it continues to evolve. In addition, partial validation
3-6
-------
can be conducted by addressing only those components of the model for which better
characterization information exists. Sensitivity analyses can help focus these efforts,
targeting those elements that have the greatest impact on the model outputs.
Comparing these models with other models that frame the cumulative risk question from
different perspectives (e.g., those that emphasize different elements) also can
strengthen the modeling process.
As an example of progressive model validation, many years ago the dioxin data
from the Seveso plume were combined with health effects data from the local
community in order to reconstruct doses and improve the extant scientific model of
human toxicity. Early conceptual models have since been updated with human data,
including those from recent individual exposure incidents (CNN, 2004; Edmond et al.,
2005; also see model development description in NRC, 2007). Thus, model
representation of human exposure, toxicity and risk continues to be refined as new data
become available.
3.3. CUMULATIVE EXPOSURE ASSESSMENT: ANALYSIS PHASE
As described in U.S. EPA risk assessment guidance documents, the analytic
phase of an exposure assessment begins after the analyst has developed a preliminary
list of chemicals of potential concern and has identified the population and
subpopulations of concern. The materials presented in Chapter 2 identify data sources
and approaches that can be considered when conducting a cumulative exposure. The
applicable procedures differ depending on the initiating factor. If the initiating factor is
sources, typically there will be a well characterized list of chemicals known to be
released from the identified sources under both typical and abnormal operating
conditions. If the initiating factor is chemical concentrations then the initial list of
relevent chemicals will be identified during the characterization of the initiating factor. In
a cumulative risk assessment, the exposure analyst will typically consult with the toxicity
analyst to determine if other chemicals need to be considered for inclusion in the
analysis because compounds of concern are known to interact toxicologically with the
chemicals that initiated the cumulative risk assessment. If the initiating factor is
population illness, the toxicity and exposure analysts will colaborate closely to identify
the types of chemical exposures that could be associated with the illness and to
determine if such exposures could occur within the population (e.g., given the emission
sources in the area or past land uses).
As described in Chapter 2, the linkages between relevant aspects of the analysis
can be depicted using a conceptual model. Figure 3-1 provides an example conceptual
3-7
-------
Primary
Primary Release Environmental
Source Mechanism Transport Medium
Surfac
Conta
Subsu
._._ Fmission
r
Contact
i
N 1 i
Water
diment
Primary
Exposure
Route
Potential Receptor (Current or
Future)
Worker
Resident*
Recreational
Visitor
^| Inhalation
minated ,-, , * W Groundwater w
miridieu ^ Percolation 1 1
' .
Contaminated
Surfac
Slud
Sec
e Water, fc ni, t C
iment
Contact
Ingestion
Dermal Absorption
Ingestion
Dermal Absorption
Ingestion
Inhalation
Dermal Absorption
Ingestion
Dermal Absorption
| Bulk
ings | ^
Particulate or
Gaseous Emission
H Air 1 H Inhalation
Ingestion
Dermal Absorption
-K:
Air
Inhalation
Drinking Water
Ingestion
Inhalation
Dermal Absorption
Ingestion
Inhalation
Dermal Absorption
* The resident or visitor scenario may be expanded for cumulative assessments to consider unique exposures
of specific sensitive populations (e.g., subsistence fishers).
Grey fill boxes indicate complete exposure pathways.
Open boxes indicate exposure pathways that are not complete.
FIGURE 3-1
Conceptual Model for Hypothetical Cumulative Exposure Assessments
Illustrating Pathways Considered and Complete Pathways
3-8
-------
model for a contaminated site. Although the initiating factors could vary across
communities, as indicated in Figure 2-1, the same exposure assessment steps are
addressed (see Text Box 3-3). In each of Sections 3.3.1, 3.3.2 and 3.3.3, cumulative
exposure issues are identified and existing approaches are shown that can be used to
address the issue. Typically, exposures are estimated for complete exposure
pathways. Complete implies
that each exposure assessment
component is present from the
occurrence of the chemical through
relevant exposure pathways and routes
to the receptor. Exposures may be
estimated for pathways that are not
currently complete but are considered
likely to be complete in the future.
3.3.1. Exposure Setting. Describing
the environmental characteristics of the
Exposure Assessment: Analysis Steps
(Text Box 3-3)
Characterize the
exposure setting
(3.3.1)
Identify potential
exposure pathways
(3.3.2)
Quantify exposures
through multiple
exposure routes
(3.3.3)
Identify environmental features
and potential receptors
Describe sources, release
mechanisms, receiving media
and locations for chemicals
Estimate medium-specific
chemical concentrations at
points of human exposure and
calculate intakes (considering
time, frequency, duration)
study area and identifying the people who were, are or could be exposed to multiple
chemicals are the two main elements of the exposure setting for a community-based
assessment. The following subsections describe cumulative risk assessment issues
related to these elements.
3.3.1.1. Environmental Features Characterizing the exposure setting
potentially involves compiling basic data on topography, surface hydrology, soil geology,
vegetation, groundwater hydrology, climate and meteorology, land use, pollution
sources and demography of the community. The analyst routinely assembles
geographic and meteorologic data when conducting an exposure assessment. Basic
geographic information about a community is available through a variety of sources
including those offered by the U.S. Geological Service and U.S. Department of
Agriculture. Climate and meteorologic data are available from the National Oceanic and
Atmospheric Administration, e.g., the National Weather Service. Analysts also identify
land uses. Land use analyses include the identification of all residential areas, work
places, recreational areas and places where foods are grown or collected as well as
relevant pollution sources inside and outside the community. For example, regionalor
outsideemission sources could impact pollutant levels in the community and in some
analyses these require identification. Community input to these identification processes
3-9
-------
is important. This includes gaining an understanding of how different locations in a
community are currently used and how they were used. Past uses may provide the
analyst with important insights in to once-common polluting practices and to potential
past exposures.
In a community assessment, in addition to examining the contaminants present,
the analyst may need to examine environmental conditions in the broader region. For
example, if there are atmospheric sources of concern for an affected community in
which there is a Superfund site, the EPA requires that the assessment include an
examination of the concentrations in the local environment from these atmospheric
sources and the potential for airborne contamination from the Superfund site (U.S. EPA,
1989a). Ambient data for such an analysis can be obtained from various organizations,
such as U.S. EPA regional offices and state, county or city environmental agencies.
Pollutant release data can be obtained from the Toxics Release Inventory,
http://www.epa.gov/tri/chemical/index.htmffchemlist: Appendix A lists additional
resources providing such data.
To illustrate how different types of data can be used, Text Box 3-4 illustrates data
sources tapped for a recent cumulative study of air toxics in an urban area. The
broader scope of a cumulative exposure assessment could include background data on
chemical concentrations in local soil and
water, both naturally occurring (such as
metals) and anthropogenic chemicals
(such as PAHs, PCBs and dioxins) as
well as concentrations of chemical
pollutants in the U.S. food supply. For
example, Volume 2, Properties,
Environmental Levels, and Background
Exposures, of the draft U.S. EPA dioxin
document (2003c) lists typical
concentrations of dioxin congeners in the
U.S. food supply. These nationally
representative samples could be
incorporated into a cumulative exposure
assessment, if relevant. Such exposure
pathways when combined with local
exposure pathways may be a significant
source of exposure.
Example Data Sources and Uses
(Text Box 3-4)
A recent air screening hazard assessment
(U.S. EPA, 2004c) used data from several regional
and local sources, including emissions data from the
TRI, Cumulative Exposure Project (CEP), and
Regional Air Pollutant Inventory Development
System (RAPIDS), as well as outdoor air monitoring
data. These data were combined and compared to
identify any consistently higher hazard areas,
pollutants and sources. Two methods were used to
estimate relative inhalation hazards of outdoor air
toxics: one for emissions mass (using TRI and
RAPIDS data) and the other for outdoor
concentrations (using CEP and monitored data).
Emissions data enabled sources and release
locations to be identified which improved the
exposure assessment. (Note that TRI and RAPIDS
emissions databases differ: TRI data are self-
reported by facilities, while RAPIDS data are
estimated by states from permits and other
information sources.) Ambient data provided limited
information on spatial distribution, without regard to
specific sources. A WOE approach was used to
assess data among different sources.
3-10
-------
The analysis of environmental features identifies potentially vulnerable
populations (see Section 3.3.1.2) and the location of sites where people in a community
could be exposed. Community members may provide valuable input into the
identification of such sites, the relevant activities that may occur there and the frequency
with which the site is or may have been used. This information can provide insights into
potential exposures and potential subpopulations being exposed through use of the
location. When performing cumulative exposure assessments, the analyst may need to
evaluate exposures where community members gather. For example, community
members gather in schools and at playgrounds, and the analyst may need to evaluate
exposures in susceptible populations (e.g., asthmatic children) at these locations. The
analyst also may want to examine exposures that occur in and around facilities that care
for the elderly and disabled members of a community. Depending on the chemicals
being evaluated, the exposure and toxicity analysts may wish to colaborate closely to
identify other settings where chemical exposures could occur in vulnerable populations
(see Section 3.3.1.2 for further discussion).
3.3.1.2. Receptor Characteristics Considered in Cumulative Risk
Assessments During characterization of the exposure setting, the analyst identifies
individuals and population groups that could be exposed to contaminants. Then,
information on the residential locations, activity patterns and workplaces is collected.
Cumulative risk assessments also examine exposures among both "typical"
members of a community and vulnerable populations. Usually, the exposure analyst
and the toxicity analyst work together to identify the potentially vulnerable populations.
U.S. EPA's Framework for Cumulative Risk (2003a) adopts "vulnerability" concepts
described by Kasperson that encompass the topic of receptor characteristics. The EPA
document details four areas of vulnerability:
Differential exposure
Susceptibility/sensitivity
Differential preparedness
Differential ability to recover
Typical exposure assessments routinely identify some subpopulations that are,
or may be, differentially exposed due to close proximity to a source or contaminated site
and some exposure assessments also may identify subpopulations that exhibit activity
patterns that may result in elevated exposures to pollutants (e.g., subsistence fishing).
In these cases, detailed recreational uses and activity patterns are based on survey
data, especially for fishing and hunting. Such data may be obtained from state or
3-11
-------
county departments of environment, conservation, natural resources or parks and
recreation. The analyst may conduct community-specific surveys to fill important gaps
(U.S. EPA, 1998a). The analyst also may meet with specific groups that are or could be
affected, in order to assess possible unique exposures. For example, Native Americans
may gather special vegetation or wildlife for food, medicine or ceremonies or visit lands
that are sacred.
Exposures in subpopulations exhibiting susceptibility/sensitivity, differential
preparedness and differential ability to recover are not always considered in typical
exposure assessments but are given special consideration in a cumulative risk
assessment. Exposures may be calculated separately for identified subpopulations with
specific receptor characteristics to yield more realistic exposure estimates for those
subpopulations. The receptor population characteristics considered in a cumulative risk
assessment may include diverse factors such as genetic susceptibility, age, stress,
disease state, economic status, ethnicity, health status, availability of health care, etc
(see Figure 1-6). It is particularly important that the analyst evaluate whether certain
potentially susceptible populations are exposed to high levels of pollutants. Examples
of information the analyst can use to support this evaluation are highlighted in
Text Box 3-5. Pregnant women can represent a subgroup of special concern due to the
fetus's sensitivity for potential
effects under some types of
chemical exposures. For
example, the fetal nervous
system is considered the most
sensitive target of
methylmercury and the EPA's
reference dose (RfD) has been
developed based on
neurological effects associated
with intrauterine exposures
(U.S. EPA, 2001 b).
Young children can be more biologically sensitive to many chemicals because
certain protective body functions (e.g., liver enzyme production) are developing during
the early stages of life and not yet fully protective. They also can incur higher
exposures than the general population because of their different behaviors (e.g., pica or
recreational swimming) and because their doses per unit body weight are higher than
those of adults. Following the 1997 Executive Order for the protection of children from
Information for Susceptibility Assessment
(Text Box 3-5)
Type of Information
Demographic data
Subpopulation groups
Locations (e.g., schools,
hospitals, nursing homes)
Exposure data (e.g., blood
lead levels)
Cancer registries
Other health effect
registries
Resources
U.S. Census Bureau (www.census.gov)
EPA report: Sociodemographic Data Used
for Identifying Potentially Highly Exposed
Populations (U.S. EPA, 1999c)
Plat maps, city and county health
departments
State registries, county and city health
department reports
Centers for Disease Control (national data
and links to state cancer registries,
www. cdc. go v/can cer/npcr/statecon .htm)
State registries of birth defects, asthma
registry
3-12
-------
environmental health and safety risks, the EPA continues to develop approaches to
account for differences such as body weights and toxicokinetics so risks to infants and
children can be evaluated in further detail whenever there appears to be a greater
concern for adverse health effects than for the general population.
Other people with higher than average biological sensitivity to environmental
stressors include those with allergies and with pre-existing medical conditions (e.g.,
asthma). Some state health departments have established health registries for health
conditions, such as asthma, and for exposure measurements, such as blood lead
levels. The analyst also may contact these agencies to determine if any clusters of
affected individuals live in the community. Elderly and immunocompromised
populations can be more susceptible to environmental exposures due to their health
status. Other factors, like socioeconomic status, can affect access to health care or
contribute to poor diet. Thus, poverty could indicate a potential increased susceptibility
or biological sensitivity.
3.3.1.3. Cumulative Exposure Assessment Practices for Receptors Once
the land uses and sources of pollutants in the community have been identified (Section
3.3.1.1), it is common practice in exposure assessments to identify representative
default receptors, such as a current or future resident, trespasser, home gardener and
recreational angler. Exposures among these default receptors are subsequently
estimated. The Exposure Factors Handbook (U.S. EPA, 1997c) provides factors
associated with these receptors (e.g., quantities of homegrown vegetables consumed
daily).
In typical assessments, the individual receptors are located in close proximity to
a pollution source (e.g., at the fence line, the nearest housing development or the
closest fishable lake). The analysts may have to use atmospheric dispersion models to
identify sites proximal to multiple pollution sources. The analyst may evaluate other
receptors who are, or could be, subjected to higher than average exposures, including
people living near multiple sources of pollution (e.g., waste facilities, urban industrial
areas or transportation corridors), residents of older homes with lead-based paint and
people whose jobs or recreational activities can cause specific chemical exposures or
increased opportunities for exposure. The exposure analyst also evaluates exposures
in vulnerable populations (Section 3.3.1.2). If these screening practices do not reveal
exposures of concern, the exposure analyst can drop the receptors from the analysis,
after consultation with the dose-response analyst.
3-13
-------
If the exposure levels are deemed to be of concern, then the analyst can use
demographic data to estimate the typical ages and ethnicities of these hypothetical
community members who may be differentially exposed to pollutants from a source.
These data may be used to refine the exposure estimate (see Section 3.3.3).
3.3.2. Exposure Pathways and Routes. An exposure pathway describes how
chemicals are transported from a source to a person or subpopulation. An exposure
route identifies the way the contaminant actually enters the body. For environmental
pollutants, the major exposure routes are inhalation, ingestion and dermal absorption.
This section identifies considerations for how exposure pathways can be evaluated in
an assessment. The basic process elements are summarized in Text Box 3-6.
The overall analysis plan for a risk
assessment typically describes the general
data, models and assumptions that will be
used to characterize exposure (Chapter 2).
A main emphasis for cumulative risk
assessments is on how sources, chemicals,
media and receptors can be grouped for
joint pathway analyses. Various examples
are offered in this section, with additional
detail for one pathway (air) offered in
0 ,. _ , ... , , . ... . . chemicals or drinking water)
Section 3.4 to illustrate how cumulative risk
Exposure Pathway Elements (Text Box 3-6)
Locations of sources, mechanisms by which
chemicals could be released from sources, and
identification of receiving environmental media
Transport of chemicals in the receiving media and
movement from receiving media into other
environmental media (e.g., from soil to air or
water), degradation and transformation (change in
speciation, sorption, etc.)
Estimated concentrations of contaminants at
points of potential human contact (i.e., exposure
points) and associated routes of exposure (e.g.,
incidental ingestion of soil, inhalation of airborne
assessment issues can be considered. If the initiating factor is population illness, then
the exposure analyst would have to collaborate with the toxicity analyst to identify
possible relevant chemicals and then determine if there are possible sources of such
contaminants in the local environment. If the initiating factor is elevated chemical
concentrations or multiple sources, then the sequence of steps is similar to that which
follows.
3.3.2.1. Sources and Fates of Chemicals and Chemical Mixtures When
performing a cumulative risk assessment initiated by environmental contaminants, the
analyst would want to identify all sources being considered and all potential exposure
pathways for each medium of exposure. The analyst then reviews the pathways to
determine if they are relevant. The completeness of each exposure pathway is then
evaluated. A pathway is complete when these four components are present:
3-14
-------
A source and a mechanism of contaminant release
An environmental transport medium
A point of human contact with the contaminated source or transport medium
A route of exposure at that point
The exposure analyst develops criteria for inclusion in the cumulative risk assessment
after discussion with the toxicity analyst; then resources can be efficiently focused on
toxicologically relevant exposures. The pathways selected for inclusion are then
characterized, and the exposures from all relevant pathways are jointly evaluated for the
cumulative risk assessment.
In cumulative exposure assessment, an evaluation of environmental
transformation of each chemical under consideration is a critical component for each
selected pathway. While environmental transformation is recognized as a major factor
for organic compounds, some metals can be altered in the environment, e.g., via
methylation by biological processes, which can change bioavailability and toxicity.
For example, BIOCHLOR is a numerical screening model developed by the Air
Force and maintained by the EPA to assess monitored natural attenuation (MNA) for
sites contaminated with chlorinated solvents. The code simulates solute transport for
sequential reactions from a single parent chemical, involving up to four fate products
(Jones et al., 2006). It can be run either with or without biotransformation assuming
sequential first-order decay (notably for reductive dechlorination, which is generally the
main biodegradation process at these sites). The model has been applied for a number
of projects to integrate site-specific information into evaluations of solvent degradation.
These projects extend beyond industrial Superfund sites (Clement et al., 2002) and U.S.
federal sites (U.S. DOE, 2005) to international sites (Nakashima et al., 2005), with a
variety of case studies reviewed by McGuire et al. (2004). Where site-specific data are
unavailable, the User's Manual provides ranges for a number of input parameters
(available from U.S. EPA, 2002k), which can be used to help focus site investigations in
support of MNA suitability or performance assessments as indicated.
Environmental transformation is a critical consideration when addressing
exposures to environmental mixtures. For organic chemicals such as common
solvents, environmental transformation or degradation can produce a number of new
chemicals of potential concern in addition to those originally released. While some
degradation products are less toxic than their parent compounds, this is not always the
case. Thus, it is helpful to review historic operations records and other readily available
data to consider additional contaminants that might warrant consideration. To illustrate,
the solvent tetrachloroethylene is a common groundwater contaminant, and this volatile
3-15
-------
organic compound (VOC) is converted over time to the more toxic vinyl chloride. Key
properties of selected organic chemicals and degradation products are illustrated in
Table 3-1 to show that data are available to characterize environmental fate of multiple
pollutants and cumulative exposures.
When evaluating environmental fate and transport across media for these
assessments, it is important4 that mass be maintained when predicting concentrations
of parent chemicals and degradation products. Chemical speciation can also be
important for cumulative risk assessments. Different oxidized or reduced forms of
metals react differently in the environment and have different toxicities; trivalent and
hexavalent chromium provide a good example, with the latter being much more toxic.
Thus, it is important to characterize the soil and water chemistry at sites to assure that
appropriate physicochemical characteristics are being reflected in the assessment. In
evaluating combined chemicals, care must be taken to assure that assumptions are
internally consistent among all chemicals within a given setting. For example, assuming
the presence of a reduced form of a metal may be incorrect, especially in an aerated
environment where other chemicals are assumed to be in the oxidized form.
When evaluating contaminant fate, it is important to consider setting conditions
that can contribute to chemical-chemical interactions. This applies to natural systems
such as fields or ponds as well as manmade systems such as drinking water distribution
networks, where chemical interactions (often combined with microbial processes) can
convert introduced compounds to other forms. Main environmental reaction processes
are oxidation, complex formation with various ligands and biologically mediated
reduction (methylation).
The environmental fate of mercury (Hg) illustrates the importance of considering
setting conditions. Hg could be released from a combustor as an elemental vapor and
converted in the local atmosphere to a reactive gaseous form. Reactive gaseous Hg is
thought to deposit rapidly to the surface of the earth. In aqueous environments and in
wetlands, mercury can be transformed to methylmercury, which bioaccumulates in fish.
Setting conditions, including wind direction; wind speed; local atmospheric chemistry;
proximity of Hg releases to wetlands, lakes or rivers; the aquatic chemistry in these local
bodies of water; and the size of the watershed, influence methylmercury levels in local
fish.
4 Although models may not explicitly conserve mass, post-processing can be applied to assure that this is
maintained.
3-16
-------
TABLE 3-1
Properties of Selected Organic Chemicals and Degradation Products to Demonstrate Availability of Information*
Chemical
Key
Degradation
Products
General Fate/
Persistence
Environmental Half-Life
and Reaction Mechanism
Log Kow
(unitless)
Log Koc
(unitless)
Illustrative U.S.
Concentration
EPA IRIS
Toxicity
Value
Toxicity
Relative
to Parent
Pesticides
Aldrin
(see below)
Dieldrin
Binds tightly to
soil and does not
leach readily, so
is not usually
found in
groundwater;
moderately
persistent;
bioaccumulates
As for aldrin, but
very persistent
Air
36 minutes
(Howard, 1989)
Water
5 weeks
(Howard, 1989)
Soil:
20-109 days
(ATSDR, 2002a;
Howard, 1989)
Air
not specified
(Howard, 1989)
Water:
hours to months
(Howard, 1989)
Soil
2.5-7 years
(ATSDR, 2002a;
Howard, 1989)
Hydroxyl
radical
oxidation
(not specified)
(not specified)
Photo-
degradation
Evaporation
(not specified)
6.5 (5.7-7.4)
(bioaccumulation
likely)
(Howard, 1989,
ATSDR, 2002a)
6.2 (4.3-6.2)
(bioaccumulation
likely)
(ATSDR, 2002a)
7.7(5.4-7.7)
(expected to
strongly adhere
to soil)
(ATSDR, 2002a)
6.7
(expected to
strongly adhere
to soil)
(ATSDR, 2002a)
Air 0.00003 ppb
(mean for 1970-72
from 13.5% positive
samples, 16 states)
Water 0.001 ppb
(STORET median,
ambient water,
40% detects; 1985)
Sediment 0.1 ppb
(STORET median,
33% detects; 1985)
(ATSDR, 2002a)
Air 0.0001 ppb
(mean for 1970-72
from 94% positive
samples, 16 states)
Water 0.001 ppb
(STORET median,
ambient water,
40% detects; 1985)
Soil
1-49 ppb (mean)
Sediment 0.8 ppb
(STORET median,
33% detects; 1985)
(ATSDR, 2002a)
RfD
0.00003
mg/kg-d
DWUR:
0.49 per
mg/L
IUR:
4.9 per
mg/m3
RfD
0.00005
mg/kg-d
DWUR
0.46 per
mg/L
IUR
4.6
per
mg/m3
NA
Noncancer
(oral)
60%
Cancer
(oral-inhln)
94%
3-17
-------
TABLE 3-1 cont.
Chemical
Chlordane
Key
Degradation
Products
NA
(this
compound is
not typically
transformed
in the
environment)
General Fate/
Persistence
As for dieldrin,
and in surface
water will
volatilize and
adsorb to
sediments
Environmental Half-Life
and Reaction Mechanism
Air
1.3 days
(ATSDR, 1994a)
6.2 hours
(Howard, 1989)
Water
240 days
(U.S. EPA, 2000b)
7.3-7.9 hours
(Howard, 1989)
Soil
3.3 years
(Howard, 1989)
(not specified)
Hydroxyl
radical
oxidation
(not specified)
Volatilization
(not specified)
Log Kow
(unitless)
5.5
(estimate for
pure chemical)
(bioaccumulation
likely)
(U.S. EPA,
2006b; ATSDR
1994a; Howard,
1989)
Log Koc
(unitless)
3.5-4.6
(4.2 -4.4
estimated)
4.1 (mean)
expected to
adhere to soil)
(U.S. EPA,
1996a, 2006b;
ATSDR, 1994a)
Illustrative U.S.
Concentration
Surface and
groundwater
0.1 ppb
(mean in selected
areas)
Soil
<1-140 ppm
(ATSDR, 1994a)
EPA IRIS
Toxicity
Value
RfD
0.0005
mg/kg-d
RfC
0.7 ug/m3
DWUR
0.01 per
mg/L
IUR
0.1
per mg/m3
Toxicity
Relative
to Parent
NA
3-18
-------
TABLE 3-1 cont.
Chemical
Key
Degradation
Products
General Fate/
Persistence
Environmental Half-Life
and Reaction Mechanism
Log Kow
(unitless)
Log Koc
(unitless)
Illustrative U.S.
Concentration
EPA IRIS
Toxicity
Value
Toxicity
Relative
to Parent
Solvents
Carbon
tetrachloride
(see below)
Chloroform
Stable in air;
volatilizes rapidly
from soil and
surface water;
little binds to soil
(moderately
soluble so can
leach to
groundwater);
does not
bioaccumulate
Persistent in
groundwater;
does not
bioaccumulate
Air
330 years
Groundwater:
0.4-4.5 days
Surface water:
0.5-1 years
7-28 days
7000 years
Soil
0.5-1 years
(data from ATSDR
2003a)
Air
80 days
(Howard, 1989)
Water
36-40 hours
(Howard, 1989)
Surface water
44 days
(ATSDR, 1997a)
Oxidation
Reaction
with minerals
Aerobic
biodegradation
Anaerobic
biodegradation
Hydrolysis
(Based on
aerobic
conditions)
Hydroxyl
radical
oxidation
Volatilization
Hydrated
electrons
2.6-2.8
(ATSDR, 2003a;
Howard, 1989)
2.0
(not likely to
bioaccumulate)
(ATSDR, 1997a;
Howard, 1989)
2.0
(expected to
move with
groundwater)
(ATSDR, 2003a)
2.0 (mean)
(expected to
move with
groundwater)
(ATSDR, 1997a)
Air
0.2 ppb (mean)
Drinking water
0.5 ppb (mean for
the 3% of samples
with detectable
levels)
(ATSDR, 2003a)
Drinking water:
23 ppb (mean)
(ATSDR, 2003a)
RfD
0.0007
mg/kg-d
DWUR
0.0037
per mg/L
IUR
0.015
per mg/m3
RfD
0.01
mg/kg-d
IUR
0.023
per mg/m3
NA
Noncancer
(oral)
7%
Cancer,
(inhaln)
150%
3-19
-------
TABLE 3-1 cont.
Chemical
Tetrachloro-
ethylene
Key
Degradation
Products
Chlorine
(see below)
General Fate/
Persistence
Reacts with water
to form
hypochlorous and
hydrochloric
acids; volatilizes
from soil; persists
in groundwater;
does not
bioaccumulate
Volatilizes rapidly
from surface
water and soil;
can leach slowly
to groundwater,
(only slow soil
biodegradation);
does not
bioaccumulate
Environmental Half-Life
and Reaction Mechanism
Air
seconds to
minutes
(NPI, 2005)
Surface water
seconds to
minutes
(U.S. EPA, 1994b)
Air
70-250 days
(ATSDR, 1997b)
110 days
(Mackay et al.,
2006)
Water
0.8-6 years
(ATSDR, 1997b)
4-4.5 hours
(ATSDR, 1997b)
180 days
(TOXNET, 2005)
98 days
(HSDB, 2006)
Soil
2-16 days
(ATSDR, 1997b;
Mackey et al.,
2006)
Hydrolysis
Hydroxyl
radical
oxidation
Tropospheric
reaction
Hydrolysis
Volatilization
Aerobic
biodegradation
Anaerobic
biodegradation
Volatilization
Log Kow
(unitless)
0.9
(not likely to
bioaccumulate)
(TCEQ, 2003)
2.5-3.4
(not likely to
bioaccumulate)
(ATSDR, 1997b;
Mackay et al.,
2006,
Howard, 1989)
Log Koc
(unitless)
Not identified
(organic carbon
in soil does not
appear to play a
major role)
1.8-3.6
(expected to
moderately bind
to soil and can
leach to
groundwater)
(ATSDR, 1997b;
Mackay et al.,
2006)
Illustrative U.S.
Concentration
Air
0.0006-0.02 ppm
(HSDB, 1991;
ARE, 1997)
Drinking water
0.2-1 mg/L
(WHO, 1996)
1.4-2. 7 mg/L
(finished water in
several U.S. cities)
(U.S. EPA, 1981)
Air
0.50 ppb (mean,
including areas
close to emission
sources)
Drinking water
0.75 ppb (median,
from ground water,
for the 8% of
samples with
detectable levels)
Sediment
5 ppb (median)
(ATSDR, 1997b)
EPA IRIS
Toxicity
Value
RfD
0.1
mg/kg-d
RfD
0.01
mg/kg-d
Toxicity
Relative
to Parent
Noncancer
(oral):
0.7%
NA
3-20
-------
TABLE 3-1 cont.
Chemical
Tetrachloro-
ethylene
(cont.)
Key
Degradation
Products
Trichloro-
ethylene
1,1-Dichloro-
ethylene
General Fate/
Persistence
Volatilizes quickly
from surface
water; binds to
soil; persistent in
groundwater;
does not
bioaccumulate
Volatilizes
relatively quickly
from surface
water and soil;
moves with
groundwater;
stable in water;
degradation
expected to be
slow; does not
bioaccumulate
(U.S. EPA,
2006c)
Environmental Half-Life
and Reaction Mechanism
Air
6.8 days
(ATSDR, 1997c)
Water
100 days
(Mackay et al.,
2006)
400 days
(Mackay et al.,
2006)
320 days
(Mackay et al.,
2006)
Months to millions
of years
(ATSDR, 1997c)
Air
2.3 days
(ATSDR, 1994b;
Mackay et al.,
2006)
11 hours
(Mackay et al.,
2006)
Water
4 days (mean)
(ATSDR, 1994b)
670-4300 hours
(Mackay et al.,
2006)
Hydroxyl
radical
oxidation
Aerobic
biodegradation
Anaerobic
biodegradation
Hydrolysis
Hydrolysis
Hydroxyl
radical
oxidation
Photooxidation
Volatilization
Aerobic
Log Kow
(unitless)
2.3-2.6
(not likely to
bioaccumulate)
(ATSDR, 1997c;
Howard, 1989;
Mackay et al.,
2006)
1.3-2.1
(not likely to
bioaccumulate)
(ATSDR, 1994b;
Mackay et al.,
2006)
Log Koc
(unitless)
2.0-2.7
expected to
moderately
bind to soil and
move with
groundwater)
(ATSDR, 1997c;
Howard, 1989;
Mackay et al.,
2006)
1.8-2.2
(not expected to
bind to soil;
expected to
move with
groundwater)
(ATSDR, 1994b;
Mackay et al.,
2006)
Illustrative U.S.
Concentration
Air
0.56 ppb (mean,
including areas
close to emission
sources)
Drinking water
1 ppb (median,
from groundwater,
for the 10% of
samples with
detectable levels)
Sediment
<5 ppb (median)
(ATSDR, 1997c)
Air
4.6 ppb (mean);
Drinking water
0.6 ppb (mean, for
the 3% of samples
with detectable
levels)
(ATSDR, 1994b)
EPA IRIS
Toxicity
Value
NA
RfD:
0.05
mg/kg-d
RfC:
0.2 mg/m3
Toxicity
Relative
to Parent
NA
Noncancer
(oral):
20%
3-21
-------
TABLE 3-1 cont.
Chemical
Key
Degradation
Products
1,2-Dichloro-
ethylene,
trans-
(the cis- form
is also
produced but
no toxicity
value exists)
Vinyl chloride
General Fate/
Persistence
Volatilizes quickly
from surface
water and soil;
moves with
groundwater;
does not
bioaccumulate
Volatilizes quickly
from surface
water and soil;
moves with
groundwater;
does not
bioaccumulate
Environmental Half-Life
and Reaction Mechanism
Soil <10 days
(Mackay et al.,
2006)
Air
3.5-5 days
(ATSDR, 1996;
Howard, 1989;
Mackay et al.,
2006)
Surface water
3-6.2 hours
(ATSDR, 1996;
Howard, 1989)
Air
1.5 days
(Howard, 1989;
Mackay et al.,
2006)
1 1 weeks
(Mackay et al.,
2006)
Water
0.81 hours
(Howard, 1989;
Mackay et al.,
2006)
Soil
30-180 days
(Mackay et al.,
2006)
(not specified)
Hydroxyl
radical
oxidation
Volatilization
Hydroxyl
radical
oxidation
Tropospheric
reactions
Volatilization
(not specified)
Log Kow
(unitless)
2.1
(not likely to
bioaccumulate)
(ATSDR, 1996;
Howard, 1989;
Mackay et al.,
2006)
1.4-2.8
(not likely to
bioaccumulate)
(ATSDR, 1997d;
Howard, 1989;
Mackay et al.,
2006)
Log Koc
(unitless)
1.6-1.8
(not expected to
bind to soil;
expected to
move with
groundwater)
(ATSDR, 1996;
Mackay et al.,
2006)
0.5-2.0
(not expected to
bind to soil;
expected to
move with
groundwater)
(ATSDR, 1997d;
Mackay et al.,
2006)
Illustrative U.S.
Concentration
Air
0.037 ppb (median)
Drinking water and
groundwater
173 ppb (mean)
(ATSDR, 1996)
Air
Oto 0.04 ppm
(generally not
detected, but can
be elevated near
landfills or industrial
facilities with this
chemical or
parents)
Drinking water
1, 8.4 ppb
(maxima for
random and
nonrandom sites,
respectively;
detected in 0.74%
of groundwater
supplies)
(ATSDR, 1996)
EPA IRIS
Toxicity
Value
RfD
0.02
mg/kg-d
RfD
0.003
mg/kg-d
DWUR
(from
birth)
0.0021
per mg/L
IUR
(from
birth)
0.0088
per
mg/m3
Toxicity
Relative
to Parent
Noncancer
(oral)
50%
Noncancer
(oral)
333%
3-22
-------
TABLE 3-1 cont.
Chemical
Key
Degradation
Products
Carbon
tetrachloride
General Fate/ Environmental Half-Life
Persistence and Reaction Mechanism
Log Kow
(unitless)
Log Koc
(unitless)
Illustrative U.S.
Concentration
As described above (for listing its as a primary chemical solvent)
EPA IRIS
Toxicity
Value
RfD
0.0007
mg/kg-d
(and other
toxicity
values)
Toxicity
Relative
to Parent
Noncancer
(oral)
1400%
*Organic compounds illustrated here are often found at Superfund sites; others also commonly found include acetone, 2-butanone and methylene
chloride; polycyclic aromatic hydrocarbons (PAHs)/naphthalene, pentachlorophenol and polychlorinated biphenyls (PCBs); and benzene, toluene
and xylene (designated by U.S. EPA as "pending" for this list). (Source: U.S. EPA's Common Chemicals Found at Superfund Sites,
http://www.epa.gov/superfund/resources/chemicals.htm.) General fate content is highlighted from the ATSDR toxicological profiles as also
supported by various EPA facts sheets (as indicated), and many properties and environmental levels are also from the toxicological profiles.
NA = not available. Gray shading indicates the entry is not applicable because this is the parent compound. Toxicity values are highlighted from
the EPA IRIS database (U.S. EPA, 2007); the RfD (reference dose) and RfC (reference concentration) address the noncancer endpoint, while the
DWUR (drinking water unit risk) and IUR (inhalation unit risk) address the cancer endpoint; inhln = inhalation.
The environmental half-life represents the time it takes for the initial amount of a chemical to be reduced by half in that medium. To
streamline the presentation (with many of the values having been calculated or estimated), numbers are generally rounded to two significant
figures or one decimal point. The Kow indicates whether a chemical is hydrophilic and will be predominantly found in water, or is lipophilic and will
be found in fatty tissue of animals or associated with other organic materials in aquatic systems. The Kow values are presented as logarithms
because this measure varies widely across compounds. A log Kow of 0 indicates an equal affinity for lipids and water. A high log Kow indicates
the chemical is not very soluble and will not move with water; a low log Kow indicates the chemical is very soluble and will move with water (it also
indicates the chemical will be readily absorbed from the gastrointestinal tract after being ingested or from the lungs after being inhaled). As the log
Kow increases, the solubility in lipids increases, which means the potential to bioconcentrate in aquatic organisms increases; when the log Kow
reaches 5 to 6 it indicates the chemical can bioconcentrate significantly in aquatic organisms. As it increases above 6, the chemical is less likely
to bioconcentrate, approaching no bioconcentration at a log Kow of 12. The Koc indicates how the organic compound will partition between water
and the organic carbon portion of soil/sediment and biota. The Koc indicates whether or not a chemical will move with ground water. These
values are also presented as logarithms because, like Kow, this measure varies widely across compounds. A high log Koc (e.g., 3.5 or higher)
indicates the chemical is likely to sorb to soils, sediments, or sludges and is less likely to move with surface water or groundwater. A low log Koc
(e.g., 2.4 or below) indicates the chemical is not likely to sorb to soils, sediments or sludges and thus is more is likely to move with water.
Contaminants with a log Koc between 2.4 and 3.5 likely partition to soils, sediments or sludges and surface water or groundwater. (Source: U.S.
EPA Pollution Prevention (P2) Framework, Environmental Fate Models, see http://www.epa.qov/oppt/p2framework/docs/envfate.htm).
Note that these chemicals were selected to illustrate fate links for pesticides of historical interest and for solvents commonly found at
contaminated sites. The level of information highlighted in this table these will typically not exist for all compounds for a given cumulative risk
assessment. The unavailability of key data for one or more chemicals being assessed can represent a main source of uncertainty for the analysis,
and it is important to address this as part of the risk characterization discussion.
3-23
-------
Major ions in the environment with which introduced chemicals can react include
iron and manganese cations and anions of sulfur and phosphorous anions. Naturally
occurring metals such as arsenic (As), Cd, chromium (Cr), Hg, nickel (Ni), Pb and zinc
(Zn) are also common, introduced contaminants in terrestrial and aquatic systems, from
releases such as combustor emissions and effluent discharges. Metal cations can exist
as potentially toxic uncomplexed species or as relatively nontoxic complexed forms,
usually with organic ligands or non-metallic inorganic anions such as oxides, sulfates or
phosphates.
The potential for chemical-chemical interactions depends on many factors related
both to the chemical (e.g., for metals, the activity, solubility, electronegativity,
coordination number and density) and the nature of the medium (e.g., for aquatic
systems, the pH and temperature, oxidizing and photolysis potential; organic, particulate
and microbial content and salinity and presence of other chemicals). To illustrate for
one of these factors, the pH of natural or treated waters affects both the type of metal
complexes that form and the fraction of various species that would precipitate. For
example, metal carbonates are expected to precipitate as the pH rises above 8, while
the cation and anion would stay in solution at an acidic pH below 6. Thus, when
assessing the joint fate of contaminants to estimate exposure levels for a cumulative
risk assessment, characterizing the setting well can be key to a realistic analysis.
The various chemical interactions in a drinking water distribution system also
illustrate the types of interactions that analysts may encounter when conducting a
cumulative risk assessment. Free chlorine (Cb), which can be represented by
hypochlorous acid (HOCI) or hypochlorite (OCI), is a common disinfection residual. CI2
is a potent oxidizer that is a strongly electronegative and acts as an electron acceptor in
forming complexes with a wide variety of both inorganic and organic chemicals that
could be present in finished water. For example, it can combine readily with
(1) ammonia, to form chloramines (2) reducing agents such as ferrous ion (Fe2+), to
form the chloride and (3) humic material, to form trihalomethanes. In aerobic systems,
chlorine can also rapidly convert the trivalent form of As (arsenite) to the pentavalent
form (arsenate), which is less toxic when based on environmental exposure levels (due
to less cellular uptake than the trivalent form, while equivalent intracellular levels are
equipotent; ATSDR, 2000d). Biological processes can also combine to produce organic
forms of As, which are generally less toxic than inorganic forms that may have been
introduced to the system. Thus, for both natural and manmade systems, a number of
chemical-chemical interactions can influence the exposure profile for a cumulative risk
assessment.
3-24
-------
For radioactive compounds, the natural physical decay process causes
radionuclides to change over time. For these contaminants, natural attenuation
(radioactive decay) will reduce contaminant levels over time. The basic concepts of
half-life and natural attenuation over time are illustrated in Figure 3-2 (from Brown,
1999). Table 3-1 shows that the half-life for tritium is approximately 12.3 years.
Figure 3-2 illustrates natural attenuation over time showing that ambient levels of tritium
are predicted to be approximately 10% of original levels after 50 years. The parallel
evaluation for non-radioactive chemicals reflects environmental half-life.
Once released from different sources in various forms, chemicals can migrate to
other locations and media. The degree to which a particular chemical substance favors
a given transport path depends on the form of the chemical released, its physical state
and the nature of any particulate matter to which it might adsorb upon or following
release. These pathways are generally predictable from the known release processes
and expected physical forms of the chemicals.
The transport and fate of mixtures of chemicals released to the environment are
not random but can be predicted to varying degrees by considering a number of factors
related to the release, migration and persistence of their constituents. Following release
from a source, mixture components are typically differentially transported through the
environment. These chemical mixtures are subject to transformation reactions in the
environment, which can change their composition. Some chemicals are degraded,
while others are formed through various environmental reactions. Changes in the
mixture composition can be specific to the environmental medium. It is important to
document these changes in the mixture composition. The differential nature of transport
can be an important consideration in the toxicity of a mixture because the composition
of the mixture to which a community is exposed could be very different from the mixture
that has undergone toxicological testing. Sufficient similarity is a key concept for
evaluation of a complex mixture. It is applied when inadequate toxicity data are
available directly on a mixture of concern, but toxicity data can be acquired on a mixture
composed of similar chemical components in similar proportions. If the two mixtures
are judged to be sufficiently similar, then the toxicity data for the latter can be used as
surrogate data in conducting a quantitative risk assessment for the mixture of concern.
The EPA has proposed this general concept for the evaluation of complex mixtures in
its risk assessment documentation (U.S. EPA, 2000a). The exposure analyst and dose-
response analyst logically would jointly discuss this issue. It can be helpful for the
exposure analyst to consider three broad categories of transfers that can occur between
environmental compartments:
3-25
-------
3500
3000
2500
Global
tritium 2000 i
inventory
(MCi)
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95100
Elapsed time from 1963 (years)
FIGURE 3-2
Illustration of Global Background from Atmospheric Fallout of Tritium
Source: Brown (1999)
3-26
-------
Differential transfer between different abiotic media (e.g., soil and surface water)
Differential transfer between abiotic and biotic media
Differential transfer between different biotic media
Mixture components can be differentially transferred between abiotic media. For
example, DBFs, such as chloroform and bromodichloromethane, are highly volatile;
others, such as monochloroacetic acid, are not (U.S. EPA, 2003b). Consequently, the
composition of a DBP mixture in the indoor air differs considerably from the DBP
mixture in a glass of water. The insecticide toxaphene provides a second example.
Technical grade toxaphene, which contains over 670 chemicals, was one of the most
heavily used insecticides in the U.S. until 1982 when it was canceled for most uses. It
was used primarily in the southern U.S. to control insect pests on cotton and other
crops. Some components of technical toxaphene may volatilize to air; others do not
dissolve well in water. The composition of the toxaphene mixture will differ depending
on whether it is measured in soil at a hazardous waste site, the air around the site or
sediment at the bottom of lakes or streams near the site (ATSDR, 1996).
Mixture components can be differentially transferred between abiotic and biotic
media. For example, the Site-Specific Assessment Procedures volume in the review
draft Exposure and Human Health Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-
Dioxin (TCDD) and Related Compounds (U.S. EPA, 2003c) provides methods for
predicting differential uptake of different dioxin congeners from the atmosphere into
plant tissue and the selective retention of dioxin congeners in fish adipose tissues.
Some components of technical toxaphene have been measured in shellfish and fish
(ATSDR, 1996).
Mixture components can be differentially transferred between biotic media. For
example, the Site-Specific Assessment Procedures volume in the review draft Exposure
and Human Health Reassessment of 2,3,7,8- Tetrachlorodibenzo-p-Dioxin (TCDD) and
Related Compounds (U.S. EPA, 2003c) provides methods for predicting the selective
uptake and retention of different dioxin congeners from grass into the adipose tissues of
grazing cattle.
3.3.2.2. Grouping Chemicals for Cumulative Risk Analysis Mixtures
occurring in a community may originate from different sources. This section provides
six tables that illustrate how information about sources of chemical pollutants, chemical
properties and fate can be organized to guide chemical groupings for cumulative risk
assessments in contaminated communities. These tables provide context regarding the
normal uses of chemicals often found in mixtures and their behavior in the environment
3-27
-------
that leads to their coexistence in media to which people can be exposed. The grouping
of the chemicals could be based on the potential for their co-occurrence in each
compartment/medium, potential for interactions affecting transformation and potential
for co-occurrence and interaction along each transport pathway between media.
Figure 3-3 provides an overview of how this information might be organized according
to media and the processes of fate and transport.
While chemicals can be easily grouped based on common sources and releases
(e.g., chemicals in diesel exhaust), the usefulness of groupings for various chemical
classes can be improved based on typical primary release mechanisms that would be
expected to control initial contamination and migration behavior in the environment as
illustrated in Table 3-2. Released chemicals can disperse quickly over a fairly wide area
by convection (such as via wind or surface water flow), and they can also migrate
following waste placement. The dominant processes at a given location determine what
will be the "receiving medium" into which a particular class of chemicals is introduced
and from which they can migrate.
Contaminant properties relevant to fate and transport include volatility, water
solubility and partition coefficients for
water and available organic phases (as represented by the octanol-water
partition coefficient, Kow);
water and solid phases (soil-water partition coefficient, Kd); and
water and air (Henry's constant, KH).
Additional properties for soil and sediment include the fraction of organic carbon
(foe) and the clay content, which provide an indication of the amounts and types of
sorption sites are available. Table 3-3 can be used to group chemicals per their
expected general partitioning in media based on well-known physical constants for the
chemicals and media. Chemical-specific soil-water partition coefficients in various soil
textures can be displayed to help evaluate possible chemical grouping based on similar
mobility as shown in Figure 3-4.5 The analysts may evaluate the soil type,
geochemistry and other data to determine generally appropriate values and site-specific
studies important to the selection of the actual values for key contaminants. This
concept is illustrated in Text Box 3-7. Table 3-5 gives examples of Kow and solubility
values for selected chemicals to support these types of groupings.
5 Note that the Kd values overlap given the wide range of soils used to develop the figure. Kd values for
specific types of soil or additional data may be needed to implement this grouping step.
3-28
-------
Pollution Source
1
transformation
Transformation
transformation refers to a group of
processes that can act to change the
composition of a mixture.
blntracompartment transport refers to the
processes that move a mixture through an
individual compartment (e.g., turbulence and
wind will move a mixture through the
atmosphere) and intercompartment transport
refers to processes that move a chemical
mixture from one medium to another.
Pollution Source-
Pollution Source,
Receiving Media
*\ lntracompartmentb
Transport
Intercompartment Transport6
Other Media
Concentrations at
Points of Exposure in
Multiple Media
Human Activity Patterns
Exposed
Subpopulations
Toxicokinetics
Target Tissue Doses
FIGURE 3-3
Approach for Estimating Exposure in Cumulative Risk Assessments
3-29
-------
TABLE 3-2
Grouping Chemicals by Common Migration Behavior
Migration Initiation
Process
Volatilization to air
Dissolution in
groundwater
Dissolution in surface
water
Particulate emissions
from combustion
(stacks)
Gaseous emissions
from combustion
(stacks)
Dust-blown migration
Waste placement
Leaching to
groundwater
Organic Chemicals
Chlorinated solvents,
Petroleum-based solvents,
Fuels
Chlorinated solvents,
Aromatic hydrocarbons
(benzene, toluene,
ethylbenzene, xylene),
Pesticides
Phenols, amines, ethers,
alcohols, organic acids
Products of incomplete
combustion (PICs)
- PCBs, PAH, dioxins,
furans
Light hydrocarbons
Nonvolatile organics
- PAHs, PCBs, dioxins
All listed above
Chlorinated solvents
(DNAPLs)
Inorganic Chemicals and
Gases
Cb, ammonia, tritium, 862,
NOX, CO, C02
Cations and Anions
Cations and Anions (e.g.,
perchlorates)
Heavy metals
802, NOX, CO, ammonia
Heavy metals
All listed above
NA
Heavy metals are as indicated in Table 3-1 . Acronyms not previously defined (in
Table 3-1) are CO = carbon monoxide; C02= carbon dioxide; DNAPLs = dense non-
aqueous phase liquids; NOX = nitrogen oxides; and 802 = sulfur dioxide.
3-30
-------
TABLE 3-3
Grouping Chemicals by Environmental Fate Measures3
Environmental
Compartment
Persistence
(environmental half life)
Environmental Partitioning
(equilibrium-based)13
Mobility
(convection- and dispersion-based)0
Organic matter in soil and
sediments,
soil organisms
High for
High Kow/Kd
Low biodegradability
Low for
High Kow/Kd
High biodegradability
Presence favored by
High Kow/Kd
High persistence
High binding for
High-Kow/Kd organics and inorganics
Low binding for
Low-Kow/Kd organics and inorganics
Soil inorganic phase
High for
High-Kd inorganics
Low-Ksp inorganics
(including metals that form
complexes in soil)
Low for
Low Kow/Kd organics/inorganics
Presence favored by
High-Kd and low-Ksp inorganics
High mobility for
Cations, anions, water- soluble organics
(low Kow/Kd)
High-Ksp colloids
Low mobility for
High-Kow/Kd organics
High-Ksp solids
Surface water
Higher for
Insoluble (high Kow)
Non-photodegradable
Non-biodegradable
Lower for
Water soluble (low Kow)
Volatile (low KH)
Photodegradable
Biodegradable
Presence favored by
Low Kow/Kd
High KH
(low volatility to air)
High-Ksp inorganics
High transport for
High solubility
Low volatility
Low transport for
Precipitates (low Ksp)
Low solubility
(high Kow)
Biodegradable
Photodegradable
3-31
-------
TABLE 3-3 cont.
Environmental
Compartment
Persistence
(environmental half life)
Environmental Partitioning
(equilibrium-based)13
Mobility
(convection- and dispersion-based)
Groundwater
Higher for
Low biodegradable
DNAPL-forming
Lower for
Biodegradable
Highly soluble (low Kow/Kd)
LNAPL-forming
Presence favored by
High solubility
(low Kow/Kd)
Ionic forms
(cations and anions)
High-Ksp inorganics
High mobility for
Low Kow/Kd organics and inorganics
Ionic forms
Low mobility for
High Kow/Kd organics and inorganics
Inorganic solids
Air
Higher for
Low photodegradable
Low reaction rate with hydroxyl
radical and other free radicals
Low wash out rate (low KH)
Gas phase
Lower for
Photodegradable
High reaction rates
High wash out (high KH)
Particulate phase
Presence favored by
High volatility substances (gases
and low boiling point liquids)
High volatility from water
(low KH)
High mobility for
Gas phase
High persistence
Small-particle bound
Low mobility for
Low persistence
Large-particle bound
Aquatic and terrestrial
biota
Higher for
Lipid soluble (high Kow)
Non-biodegradable
Low depuration rates
Lower for
Water soluble (low Kow)
High depuration rates due to
enzyme-oxidizable and/or
forms complexes with GHS,
other agents
Presence favored by
High organic solubility (high Kow)
High BCF
Persistence in biota/prey (high
BAF)
Mobility enhanced by
High persistence in biota
High vegetative uptake factors (high Kow),
specific binding factors)
Mobility reduced by
High degradation rates
High elimination rates
Low uptake factors
3-32
-------
aBAF = bioaccumulation factor, BCF = bioconcentration factor, GHS = glutathione, LNAPL = light non-aqueous phase liquid, Kd = soil/water
partition coefficient, KH = Henry's constant (water/air distribution constant), Kow = octanol/water partition coefficient (octanol approximates soil
organic matter, or biomass), Ksp = solubility product constant for inorganic complexes.
b "Presence favored by" indicates that concentrations would be relatively higher compared to adjacent compartments, i.e., activity coefficients for
the substances are relatively low in the given compartment/medium.
c In general, advection is transport by large-scale motions and can be described as the movement of a chemical by virtue of its presence in a
medium that is flowing. Convection describes local transport phenomena and can be described by the flux of a chemical through porous media.
Diffusion is a redistribution (spreading/dilution) of a chemical mass within a phase attributable to molecular (Brownian) motion and tending toward
equilibrium (e.g., movement of a chemical from an area of high concentration to one of lower concentration), which results in the net transport of
a chemical within the liquid, solid or gas phase. Dispersion is net transport (mixing) resulting from differential advection, which can be referred to
as turbulent diffusion causing longitudinal, transverse and vertical spreading.
3-33
-------
6
5
4
logKd
(Ukg) 3
* median
2
1
0
-1
i
i
t
i
-------
Illustration of Groupings Based on Properties and Fate
(Text Box 3-7)
Chemicals can be grouped based on expected persistence or
degradation in various environmental compartments, as a general
indication of potential joint exposures to various contaminated
media.
For example, organic contaminants with high Kow and low volatility
that would be expected to be found together in sediments and in the
lipids offish would include: persistent pesticides aldrin, dieldrin,
chlordane, dichlorodiphenyltrichloroethane (DDT),
dichlorodiphenyldichloroethylene (DDE) and
dichlorodiphenyldichloroethane (ODD), as well as 2,3,7,8-TCDD
and pentachlorophenol.
While benzo(a)pyrene would be expected in sediments, in fish it
may be metabolized and excreted without significant accumulation
in lipids.
Conversely, contaminants with medium-to-low Kow and medium-to-
high solubilitysuch as toluene, trichloroethylene and phenol
would be expected to be found mainly in water phases, while
toluene and trichloroethylene would volatilize appreciably from the
water surface due to their relatively high vapor pressure and low KH.
In contrast, phenol with only a medium vapor pressure and high KH,
would not.
To illustrate how
grouping tables can be
applied to assess multiple
chemicals in different classes
fora cumulative risk
assessment, we offer the
following example for PCBs
(representing a group of
congeners). First, the
properties for PCBs are
discussed, and then other
chemicals and chemical
classes that might be
included in the PCB groups
based on their similar
physical-chemical properties
are identified. The general
grouping information in Table 3-3 can be combined with illustrative parameter
information in Tables 3-4 and 3-5, and from this information, the persistence of PCBs in
soil organic matter would be expected to be high given the high Kow values and low
biodegradability. Also, concentrations would likely be high in soil organic matter
compared to other media such as soil inorganic matter or soil pore water, again
because high Kow values indicate higher partitioning to organic phases. Their mobility
in soil would be controlled by two processes: dissolution in water (e.g., moving laterally
as surface transport or generally downward with percolating water) and retardation due
to sorption onto inorganic soil particles (assuming foe is low for subsurface soils, as the
near-surface soil horizons contain the bulk of organic matter that has not yet been
mineralized).
In this example, groundwater concentrations of PCBs are expected to be very
low based on likely partitioning of PCBs to solids in the soil. If some PCB congeners
could migrate through the soil and reach the groundwater, this would lead to dilute PCB
congener concentrations in this medium. The concentrations reaching groundwater
would likely be very low, perhaps undetectable by usual measurement methods. In
addition, the congener composition would change during transport, in accordance with
the varying solubility and sorption properties of compounds with different levels of
3-35
-------
TABLE 3-4
General Grouping Categories for Key Fate Parameters3
Parameter13
Partition coefficient Kow
Solubility product Ksp
Water solubility Sw (ppm)
Henry's constant KH (mol/L*atm)
Vapor pressure VP (mm Hg)
Melting point MP (°C)
Boiling point BP (°C)
General Categories and Examples
Low
<100
<1 xlO'50
<10
<0.01 to 1
<0.001
<0
<50
Medium
100-10,000
1 x 1Q-10to1 xlO'50
10-1000
1-000
0.001-1
0-100
50-300
High
>1 0,000
>1 x 10'10
>1000
>1000
>1
>100
>300
a General ranges indicated in this table illustrate the principles outlined in Table 3-3;
other general bounds would also be appropriate. For example, a Ksp of 10~5 could be
used as a delineator for "readily soluble" for one-molar electrolyte solutions, while
formal water solubilities <0.003 mole/liter could indicate the compound is "not readily
soluble."
b Kow is the partition constant between water and octanol, which represents a generic
"organic" phase; this coefficient applies mainly to organic chemicals (those containing
carbon). Ksp is the solubility product of inorganic compounds, which describes the
equilibrium between the (excess) solid form and dissolved (or solvated) ions and is
used to determine if a solid is readily soluble in water. The Ksp is a function of the
water solubility, Sw. KH is the distribution constant for a chemical between air and
water phases, based on the partial pressure of the gas above the solution to its
dissolved concentration; the extent to which a given gas dissolves in solution (here,
water) is proportional to its pressure (Henry's law), and KH is the proportionality
constant for this relationship. VP is the pressure exerted by a vapor in equilibrium with
its solid or liquid phase, typically used for a vapor in contact with its liquid (so it would
represent the vapor-phase pressure of the pure liquid). The melting point (MP) and
boiling point (BP), the melting and boiling point s, are simple physical constants; they
are used here to help guide the grouping of organic chemicals.
3-36
-------
TABLE 3-5
Specific Parameter Values for Example Chemicals3
Chemical"
Toluene
Trichloroethylene
Phenol
Benzo(a)pyrene
PCBs
Dioxin
(2,3,7,8-TCDD)
Pentachlorophenol
Atrazine
Mercury (Hg)
Mercury sulfide
(HgS)
Lead chloride
(PbCI2)
Kow
(unitless)
540
260
29
1,300,000
12,600,000
6,300,000
132,000
410
4.2
NA
NA
KH
(mol/L*
atm)
0.15
0.1
3000
2200
2.4
20
40,800
420,000
0.12
NA
NA
Ksp
(unitless)
NA
NA
NA
NA
NA
NA
NA
NA
NA
1.6 x10'52
1.6x 10'5
Sw
(ppm)
526
1280
83,000
0.001
0.7
0.0002
14
35
0.06
2x 10'21
3,300
BP
(°C)
111
87.2
182
311
NA
NA
309
NA
357
NA
NA
VP
(mm Hg)
28
69
0.35
5x 10~9
0.0005
1.5x 10'9
0.0001
3x 10~7
0.002
NA
NA
MP
(°C)
-95
-84.7
40.9
176.5
NA
305
174
173
-39
NA
NA
a Parameters are defined in Table 3-4. NA = not applicable. Representative values shown here
are taken from multiple sources and are offered simply for illustration; to calculate
environmental behavior for a specific case, setting-specific information may be used to
determine the appropriate value for a given parameter.
b Chemicals were selected to represent a wide range of physical properties, applications and
sources. Values for dioxin are for the tetrachlorodibenzodioxin isomer generally regarded as
most toxic.
3-37
-------
chlorination (e.g., more highly chlorinated compounds are less soluble). Additional data
show that PCBs degrade slowly in soils (ATSDR, 2000c).
Moving down Table 3-3, one would predict that while PCB concentrations would
be low to intermediate in soil inorganic phases and very low in surface water and
groundwater, some volatilization to air might occur for low-chlorinated congeners as
indicated by their relatively low boiling points and appreciable vapor pressures. Some
volatilization from water would be expected based on the relatively low KH values of
PCBs. Migration through air might be possible via adsorption to particulate matter, and
rain washout would depend on the relative fraction of PCBs in the vapor phase versus
the particulate phase as well as the partitioning between air and rain water as indicated
by Henry's constant. (This constant defines the wet removal process for soluble gases;
the effective Henry's constant is used to predict dry deposition velocity for gases and
particles, in a calculation that also includes molecular weight and surface reactivity and
diffusivity ratios.)
Further, expected levels of PCBs in aquatic and terrestrial biota (i.e., via food
web transfers) might be high relative to surrounding media (water or inorganic soil), and
these levels would be expected to persist due to high lipid solubility (high Kow) and low
biodegradability. Finally, given their persistence in fatty tissues, these levels are
expected to accumulate in the food chain; apex predators would likely have the highest
concentrations.
The analysts can then explore grouping PCBs with other chemicals by applying
concepts presented in Table 3-3 using Tables 3-4, 3-5 and 3-6. As seen from
Table 3-6, PCBs in soil organic matter could be grouped with other persistent organics
such as PAHs (see Table 3-5 for details on benzo(a)pyrene), dioxins and atrazine. The
general grouping scheme in Table 3-4 is based on relative ranges of values for a
number of important physical constants that determine the behavior of chemicals in the
environment (including constants identified in Table 3-3). These ranges have been
drawn from information on a wide variety of chemicals in order to illustrate an approach
that can be used to group chemicals. Physical properties are given for several
chemicals in Table 3-5; these example chemicals were selected to illustrate a wide
range of values for the parameters discussed above.
Groups of chemicals that might be expected to be distributed to various
environmental compartments (or media) as described above are illustrated in Table 3-6.
The implicit assumption in these examples is that sufficient time has passed for
transport and system equilibration to occur. In some cases, such as deposition in
aquatic sediments or transport through the food chain, this process can
3-38
-------
TABLE 3-6
Summary Comparison and Screening Suggestions
Media/Compartments
Soil organic phase
(upper soil horizon)
Soil inorganic phase
(lower horizons)
Aquatic sediments
Surface water
Groundwater
Air
Aquatic biota
Terrestrial biota
Suggested Chemical Grouping (for contaminated sites, overtime)
Low volatility, high Kow, persistent organics
PCBs, dioxins, PAHs; moderately persistent atrazine
High Kd inorganics
Metal oxides, hydroxides, carbonates
High Kow organics, low Ksp inorganics
PCBs, chlorinated pesticides, dioxins, insoluble metal complexes
High water-soluble organics, high Ksp inorganics
Phenols, ethers, esters, nitro- and amino-organics, soluble metal
complexes
Medium Kow, medium volatility, medium water-soluble persistent and
dense organics, medium to high water-soluble, medium to low Kd
inorganic complexes and free ions
TCE, vinyl chloride, BTEX, ethers (e.g., methyl-tert-butyl ether,
MTBE), phenols, atrazine, soluble metal complexes, colloidal metals
Volatile organics, particle-associated organics and inorganics
Chlorinated solvents, light hydrocarbons, freons, BTEX and particle-
bound PCBs, dioxins and metals
High Kow, persistent organics
PCBs, chlorinated pesticides, PAHs, methyl mercury
High Kow, persistent organics, bioaccumulated metals and
radionuclides
PCBs, DDT, mercury, lead, radium
*This table illustrates groups of chemical contaminants that may be expected to persist or be
subject to degradation in various environmental compartments or phases, sometimes referred to
as chemical sinks. These groups are based on sampling and analysis experience and are
simply intended as a general indication of chemicals that may be combined in a cumulative risk
assessment of exposures to a particular medium. (Note that chemicals are not limited to a
single compartment.)
3-39
-------
take from months to years following an initial release of contaminants. By the same
token, after an extended time, chemicals from a variety of different sources would be
expected to ultimately reach similar environmental sinks. In some cumulative risk
assessments, it may be important to examine when these chemical movements
would occur.
This concept is illustrated for an example release scenario (industrial spill) in
Table 3-7. This concept applies to any environmental release, so other scenarios can
also be considered, such as combustor emissions related to routine operations or
temporary releases (e.g., due to excursions from a continuous-operation facility or
discrete releases from a mobile facility). An approach for addressing that type of
situation is illustrated in Figure 4-8.
An example that illustrates how available information can be evaluated to
determine what release processes and receiving media are most significant,
considering past, current and possible future releases, is offered in Text Box 3-8
(U.S. EPA, 2004c). Note that both the transfer of contaminants from one medium to
another and environmental transformation are considered as part of the fate and
transport evaluation.
In this example, in order
to identify the most significant
sources leading to air
contamination, the exposure
analyst would consider
information such as chemical
form, physical-chemical
properties (such as volatility),
transformation, partitioning and
mobility, persistence and bio-
uptake (including combined
environmental fate and co-
location). The exposure analyst
would not conduct a quantitative
fate and transport analysis until
later in the process (see Section
3.3.2.3); the intent at this point is to identify what media are receiving chemicals from
the identified source (or sources). A number of tools and databases exist to support the
Example of Possible Release Sources (Text Box 3-8)
To assess cumulative hazards of urban air toxics in the
Chicago area, focusing on multiple releases to air was
determined to be most useful. Most source release data
identified in an environmental loadings profile were for point
releases; some data for area and mobile sources of air
pollution were also available. Although data on discharges to
surface waters could have been obtained, the potential for
exposure through this source was considered more limited
than for exposure through source releases to air. Similarly,
because the source of tap water for much of the Chicago area
is Lake Michigan, very limited (if any) exposure to
groundwater exists via the drinking-water pathway. Finally, if
a chemical spill occurred, cleanup was assumed to be
relatively quick (following environmental regulations) when
compared to other sources of exposure, so the potential for
exposure to soil contaminated from a recent spill was
considered very low.
One study finding was that relatively few point sources
account for a high percentage of point-source hazards,
suggesting that such sources provide a logical starting point
for hazard management actions. In summary, focusing on
suspected predominant sources can reduce the complexity
and cost of the initial exposure assessments.
3-40
-------
TABLE 3-7
Example Groupings Based on Exposure Considerations (Media and Timing)*
Exposure
Duration
Environmental Medium
Transport/Removal Process
Soil upper horizon - volatilization and leaching
from surface, biodegradation
Air-
volatilization from soil
Surface water (river) -
overland flow and particle transport
from surface soil
Aquatic sediments - precipitation from water,
adsorption on particles, deposition
Soil lower horizons - leaching from surface soil,
adsorption, biodegradation
Groundwater-
leaching from soil
Release Scenario
Industrial Spill on Soil near a River
(VOCs, SVOCs and Metals)
Acute to Short-Term
-------
* Projected intervals reflect physical-chemical properties and fate data, including half-lives; other factors also affect partitioning and timing,
including local conditions such as temperature (for volatilization); organic content (for soil and sediment sorption), which for this example is
assumed to be relatively low; and depth to aquifer (for leaching to groundwater), which is assumed to be moderate to deep.
CCI4 = carbon tetrachloride; DCA = 1,1-dichloroethane; DCE = 1,1-dichloroethylene; SVOCs = semivolatile organic compounds;
TCE = trichloroethylene; VC = vinyl chloride.
3-42
-------
evaluation of contaminant fate and transport. Selected highlights are offered in the
cumulative risk toolbox in Appendix A.
For a given set of chemicals, only one medium might be contaminated under
current conditions (e.g., site soil), but different media could be affected over time, e.g.,
as contaminants migrate to groundwater or surface water or are taken up in food
products. Thus, other time-related considerations include differential travel times for
multiple contaminants (e.g., migrating to groundwater) and for subsequent transport to
an exposure point. In addition, interactions could influence the mobility of multiple
chemicals present together, or interactions could occur among transformation products
that are formed over time. These concepts of migration and transformation are
illustrated by the differential toxicity of the degradation products of TCE, notably
1,2-dichloroethylene and vinyl chloride, as was described in Section 3.3.2.1 and as
shown in Table 3-1. The concept of migration is illustrated by an example in
Figures 3-5 and 3-6, which shows that while the exposure profile changes in the
temporal scale, so can the toxicity profile. For example, in a chlorinated plume, the
parent compound, tetrachloroethylene, degrading through TCE to vinyl chloride (Vogel
and McCarty, 1985) could actually pose greater health risk later (as the plume
contaminants gradually degrade) both in groundwater and via the passive (indoor air)
inhalation pathway as the more volatile vinyl chloride preferentially passes through the
vadose zone and could become trapped closer to the receptors at the land surface.
Cumulative risk assessments may also evaluate combined sources and joint
environmental fate and transport. Although some traditional assessments do consider
multiple sources and multiple contaminants, differential partitioning into environmental
media over time is often overlooked:
Dioxin congeners can partition differently between soil and vegetation
Site-specific soil characteristics will determine the extent of volatilization for
volatile organic compounds
The extent of vegetative cover determines soil runoff into surface water
Weathering can change the composition of an original contaminant mixture
The composition of spilled oil has changed over time, as has that of the
toxaphene mixture described in Text Box 3-9 (U.S. EPA, 1997d). Methods to account
for differential partitioning continue to evolve. For example, the EPA soil screening
guidance considers the potential for individual soil contaminants to migrate to
groundwater, based on a simple soil screening-level
3-43
-------
Indoor Air
Ambient air:
can or
nonpoint to t
air
Vadose Zone -
Soil Gas
Soil J
Contamination [_
(residual or
mobile NAPL)
Chemical Vapor Migration
Groundwater
Contamination
profile over
NAPL = non-aqueous phase liquid
T (in minutes)
First-Order GW a
Decay Constant b
PCE
TCE
VCC
T0 Concentration
Soil GW
<0.1-110(avg4) 100
<0.1-90(avg1) 30
O.2-20 (avg 0.6) 0.5
5.2
6.7
2
(ppm)
IA
T1 Concentration
Soil GW
ND 10
ND 3
ND 0.05
0.1
2.5
1.1
(ppm)
IA
T10 Concentration
Soil GW
4 ND
2.4 ND
3.1 ND
ND
0.0003
0.005
(ppm)
IA
3.5
0.5
ND
a Abbreviations as follows: avg=average. GW= ground water. IA =indoorair. ND = not detectable.
PCE = perchloroethylene (tetrachloroethene). ppm = parts per million.! =time. TCE =trichloroethylene. VC = vinyl chloride.
b U.S. EPA 1998. Technical Pro to col for Evaluating Natural Attenuation of Chlorinated Solvents in Ground Water.
Office of Research and Development, Washington DC. EPA/600/R-98/128. September.
c Assuming natural attenuation and degradation are occuring all the way through ethane, excess VC is not generated, as shown
here. However, if incomplete degradation occurs, VC may accumulate, and the reductions shown here may not occur.
FIGURE 3-5
Example Changes in Exposure Profile from Degradation and Partitioning
3-44
-------
Soil PCE source started at 100
ppm at Time0soil TCE source
started at 30 ppm at Time
Soil Change in Concentration
(Exposure) Over Time
E
Q.
Q.
-§
-fc
CD
o
c
o
o
E 5
Q.
.2 4
15
"£ 0
CD O
O
O
O o
0 L"
4567
jnmeJVears)
10
Groundwater Change in Concentration (Exposure)
Over Time
PCE
TCE
VC-Degrading
VC-Stalled
23456
Time (years)
8
10
Indoor Air Change in Concentration
PCE
TCE
VC-Degrading
VC-Stalled
0123456
Time (years)
8 9 10
FIGURE 3-6
Illustration of Changing Media Concentrations Affecting Potential Exposures
3-45
-------
partitioning equation and the use of either of two dilution attenuation factors (U.S. EPA,
1996a). This simple partitioning approach could be used for screening multiple
contaminants to support grouping for a cumulative risk assessment.
Weathering Example: Toxaphene (Text Box 3-9)
Until the 1970's, toxaphene was the most heavily used
pesticide in the U.S. It was formulated using multiple
ingredients, and their relative amounts change after the
pesticide is released because of differential partitioning and
transformation processes in air, water and soil. (The soil
half-life can be 1 to 14 years.) Overtime these
components continue to change, so the composition of
weathered toxaphene differs significantly from the original
mixture. Samples collected from different sources might
also differ, depending on the location-specific
environmental processes to which the original mixtures
were exposed. For example, weathered toxaphene in an
anaerobic soil does not resemble that in an aerobic soil,
and that in an air sample from the Arctic does not resemble
residues found in the blubber of an Arctic seal. Some
components of this environmental mixture might not be
routinely identified through standard analyses. Site-specific
partitioning and transformation processes can then be
considered to properly assess what compounds could be
present at a given time. It is also important to link this
information with the toxicity evaluation, because weathered
compounds will also exhibit different toxicities from the
original mixture components.
3.3.2.3. Exposure Points
and RoutesThe next phase of
the exposure assessment involves
identifying who is likely to be
exposed to chemical pollutants,
where and by what route(s) of
exposure. The exposure points
(the geographic locations where
people could encounter the
chemicals) and exposure routes
are identified for each exposure
pathway and then integrated for
the cumulative risk assessment.
The analyst also may consider
interactions that might enhance
exposures or associated effects
and evaluate when these exposures may occur.
Non-chemical factors can change exposures and potentially influence the
toxicokinetics (e.g., rate of disposition to a target tissue). Higher breathing rates for
joggers running near an emission source is an example of an exposure factor that
influences exposure. Higher breathing rates could result in an increased rate at which
the jogger inhales airborne chemicals. Co-exposure to toluene and noise offers an
example of synergism because this organic compound damages the auditory system
and can also potentiate additional damage by noise, a physical stressor, beyond what
would be expected by the two acting separately (U.S. EPA, 2003e).
At this point of the assessment, the analyst integrates available information to
link the sources of multiple chemicals, their releases and fate/transport, the exposure
points for likely receptors and the exposure routes (U.S. EPA, 1989a). The focus is on
exposure pathways that are currently complete or are likely to become complete. Thus,
at this point the analyst may consider relevant time frames of these exposures in order
to examine the frequency, duration, intensity and possible overlaps of exposures to
multiple chemicals as well as the sequence of those exposures. The exposure analyst
3-46
-------
would consult with the dose-response analyst so that together they can determine the
level of detail needed in the exposure assessment with respect to exposure overlaps.
The dose-response analyst can provide information to determine whether the overlap of
exposures co-occurring on the same day within a week, a month or a year matters
toxicologically.
Information on background exposure levels to common environmental
contaminants can be important to cumulative risk assessments. A key resource for this
information is available through the National Human Exposure Assessment Survey
(NHEXAS) program (U.S. EPA, 2004d). That program was designed to address some
of the limitations of single-chemical and single-media exposure studies as one of its
goals is to test and evaluate different techniques and design approaches for performing
multimedia multipathway human exposure studies. An analyst could use the NHEXAS
data as baseline information for exposure assessments to indicate if specific
populations are exposed to increased levels of environmental contaminants. These
data are available in the Human Exposure Database System, which contains chemical
measurements, questionnaire responses, documents and other information related to
U.S. EPA studies of human exposures to environmental contaminants (see
Appendix A).
To evaluate what chemicals might coexist at places where individuals are-or
could be-exposed, the analyst can group site-related contaminants by considering
when they might coexist in space and time. This grouping could reflect transport and
fate considerations, including transformation, that are appropriate for the time intervals
studied. Minimally, four groups are defined to guide this evaluation of possible
exposures to multiple chemicals in various environmental media over time as shown in
Text Box 3-10. Clearly, for analyses that evaluate multiple
chemicals, there can be multiple media and multiple time
points to evaluate. Assuming that these chemicals co-
occur in media that individuals in the community may
contact, the analyst could then link these exposure
groupings with toxicity information in order to assess joint
impacts as described in Chapter 4. The analyst could
evaluate these as potential doses. (In refined cumulative
exposure assessments, toxicokinetic and toxicodynamic information could be used to
provide a comprehensive understanding of the magnitude of tissue doses over time
[see Sections 3.3.3 and 3.3.4]).
3-47
Chemical Groupings by
Coexistence in Media/Time
(Text Box 3-10)
Media
Time Same Different
Same Group 1 Group 3
Different Group 2 Group 4
-------
The EPA identifies several Examp|es of Chemjca| Rajrs |nf|uenced by
time-course issues in the
Framework for Cumulative Risk Benzo[a]pyrene (BaP) and tris(2-ethylhexyl) phosphate
document (U.S. EPA, 2003a).
demonstrate different toxicity
depending on the sequence of
exposures, with cancer initiators
Exposure Timing (Text Box 3-11)
(TPA) are an initiator/promoter pair
TPA does not have a tumorigenic effect in mouse skin
" assays, but applying it after initiation with BaP greatly
enhanced tumorigenic activity (Verma et al., 1985).
Cadmium and Lead illustrate antagonism
Initial exposure to Cd has been shown to decrease the
and nromot<=rs b<=inn th<= rlassir absorption of Pb following subsequent exposure, which has
and promoters Demg the classic the effect Qf decreasing the b,ood Pb |eve, and causing ,ess.
than-additive hematopoetic toxicity (other data suggest
different joint toxicity, as affected by the order of exposure,
from ATSDR, 2004).
example; exposure to a promoter
has no effect if it occurs prior to
exposure to an initiator. This
illustrates the same media/different time and different media/different time concepts
indicated above. Text Box 3-11 shows examples of chemical pairs for which the
toxicological effect is influenced by exposure timing. Specific joint toxicity issues are
discussed in Chapter 4. Several commercial exposure models have been developed to
capture the time aspects of exposures, and Appendix A lists several of these.
3.3.3. Exposure Quantification. Outputs of fate and transport models, such as from
air dispersion modeling, can be used to define the temporal and spatial distribution of
chemicals needed to quantify human exposures. When monitoring data are available,
estimates of exposure could primarily be based on those measures of contaminant
concentrations in the environment, as indicated by the type and quality of the data.
Cumulative exposures to a given population could be estimated for various
exposure pathways and for contaminants of interest to the community. For this
assessment, as many of the following data as are applicable are used to determine
cumulative exposures to a given population:
Body burdens (e.g., concentrations of lead in blood)
Measured concentrations in air, groundwater, surface water, soil, sediments and
food
Modeled concentrations in the ambient environment (not linked to sources)
Prior exposures could also be considered if data are available.
Such a total exposure approach could result in certain sources being essentially
unidentifiable and might include non-industrial contaminant sources, such as consumer
products, environmental tobacco smoke, radon and pesticide residues on foods.
However, the end result could be comprehensive exposure estimates for the population,
3-48
-------
which would include environmental contaminants that are showing up in the monitoring
data. Some stakeholders might desire such an assessment, but such a comprehensive
exposure assessment would typically be beyond the scope of a contaminated site
assessment project. The assessment may highlight the need for an evaluation of
unknown sources of contaminants. The exposure analyst can use information offered in
this report and many other resources to support such complementary analyses by other
groups as desired.
3.3.3.1. Exposure Point ConcentrationsThe concentrations of chemicals to
which people areor could beexposed over the time period of interest can be
represented by a combination of monitoring data and transport and fate models. As
was discussed in earlier sections, using models is the only way to estimate future
contaminant concentrations. Models are used to fill gaps in data for current conditions.
Models can be applied at different levels during a cumulative risk analysis,
beginning with a simple screen to winnow down the list of chemicals of concern and
exposure pathways by eliminating those clearly not expected to contribute to adverse
effects. Using known (not missing) information, this screen reduces the list of chemicals
included in a more detailed analysis, thus facilitating a more focused evaluation.
Exposure analysts can use simple fugacity models to predict movement and phase
change in the environment, for example, to identify which chemicals volatilize, stay soil
bound or lodge in fat of fish or other food species. Environmental breakdown products
also could be identified as indicated by the data or acknowledged as potentially present
where those data do not exist. If resources are available, rare events that might result
in different combinations of chemicals being released to the environment at higher
levels may be considered. When describing the exposures that result from such events,
the analyst may wish to describe the likelihood of such an event occurring.
The next step could be ranking mixtures by defining the chemical and exposure
combinations of main concern and those mixtures that are unlikely to pose a problem.
Exposures to the population of concern could be quantified assuming steady state, also
indicating expected departures from steady state conditions. If needed, a final iteration
would involve applying more detailed dynamic fate and transport models to predict time-
varying concentrations in each media, also including spatial changes in exposure
concentrations.
For more precision, this kind of exposure modeling over time could consider
physiological factors as indicators of likely overlap of internal doses and of possible
damping of external exposure fluctuations (internal overlaps are discussed in Section
3-49
-------
3.3.3.4). Quantitative estimates of exposure would then be determined over these
different time periods. Selected exposure models that can be used to support these
exposure analyses are included in the cumulative risk toolbox in Appendix A.
3.3.3.2. Intake Estimates Using measured and predicted estimates of the
concentrations of multiple chemicals at each exposure point of interest, the exposure
analyst could then apply exposure factors relevant to each receptor and then calculate
pathway-specific intakes. These intakes are calculated using equations that generally
include intake variables for media concentrations (over time), the contact rate, exposure
frequency, exposure duration, body weight and exposure averaging time, as indicated in
the basic EPA guidance (U.S. EPA, 1989a). The Exposure Factors Handbook (U.S.
EPA, 1997c) identifies specific intake rates for air, water and foods. These equations
are then adapted to the specific exposure route: oral, inhalation, or dermal.
The general intake equation is 0
Intake (mg/kg-day) = CxlRxEFxED (3.^
BWxAT
where:
C = concentration (i.e., exposure point concentration) (e.g., mg/L for water)
IR = intake rate (e.g., L/day for water)
EF = exposure frequency (days/year)
ED = exposure duration (years)
BW = body weight (kg)
AT = averaging time (period over which exposure is averaged, in days).
If available, the exposure analyst can use individual or community-specific
exposure factors when estimating intakes, but generic default values are typically used
in conservative screening-level analyses. The cumulative risk across all chemicals,
media and exposure routes will be estimated from these combined calculations linked
with toxicity data. For example, rare events that might result in different combinations of
chemicals could yield different exposure point concentrations that would not normally be
evaluated but would be included in the exposure assessment.
To illustrate the evaluation of multiple pathways and degradation products, we
develop and present a hypothetical example depicting current and future land use at a
fictitious contaminated site. Table 3-8 shows that at this site the receptors under current
conditions are assumed to be an on-site maintenance worker and off-site resident.
Exposure route-specific chemical intakes are illustrative only. Table 3-9 shows that at
this site the receptors under future conditions are assumed to be an on-site resident and
3-50
-------
TABLE 3-8
Example of Cumulative Exposures for Current Land Use*
Chemicals/
Transformation Products
Exposure Medium
and Location
Chemical Intakes
(mg/kg-day)
Ingestion
Inhalation
Dermal
On-Site Maintenance Worker
Tetrachloroethylene
Chlorine
Trichloroethane
Vinyl chloride
Benzo(a)pyrene
Anthracene
PCBs (as Aroclor 1254)
Aldrin
Dieldrin
Site soils
Ambient air
Ambient air
Site soils
Ambient air
Ambient air
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
2x10'5
4x10"8
8x10"4
1 x10'6
2x10'7
9x10-10
2x10'5
7x10"7
2x10'3
5x10'5
1 x10'6
2x10'6
5 x 1 0'6
7 x 1 0'7
3 x 1 0'9
6x10-10
2 x 1 0'5
6 x 1 0'8
6 x 1 0"6
1 x 1 0'5
4 x 1 0"7
5x10'7
8x10-10
7x10"6
2x10"8
4x10'9
3x10"11
2x10'7
5x10-9
4x10'5
4x10'7
1 x10'8
4x10-10
3-51
-------
TABLE 3-8 cont.
Chemicals/
Transformation Products
Arsenic
Chromium
Lead
Mercury
Exposure Medium
and Location
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Chemical Intakes
(mg/kg-day)
Ingestion
8x10"6
5x10"7
8x10"7
7x10-9
3x10'6
9x10"9
4x10'5
6x10"7
Inhalation
3 x 1 0'7
5 x 1 0'8
2 x 1 0'7
8x10'6
Dermal
2x10"8
9x10-10
2x10"9
3x10"11
8x10'8
1x10-10
3x10'7
2x10"9
Off-Site Resident
Tetrachloroethylene
Chloroform
Chlorine
Trichloroethane
Vinyl chloride
Aquifer - tap water
Vapors from shower
Aquifer - tap water
Vapors from shower
Aquifer - tap water
Vapors from shower
Vapors from shower
1 x 1 0"5
9 x 1 0"6
7 x 1 0'8
6x10'8
5 x 1 0"7
4x10-9
9x10-10
2x10"7
3x10"7
2x10-10
* The example scenarios assume exposures at the site under current conditions, e.g., degradation
products are identified for chemicals that undergo conversion on the order of hours or days. The source
release is assumed to be a spill to surface soils with subsequent leaching to subsurface soils and
groundwater. The exposure media are site soils at or beneath the spill location, ambient air from
resuspended particulate matter, surface soils from deposition of resuspended particulate matter, in
groundwater at the tap, and water vapors from showering. Estimates will depend on the default and/or
site-specific exposure factors used in the intake equations.
3-52
-------
an off site visitor. In the current exposure scenario, exposures are analyzed following a
chemical spill to surface soils. The spill subsequently leaches to subsurface soils and
groundwater. The exposure media are site soils at or beneath the spill location,
ambient air from resuspended particulate matter, surface soils from deposition of
resuspended particulate matter, in groundwater at the tap, and water vapors from
showering. Exposure estimates will depend on the default and/or site-specific exposure
factors used in the intake equations. To account for changes over time, cumulative
intakes are calculated for exposures to original chemicals as well as to degradation
products that can result from relatively rapid chemical reactions in the environment.
Intakes for ingestion, inhalation and/or dermal contact are calculated for applicable
media and are then used to calculate cumulative risk estimates in the Risk
Characterization phase.
For a future land use scenario at the hypothetical contaminated site, an exposure
analyst might identify two receptors: on-site residents and off-site recreational visitors.
As presented in Table 3-9, exposures occur by several pathways that reflect a longer
time frame (e.g., 20 years) than the current scenario. To account for changes over
time, cumulative intakes are calculated for exposure to chemicals plus conversion
products that result from relatively slow degradation (on the order of months or years).
Volatile organics in surface or near-surface soils are assumed to have dissipated so are
not considered in future exposure assessments. Intakes for the exposure routes of
ingestion, inhalation and/or dermal contact are calculated for applicable media and are
then used to calculate cumulative risk estimates in the Risk Characterization phase.
3.3.3.3. Calendar Approach While no EPA-wide standardized procedure
exists for detailed consideration of exposure timing in dose/response assessment, the
Office of Pesticide Policy provides an approach, identified as the calendar approach, in
General Principles for Performing Aggregate Exposure and Risk Assessments (U.S.
EPA, 2001 a). Figure 3-7 provides an overview of the steps entailed in this approach.
The calendar approach estimates sequential, daily chemical exposures by linking
episodic exposures (e.g., seasonal exposures to pesticides through surface water
contact following residential lawn applications of pesticides in the spring and summer)
with routine exposures (e.g., contaminants in the food supply). Figure 3-8 illustrates a
hypothetical pattern of results that could be predicted using such an approach. The
discussion that follows adapts this approach, which covers aggregate exposures, to
cumulative exposure practices. This discussion focuses on Steps 1-6, followed by
additional information about the calendar approach.
3-53
-------
TABLE 3-9
Example of Cumulative Exposures for Future Land Use*
Chemicals/
Transformation Products
Exposure Medium
and Location
Chemical Intakes (mg/kg-day)
Ingestion
Inhalation
Dermal
On-Site Resident
Benzo(a)pyrene
Anthracene
PCBs(asAroclor1254)
Dieldrin
Arsenic
Chromium
Lead
Mercury
Benzo(a)pyrene
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Site soils
Ambient air
Surface soils
Surface runoff to lake
3x10'4
1 x10'6
2x10'3
8x10'6
2x10'6
4x10'8
1 x10'6
3x10'8
9x10'3
2x10'6
5x10'3
2x10'5
8x10'3
9x10'5
1 x10'6
2x10'7
1 x10'8
2x10'5
6x10'5
5x10'5
9x10'6
1 x10'5
7x10'4
4x10'4
6x10'6
2 x 1 0'8
2x10-10
5 x 1 0'4
2 x 1 0'7
6 x 1 0'7
2 x 1 0'9
2 x 1 0'8
2x10-10
7 x 1 0'7
6 x 1 0'9
2 x 1 0'5
8 x 1 0'7
3 x 1 0'7
2 x 1 0'9
5 x 1 0'8
5x10-10
2x10'11
3-54
-------
TABLE 3-9 cont.
Chemicals/
Transformation Products
Exposure Medium
and Location
Chemical Intakes (mg/kg-day)
Ingestion
Inhalation
Dermal
Off-Site Recreational Visitor
Anthracene
PCBs(asAroclor1254)
Dieldrin
Arsenic
Chromium
Lead
Mercury
Methyl mercury
Surface runoff to lake
Surface runoff to lake
Fish in lake
Surface runoff to lake
Surface runoff to lake
Surface runoff to lake
Surface runoff to lake
Surface runoff to lake
Fish in lake
4x10'7
9x10'9
5x10'6
2x10'9
3x10"7
8x10"8
1 x10"7
2x10"8
3x10"5
1x10-10
4x10'12
8x10'12
6x10-10
2x10"11
7x10-10
5x10"11
* These example scenarios assume exposures at the site under future conditions, e.g., degradation
products are identified for chemicals that undergo conversion on the order of months or years. In
addition, TCE and PCE in surface soils are assumed to have completely volatilized by the time the future
land use scenario begins, with aldrin having been converted fairly rapidly to dieldrin. The source release
is assumed to be a spill to surface soils with subsequent leaching to subsurface soils and groundwater.
The exposure media are site soils at and beneath the spill location, ambient air from resuspended
particulate matter, surface soils from deposition of resuspended particulate matter, surface water and lake
fish. Estimates will depend on the default and/or site-specific exposure factors used in the intake
equations.
3-55
-------
Step 1 Identify toxicological
endpoints for each potential
exposure route and duration
Step 2 Identify potential
exposures for each pathway
Step 3 Reconcile routes and durations of
potential exposures with routes and durations
of health effects
Step 4 Determine co-occurrence of chemicals
through different exposure scenarios
Step 5 Determine magnitude, frequency and
duration for all pertinent exposure
pathway/route combinations
Step 6 Determine data treatment method
Step 7 Assign route-specific risk metrics
Step 8 Conduct Aggregate Exposure and
Risk Assessment
Step 9 Conduct sensitivity analysis
Step 10 Aggregate Exposure and
Risk Characterization
FIGURE 3-7
Ten Steps in Performing Aggregate Exposure and Risk Assessment
(Adapted from U.S. EPA, 2001 a)
3-56
-------
A. Food Exposure
Acute toxicity
endpoint
Short-term toxicity
endpoint
Time (Days)
Drinking Water Exposure
Acute toxicity
endpoint
Short-term toxicity
endpoint
Time (Days)
C. Residential Exposure
Acute toxicity
endpoint
Time (Days)
Time (Days)
FIGURE 3-8
Pathway-Specific and Combined Exposure to a Single Hypothetical Chemical
3-57
-------
The dose-response analyst and the exposure analyst work together on the first
and third steps. The goal of these particular steps is to identify the health effect(s)
associated with each chemical or group of chemicals identified. Health effect(s)
identification includes an analysis of which exposure route(s) and exposure duration(s)
produced the effect(s) (Step 1) and a step to ensure that the dose-response
assessment and the exposure assessment are concordant (Step 3). A previous
document (U.S. EPA, 1999e) describes five general durations of exposure to be
considered:
Acute - in a cumulative risk assessment this could include one-day exposures
through oral (food and water pathways, which reflect distribution of daily food
consumption and daily water residue values), inhalation (atmospheric
concentrations) and dermal routes, which reflect daily water and soil residue
values)
Short-term - could include 1-30-day exposure scenarios
Intermediate-term - could include 30-180-day exposure scenarios
Chronic/long-term - could include exposures of greater than 6 months in duration
Cancer - lifetime assessment
Following the identification of the toxicological endpoint(s), duration of exposure(s),
exposure scenario(s) of concern, Step 4 requires the analyst to examine residential
exposures that might occur to potential receptors (e.g., home pesticide or herbicide)
(U.S. EPA, 2001 a). The exposure analyst accomplishes this by appropriately
combining information about a potentially exposed individual's demographic (e.g., age,
gender and racial/ethnic background), temporal (season) and spatial (region of the
country) characteristics.
A cumulative exposure assessment could involve the same steps: combining
national data to estimate background exposures with site-specific data to estimate local
exposures. This point is illustrated using a single chemical exposure. Methylmercury
exposures can result from consumption of local-caught fish and commercial fish (i.e.,
two different sources offish). An analysis could examine the correlation between
consumption rates of local-caught and commercial fish and use both average local fish
methylmercury levels and average commercial fish methylmercury levels to estimate
methylmercury exposures in individuals consuming a mix of these fish. Such an
analysis could also capture seasonal consumption patterns (and associated exposure
patterns) offish (e.g., in some areas of the U.S. there could be a decrease in local-
caught fish consumption during winter). Furthermore, U.S. EPA (1999e) indicates that
distributional data analysis (as opposed to a point estimate approach) is preferred
3-58
-------
because this tool allows an aggregate exposure analyst to more fully evaluate exposure
and resulting risk across the entire population rather than the exposure of a single, high-
end individual.
Steps 5 and 6 integrate the magnitude, frequency and duration of exposure for all
relevant pathway and route combinations. Consequently, the exposure analyst would
consider the hypothetical individual's temporal, spatial, demographic and behavioral
exposure characteristics for each relevant duration in the assessment. This results in a
calendar approach to the exposure assessment because the timing of the multi-route
exposure relative to each other is critical to the evaluation of the health endpoint.
Figure 3-8 (adapted from a figure in U.S. EPA, 2001 a) illustrates the combination of
exposure pathways over time (in this case, days) for a single chemical.
Exposures to two or more chemicals can overlap if the chemicals coexist in the
same environmental medium during the same exposure period of interest. If there are
multiple pathways that involve different chemicals, the analyst logically would not
assume independence (see Chapter 4). Instead, joint exposure can be evaluated for
potential overlap of potential doses (e.g., chemicals in local fish and air that result in
overlapping potential doses) and internal dose
(including metabolites), for potential
toxicological interactions or for potential overlap
of effects. Information on environmental fate is
important input to this evaluation. For example,
a screening-level comparison of Kd values in
soil could be used to gauge the potential for
simultaneous migration of a group of chemicals
(see Table 3-3).
People can be exposed to chemicals at
the same time but in different media, e.g.,
multiple exposures may include inorganic
mercury in soil and shellfish, DBPs in drinking
water and during showering and volatile organic
compounds in indoor air (originating from a site
or from the use of household or office products).
Such exposures could be combined in a
cumulative risk assessment. Text Box 3-12
uses the chemical groupings based on
Examples of Chemical Groupings by
Coexistence in Media/Time
(Text Box 3-12)
Media
Time
Same
Different
Same
Group 1
Co exposures
to mixture of
DBPs via
consumption
of unheated
tapwater
Group 2
Exposures via
contaminated
drinking water
to different
pesticides
with short
environmental
half-lives
Different
Group 3
Co-exposures to
volatile and non-
volatile DBPs
via inhalation
while showering
and
consumption of
unheated
tapwater
Group 4
VOC exposures
via inhalation
due to
temporary
incinerator to
remediate a site
and, years later,
exposures to
metal mixture
via consumption
of contaminated
groundwater
3-59
-------
coexistence in media and time to illustrate chemical combinations highlighted in this
paragraph and other potential combinations.
Although, in general, less information is available to assess dermal exposures,
this route can be important in cumulative risk assessments, depending on the specific
exposures and contaminants involved. Guidance for assessing dermal exposures has
continued to be refined as additional data and exposure assessment methods emerge
(see U.S. EPA2004a,h, and updates at
http://www.epa.gov/oswer/riskassessment/ragse/).
3.3.3.4. Combining the Calendar Approach with Toxicokinetic Models
The calendar approach (U.S. EPA, 2001 a) can be combined with toxicokinetic models
to estimate tissue doses for mixture components over time. U.S. EPA (2001 a)
describes a calendar approach that estimates daily exposures with occurrence up to a
full year. The calendar approach can be used to assess exposures resulting from
seasonal activities such as timing of pesticide applications over a year or the timing of
pesticide runoff during the year. Such an approach can also be used to evaluate
exposures via indoor air, which could change seasonally. The approach integrates
exposures by route using probabilistic6 input data (e.g., this approach could integrate
oral exposures that result from food intake, drinking water consumption and soil
ingestion). The approach predicts distributions of potential doses via different exposure
routes (see Figure 3-7). Clearly, this type of approach is most useful for pollutant
concentrations that vary over relatively short periods of time (daily or weekly).
Figure 3-8 illustrates the results of a multi-pathway exposure assessment using a
calendar-based approach. Panel A of Figure 3-8 shows that the potential doses of this
hypothetical pesticide through food consumption are relatively constant over the period
of time evaluated. Panel B shows that the potential doses of this hypothetical pesticide
are generally low. However, the potential doses from this exposure pathway may be
quite high during a fraction of the period of time evaluated. The high exposures through
the consumption of private drinking water might be due to runoff of this pesticide
following applications to lawns or agricultural lands. Panel C illustrates a residential
exposure. It suggests that there is no pesticide dose from this pathway during certain
periods of time (e.g., winter months) but a relatively large dose during other periods of
time. Panel D combines these three pathways of exposure showing the potential dose
of the hypothetical pesticide for each day of the exposure duration evaluated.
6 In probabilistic exposure assessments, the population's exposures are characterized by distributions of
exposure factors and contaminant concentrations.
3-60
-------
U.S. EPA (2003b) conducted research to examine the feasibility of conducting a
cumulative risk assessment for DBP mixtures by combining exposure modeling and
physiologically-based toxicokinetic (PBTK) modeling. Initially, a comprehensive
exposure modeling effort was implemented to estimate population-based exposures
and absorbed doses for 13 major DBPs, incorporating parameters for chemical
volatilization, human activity patterns, water use behaviors, ingestion characteristics,
building characteristics, physiological measurements and chemical concentrations in the
water supply. Daily exposure estimates were made for an adult female and an adult
male and for a child (age 6) of total absorbed doses inclusive of exposures via oral,
dermal and inhalation routes. Estimates were developed for 13 major DBPs,
accounting for human activity patterns that affect contact time with drinking water (e.g.,
tap water consumed, time spent showering, building characteristics) and
physicochemical properties of the DBPs (inhalation rates, skin permeability rates,
blood:air partition coefficients, etc.). Combining daily exposure information with a
toxicokinetic model provides additional insights into the exposures, including residual
concentrations in the body. Figure 3-9 provides an overview (from a biological
perspective) of the exposure metrics that can be used in different cumulative risk
assessments. Figure 3-10 illustrates how an exposure assessment model was linked
with a PBTK model for DBPs to estimate the organ-specific doses (estimated as an
area under the curve [AUC]). PBTK models provide a useful approach for integrating
exposures across multiple exposure routes.
The kinetics of toxicants, when combined with exposure information, can be an
important factor in determining whether chemicals will be present in the same target
tissue within the body at the same time. While estimates of potential doses and the
potential daily or seasonal variability in such doses are useful (based on the
concentration of pollutants encountered in the environment, activity patterns and intake
rates), toxicokinetic models can provide refinements to this measure that may be critical
to the cumulative exposure assessment. These refinements may include differential
absorption of mixture components across boundaries, differences in the distribution of
mixture components in the body, differential metabolism and differences in elimination
(e.g., clearance rates). Models can also be developed to estimate the kinetics of by-
products of metabolism.
Figure 3-11 summarizes different levels of dose specificity that the analyst may
need in order to perform a cumulative exposure assessment. Moving from level 1 to
level 4 requires additional analytic detail. Depending on the chemicals being evaluated,
levels 1 and 2 may require the use of dynamic fate and exposure models (e.g., the
3-61
-------
Source.,
Source.,
Source,,
Environmental Concentrations in Various Media
Oral Exposure
Activity Patterns
Potential
Dose
Dermal Exposure
Inhalation! Exposure
Oral Exposure
Barrier
Absorbed
Dose
T
Skin Lung Intestinal Tract
X /
x^
X
X
X
X
Dermal AbsorbddxDose
X
X
X
>
Inhalation
. Absorbed Dose
s
/
,'6ra
/
s
Absorbed Dose
Total Absorbed Dose
(Internal Dose)
Pharmac
okinetics
Tissue
Dose
Tissue/Organ Dose
FIGURE 3-9
Dose Metrics for Environmental Contaminants (Source: U.S. EPA, 2003b)
3-62
-------
Modeling of Input Data on Chemical Properties, Human Activity
Patterns, Human Intake Parameters, Building Characteristics
24-Hour Exposure Time Histories
Simulated by TEM for a human of a specified age and sex
Modeling of Data on
Physiological Parameters
Input Da
Measure
Concent
Drinking
at the Fc
i
ta
jdDBP
rations in
Water
jucet
Estimated DBP
Concentrations
in Household Air
DBP Oral
External
Exposure
Estimates
DBP Dermal
External
Exposure
Estimates
DBP
Inhalation
External
Exposure
Estimates
DBP Oral
Absorbed
Dose
Estimates
DBP Dermal
Absorbed
Dose
Estimates
DBP
Inhalation
Absorbed
Dose
Estimates
i
DBP
Multiple
Route
Total
Absorbed
Dose
Estimates
i
DBP Multiple Route
Tissue and Organ
Dose Estimates
- AUC* Kidney
- AUCTestes
- AUC Liver
- AUC Venous Blood
*AUC - Area under the curve
FIGURE 3-10
Linking Exposure Assessment Modeling with a PBTK Model for DBPs (Adapted from U.S. EPA, 2003b)
3-63
-------
Environmental
Mixture
Concentrations
Potential
Doses
Environmental
Mixture
Concentrations
Potential
Doses
Internal Doses
(Parents)
Environmental
Mixture
Concentrations
Potential
Doses
Internal Doses
(Parents)
Tissue Doses
(Parents)
Environmental
Mixture
Concentrations
Potential
Doses
Internal Doses
(Parents)
Tissue Doses
(Active Chemical Species)
FIGURE 3-11
Levels of Dose Specificity that can be Estimated in a Cumulative Exposure Assessment
3-64
-------
calendar approach). Depending on the variability of the exposures in the pathways,
undertaking an analysis as depicted in levels 3 or 4 would likely require a dynamic
exposure model that could simulate daily potential doses of multiple chemicals.
Because of the chemical-specific nature of absorption, distribution, metabolism and
elimination, chemicals contacted at the same time may not remain in the tissues of the
body for the same period of time. Thus, some compounds may be quickly eliminated
and others may be slowly eliminated resulting in prolonged tissue exposure.
Figure 3-12 builds upon Panel D of Figure 3-8. It illustrates the target organ doses that
correspond to the cumulative exposure depicted in panel A depend on whether the
chemical is rapidly eliminated (panel B) or slowly eliminated (panel C). Figure 3-13
illustrates the different retention times exhibited by Cr(lll), Cr(VI) and tritium. The
disposition of chemicals absorbed through different exposure routes may differ. The
analyst may need to undertake an analysis as depicted in level 3 or 4 (Figure 3-11) to
determine if the exposures through different routes result in overlapping internal doses.
The analyses depicted in levels 3 and 4 require a thorough understanding of
toxicokinetic conditions. Level 3 estimates concentrations of the parent compounds in
the target tissues over time. Level 4 requires input concerning whether the compounds
are toxic in their parent form or as metabolites. In turn, level 4 analysis predicts
concentrations of the toxicologically active chemical species in the target tissue over
time.
In summary, doses may be considered at different levels of specificity. Each is
potentially useful and differentially resource-intensive. The exposure analyst would
consult with the dose-response analyst to determine the level of detailed analysis
necessary (level 1, 2, 3 or 4). The dose-response analysis may provide information
demonstrating the biological longevity of contaminants to determine potential overlap of
tissue concentrations or provide important toxicodynamic information. If available,
information on the tissue dosimetry of single chemical exposures and information
identifying sensitive tissues/organs and interaction with key biochemical pathways
(whether related to metabolism/excretion or cellular function) may be combined to allow
a more complete evaluation of interactions among mixture components leading to
changes in internal exposure duration.
As illustrated in Figure 3-14, biological effects can continue even after the
chemical(s) has been eliminated from the system. Persisting biological and/or
biochemical effects can have multiple toxicodynamic effects including those based on
chemical distribution and tissue effects. These effects can relate to subsequent
3-65
-------
A. Total Exposure
0)
E
in
8.
X
LU
0)
E
8.
Time (Days)
B. Target Organ Dose: Rapid Elimination
D)
in
o
CL
Time (Days)
C. Target Organ Dose: Slow Elimination
Time (Days)
FIGURE 3-12
Multipathway Potential Doses and Target Organ Doses
3-66
-------
Whole
Body
Cr(lll)
Cr(VI)
Tritium
Dose
Absorption
Dose
Metabolism to Cr(lll)
Absorption
Feces
Absorption and Incorporation
+=100%
3H - body water
3H - organically bound
10
i
100
I
1000
I
10000 hours
i
0.42
4.2
42
420 days
FIGURE 3-13
Human Residence Time for Selected Contaminants
3-67
-------
0)
o
Q
"03
i_
CD
Persistence of
Compound
Persistence of
Biological Effect
o
x
o
Exposure
FIGURE 3-14
Conceptual Illustration Showing the Persistence of a Biological Effect Exceeds the Duration of the Exposure
3-68
-------
exposures to the same chemical and to other chemicals, depending upon the extent to
which multiple chemicals interact with the same biochemical or cellular targets.
Finally, even a qualitative description of the possible alteration of effects based
on exposure sequence and pattern constitutes a step forward. The exposure sequence
could be an issue for chemicals in different media at different times. For example,
combined exposures from multiple routes could have occurred if an individual's past
exposure history is considered. These current and past exposures via the same or
different exposure routes/media may increase an individual's susceptibility to a chemical
(U.S. EPA, 2003e). A database of chemical pairs for which exposure timing may be
considered could be useful for cumulative risk assessments. The EPA has developed
initial information in its Mixtox database, which is described in Chapter 4. Some
information related to exposure is included in the interaction profiles that have been
drafted by ATSDR for a limited set of chemical combinations (see Appendix A). Further
discussion of toxicity as influenced by exposure sequence is presented in Chapter 4.
3.4. ILLUSTRATION OF CUMULATIVE CONCEPTS FOR THE AIR PATHWAY AT
A CONTAMINATED SITE
Local communities are understandably concerned about possible exposures to
chemicals from contaminated sites, with air and groundwater being two main transport
pathways. When the water table is reasonably shallow and local citizens are using
nearby wells, the groundwater pathway can be a main concern. The air pathway can be
an issue, for example, when the surface is still contaminated with volatile compounds,
when wind speeds are high enough to carry contaminants in surface soil off-site or
when operating facilities with stacks are present.
Sites without operating facilities are not usually of concern for ambient air quality
or public health under baseline conditions. However, cleanup of these sites can be a
much different story. Air is considered the principal pathway by which the public could
be exposed to site contaminants during the cleanup period. To emphasize the
importance of evaluating risks associated with possible cleanup measures for both
workers and the public, the following discussion illustrates cumulative considerations for
the air pathway during the cleanup period for a contaminated site. Many of the same
general concepts discussed here would also apply to the assessment of the
groundwater pathway. Tables within Appendix A include a number of tools that may
help evaluate the groundwater pathway.
3-69
-------
Several cleanup alternatives are typically evaluated for contaminated sites,
ranging from no action (the baseline) to various actions that can include excavating soil
and waste, decontaminating and
demolishing buildings, treating
wastes and transporting them for
disposal, all of which involve
Basic Steps for Cumulative Air Analysis
(Text Box 3-13)
1. Create an emissions inventory for multiple sources
2. Model air dispersion for multiple chemicals
airborne releases. Thus, for the 3 Estimgte exposures for receptors (to trans|ate to risks)
cleanup period, air contamination is
typically a community's major environmental concern. The basic steps of an air
pathway analysis for a cumulative risk assessment are summarized in Text Box 3-13.
Results are ultimately used to guide emission control strategies to minimize impacts. In
assessing this pathway, emission rates are estimated for site-related sources, and air
dispersion is modeled to predict the amounts and possible distributions of multiple
contaminants at locations of interest, which typically include the site boundary and
representative receptor locations such as homes or schools.
Of course, actual measurements of particulate and multiple airborne chemicals
would best characterize current site conditions. However, a comprehensive air
monitoring program is extremely expensive, and accuracies decrease near the
threshold of detectability, which is often the level of interest for environmental projects.
Thus, measured data usually are limited and air quality models can be applied to
assess impacts. Uncertainties related to air modeling are thought to be acceptable
when considering the high cost of monitoring.
These models combine relevant meteorology data with site emission estimates to
mathematically simulate atmospheric conditions and calculate where and when
released contaminants will reach receptor locations as well as where and how much
particle deposition will occur. Even when some data are available, monitoring will never
be able to measure concentrations for all chemicals at all locations. Therefore, modeled
estimates will be needed to fill those gaps. With modeling results, analysts can also
determine impacts of one source from among many (source attribution) and forecast
how concentrations will change if a given emissions source is modified. In addition, air
dispersion modeling is the only tool available to help analysts assess impacts from
hypothetical sources. They are valuable tools for assessing potential impacts
associated with both existing emission sources and those projected during the cleanup
period. Text Box 3-14 summarizes their benefit. Section 3.4.1 offers Illustrative
information to guide the development of emission inventories for a cumulative risk
3-70
-------
Benefits of Dispersion Models (Text Box 3-14)
Fills gaps in monitoring data to predict levels and co-
locations of combined chemicals from site releases
Avoids detectability constraints, high monitoring costs
Identifies contributing sources to joint concentrations
Projects impacts from new facilities being considered
assessment at a contaminated site, and
Section 3.4.2 gives information to guide
dispersion modeling for these sites.
3.4.1. Emission Inventories. Cleanup
of a contaminated site can involve many
different sources of emissions. Various source configurations and examples are point
(incinerator stack), area (waste impoundment or pile), volume (water treatment facility)
and line (road). Some sources are stationary while others are mobile. Common
emission sources at these sites are summarized in Text Box 3-15. At many sites,
distinct areas of contamination can contain different combinations of chemicals at
different concentrations.
Multiple Emissions During Cleanup
(Text Box 3-15)
Fugitive dust from mechanical disturbance of soil by
heavy construction equipment during excavation
(scaled to chemicals/concentrations at each area)
Dust emissions from construction and material/waste
transportation vehicles
Contaminant emissions from on-site treatment systems
(such as an incinerator or air stripper)
Windblown dust from cleared areas (when threshold
wind speed is exceeded)
Emissions of volatile and semivolatile organic
compounds due to soil disturbance (otherwise trapped
in subsurface soil pore spaces, migrating slowly)
Particulates and mixtures exhaust from diesel-burning,
heavy construction equipment (bulldozers, front-end
loaders, field generators) and transport vehicles
For cumulative risk assessments,
clearly grouping the chemicals at each
source area is important so that they can
be appropriately scaled to the fugitive
emissions estimated for that source.
Proper grouping will assure that the
model projects the appropriate
chemicals and concentrations from that
source at the receptor locations, and it
will enable the combined chemicals at
those receptor locations from multiple
sources to be back-tracked to the
originating source and activity.
Emission factors are developed
for these activities, but they do not provide any information on the temporal or spatial
patterns of releases nor on the greatest potential emission source, which is needed to
develop effective control measures. That information is developed at the next step
when emission estimates are used in the air dispersion models. To guide the
development of emissions inventories for many situations including contaminated sites,
the EPA has developed a number of databases and methods. The Air/Superfund series
provides considerable coverage of topics and methods, including an overview of air
assessments, estimation of emissions from baseline and cleanup activities and ambient
air monitoring and modeling. Specific types of emissions that would be grouped in a
cumulative risk assessment are also discussed, such as emissions of volatile and semi-
3-71
-------
Emission Factors for Multiple Sources
(Text Box 3-16)
volatile compounds from disturbed soil. Text Box 3-16 highlights key resources. When
using these and similar information sources, the analyst then can characterize whether
they likely lead to an overestimate, underestimate or central tendency estimate of the
emissions from these sources.
Of special interest for
cumulative risk assessments are
exposures to chemical mixtures.
Notably for site workers, engine
emissions from equipment and
vehicles represent such a
chemical mixture since diesel
Information
Emissions
from point and
area sources
Resource
U.S. EPA Technology Transfer Network, AP-42
(www.epa.gov/ttn/chief/ap42/index.html)
Methods to
assess
specific
emissions
Air/Superfund National Technical Guidance Study
Series
(www.epa.gov/ttn/amtic/files/ambient/other/airsuper/
superfnd.txt)
Estimation
software
U.S. EPA Clearing House for Inventories and
Emission Factors (CHIEF) (www.epa.gov/ttn/chief/)
exhaust is considered a chemical
mixture for which some toxicity
information exists (see Chapter 4). An analyst can use tools developed by the EPA,
such as those summarized in Text Box 3-17, to evaluate these and other mobile source
emissions. As noted for Text Box 3-16, users can characterize their confidence in
emissions estimates developed from sources, such as those cited in Text Box 3-17.
Although these tools do not consider interactions among chemicals, hydrocarbon
fractionation is included. By accounting for that specific input in the exposure
assessment, component toxicities can be assessed with mixtures approaches that
consider relative potencies (discussed in
Chapter 4).
In many cases the particulate releases
will dominate and other criteria pollutants will
be negligible. For that situation the analyst
could conduct a screening worst-case
analysis for those other pollutants to assure
that estimated maximum impacts are
captured in the analysis, integrated with the
other projections and presented to decision
makers and stakeholders. If this worst-case
analysis showed that the non-particulate
pollutants likely posed little risk to the
population, then this approach would lead to
an increase in the attention given to the
particulates.
Mobile Sources and Multiple Chemicals
(Text Box 3-17)
Source TypeiModel Emissions Estimated
On-road mobile
MOBILE62:
www. epa. gov/otaq/
m6.htm
Non-road mobile
(vehicle/ equipment
engines): NONROAD
www. epa. gov/otaq/
nonrdmdl.htm
Mobile, toxic
fractions of
hydrocarbons (e.g.,
engine exhaust)
www. epa. gov/ttn/chief/
net/1999inventory.html
Criteria pollutants (sulfur
dioxide, nitrogen oxides,
carbon monoxide, PM10,
PM2.5, lead);
hydrocarbons; carbon
dioxide; ammonia; & six
toxics (benzene; methyl
tertbutyl ether; 1,3-buta-
diene; formaldehyde;
acetaldehyde; acrolein).
Criteria pollutants and
hydrocarbons.
Fraction-specific
emissions forspeciated
hydrocarbons
ftp'Mp. epa. gov/pub/EmisIn v
entory/finalnei99ver3/criteria
/documentation/nonroad/99n
onroad_vol1_oct2003.pdf.
3-72
-------
Both contaminated and uncontaminated participate matter (PM) may be released
during site cleanup activities. The former can be released when contaminated materials
are excavated and staged in stockpiles and then treated in an on-site operation or
placed for transport or disposal. Uncontaminated emissions can be associated with
excavating local borrow soil (used for filling, mostly sand and gravel) and backfilling and
re-grading areas that are excavated on-site or with transporting project materials
(including treatment supplies) on paved or unpaved roads.
A cumulative risk assessment
could include both types of releases.
Text Box 3-18 summarizes the
characteristics to consider in grouping
PM and associated chemicals for
these assessments. Contaminated or
not, inhaled particles can affect human
health (as with asthma) (see
Chapter 4 for the toxicity discussion).
Of course the multiple chemicals such
as metals or organic compounds
attached to particle surfaces or
incorporated into the matrix are of
specific interest for their joint toxicities.
Fugitive PM emissions during cleanup can be estimated by considering three
factors: (1) total mass of material handled (based on the estimated volume and density),
(2) total number of activity hours (e.g., for bulldozing or scraping) and (3) total number
of vehicle miles traveled (e.g., by dump trucks). In defining the mass handled, for
cumulative risk assessment an analyst may consider what materials are being
combined together in order that representative concentrations of those materials can be
appropriately grouped and scaled to the estimated emissions. For the second factor,
production rates for each type of equipment are taken from standard reference sources
(such as the Caterpillar handbook) then combined with the mass handled (determined
for the first factor) to estimate the activity hours. Examples of additional factors used to
estimate the emissions inventory for fugitive dust are given in Text Box 3-19; many
State environmental agencies have also developed standard approaches for examining
such emissions.
Further, at many sites the contaminated source areas will be widely scattered.
Thus, in estimating fugitive emissions for cumulative risk assessments, an analyst also
Comparison of PM Properties (Text Box 3-18)
Characteristic
Relative weight
Airborne time
Travel distance in air
(depends on wind
speed atmospheric
stability)
Movement in airway
after being inhaled
Ratio of surface area
to volume, relative
potential for adsorbed
toxics
Associated toxicity
PM10(<10 |am)
Heavier
Minutes to hours
1 00 yards to
30 miles
Impinge on
sides, wedge in
narrow
passages
Lower
Generally lower
PM2.5 (<2.5 (o,m)
Lighter
Days to weeks
Farther, to
100s of miles
(~like a gas)
Pass through
small airways,
deeper in lung
Higher
Often higher
3-73
-------
could consider when different areas will be Examp|e Particulate Factors (Text Box 3-19)
Fugitive dust emissions can be estimated using a
lumped emission factor for heavy construction
activities, which is given as 1.2 tons total
suspended particulates (TSP) per acre per month
remediated in order to properly group the
emissions estimated for activities
conducted in the same time period. Then
thP disnprsinn mndplinn ran PvaluatP of activity. To estimate PM10 and PM2.5
the dispersion modeling can evaluate emissions, respective particle size multiplication
factors of 26% and 3.8% can be applied to the TSP
for unpaved roads, considering that equipment
traffic over temporary roads at construction
(cleanup) sites are major dust emission sources
(U.S. EPA, 1995a, Chapter 4). A similar lumped or
grouped approach could also be considered for
emissions from contaminated areas.
these sources jointly.
Site-specific information to support
such temporal exposure analyses are
usually presented in the general contractor
plans (i.e., these plans present a schedule
for cleanup activities and list expected equipment, based on preliminary engineering
estimates). These data can be used to select emission factors for those specific unit
operations per construction phase (see U.S. EPA, 1995a, Chapter 4).
3.4.2. Dispersion Modeling. The EPA has developed guidelines for air quality
modeling and has made many air dispersion models available within two general
categories: screening and refined. (These can be obtained via the EPA Support Center
for Regulatory Air Modeling http://www.epa.qov/scram001 as indicated in Appendix A.)
Screening models involve relatively simple estimation techniques and generally use
preset, worst-case meteorologic conditions to produce conservative estimates of the air
quality impact of a specific source or source category. Analysts use these instead of
more detailed (and more expensive) models to assess sources that clearly will not
cause or contribute to ambient concentrations above any of the following:
Ambient standards (such as the National Ambient Air Quality Standards
[NAAQS] or Prevention of Significant Deterioration (PSD) levels)
Health criteria (such as threshold limit values (TLVs) or permissible exposure
limits (PELs)) developed for daily workplace exposures or
Risk-based public health guidelines
If results of conservative screening analyses indicate that multiple chemical
concentrations from one source or a combination of sources might not meet ambient
standards and health criteria, then the analyst would apply refined models for a more
representative assessment. NRC (1994) discusses tiered analytic approaches
extensively.
Refined models include methods to address physical and chemical atmospheric
processes, and more detailed input data produces more site-specific estimates. These
two levels of modeling are often paired, with a conservative screening approach used
3-74
-------
first to eliminate contributors that clearly do not pose a concern in the cumulative
context, followed by a more refined analysis. However, for many situations the
screening models are practically and technically the only viable option for estimating
impacts of multiple sources with multiple chemicals. In those cases, it is especially
important to ensure that input data are sound. (These issues are discussed a bit later
when specific models are discussed.) Text Box 3-20 summarizes inputs to the model.
Air Dispersion Model Inputs (Text Box 3-20)
Source characteristics - Emission data scaled for
multiple chemicals by source, location, type and
geometry (for type and geometry, (1) point - stack
height and diameter, stack exit temperature, and exit
velocity; (2) area - length and width, release height,
and initial vertical dimensions; (3) volume - release
height and initial lateral and vertical dimensions)
Data for nearby buildings, to address downwash effects
Meteorologic data, for both surface and upper air
Topographic information for sources and receptors
Model control options (e.g., for dust control efficiency)
Air dispersion models are not
designed to address certain cleanup
activities. For example, they do not
directly model dispersion from specific
contaminated soil excavations as
emissions can only be estimated for a
select set of standard source types
(point, area, volume and line). For this
reason, some simplifications and
modifications are usually needed to
approximate characteristics of emission sources using engineering judgment so they
can be considered generally representative of actual site conditions.
Before beginning the calculations for a cumulative risk assessment, the analyst
can identify and group emission sources into a manageable number of sources and
types for the modeling effort. To illustrate, air strippers, incinerators and in-situ vapor
extraction units would be grouped as point sources while lagoons or surface
impoundments would be grouped as area sources. Conveyor belts or material dumping
would be volume sources and mobile (vehicle) emissions along haul roads would be
line sources. The geometries of these emission sources also serve as inputs to the
model.
The analyst may wish to consider the presence of nearby buildings when
performing a cumulative risk assessment, notably when addressing stack releases from
existing facilities or those predicted from a facility being considered (e.g., incinerator for
site wastes). Turbulent wakes downwind of structures can affect concentrations of
stack releases in the vicinity, especially when the stack height is not much taller than
the building. This phenomenon, referred to as building downwash, generally tends to
increase maximum ground-level concentrations of pollutants because it brings part of
the stack effluents to the ground near the source (instead of their being carried at a
height to a farther distance from the stack). Compared to when no buildings are nearby,
downwash changes the location of the maximum pollutant concentrations as well as the
3-75
-------
spatial distribution of the concentrations, in particular for near-field receptors (e.g.,
within several miles). Thus, estimated pollutant levels can differ considerably
depending on whether the model considers nearby buildings, and this can affect
estimates for nearby receptors. Text Box 3-21 indicates additional considerations for
modeling releases of multiple chemicals from a stack and for assessing impacts of
multiple sources at multiple receptor locations (from U.S. EPA, 1985).
For the air dispersion model to
produce relevant results, the
meteorologic data inputs logically
would represent site conditions. Some
sites have meteorologic towers (such
as larger federal research/industrial
sites), but in many cases meteorologic
data are taken from National Weather
Example Model Input Considerations
(Text Box 3-21)
When the height of a stack for an existing or planned
facility is lower than suggested by good engineering
practice (GEP), building downwash can be considered.
(The GEP stack height is 2.5xthe building height for
common configurations, i.e., for buildings wider than
they are tall; the actual formula is the height plus 1.5x
the lesser of the structure height or projected width.)
To account for terrain elevation effects, elevation data
for multiple emission sources and receptors are also
needed.
Service stations. To define the array of
receptor points for which concentrations of released contaminants will be predicted, a
receptor grid is developed for the model. Text Box 3-22 highlights these inputs.
Also important is the nature of the input data used to define the concentrations of
multiple chemicals at the receptor locations of interest. In some studies, data from an
emissions database are used (e.g.,
TRI data). Because these do not
represent ambient levels from which
exposures can be estimated, the
analyst could indicate what
proportion of input data is from that
database versus other information
sources that are more relevant to
exposure concentrations.
Implications for the results may be
addressed in the uncertainty
discussion (see Chapter 5).
Similarly, when the analyst uses
actual monitoring data, it is helpful to indicate their relevance to exposure point
concentrations, for example to identify what subset reflects ambient measurements and
at what height those measurements were made, e.g., on rooftops, at ground level or
Meteorologic and Receptor Data (Text Box 3-22)
Meteorologic data: the station selected to represent the
site is based on similar spatial characteristics regarding
terrain features, land use and synoptic flow patterns.
Typically, hourly surface and twice-daily upper air data are
available from the National Climatic Data Center, NCDC
(www.ncdc.noaa.gov/oa/ncdc.html); data for 1984-1992 for
selected National Weather Service stations are available
from the EPA's Support Center for Regulatory Air Models,
SCRAM (www.epa.gov/scram001/H24.htm).
Two types of receptors are assessed: discrete and gridded.
Discrete receptors generally represent where people
actually are (e.g., in homes or schools), or monitoring
stations, or places on the site boundary or property line
that could be accessed by the public. Hypothetical gridded
receptors are used to identify where maximum
concentrations of multiple chemicals are predicted.
3-76
-------
within the breathing zone (on the order of 2 m), along with some discussion of data
quality.
A model commonly used for conservative screening analyses is the steady-state
Gaussian model SCREENS (available at www.epa.gov/scram001 /tt22.htm#screen).
This model estimates 1-hour ambient concentrations from only one source (point, area
or flare), but it can address many combinations of wind speed and atmospheric stability
class. Its main benefit is that it is quick and easy to use. It runs interactively on a
personal computer to calculate 1-hour maximum ground-level concentrations (and
maximum concentrations for other time frames but not 24-hour estimates for complex
terrain) (NRC, 1994). It also calculates the distance to the maximum concentration from
the single source.
In order to apply this model for multiple release points, some analysts combine
these multiple emission sources to be represented by a single theoretical point. In that
case, the analyst likely would justify the basis with setting-specific information, including
relative proximity to other sources and to receptors and relative impact (insignificance)
for predictions at those receptor locations. While this simplifying approach is quite
appropriate when emission sources are far from potential receptors, it can lead to
inaccurate results if the site is near a populated area.
A key disadvantage of assuming the emissions are released from a single point
is that because of its conservative assumptions, it can generate quite unrealistic results,
e.g., highly conservative values that expectedly would never be measured. Another
disadvantage is the fact that this model for cumulative risk assessments cannot
consider multiple sources, actual meteorologic data or averaging periods other than an
hour is another disadvantage. Predicted short-term concentrations are used to assess
acute effects, while long-term concentrations are input to assess chronic effects. Thus,
SCREENS results for the 1-hour period would need to be manually converted to other
averaging times, and contributions from multiple sources would need to be combined to
address cumulative issues.
To illustrate how this averaging time adjustment is made, U.S. EPA (1992b)
provides scaling factors that are recognized as conservative and could overestimate
impacts by 2-10 times. (The actual magnitude of the overestimation is unknown and
likely depends on site and source characteristics.) When a model produces unrealistic
estimates, the generalizing assumptions can be revisited and replaced with more
situation-appropriate inputs (for example, releases might initially have been assumed to
be ground-level rather than stack or exit height from the building). In this way the
assessment is iterated from an overly conservative but quick and cheap screening
3-77
-------
approach to a more representative but resource-intensive approach as warranted to
produce realistic results that can be used for the decisions (also see discussion in NRC,
1994).
When more detailed analyses are needed, the analyst can use refined dispersion
models. These include steady-state Gaussian plume models such as ISC3-PRIME or
AERMOD. (They are available at www.epa.gov/scram001 /tt26.htm#iscprime,
www.epa.gov/ttn/scram/tt26.htmffaermod.j These models require relatively intensive
efforts and computer resources. The main advantage of these models for cumulative
risk assessments is that they can simultaneously evaluate a large number and different
types of emission sources to estimate particulate (and scaled multiple-contaminant)
levels over a wide range of averaging times, to address exposure periods from acute
(e.g., for 1, 3, 8 and 24 hours) to annual time frames. Concentrations of multiple
chemicals at different receptor locations can be attributed to specific sources by setting
up source groups for each model run and identifying contributions from a given source
within that group.
These refined models improve upon the screening models for cumulative risk
assessments by including dry and wet deposition algorithms, thus producing estimates
an analyst can use to assess
multiple pathways (by providing
deposition estimates rather than
being limited to inhalation).
However, they still do not account
for chemical reactions because
chemicals are essentially
assessed one at a time and then
results are combined. However,
some models do account for
changing concentrations for an
individual chemical overtime by
incorporating exponential decay.
Text Box 3-23 provides a general
comparison of the capabilities of
screening and refined models for
cumulative risk assessments.
In general, steady-state Gaussian models are not used for areas beyond 50 km
(30 mi.) because the steady-state assumption does not hold. For large study areas, an
Model Capabilities for Cumulative Air Analyses
(Text Box 3-23)
Scope
Multiple
chemicals
Multiple
sources
Multiple
pathways
Multiple time
periods
Source
attribution at
receptors
Changes
overtime
Chemical
interactions
Realistic
predictions
Screening Model
One at a time
(individual runs)
One at a time
(individual runs)
No, just provides
estimates for air
No, only 1-hour
averages
No
No
No
No, conservative
concentrations
Refined Model
Yes, combined, and as
scaled to particulates
Yes, many of different
types, simultaneously
Yes, because also
estimates deposition
Yes, 1-hour to annual
averages
Yes, from the grouped
sources contributing to
pollutants at those points
Some cover attenuation
(for individual chemicals)
Not for metals and
organics at sites
(only ozone, acid rain)
Yes, as constrained by
relevant data availability
3-78
-------
analysts typically estimate dispersed concentrations using models that can simulate
regional-scale, long-range dispersion as well as local-scale, short-range dispersion,
e.g., the non-steady-state Lagrangian puff models such as CALPUFF (available at
www.src.com/calpuff/ calpuffl .htmj. For areas covering thousands of kilometers,
Eulerian models such as the Community Multi-scale Air Quality (CMAQ) modeling
system would be used (see www.epa.gov/asmdnerl/models3/J. This model was
designed to address overall air quality considering multiple inputs, but it is very labor-
and resource-intensive. The CMAQ modeling requires much more computer time than
the steady-state Gaussian models so CMAQ models would probably not be appropriate
for most site assessments.7 As a note, CMAQ does address chemical reactions, but
these are only for ozone and acid rain, not air toxics. The source code would have to
be modified to add algorithms for chemical processes for the contaminants of interest at
a given site to account for those potential interactions. To date, the Agency only has
released results from the CMAQ model for mercury (U.S. EPA, 2005g).
Certain site studies might consider other point sources that could contribute to
cumulative air impacts, either as assessed by the project team or in a complementary
assessment. Some analyses have considered generic distances within which
dispersion is to be assessed; some recent studies have indicated a distance of 20 km
(12.5 mi.); a generic radius of 80 km (50 mi.) has historically been used in
environmental impact assessments. However, this potential impact radius also could be
determined from setting-specific features (including meteorology, terrain and nature of
emissions) that affect the area over which airborne releases will travel. The dispersion
model itself can be used to define an appropriate study distance, by identifying a target
level and determining at what distance that target would be reached. This could be
some fraction or percent of background (e.g., 10%) or of the initial release, considering
associated health effects.
7 In addressing fate and transport overtime, time-dependent models can yield better estimates of
exposure point concentrations than steady-state models. However, computational and resource
requirements can be much more extensive as these models are not amenable to a simple spreadsheet
approach, relying on programming language or solved using special macros. In any case, even these
models cannot generally represent truly accurate calculations, with inherent uncertainties and unknowns
as for all models. A model is considered very good when it can predict concentrations within a factor of 2
of measured results. For time-dependent models, while the value of the predicted maximum
concentration across all locations can be reasonably accurate, its predicted location is not generally as
accurate. The need for accuracy could be weighed against the availability of resources.
3-79
-------
3.5. RETROSPECTIVE STUDIES
In many cases, a cumulative risk assessment will involve predicting health
outcomes from combined estimates of exposure based on a combination of point
concentrations (e.g., site-specific measures of chemical concentrations in specific
environmental media), exposure factors and toxicity data. However, in other cases,
observed health effects could initiate the cumulative risk assessment. Dose
reconstruction studies, or retrospective exposure assessments, can be used to support
risk analyses when health outcomes are observed. The aim of these studies is to
reconstruct the doses that occurred to assess potential contributions of past exposures
to the indicated health effects.
Dose-reconstruction studies are typically constrained by data limitations for
location-specific concentrations for the relevant chemical forms in the given media over
the period of interest as well as specific exposure patterns. Further complications
include the inability to control for lifestyle factors such as smoking because those data
are often not available in the historic records. Nevertheless, a number of dose
reconstruction studies have been conducted to support occupational and environmental
health risk analyses. These include dose reconstruction studies for people at U.S.
Department of Energy (U.S. DOE) facilities (e.g., see Stange et al., 2001; also
http://www.cdc.gov/niosh/ocas/ocasdose.html) as well as for people exposed to
radiation from nuclear weapons tests and explosions that began more than 60 years
ago (NRC, 2003a). To illustrate how difficult it can be to reconstruct doses in the face of
considerable data limitations, in some cases the concentrations in the locations of
interest (where people lived) had to be modeled from measurements reported for
monitoring stations hundreds of miles away in complex terrain.
Similarly, historic occupational data also have limitations, with specific exposure
patterns and concentrations documented poorly if at all, such that average and peak
concentrations and durations are difficult to estimate in order to assess cumulative daily
and repeat exposures. As a reflection of the increasing emphasis on this tool, the
National Academy of Sciences (NAS) recently reviewed previous dose-reconstruction
studies to assess whether the methods and data used were accurate, the reconstructed
doses were accurately reported and the exposure assumptions were credible. The NAS
concluded that although the methods were generally valid, resultant estimates were
highly uncertain because specific data were either sparse and highly variable or simply
lacking. A key conclusion was that the review and oversight of dose-reconstruction
studies should be commensurate with the anticipated scope of the compensation
program (NRC, 2003b). This same principle applies to the level of cumulative risk
3-80
-------
analysis, i.e., that resources applied are commensurate with the needs of the decision
to be informed.
3.6. SUMMARY COMPARISON AND SCREENING SUGGESTIONS
Text Box 3-24 summarizes a general comparison of the exposure assessment
process conducted for basic health risk assessments and for cumulative exposure
assessments.
Comparison of Exposure Assessment Processes (Text Box 3-24)
Basic Assessment
Cumulative Assessment
What general question is being addressed?
How could people be exposed to chemicals,
what would the amount of exposure be?
Similar, but emphasizing combined source
contaminants and cumulative exposures
What is evaluated?
Individual Sources/releases of chemicals
Behavior of individual chemicals in the
environment (transport/fate)
Concentrations of chemicals at points of
human contact
People who "represent" current conditions
and likely future land use
Routes by which people could be exposed to
each chemical
Amount of each chemical taken in over time
Emphasis on combined sources/releases (sources may
not be located in community)
Emphasis on joint behavior, considering environmental
interactions, differential transformation and grouped
sets of chemicals
Emphasis on sets of chemicals that coexist initially and
those that move together
Representative receptors as for the basic case, paying
attention to sensitive subgroups and unique exposure
activities (e.g., per cultural practices)
Emphasis on combined chemicals and routes over
time, considering sequencing
Emphasis on combined amounts of various forms
(potential impact on toxicokinetics)
How are results used?
Estimated intakes are linked with toxicity
information to assess potential harm
Estimated intakes are considered in groups to guide
more explicit evaluation of joint toxicity to assess
potential health harm
As this summary shows, the basic topics and outcomes are the same. The
cumulative column simply highlights additional attention that would be paid to certain
features in explicitly considering cumulative risk issues. Cumulative risk assessments
evaluate aggregate exposures by multiple pathways, media and routes over time, plus
combined exposures to multiple contaminants from multiple sources.
Practical suggestions that can be considered in conducting the exposure
assessment for cumulative risk assessments at these sites are offered below, with an
emphasis on screening for grouped evaluation.
3-81
-------
Implementing existing guidance, which identifies many cumulative risk issues, is
enhanced by more explicitly acknowledging joint evaluations and at least
qualitatively indicating the potential for interactions to define groupings. An initial
conservative screening of relative risks can be conducted to identify the sets of
contaminant sources, receptor locations and pathways to be analyzed in detail.
Focus on grouping the chemicals, affected media and exposure points that are
expected to contribute to combined pathway exposures for those receptors,
considering media and time frames.
Because relatively few major sources might account for most of the hazards
associated with a site, focus first on the main sources, especially when resources
are constrained. However, following that initial focus, iterate through the
assessment process to assure that cumulative exposure issues have been
appropriately considered.
In modeling chemical transport and fate, account for environmental
transformation overtime (including mixtures) and adapt transport/dispersion
models to account for multiple chemicals, e.g., scaling to source concentrations
for those chemicals moving together and defining source attributions at multiple
receptor locations.
In developing groupings for chemicals and exposure pathways, focus on (1) the
potential for relatively high exposures to sensitive populations and possible
contribution to induction of health effects that already exist at relatively high
levels in the study population, (2) in addition to those with high inherent hazard
(toxicity) in combination with (3) the amount present; (4) potential interactions
with other chemicals; and (5) tendency to persist, bioaccumulate and/or be
transported between environmental media.
To screen potential vulnerable or susceptible subgroups into the enhanced
cumulative risk assessment process, pursue existing data such as indicator
information in demographic studies and health registries.
Consider the total exposure context to evaluate whether contributions from site
contaminants combined with existing body burdens might exceed levels that are
expected to be safe. For stakeholders desiring a more explicit assessment of
total exposure, to cover chemicals not related to the site, indicate information
resources that can be used to guide such a complementary assessment.
3-82
-------
4. TOXICITY ASSESSMENT OF MULTIPLE CHEMICALS, EXPOSURES AND
EFFECTS
This chapter provides detailed information on the toxicity assessment of multiple
chemicals, exposures and effects, a subset of cumulative risk issues that are described
in Chapters 1 and 2. The goals of Chapter 4 are to
define cumulative toxicity assessment as conducted in this chapter (Section 4.1);
summarize existing EPA guidance for conducting toxicity assessments, including
chemical mixtures risk assessments (Section 4.2); and
expand those ideas to include multiple route exposures at various time frames
(Section 4.7), the value of pharmacokinetic information in evaluating internal co-
exposures (Section 4.3), consideration of secondary and tertiary effects (Section
4.5) and the impact of chemical interactions on cumulative risk (Section 4.6).
Section 4.4 presents a flow chart for the purpose of facilitating and organizing the
analyst's effort to evaluate toxicity groups for cumulative toxicity assessment. The
approach presented in this chapter provides a method for grouping chemicals by their
potential for joint toxic action as a refined classification of the cumulative exposure
groups (developed in Chapter 3) and then to provide a set of cumulative risk
assessment methods for addressing multiple toxic effects, multiple exposure routes and
toxicological interactions for chemical mixtures. These methods may be used in
cumulative risk assessment in several different ways depending on data availability and
on the goals of the assessment. They may be applied as screening tools (e.g., to
decide whether or not toxicological interactions are of importance for a certain group of
chemicals) or as tools for estimating quantitative risk numbers (e.g., estimating the risk
of an adverse level of cholinesterase inhibition by applying a Relative Potency Factor
[RPF] approach to a group of pesticides). In some cases all of the methods shown in
this chapter might be applied, and in other cases, only a select few methods would be
useful depending on the exposure scenario.
This chapter presents a number of approaches, some of which can be easily
implemented with existing data and published methods and some of which would be
resource intensive in terms of data collection and analysis. They are all shown here in
the interest of advancing the field of cumulative risk assessment and for the purpose of
providing the EPA with data sources and methodology for conducting such
assessments.
4-1
-------
4.1. DEFINING CUMULATIVE TOXICITY ASSESSMENT
Toxicity assessments developed in this chapter in support of cumulative risk
assessments evaluate a population's potential to develop adverse health effects from
exposures to multiple chemicals through multiple routes of exposure over time. In
addition, such assessments may consider the potential for multiple health effects and
for joint toxic action from multiple route exposures to chemical mixtures. Timing and
intensity of exposures to different chemicals may be evaluated, including the
examination of internal co-occurrence of multiple chemicals and toxicological
interactions in the target tissue(s).
The development of methods in this chapter are narrowly constrained to multiple
chemicals, exposures and effects; thus they will aid the analyst in conducting a
cumulative toxicity assessment, but they may be augmented with additional information
and analyses in order to produce a more comprehensive, community focus, where the
population may be exposed to stressors other than chemicals, potentially from multiple
sources. In addition, information developed during data collection and organization
regarding the population profile may be further incorporated, including considerations
related to vulnerability (i.e., susceptibility/sensitivity, differential exposure, differential
preparedness and differential ability to recover).
4.2. TOXICITY ASSESSMENT
GUIDANCE AND METHODS
The general methods the EPA
uses for toxicity assessment are detailed
in a number of risk assessment guidelines
and guidance documents, as illustrated in
Text Box 4-1. The EPA's Program Offices
use these various documents to conduct
assessments and also to develop
additional guidance and tools specific to
their respective media and sites.
Information regarding toxicity assessment
and many other aspects of risk
assessment can be found within EPA's
Web site (www.epa.gov). For example, to
supplement its primary guidance for site
assessments (U.S. EPA, 1989a),
Selected Information Guides for Toxicity
Assessment (Text Box 4-1)
Risk Assessment Guidelines of 1986, including
chemical mixtures, mutagenicity, cancer, exposure
assessment, developmental effects (U.S. EPA,
1986b, 1987)
Risk Assessment Guidance for Superfund
(U.S. EPA, 1989a)
Guidelines for Developmental Toxicity Risk
Assessment (U.S. EPA, 1991 a)
Reproductive Toxicity Risk Assessment Guidelines
(U.S. EPA, 1996a)
Guidelines for Neurotoxicity Risk Assessment (U.S.
EPA, 1998b)
Guidelines for Ecological Risk Assessment (U.S.
EPA, 1998c)
Supplementary Guidance for Conducting Health
Risk Assessment of Chemical Mixtures (U.S. EPA,
2000a)
Guidance on Cumulative Risk Assessment of
Pesticide Chemicals That Have a Common
Mechanism of Toxicity (U.S. EPA, 2002c)
Guidelines for Carcinogen Risk Assessment (U.S.
EPA, 2005d)
Supplemental Guidance for Assessing
Susceptibility from Early-Life Exposure to
Carcinogens (U.S. EPA, 2005e)
4-2
-------
Superfund provides a set of tables to be used as templates for conducting HI
calculations (online at http://www.epa.gov/oswer/riskassessment/ragsd/tables.htm).
Most of the documents providing risk assessment guidance (see Text Box 4-1)
focus on specific health endpoints such as cancer, mutagenicity, reproductive and
developmental effects and neurotoxicity. These documents can be used in a
cumulative toxicity assessment to evaluate their respective health endpoints; the
resulting information can then be combined using guidance that deals with cumulative
risk issues such as the 2000 Supplementary Mixtures Guidance (U.S. EPA, 2000a) or
the Methodology for Assessing Health Risks Associated with Multiple Pathways of
Exposure to Combustor Emissions (U.S. EPA, 1998a). Guidance also is available for
evaluating toxicological mechanisms of action, including those related to cumulative risk
for pesticide exposures (U.S. EPA, 2002c) and for mechanisms of carcinogenicity (U.S.
EPA, 2005d). The assessment of vulnerable subpopulations is also addressed by
Superfund in their site assessment guidance (1989a). Children are specifically
addressed in a supplemental guidance to the 2005 carcinogen risk assessment
guidelines (U.S. EPA, 2005e). In summary, there are many EPA resources that
describe methods and approaches that can be used to address various aspects of
cumulative toxicity assessments for community-based cumulative risk assessments.
4.2.1. Practices for Evaluation of Toxicity for Various Durations. In toxicity
assessment, Reference Values (RfVs)1 are often used as target levels that are
protective of human health. The focus of most site assessments is on evaluating health
effects from chronic exposures. However, shorter-duration exposures can also play a
key role in risk assessments, such as the evaluation of remediation activities at
contaminated sites. For example, health effects are assessed for workers and the
public from short-term exposures to releases associated with cleanup measures, such
as excavation or treatment processes for contaminated materials.
The Risk Assessment Guidance for Superfund (U.S. EPA, 1989a) outlines
approaches for evaluating potential health effects associated with different time frames,
using RfVs developed for exposure duration. More recently, the NRC discussed the
issue of varying exposure durations and selection of corresponding RfVs in its Review
of the Army's Technical Guides on Assessing and Managing Chemical Hazards to
Reference Value (RfV): EPA's estimate of an exposure for a given duration to the human population
(including susceptible subgroups) that is likely to be without an appreciable risk of adverse health effects
over a lifetime. Durations include acute, short-term, subchronic and chronic. EPA develops numerical
toxicity values for the oral RfD and inhalation RfC. (See the Glossary in Chapter 7 for complete
definitions.)
4-3
-------
Deployed Personnel (NRC, 2004). RfVs have been and continue to be developed for
chronic exposures. However, RfVs for shorter durations are also available for a more
limited number of combinations of chemicals and exposure durations, some of which
might deviate from the assumed constancy of the concentration * time product (see the
glossary in Chapter 7 for complete definitions). Table 4-1 highlights selected, additional
RfVs.
As noted by NRC (1994), chronic RfCs and RfDs can also be examined to
determine if an Uncertainty Factor (UF)2 of 10 was applied in the original derivation for
subchronic to chronic extrapolation. In this case, it may be appropriate to multiply the
chronic RfC or RfD by a factor of 10 for evaluating less than chronic exposure durations.
Further, some chronic RfVs may be appropriately applied to shorter exposure durations
(in the absence of an RfV derived for the duration of interest), particularly for chemicals
whose toxicity is more a function of concentration than cumulative exposure. It is
important to discuss the uncertainty or confidence in the values used within a risk
assessment, giving consideration to the correspondence between the context and
exposure duration for which the RfVs were developed and then applied in the risk
assessment and their source and the nature of their RfV development process.
4.2.2. Practices for Evaluating Chemical Mixtures. The EPA evaluates risks from
exposure to chemical mixtures using peer-reviewed Guidelines and Guidance
documents (U.S. EPA, 1986b, 1989a, 2000a) that identify both component-based and
whole mixtures methods. The flow chart from U.S. EPA (2000a), shown in Figure 4-1,
illustrates that the selection of a method (e.g., HI, RPF) depends on the availability and
interpretation of information on toxicological joint action and chemical composition of the
mixture.
Whole mixture methods (e.g., mixture RfDs, RfCs and cancer slope factors)
account for unidentified chemicals in a complex mixture and inherently incorporate joint
2 Uncertainty/Variability Factor (UFs): One of several, generally 10-fold, default factors used in
operationally deriving the RfD and RfC from experimental data. The factors are intended to account for
(1) variation in susceptibility among the members of the human population (i.e., inter-individual or
intraspecies variability); (2) uncertainty in extrapolating animal data to humans (i.e., interspecies
uncertainty); (3) uncertainty in extrapolating from data obtained in a study with less-than-lifetime exposure
(i.e., extrapolating from subchronic to chronic exposure); (4) uncertainty in extrapolating from a lowest-
observed-adverse-effect level (LOAEL) rather than from a no-observed-adverse-effect level (NOAEL);
and (5) uncertainty associated with extrapolation when the database is incomplete.
4-4
-------
TABLE 4-1
Selected Reference Values for Different Exposure Durations
Exposure
Duration
Acute
(<24
hours)
Short
Term
(1-30
days)
Longer
Term
(>30 days
to 7 years)
Chronic
(>7 years)
Toxicity Value or
Guideline
1-day drinking water
health advisory
Acute exposure
guideline level
(AEGL)
Acute minimal risk
level (MRL)
(1-14 days)
10-day drinking water
health advisory
Intermediate MRL
(15-364 days)
Chronic reference
dose and
concentration
Chronic MRL
(>1 year)
Source
EPA Office of Water
http://www.epa.gov/waterscience/criteria/drinking/
National Advisory Committee, National Research
Council
ATSDR
(http://www.atsdr.cdc.gov/mrls.html)
EPA Office of Water
http://www.epa.gov/waterscience/criteria/drinking/
ATSDR
(http://www.atsdr.cdc.gov/mrls.html)
Integrated Risk Information System (IRIS)
http://www.epa.gov/iris/
ATSDR
(http://www.atsdr.cdc.gov/mrls.html)
Notes
Based on oral toxicity values derived by
the EPA Office of Water
Derived for inhalation exposures for
exposure times ranging from 10 minutes
to 8 hours
Based on oral or inhalation toxicity values
derived by ATSDR
Based on oral toxicity values derived by
the EPA Office of Water
Based on oral or inhalation toxicity values
derived by ATSDR
For oral and/or inhalation exposure
Based on oral or inhalation toxicity values
derived by ATSDR
4-5
-------
c
Assess Data Quality
inadequate
» Only Qualitative Assessment
adequate
0 Mixture |
f Concern I
Mixture
RfD/C;
Slope
Factor
Whole Mixture
s
^
\ \
/SufficientlyX
Similar
\ Mixture /
/
/ Group of
Similar
V Mixtures
\!
Components
I Toxicologically i (Toxicologically j
I Similar J I Independent J
Interactions
Comparative
Potency
Environmental
Transformation
Hazard
Index
Relative
Potency
Factors
Response
Addition
Interactions
Hazard
Index
Compare and Identify Preferred Risk Assessment,
Integrate Summary with Uncertainty Discussion
The different types of mixtures assessments based on the availability and quality of the data.
All possible assessment paths should be performed.
FIGURE 4-1
Approach for Assessing Mixtures Based on the Available Data (U.S. EPA, 2000a)
4-6
-------
toxic action among chemicals (Figure 4-1 ).3 Dose response assessments based on
tests of whole mixtures or on epidemiologic data determine combined effects
empirically. Examples of these (U.S. EPA, 2007) include (1) RfDs on commercial PCB
mixtures (Aroclors 1016 and 1254) based on primate data and (2) a cancer slope factor
for coke oven emissions based on human occupational exposures.
The usefulness of toxicological data on a whole mixture depends strongly on how
similar the studied mixture is to the environmental mixture of concern (U.S. EPA,
2000a). The fundamental requirement for what is called sufficient similarity is that the
complex mixture that is being considered as a surrogate has roughly the same major
chemical components in approximately the same proportions as the environmental
complex mixture that is being evaluated. Any additional information on toxicological
similarity, i.e., data on similar health effects and dose-response relationships for the two
complex mixtures or their common components, may also be useful in establishing
overall similarity. The EPA's 2000 Supplementary Mixtures Guidance discusses several
issues with determining toxicological similarity of two complex mixtures (U.S. EPA,
2000a). For example, the RfD, RfC or cancer potency for a complex mixture can be
determined by treating the mixture as if it were a single substance and using the dose-
response data on that substance in the same fashion that single chemical dose-
response data are used. The main challenge for an analyst to ensure that the mixture
composition (relative proportions of the component chemicals) remains fairly constant.
The simplest component-based methods utilize single chemical exposure and
dose response information to form a mixtures assessment and are useful in comparing
mixtures containing the same chemicals but with varied concentrations and proportions.
Component-based methods include those based on assumptions of response addition
(toxicological independence) and dose addition (toxicological similarity). These
methods, however, do not directly address interaction effects among components (i.e.,
effects greater than or less than those observed under a definition of additivity). To
address the latter concern, the Interaction-Based HI method may be applied, using
information on binary (pairwise) interactions among chemicals in a mixture to modify its
HI (see Section 4.6.2 for details on this method). The main toxicological considerations
for the component-based risk assessment methods used by U.S. EPA are then
toxicological independence, toxicological similarity and pairwise interaction.
Dose addition and response addition are fundamentally different methods, relying
on different toxicity assumptions. The two additivity assumptions are briefly described
3lt may be noted here that this chapter does not employ the comparative potency or environmental
transformation methods shown in Figure 4-1; thus, they will not be described further.
4-7
-------
in the following text. Extensive discussion of these mixture methods is given in the
EPA's 2000 Supplementary Mixtures Guidance (2000a).
Dose addition sums the doses of the components in a mixture after they have
been scaled for toxic potency relative to each other. The predicted mixture
toxicity is determined from this summed dose. Dose addition requires the
component chemicals to be toxicologically similar (i.e., to share a common toxic
mode of action [MOA]). If dose addition is applied using an index chemical to
estimate risk, the mixture components are required to have similarly shaped
dose-response curves for the endpoint being evaluated.
Response addition first estimates the probabilistic risk of observing a toxic
response for each chemical component in the mixture. Then, the component
risks are summed to estimate total risk from exposure to the mixture, assuming
independence of toxic action (i.e., the toxicity of one chemical in the body does
not affect the toxicity of another chemical). This can be thought of as an
organism receiving two (or more) independent insults to the body, so the risks
are added under the statistical law of independent events.
4.2.2.1. Dose Addition Superfund site assessments have applied dose
addition in the form of a HI to evaluate sites for indications of health risk (U.S. EPA,
1989a). The HI is calculated as the sum of HQs for the chemical components of the
mixture. (Note the HI is not dependent on using an index chemical to assess risk, so
the components are not required to have similarly shaped dose-response curves.) An
HQ is typically calculated as the ratio of a chemical's exposure level to its safe or
allowable level, such that values larger than 1 are of concern. For a group of n
chemicals in a mixture and using the RfD as a safe, allowable level, the HI for oral
exposure is calculated:
where:
EJ = exposure level of the /th chemical
RfDj = Reference dose of the /th chemical.
A similar index for inhalation exposure uses the RfC for the allowable level. The HI is
usually calculated for groups of chemicals whose effects are observed within a common
target organ. The HI is interpreted similarly to the HQ: the more HI exceeds 1 , the
greater is the concern for mixture toxicity. Note that the HI provides an indication of risk
but is not an explicit risk estimate.
To estimate actual risk, a slightly different approach, also based on dose
addition, uses RPFs for the dose scaling. Because the total dose of the chemicals in
4-8
-------
the mixture is of importance, the chemical components of a mixture are scaled for
relative toxicity to an index chemical and then summed to produce a total index
chemical equivalent dose. In this method, the total index chemical equivalent dose is
evaluated using the index chemical's dose response curve to estimate risk (see Section
4.7.1.2 for details). Note that the toxicity equivalence factors (TEFs), developed for
dioxin assessment, are a special case of the RPF approach (U.S. EPA, 1989b).
As an expression of dose addition, the formula for HI has three important
uncertainties (U.S. EPA, 2000a):
1) The assumption of common MOA might not apply because only commonality of
the target organ is considered.
2) The use of a safe level, such as a lower bound on the toxicity threshold, might
not be an accurate measure of toxic potency. Weak toxicity data usually result in
a lower safe level because of larger uncertainty factors or use of lower
confidence bounds on dose.
3) The use of RfDs as safe levels may result in an overestimate of the degree of
concern because the RfD is based on one critical or most sensitive effect. Thus,
when a chemical causes multiple effects and is to be included in more than one
HI calculation, the general use of its RfD is problematic. A solution is to generate
Target organ Toxicity Doses (TTD) (derived for each target organ of concern
using RfD methodology for noncancer endpoints only) for use in target organ
specific HI calculations (Mumtaz et al., 1997; U.S. EPA, 2000a).
Appropriate interpretation of the HI requires detailed understanding of the
individual chemical's dose-response curves, the nature and commonality of the toxic
effects and the quantitative relationship between the effect of concern and the critical
effect.4
4.2.2.2. Response Addition Toxic effects described by the proportion of
exposed animals showing toxicity are often determined for mixtures using response
addition. For example, the probabilistic risk of cancer in a given dose group is typically
estimated by the proportion of responders in that group. Total cancer risk is then
estimated for a mixture by summing the individual cancer risks for the carcinogens in
the mixture (U.S. EPA, 1989a). This calculation is derived using the statistical law of
independent events, where, for a two chemical mixture, the mixture risk (Rm) is equal to
one minus the probability of not responding to either chemical 1 (r?) or chemical 2 (r2):
4The critical effect is defined as the first adverse effect, or its known precursor, that occurs to the most
sensitive species as the dose rate of an agent increases.
4-9
-------
(4-2)
Simplification of this equation shows that Rm is the sum of the risks for chemical 1 (r?)
and chemical 2 (r2) minus the probability that the toxic event from exposure to chemical
1 would overlap in time with the toxic event from exposure to chemical 2, as expressed
in the following equation:
Rm=r1+r2-(r1xr2) (4-3)
When risks are very low, the subtracted term is so small that its impact on Rm is
negligible (e.g., for r* = 0.01 and r2 = 0.02, Rm:= 0.01 + 0.02 - 0.0002 = 0.0298 or
~0.03); thus, low risks can simply be summed. Risks are appropriately aggregated for
cancers across various target organs because the result is interpreted as the risk of any
cancer, and the cancers from each chemical component are considered to be
independent events in the body.
The applicability of both dose addition and response addition can be evaluated
by appropriate toxicity testing that produces dose-response data for the whole mixture
and its component chemicals. Any use of the additivity formulas to obtain estimates of
mixture toxicity extrapolated beyond the range of actual mixture data are typically
accompanied by a description of the evidence supporting the additivity assumptions,
i.e., commonality of toxicity for dose addition and toxicological independence for
response addition.
4.2.3. Old, New and Enhanced Approaches for Cumulative Toxicity Assessment.
Cumulative risk assessments add layers of complexity to evaluation of chemical
mixtures. Figures 4-2a and 4-2b take the concepts developed in Figure 4-1 and expand
them by presenting both established methods along with new or enhanced methods that
may be used to evaluate various aspects of cumulative risk. For example, Figure 4-2a
shows the same development of toxicity values (i.e., RfDs, RfCs and slope factors) as
presented before for whole mixtures and sufficiently similar mixtures, but Figure 4-2a
now includes additional epidemiologic evaluations that may be conducted when
illnesses in the population initiates a cumulative risk assessment (discussed in
Section 2.5). Figure 4-2b also maintains previously used component-based chemical
mixtures methods (i.e., RPFs, HI, Response Addition and the Interaction-Based HI), but
several other approaches are also reflected in this figure, and these will be presented
and discussed later in this chapter. Further, Figure 4-2b handles not only toxicologically
similar and dissimilar mixtures, but the figure also addresses mixes of these, as well as
addressing the case of multiple toxicological effects. Finally, additional methods are
4-10
-------
r Whole Mixture |
I Data Available J
/ \
/ Whole \
Mixture
\pf Concern/
'Sufficiently^
Similar
k Mixture )
V V
1
Toxicological
Evaluations
1
Epidemiologic
Evaluations
I
[
I
1
Derive RfDs/RfCs;
Slope Factors
Exposure Assessment of Whole Mixture
of Concern; Assessment of Similarity
I
I
Hazard Quotient;
Risk Estimate
Epidemiologic
Risk Measures
FIGURE 4-2a
Flow Chart Showing Approaches for Evaluating Whole Mixtures
4-11
-------
Multiple
Toxicological
Effects for Each
Component
Multivariate
Statistical
Models, e.g.,
Categorical
Regression
f Component Data Available 1
lexicologically
Similar
Components
Dose Addition
Available
Interactions
Data
Mix of \
lexicologically |
Similar & Independent I
Components /
/ Toxicologically
Independent
k Components
Relative
Potency
Factors
PBPK*
Modeling
\
Integrated
Additivity
Methods
Response
Addition
Component Exposure Assessment
\ \ \
Risk
Estimates;
Hazard
Index
Binary
Weight of
Evidence;
Interaction
Profiles
Interaction-
Based
Hazard
Index
Hazard
Index;
Cumulative
Hazard
Index
Summing
of Route-
Specific
Index
Chemical-
Based
Risk
Estimates
Internal
Dose
Hazard
Index;
Multiple
Route
Internal
Doses
*PBPK = physiologically-based pharmacokinetic
FIGURE 4-2b
Index
Chemical-
Based Risk
Estimate;
Hazard
Quotient
Risk
Estimate
Flow Chart Showing the Component Based Approaches for Evaluating Multiple Chemicals, Exposure Routes, Effects and
Toxicological Interactions
4-12
-------
discussed that include the use of PBPK models to estimate internal doses of chemicals
and examine the potential for toxicological interactions.
4.3. TOXICOLOGY OF INTERNAL CO-OCCURRENCE
This section communicates the importance of understanding tissue dosimetry of
compounds, as opposed to understanding the human exposure to them in the
environment. Toxicity is a function of the contact between a contaminant chemical and
its biological receptor, located in target tissues. Because of the complex nature of
biochemical and physicochemical factors governing chemical disposition in the body,
measures of environmental contact are insufficient to completely describe internal
disposition of chemicals in the human body and the temporal description of the toxic
sequella, including events that may modify the internal dosimetry of subsequently
encountered contaminants. At present, there is no EPA guidance on best practices of
this type of activity, though several related efforts are underway.
Toxicity assessment involves understanding and mathematically describing the
relationship between exposure (dose) and effect (response). This relationship may be
quantified at several levels of specificity (Figure 4-3). At its most fundamental level, the
end result may only be hazard identification: the ability to link an exposure with an
adverse outcome, where the data are insufficient to inform an understanding of the
dose-response relationship. The next level of detail involves knowledge of the
concentration encountered in the environment, or in the cases of most toxicity studies,
the administered (not the internal) dose. Increasing the level of sophistication requires
knowledge of the internal dose of the parent compound and is the first level at which
consideration of pharmacokinetic principles must be employed. The final two levels of
complexity require solid understanding of pharmacokinetic conditions and allow the
internal dose to be translated first to concentrations of the parent compound in the
target tissues and ultimately to concentrations of the toxicologically active chemical
species (parent or metabolite) in the target tissue. This final level of specificity requires
knowledge of whether the compound is toxic in its parent form or as a metabolite. Thus,
doses, and specifically internal doses, may be considered at different levels of
specificity; each is useful and differentially resource-intensive.
Metabolites can have a different, or even opposite action, from the parent
compound, further complicating an assessment. For example, Gierthy et al. (1997)
report that the PCB 3,4',5-trichlorobiphenyl shows antiestrogenic activity in an in vitro
assay, whereas its hydroxylated metabolite shows strong estrogenic activity. As more
is learned about mixtures of the same general class (e.g., dioxins/furans, PCBs,
4-13
-------
1 Compound
Toxicity
_ _ .
2 Compound
Environmental
Toxicity
3 Compound
Environmental Internal Dose
Concentration (Parent)
*> Toxicity
Environmental Internal Dose Tissue Dose
4 Compound + concentration * (Parent) * (Parent)
Toxicity
5 Compound
Environmental
Concentration
Internal Dose
(Parent)
Tissue Dose Toxicity
(Active Chemical Species)
FIGURE 4-3
Level of Specificity for Dose-Response Relationships
4-14
-------
polybrominated diphenyl ethers, and toxaphene) and their specific biological effects,
further refinements may be incorporated into an assessment.
Because of the compound-specific nature of their disposition in and elimination
from the body, not every compound contained in the same contacted environmental
medium will remain in the tissues of the body for the same duration. Thus, for one
chemical, a given exposure may result in prolonged retention and protracted tissue
exposure whereas a different compound encountered in the same environmental
medium may be quickly eliminated following exposure. The toxicity analysis
summarizes information demonstrating the biological longevity of contaminants to
determine potential overlap of tissue concentrations (Figure 4-4, also discussed from an
exposure perspective as Figure 3-13 in Chapter 3), again focusing on doses or
exposures most similar to the anticipated environmental exposure. Compounds
encountered at the same time from different media and through different routes may
have similar or markedly different internal exposure profiles, depending on the
compound. It is important to relate either of these situations to the potential for
overlapping internal dose as each defines a concurrent exposure. Information on the
tissue dosimetry of single chemical exposures and information identifying sensitive
tissues/organs and interaction with key biochemical machinery (whether related to
metabolism/excretion or cellular function) are combined to allow a more complete
evaluation of interactions among mixture components leading to changes in internal
exposure duration. Thus, there are advantages of evaluating exposures at the tissue
level rather than at the level of the environmental contact.
Biological effects can continue even after the chemical is removed from the
system. Persisting biological and/or biochemical effects can have multiple effects
including those based on chemical distribution and tissue effects. These effects can
relate to subsequent exposures to the same chemical, or other chemicals, depending
upon the extent to which multiple chemicals interact with the same biochemical
machinery. For example, exposure may induce, or increase the liver's content of an
enzyme (Figure 4-5, also discussed from an exposure perspective as Figure 3-14 in
Chapter 3). This can result in increased bioactivation and detoxication potential when
that enzyme is responsible for the metabolism of additional encountered compounds
(Figure 4-6). In this example (top panel), chemical A induces the expression and
subsequent metabolic capacity of the enzyme responsible for metabolizing (here,
hydroxylating) not only chemical A, but chemical B as well. With the increase in
metabolic capacity (lower panel), increased metabolism may result in a higher toxic
potential when metabolism results in a bioactivation process or lower toxic potential
4-15
-------
Whole
Body
Cr(lll)
Cr(VI)
Tritium
Dose
Absorption
Dose
Metabolism toCr(lll)
Absorption
Dose
Absorption and Incorporation
^+=100% M
3H - body water
3H - organically bound
10
i
100
I
1000
I
10000 hours
i
0.42
4.2
42
420 days
FIGURE 4-4
Human Residence Time for Selected Contaminants
4-16
-------
I Exposure |
FIGURE 4-5
Conceptual Illustration of Persistence of Mixture Components
when metabolism represents a
detoxication process. However,
enzyme induction does not always
increase chemical metabolism in vivo
(Kedderis, 1997; Lipscomb, 2003,
2004). When metabolic capacity of
the liver already surpasses the rate at
which a chemical may be delivered to
the liver via hepatic blood flow (a
condition known as flow-limited
metabolism), further increases in
metabolic capacity, e.g., through
enzyme induction does not always increase chemical metabolism in vivo (Kedderis,
1997; Lipscomb, 2003, 2004). When metabolic capacity of the liver already surpasses
the rate at which a chemical may be delivered to the liver via hepatic blood flow (a
condition known as flow-limited metabolism), further increases in metabolic capacity
(e.g., through enzyme induction) will
not increase the rate or extent of
chemical metabolism. The extent
and duration of persistent biological
effects is determined, and its impact
on the toxicity of other compounds is
investigated on a compound by
compound basis.
The timing of compound
exposure and the duration of
biological effects is to be carefully
considered. One well known
initiation-promotion chemical interaction occurs when the prior events associated with
the toxicity of benzo[a]pyrene (DMA damage) persist beyond the chemical's residence
time on the body. These effects are transformed into tumors by the subsequent
exposure to a second compound, TPA (see Text Box 3-11). Tumors are not produced
when the sequence of the exposures is reversed. This is due to the short biological
residence time of TPA (compared to B[a]P) and the short biological persistence of
TPA's effects. Mehendale and colleagues provide another example of the biological
effects persisting beyond chemical residence time (Mehendale, 1995; Soni et al., 1999).
Enzyme
1
Enzyme Induction
Increased Metabolism
A A-OH
Increased Metabolism
B ^^ B-OH
Increased
Detoxification
or Bioactivation
FIGURE 4-6
Conceptual Illustration of Effects of Metabolism on Toxicity
4-17
-------
Their results demonstrate that low levels of tissue damage can result in stimulations of
cellular repair, which are themselves protective against subsequent chemical exposure
and insult occurring during the time of increased repair. Co-exposure to agents that
inhibit repair capacity (e.g., chlordecone) potentiates the toxicity of the original
compounds at least during the time that the biological effect (inhibition or repair)
persists. This information is summarized and considered as the toxicity assessment
proceeds through the evaluation of chemical interactions.
4.3.1. Use of Internal Doses in the Hazard Index. Internal dose measurements are
becoming more common in chemical mixtures risk assessment and have been applied
in the calculation of the HI and to investigate the potential for pharmacokinetic
interactions among the chemical constituents (Haddad et al., 1999, 2001). In Haddad et
al. (2001), the authors use PBPK models to calculate an interaction-based HI using
tissue doses that account for "multiple pharmacokinetic interactions occurring among
the mixture constituents." The equation used for a mixture of n chemicals is:
where:
th
TMj = tissue dose of the i mixture constituent estimated by the PBPK model for
the human exposure level
TRj = tissue dose of the ith mixture constituent estimated by the PBPK model for
a human "safe level."
The authors compared the interaction-based HI computed for central nervous system
effects using Equation 4-4 with the conventional HI (computed using internal doses)
over a range of exposure concentrations for different mixtures of dichloromethane,
benzene, toluene, ethylbenzene and m-xylene, showing greater than additive effects at
the higher total dose levels of the mixture. Such uses of PBPK models can improve the
way chemical mixture risk assessments are conducted.
4.4. CHEMICAL MIXTURES GROUPING AND TOXICITY ASSESSMENT SCHEME
The object of grouping chemicals for toxicity assessment is to take advantage of
established chemical mixtures risk assessment approaches that rely on groups made
up of individual chemicals that act through a common toxic mode of action or,
conversely, are toxicologically independent of one another (while sharing a common
toxic endpoint). In cumulative risk assessment, the initial four exposure categories
group chemicals by exposures in the same or different media and at the same or
4-18
-------
different point in time (see Section 3.5.2.2). This chapter begins with those rough
exposure groupings and further evaluates them to form revised groups based on
toxicological similarity based on common mode of action or, in cases where data are
sparse, on common target organ. A systematic approach is presented to evaluate
these chemical groups using cumulative risk assessment methods.
Grouping chemicals by the potential for co-occurrence and joint toxic action is a
key simplifying concept for the conduct of cumulative risk assessments. Chemical
components of mixtures can be screened for inclusion in a cumulative risk assessment
using the elements of component-based methods. Figures 4-7a, 4-7b, 4-7c and 4-7d
outline a process for classifying chemicals into groups suitable for analysis and then
applying the methods shown in Figures 4-2a and 4-2b. These steps are
1) Figure 4-7a (same as Text Box 3-10) - Classify all chemicals of concern
into initial groups by their potential to occur in the same or different media
and at the same or different time. (See Chapter 3 for details on exposure
assessment; Section 3.3.2.2 for information on exposure grouping.)
2) Figure 4-7b - Divide these exposure/time groups further into subgroups in
which chemicals are thought to cause toxicity by the same mode of action
or affect the same target organ. Include all target organs or effects for
which positive evidence exists of adverse health effects. An initial step
here is to collect toxicological and pharmacokinetic data on each of the
individual chemicals to be considered in the risk assessment. Factors to
consider in forming these toxicity groups include pharmacokinetic
parameters, persistence of the chemicals in the body and the formation of
metabolites. Note that common toxic mode of action is the preferred way
to categorize chemicals into groups for analysis of combined toxicity.
However, when such data are not available, common target organs can be
used, but with less confidence in the results. A discussion of the data and
decisions used to group chemicals is included in the risk characterization.
3) Figures 4-7c and 4-7d - Assess the toxic potential of the chemicals/whole
mixtures of concern using methods in Figures 4-2a and 4-2b. Figure 4-7c
shows a flow chart that first evaluates the whole mixtures and single
chemicals for toxicity potential, ensuring that those with the greatest
potential to cause toxicity are maintained in the cumulative risk
assessment. Then, the chemical groups formed in Figure 4-7b are
evaluated for joint toxicity, addressing multiple effects, interactions and
exposure routes; these groups are then screened into or out of the
cumulative risk assessment. Figure 4-7d provides additional detail on the
processes shown in Figure 4-7c, indicating the methods and outputs from
this data analysis.
4-19
-------
Chemical Groupings by
Co-occurrence in Media/Time
Time
Same
Different
Media
Same
Group 1
Group 2
Different
Group 3
Group 4
FIGURE 4-7a
Chemical Grouping by Co-occurrence in Media and Time
Because of
Exposure
Group
Consider
These
Factors to
Form
Toxicity
Groups
Exposure Groups
Same Media;
Same Time
Similar effects
or metabolites
Same Media;
Different Time
Similar effects or
metabolites; Body
burden;
Persistence of
effects
Different Media;
Same Time
Similar effects or
metabolites;
Pharmacokinetics;
Multi-route
exposures
Different Media;
Different Time
Similar effects or
metabolites; Body
burden,
Pharmacokinetics;
Persistence of effects;
Multi-route exposures
Chemicals in Exposure Groups (Above) Further Grouped Based on Similar Toxicity
Kidney
Liver
.
.
.
Lung
Group 1,1
Group 1,2
.
.
.
Group 1,n
Group 2,1
Group 2,2
.
.
.
Group 2,n
Group 3,1
Group 3, 2
.
.
.
Group 3,n
Group 4,1
Group 4,2
.
.
.
Group 4,n
FIGURE 4-7b
Chemical Groupings by Common Target Organs and Effects. Each exposure group is
subdivided based on commonality or overlap of toxic effects, metabolic pathways or
tissue concentrations. Chemicals are retained for assessment if information exists on
their toxicological interactions.
4-20
-------
Component or
Whole Mixtures
Data for Toxicity
Group(s)?
Component
Calculate Single
Chemical
HQ's, Cancer Risks,
Public Health Data
Apply Component Mixture Risk Assessment
Methods to Toxicity Group(s)
Apply Whole
Mixture Risk
Assessment
Methods
to Toxicity
Group(s)
Optional Evaluation of Multiple Effects
Optional Evaluation of Multiple Route Exposures
Optional Evaluation of Interaction Effects
Any HQ>1 or
cancer risk >10-6?
Or Exceeding of
Public Health
Levels?
Continue with
Group(s)
Add Single
Chemicals to
Cumulative
Risk
Assessment
Any Hl>1,
Health Risks >10~
Odds Ratios >1?
Screen out Group(s)
from Cumulative
Risk Assessment
Conduct Cumulative Risk
Assessment for Single
Chemicals and Group(s)
FIGURE 4-7c
Grouping Chemicals for Cumulative Risk Assessment. The mixture risk methods are applied to each group, with
"concern" judged by the appropriate screening value (e.g., mixture RfD for whole mixture oral exposure). Groups can be
screened out only if both whole mixture and component methods indicate no concern.
4-21
-------
Apply Component Mixture Risk Assessment
Methods to Toxicity Group(s)
Optional Evaluation of Multiple Effects
Methods: Multivariate Modeling (e.g., Categorical
Regression, Multivariate Normal Linear Regression)
Outputs: Use Results in Hazard Index or Response Addition
Apply Whole Mixture Risk Assessment Methods to
Toxicity Group(s)
Methods: Calculate Mixture RfD/C or Estimate
Risks, Conduct Epidemiologic Study
Outputs: HQ, Risk Estimate, Odds Ratios or Other
Epidemiologic Relative Risks Measures
Optional Evaluation of Multiple Route Exposures
Methods: Cumulative Hazard Index (CHI),
Sum of Risks from Route Specific RPF's,
Cumulative Relative Potency Factors (CRPF),
PBPK Model Estimates of Internal Doses;
Outputs: Use Results to Indicate Risk Potential (CHI)
or Estimate Health Risks (CRPF, PBPK, RPFs)
Optional Evaluation of Interaction Effects
Methods: Locate Interactions Data (e.g., ATSDR
Interaction Profiles, ARCOS & MIXTOX Databases,
Journal Articles on Toxicological Interaction Studies);
Outputs: Use Results to Qualitatively Assess Potential
Interactions or Calculate the Interaction-Based Hazard Index
Outputs:
Risk Indicator
of Concern?
e.g., CHI, HI or
HI|NJ>1,
Qualitative
Judgment of
Interaction
Potential
Outputs:
Risk Estimate
of Concern?
e.g., Health
Risks >10-6,
Elevated
Odds Ratios
No
Yes
Screen out
Group(s)
from Cumulative
Risk Assessment
Conduct
Cumulative
Risk Assessment
for Single
Chemicals
and Group(s)
Conduct Thorough
Risk Characterization,
Uncertainty Analysis
FIGURE 4-7d
Grouping Chemicals for Cumulative Risk Assessment (cont). Specific mixture risk methods are applied depending on
which multiples are being evaluated, with "concern" judged by the appropriate screening value as determined during the
Problem Formulation phase of cumulative risk assessment.
4-22
-------
4.4.1. Chemical Groupings by Common Effects. The groupings developed in the
exposure analysis (Figure 4-7a) categorize multiple chemicals into groups comprised
roughly of exposures in the same or different media at the same or different exposure
time (see Section 3.3.2.2). Note that many exposure groups could be formed when
multiple exposure media and timeframes are found to be important to the assessment.
Figure 4-7b shows that for each media/time combination, the occurring chemicals are
grouped by common target organ or effect, which does not necessarily imply a common
toxic mechanism or MOA. Because the exposure scenarios vary with media and time,
factors relating to exposure routes and fate within the body are then considered to
further refine the subgroups for the toxicity assessment (see Figure 4-7b). Through
consultations among exposure analysts and toxicity analysts, several different
groupings can be developed based on available exposure and toxicity data. In addition,
most chemicals are likely to end up in several different groups because they can exist in
more than a single medium, and they cause more than one toxic effect in different target
organs. (Text Box 4-2 discusses the availability of EPA toxicity information beyond IRIS
values for use in the cumulative toxicity assessment.)
An example of the grouping process
can be seen using the information shown in
Figures 4-8 and 4-9. In Figure 4-8, several
organ systems are represented (i.e., the
nervous, renal, cardiac, developmental,
respiratory systems), with specific target
organs indicated in the second row. The
third and fourth rows list chemicals causing
primary or secondary effects in those
systems, respectively (see Tables B-1 and
B-3 of Appendix B for chemical toxicity
information). A primary effect is the adverse
effect observed at the lowest dose on the
dose-response relationship developed for
each adverse effect noted from single
chemical exposures. Secondary effects can
be thought of in several ways: effects
mediated by chemical metabolites, effects
that follow from chemical insult but do not result in adversity (e.g., enzyme induction), or
adverse effects that occur at doses higher than those producing the critical effect.
Target Organ Toxicity Doses (TTDs)
(Text Box 4-2)
The EPA's IRIS database generally derives
an oral RfD based on a single critical effect
(i.e., the first adverse effect, or its known
precursor, that occurs to the most sensitive
species as the dose rate of an agent
increases) for a single chemical. Thus,
cumulative toxicity assessments using
secondary effects require the development of
additional dose response information beyond
readily available EPA values. EPA (2000a)
suggests the development of TTDs for use in
these situations. TTDs are developed for
secondary effects using the same
methodology as applied in the derivation of
an RfD (Mumtaz et al., 1997). At this point in
time, the TTD methodology has only been
proposed for noncancer endpoints and for
oral exposures. TTDs can then be used in
HI calculations instead of using an RfD to
represent a safe level for all target organs.
The alternative is to use the IRIS RfD
regardless of target organ, resulting in a
likely overestimation of the HI.
4-23
-------
Brain
Hg, PCB,
TCE
As,
Hg
DCA
CCI4,
, Ni,
, TCE
Spinal
cord
CCI4,
Hg, TCE
Organ
System
Specific
Target
Organ
Chemicals
Causing
Primary
Effects
Chemicals
Causing
Secondary
Effects
Sources:
Municipal Waste Combustor: Hg, Cd
Fish Consumption: Hg, PCB
Drinking Water Disinfection By-Products (DBPs): BDCM, DCA
Source Water Contaminants: TCE, Ni, As, CCI4, Cr
Contaminated Groundwater: U
Temporary Combustor for Site Remediation: Cd, Cr, Ni
As = Arsenic (inorganic)
BDCM= Bromodichloromethane
Cd = Cadmium
CCI4 = Carbon tetrachloride
Cr III = Chromium III (insoluble salts)
CrVI = Chromium VI
DCA = Dichloroacetic Acid
Hg = Mercury (based on mercuric chloride)
Ni = Nickel (soluble salts)
PCB = Polychlorinated Biphenyls (Arochlor 1016)
TCE = Trichloroethylene
U = Uranium (soluble salts)
FIGURE 4-8
Information on Primary and Secondary Effects Linked with Hypothetical Exposure Sources to
Show Example Chemical Groups
(see Appendix B, Tables B-1 and B-3 for chemical information sources)
4-24
-------
Exposure
Group:
Exposure
Scenarios:
Exposure Groups
Same Media;
Same Time
Air: Daily
Exposure to
Municipal Waste
Combustor
Emissions
Air: Daily
Inhalation
Exposure to
Disinfection By-
products via
Showering
Same Media;
Different Time
Drinking Water:
Acute Accidental
Exposure to
Source Water
Contaminants
Drinking Water:
Exposure to
Uranium
Contaminated
Ground Water,
Years Later
Different Media;
Same Time
Drinking Water:
Daily Exposure to
Disinfection By-
products via
Ingestion and
Showering
Fish: Daily
Exposures via
Local Fish
Consumption
Different Media;
Different Time
Air: Short Term
Exposure to
Emissions from
Temporary
Combustor
Drinking Water:
Acute Accidental
Exposure to Source
Water
Contaminants,
Months Later
Chemicals in Exposure Groups (Above) Further Grouped Based on Similar Toxicity
Kidney
Brain
Fetus
Heart
Lung
Hg, Cd, BDCM
Hg, DCA
Hg, BDCM, DCA
Hg, Cd
Hg
Ni, TCE, U, Cr
TCE, As, Ni, CCI4
TCE, Ni, Cr
TCE, Ni, As, Cr
Ni, Cr
Hg, BDCM
Hg, DCA, PCB
Hg, BDCM, DCA,
PCB
Hg
Hg
Cd, Ni,TCE, Cr
TCE, As, Ni, CCI4
TCE, Ni, Cr
Cd, TCE, Ni, As, Cr
Ni, Cr
FIGURE 4-9
Hypothetical Example of Chemical Groupings by Co-occurrence in Media and Time, Similar Toxicity
4-25
-------
Following these rows is a list of six hypothetical exposure sources under consideration
for a cumulative risk assessment and a list of the associated contaminants to which the
population is exposed. This information is then used to form initial toxicity groups in
Figure 4-9, which begins by setting up hypothetical exposure scenarios for each
combination of same/different media and same/different time. The target organ specific
toxicity groups in Figure 4-9 are developed by distributing the chemicals associated with
the hypothetical exposure sources (Figure 4-8) into the five bottom rows that designate
specific target organs, according to the combinations of these sources shown in the
media/time exposure scenarios. In this way, contaminants that are expected to co-
occur in media and time are grouped by common target organ for analysis. For
example, in the first column, the population is exposed via inhalation to municipal waste
combustion emissions and drinking water DBFs through showering, so the chemicals
associated with these two sources are grouped by common target organ.
4.4.2. Refinement of Toxicity Groups. Once these initial groups are formed, then
several other factors are accounted for before the groups are subjected to a risk
assessment procedure. At this point, the chemicals within each group do not
necessarily act by the same toxic mechanism or mode of action and have not been
considered yet in terms of whether the exposure levels are within ranges that may
cause toxicity, additive joint toxic action or toxicological interactions. These groups are
refined using considerations of appropriate exposure routes, timing of exposures and
effects, persistence of chemicals within the body and the potential for joint toxic action.
This refinement results in final chemical groupings that are ready for analysis using
chemical mixture risk assessment methods. The following issues are considered:
Given the exposure routes and health effects of concern, are the chemicals in the
toxicity groups appropriate?
Example: For the Same Media/Same Time exposure scenario, DCA is a non-
volatile DBP that would not volatilize, but would be found in aerosol (water
particles) during showering. Because of the relatively low level of exposure via
inhaled aerosols during showering, it could be removed from the toxicity groups.
Also, BDCM is known to cause renal effects via inhalation, but the toxicity data
on fetal loss are from oral exposures, with no developmental data available for
inhalation exposures; thus, because of the potential for a large inhalation
exposure to BDCM during showering and because fetal loss is a severe effect, it
would be reasonable to retain BDCM in the "fetus" grouping, but this uncertainty
is then discussed in the Risk Characterization phase.
4-26
-------
Do data exist on toxicological interactions between chemicals in the groups that
would raise concerns for increased (or decreased) toxicity from the joint
exposure?
Example: Data exist that show a synergistic interaction effect in the brain for joint
exposures to TCE and CCI4 (ATSDR, 2003a). This relationship is only
documented for this one toxic effect. It is reasonable, however, to keep both
chemicals listed within all toxicity groups when the exposure scenario indicates
they will co-occur. Thus, in Figure 4-9, both TCE and CCU would be added to all
toxicity groups under exposure scenarios involving the contaminated ground
water source.
Are there metabolites that should be added to the groups and, if so, should the
parent compound be retained or removed?
Example: Although this exposure scenario is not shown in Figure 4-9, suppose a
same media/same time scenario involves co-exposures to the DBP, DCA and the
source water contaminant, TCE. Because DCA is a metabolite of TCE in the
body and both chemicals are known to cause effects in the brain, exposures to
both chemicals could result in elevated levels of DCA for consideration in the risk
assessment. If it cannot be determined whether or not TCE would still be
present or instead be completely metabolized, it may be reasonable to also retain
TCE in the risk assessment, but this uncertainty is then noted in the Risk
Characterization discussion.
When the population is exposed to sources at different times, do the chemicals
from the first exposure remain in the body long enough to be of concern when
the second exposure occurs?
Example: The potential for toxic interactions of Cd and TCE on the
cardiovascular system may be based on direct interactions in the heart itself, and
by additional, indirect, effects of Cd and TCE on kidney function related to blood
pressure regulation. Both TCE and Cd are readily absorbed into the body. TCE
is eliminated from the body with a half-life measured in hours, whereas Cd is
eliminated from the body with a half-life measured in decades; thus an earlier
exposure to Cd may result in persistent body burdens, and internal co-exposure
with TCE in tissues. The tissue concentrations and the effects of Cd in the heart
and kidney may persist beyond the initial exposure period, making these organs
more susceptible to the injury produced by TCE.
When the population is exposed to sources at different times, do the health
effects resulting from the first exposure last long enough to be of concern when
the second effect from the subsequent exposure occurs?
Example: As shown in Text Box 3-10, benzo[a]pyrene (BaP) and TPA are an
initiator/promoter pair. TPA does not have a tumorigenic effect in mouse skin
assays, but when it is applied after initiation with BaP tumorigenic activity is
greatly enhanced (Verma et al., 1985).
4-27
-------
Figure 4-10 illustrates a few of the changes (not comprehensive) that would be
made in Figure 4-9 based on the points raised in this section. Considerations of body
burden, pharmacokinetics, exposure route, persistence of effects, metabolites and
multi-route exposures may be used to alter and refine the toxicity groups. When the
groups are finalized then the analyst can move forward to conducting the cumulative
toxicity assessment.
4.4.2.1. Uncertainties and Data Gaps in Grouping Chemicals The amount
of data needed for grouping chemicals and analyzing risks for cumulative risk
assessment may be significant, particularly when multiple toxic effects and exposure
routes are of concern. However, lack of data for certain chemicals or exposure
durations is not unique to cumulative risk assessment. Furthermore, uncertainty due to
extrapolations (e.g., from animals to humans), potential variability in response due to
differential susceptibility of some individuals or subgroups, and knowledge gaps, are
incorporated into IRIS or other (e.g., AEGL) reference value determinations. Options for
addressing uncertainties associated with extant toxicity values as well as uncertainties
associated with lack of toxicity values for key chemicals (i.e., data gaps) are similar to
those used in site risk assessments conducted under existing guidance. EPA (e.g.,
U.S. EPA, 1989a), NRC (e.g., NRC, 1994) and OMB (2006) clearly state the importance
of providing a full and open discussion of uncertainties in a risk assessment, including
identification of the sources and magnitude of uncertainty associated with the risk
estimates. For chemicals that are critical to an analysis, EPA (or other agencies) can
be asked to develop toxicity values for the route of exposure(s) and exposure
duration(s) of interest. For cumulative assessments, it is also important to assess
secondary effects, i.e., those seen at doses above that producing the critical effect on
which the standard toxicity value is based. Toxicity values can be derived if the
underlying toxicity studies needed to complete the evaluation are available. In the
absence of an adequate toxicological database, expert judgment may be used, along
with quantitative structure activity analysis, with associated identification and
characterization of uncertainty. Finally, as a long range strategy, chemicals may be
prioritized for toxicological testing.
4.4.3. Cumulative Toxicity Assessment Scheme. After the joint exposure and target
organ groups are determined, the toxicity assessment for each group can then follow
the schematic shown in Figure 4-7c. This flow chart begins in the same way as Figure
4-2a and 4-2b in that the risk analyst examines the available data for toxicity information
4-28
-------
Exposure Group:
Exposure
Scenarios:
Exposure Groups
Same Media;
Same Time
Air: Daily Exposure to
Municipal Waste
Combustor Emissions
Air: Daily Inhalation
Exposure to Disinfection
By-Products via
Showering
Same Media;
Different Time
Drinking Water: Acute
Accidental Exposure to
Source Water
Contaminants
Drinking Water:
Exposure to Uranium
Contaminated Ground
Water, Years Later
Different Media;
Same Time
Drinking Water: Daily
Exposure to Disinfection
By-Products via Ingestion
and Showering
Fish: Daily Exposures via
Local Fish Consumption
Different Media;
Different Time
Air: Short Term Exposure to
Emissions from Temporary
Combustor
Drinking Water: Acute
Accidental Exposure to
Source Water Contaminants,
Months Later
Exposure-Toxicity Groups Refined Based on Interactions, Metabolites, Exposure Routes
Kidney
Brain
Fetus
Heart
Lung
Hg, Cd, BDCM
Hgb
Hg, BDCMb
Hg, Cd
Hg
TCE, Ni, U, Cr,
CCI4a
TCE, As, Ni, CCI4,
DCAC
TCE, Ni, Cr,
CCI4a, DCAC
TCE, Ni, As, Cr,
CCI4a
Ni, Cr
Hg, BDCM
Hg, DCA, PCB
Hg, BDCM, DCA,
PCB
Hg
Hg
Cd, TCE, Ni, Cr,
CCI4a
TCE, As, Ni, CCI4,
DCAC
TCE, Ni, Cr, CCI4a,
DCAC
Cd, TCE, Ni, As,Cr,
CCI4a
Ni, Cr
a CCI4 added to account for potential interaction effects between CCI4 and TCE.
b DCA removed because it is not a volatile compound; inhalation exposures are not a concern.
c DCA added as a metabolite of TCE.
FIGURE 4-10
Examples of Toxicity Group Refinements
4-29
-------
on the whole mixture and on the mixture components. The whole mixtures and single
chemicals are first evaluated for toxicity potential; those with the greatest potential to
cause toxicity are maintained in the cumulative risk assessment. The whole mixtures
may be evaluated according to the methods in Figure 4-2a (Section 4.3.3.1). Then, for
the toxicity groups in Figure 4-9, it is not likely that toxicity data would be available for
those specific chemical combinations, so the risk analyst would follow the flow chart in
Figure 4-2b for evaluation of component data. Initially, if data are available for each of
the single chemicals in a toxicity group, then the single chemical hazard quotients and,
if applicable, cancer risks are calculated. Public health levels for these chemicals are
also collected and checked against environmental levels. If calculations show any HQ
>1 or cancer risk >10~6 or if a public health level is exceeded, then that single chemical
is designated to remain in the cumulative toxicity assessment. (It is not removed from the
toxicity group.) The next step is to apply the component-based chemical mixture risk
assessment methods (flow chart in Figure 4-2b) to each toxicity group, using the HI
(Section 4.2.1), response addition (Section 4.2.1) or RPF (Section 4.7.1.2) approaches
as appropriate, according to the judgments made regarding toxicological similarity of the
component chemicals (see U.S. EPA, 2000a, for details on applying these methods).
Finally, the optional quantitative methods detailed in Figure 4-7d may be undertaken to
evaluate multiple effects (Section 4.5), toxicological interactions (Section 4.6) and
multiple route exposures (Section 4.7). If quantitative data are not available to conduct
the analysis, but qualitative toxicity information exists, then some qualitative discussion
of these issues may be possible. If none of these mixtures assessments raises concern
for population health risks, then the toxicity group may be screened out of the
cumulative toxicity assessment. Otherwise, the risk analyst retains both the toxicity
group(s) and the single chemicals with elevated HQs, cancer risks or public health
levels, as well as any whole mixtures that may have the potential to cause toxicity and
finalizes the cumulative risk assessment, including a complete Risk Characterization
(Chapter 5).
4.4.3.1. Evaluation of Whole Mixtures DataWhen data on the toxicity group as a
whole mixture are available, the risk assessment can use that information to estimate
health risks for the toxicity group. Also, within the toxicity group, there may be a
complex mixture with a chemical composition that is not fully characterized (e.g.,
complex disinfection by-product mixtures typically contain ~50% of unidentified total
organic halide material). Toxicity may be estimated for the whole mixture as shown in
Figure 4-2a (see procedure in Text Box 4-3) and compared with environmental
4-30
-------
exposure levels. For example, an
RfD can be calculated for the whole
mixture (RfDm) as shown for the
general case and compared to the
IRIS value for Araclor 1016 in Figure
4-11. TheArachlor1016RfDm
represents that particular PCB
mixture and could be used in the
cumulative toxicity assessment as a
surrogate value for the PCB exposure
via fish consumption with the relevant
toxicity groups for effects in the brain
and fetus. Returning to Figure 4-7c, if
the whole mixture toxicity is shown to
be of concern, then it remains in the
cumulative toxicity assessment.
Procedure for Estimating Whole Mixture Toxicity
Values (Text Box 4-3)
1) Collect and Evaluate Data
Epidemiology/human data preferred, supporting
toxicology data
2) Evaluate Stability within a Mixture
Variability in components and their relative
proportions
3) Assess Sufficient Similarity Across Mixtures
(if applicable)
Similarity across mixtures' components and relative
proportions
Similar toxicity of two mixtures or of common
components
Common sources or produced by similar process
4) Conduct Dose-Response Assessment
Use same procedures as for single chemicals (e.g.,
RfD, slope factors)
5) Characterize Uncertainties
Relevance of health effects data to environmental
exposures
Stability of the mixture and environmental fate
(U.S. EPA, 2000a)
4.4.4. Evaluating Subpopulations. Information on vulnerable subpopulations may be
collected and included in the cumulative risk assessment when such information is
available. An extensive treatment of how to incorporate such information into the
cumulative risk assessment will not be described in this report, but future research on
this aspect of cumulative risk assessment may contribute insights and is encouraged.
The Agricultural Health Study and other literature on mixture exposures and potential
susceptibilities related to environmental exposures (see Chapter 1) will become useful
data sources in the future for identifying vulnerable subpopulations of concern when
conducting a cumulative risk assessment. In the development of chemical groups for
evaluation at a site, the characteristics of the potentially exposed population may be
evaluated (Chapter 2). In order to conduct an initial evaluation of the subpopulation
health impacts, chemical mixture exposure and risk estimates for vulnerable
subpopulations may be calculated separately from risk assessments on the general
population and presented in a separate section of the Risk Characterization.
4.5. EVALUATING MULTIPLE EFFECTS
The hazard identification phase of a cumulative risk assessment is broadened to
include factors beyond those considered for single chemicals. An important difference
between cumulative risk assessment and traditional single-chemical assessments is the
4-31
-------
General Case (U.S. EPA, 2000c)
Aroclor 1016 (U.S. EPA, 2005c)
NOAEL LOAEL or BMDL
UF
m
where:
NOAEL/LOAEL = No/Lowest-Observed-
Adverse-Effect Level
BMDL = Lower 95% confidence limit on an X%
Effective Dose (e.g., ED10)
UFm = Uncertainty Factors for the mixture
(e.g., interspecies, intraspecies,
exposure duration, NOAEL to
LOAEL, data base deficiencies)
NOAEL, LOAEL or BMDL from experimental
toxicity data on the complex mixture dose-
response. Uncertainty factors are derived using
expert judgment, as is the case for single
chemicals. The uncertainty characterization
should include the relevance of the experimental
mixture from which the RfDm is derived to the
chemical composition of environmental mixtures.
IE-5 =
where:
NOAEL =
UF =
NOAEL = 0.007 mg/kg/d
UFm = 100
Reduced birth weight in monkey
reproductive study
3 for rhesus monkey to human
extrapolation
3 for infants as a sensitive
subpopulation
3 for subchronic to chronic
exposure duration
3 for missing 2 generation repro &
adult male repro studies
(i.e., 100 = 3x3x3x3, rounded up)
Confidence in RfD is medium when PCB
mixtures in the environment do not match the
pattern of congeners found in Aroclor 1016; high
if the environrmental mixture is Aroclor 1016.
FIGURE 4-11
Complex Mixture Reference Dose
4-32
-------
number of health effects evaluated. In the assessment of chemical mixtures, secondary
health effects may be observed as a result of combined chemical exposures that are
different in phenotype or magnitude from the critical (primary) effects caused by the
chemicals individually. These secondary health effects may occur at doses or
exposures higher than those causing the critical effect. Conversely, observed toxic
threshold(s) or effect level(s) (e.g., a LOAEL) for single chemicals may be altered such
that the dose(s) required to elicit the same and/or additional effect(s) may be less when
the exposure is to a mixture (e.g., Nickel causes increased sensitivity to Cobalt-induced
dermal allergy). Thus, it is important to evaluate secondary effects for those chemicals
to which humans may be exposed in combination. In these cases, the doses of the
chemicals in the mixture may act in an additive manner to cause one of these
secondary or higher level effects, or the responses (effects or risks) themselves may be
additive. In addition, co-exposure to these chemicals may result in toxicological
interactions (e.g., synergism or antagonism) related to a secondary or higher level
effect. The method described in Figures 4-7a, 4-7b, 4-7c and 4-7d shows that the
cumulative risk assessment includes an evaluation of all adverse effects, as evidenced
by available health effects data (e.g., toxicology data, public health data and
epidemiology studies). Finally, the set of identified effects takes into account the
potential routes of exposure.
The application of the toxicity assessment to actual site exposures will often
require extrapolation beyond the range of concentrations (exposures) used to develop
toxicity data in test animals. When external exposure levels are used in the risk
assessment, then inferences about multiple effects may be highly uncertain. When data
are available and resources permit a more extensive investigation, internal chemical
doses may be developed or inferred from pharmacokinetic and mechanistic
information. The issue of whether to express exposure in external or internal terms
becomes important when the relationship between exposure concentration and
internal concentrations (tissue dose) is nonlinear or has not been characterized.
Another level of complexity can be avoided when internal doses are used to evaluate
the response in multiple-organ systems, such as the immune system. Here, tissue
concentrations may vary appreciably among the multiple organs involved, and those
tissue concentrations may or may not be linearly related to external exposures.
Chemicals that affect organs or tissues that are parts of a larger biological system
may be considered as affecting the same target system. Finally, exposure to chemical
mixtures may result in toxicokinetic alterations in the body that may alter the internal
4-33
-------
dosimetry and tissue distribution of chemical components. In this way, the assessment
of multiple effects can be simplified by grouping the effects.
4.5.1. A Quantitative Method for Evaluating Multiple Effects. One of the goals in a
cumulative toxicity assessment is to account for the joint impact of all of the major
health impacts from exposure to multiple stressors. The approach demonstrated in this
report involves a three step process: a dose-response model for multiple effects, hazard
calculations using both dose-addition (HI) and response-addition approaches and a
comparison of the results. This approach begins by analyzing dose response
relationships for each single chemical and incorporating all toxic effects in the same
modeling procedure. Various statistical models could be applied (e.g., multivariate
normal linear regression or ordinal categorical regression) to predict the probability of
observing an array of toxic effects for a given dose. For many chemicals, the available
data on multiple effects differ across effects as well as across chemicals in terms of
completeness, range of doses covered and level of detail, making multivariate
approaches difficult. In this report, a simpler categorical regression model based on
toxicological judgment will be used to illustrate estimating the probability of a certain
severity level of (non-specific) response that can represent a number of different toxic
effects, given exposure to a single chemical. From the modeling results, a risk estimate
for the exposure of interest can be made for that single chemical, or a benchmark dose
(BMD) can be estimated (e.g., a 5% effective dose or ED0s). To apply dose addition,
this modeling approach is conducted for each of the chemicals in the mixture, and a HI
is calculated by summing the ratios of each chemical's exposure to its BMD, which
provides an indication of risk for the mixture. To apply response addition, the
categorical regression model can be used to predict the risk of an adverse effect for
each individual chemical at its environmental exposure level; these risk estimates can
then be summed across chemicals to calculate the mixture risk. These results can be
compared in the Risk Characterization phase (see Chapter 5) to evaluate the potential
health impacts for the exposure scenario of interest.
Ordinal categorical regression is a statistical modeling procedure that allows for a
dose-response assessment of several toxicological effects at once. The use of a
categorical regression procedure to express the risk of adverse health effects for
toxicological data was first proposed by Hertzberg and Miller (1985) and Hertzberg
(1989) and then demonstrated with several chemicals (Dourson et al., 1997; Farland
and Dourson, 1992; Guthetal., 1991; Raoetal., 1993; Teuschler et al., 1999;
Strickland and Guth, 2002).
4-34
-------
In this procedure, toxicity data, regardless of the type of effect, are interpreted
using toxicological judgment in terms of pathological staging. Toxic effects, which may
include both quantal and continuous data, are classified into ordered categories of total
toxic severity, e.g., categories 1-4 refer to none, mild, moderately adverse and severe
effects, respectively (see Appendix C for further discussion of severity of toxic effect).
The model reflects a regression of dose on the category of effect, yielding the
probability that a given dose will result in a level or category of response (e.g., the
probability of observing a level 3 adverse effect, given dose). The EPA's software,
CATREG, is useful for conducting this procedure (U.S. EPA, 2000c,d). In addition,
CATREG has the ability to incorporate other factors in the analysis, including duration,
study effects, species and censored data (Guth, 1996; Guth et al., 1991, 1997). Thus,
models may be developed to describe dose-risk relationships for a variety of exposure
scenarios.
To illustrate the modeling procedure, an example is shown here from Dourson et
al. (1997), where categorical regression analysis was used to model human clinical data
to describe the relationship between the logarithm of doses and severity levels of
cholinesterase inhibition for the pesticide, aldicarb. Table 4-2 shows the four, ordered
categories of toxic severity for cholinesterase data that were used to classify the
response data, along with the clinical effects expected to be observed at each severity
level. In Dourson et al. (1997), results from two human clinical studies available on
aldicarb, Haines (1971) and Wyld et al. (1992), were evaluated using the criteria in
Table 4-2, and each subject was placed in a severity category as summarized in Table
4-3. Both studies had similar experimental designs (see summaries in Dourson et al,
1997). A categorical regression model was developed to predict the probability that a
subject would exhibit a certain severity level of cholinesterase inhibition, given the
exposure dose.
In the catergorical regression model, the toxic response was related to the
explanatory variable, logdose, using a logistic function and Pwas defined as the
probability of observing a response of a certain severity or a lesser response. The
logistic function used to express the relationship between P and the explanatory
variable, dose, is given below:
4-35
-------
TABLE 4-2
Severity Assignments for Cholinesterase Inhibition Data
(Adapted from Dourson et al. ,1997)
Severity
Category
4
Frank
Effects
3
Adverse
Effects
2
Non-
Adverse
Effects
1
No Effects
Site
Cholinergic effects
Cholinergic effects
Whole Body
Brain, whole blood or red
blood cell (RBC)
acetylcholinesterase
Cholinergic effects
Cholinergic effects
Nervous system
Plasma, whole blood or
RBC acetylcholinesterase
All
Effect
Severe abdominal pain, nausea and/or
vomiting, diarrhea
Seizures, severe disorientation or
confusion, excitation
Mortality
Inhibition (e.g., of 20% or greater)
Mild: Muscular weakness or twitching
Mild: Blurred vision and/or watery eyes,
pinpoint pupils, excess salivation,
sweating or clamminess
Hyperactivity or altered patterns of
locomotion
Inhibition (e.g., observed, but less than
20%)
No effect
4-36
-------
TABLE 4-3
Frequency of Categories of Effect Associated with Aldicarb Exposure in Humans
(Adapted from Dourson et al. ,1997)
Study
Haines,
1971
Wyldetal.,
1992
Dose
(mg/kg/day)
0.025
0.050
0.10
0.0
0.010
0.025
0.050
0.075
Group
Size
4
4
4
22
8
12
12
4
Frequency of Responders within Categories of*
No
Effects
0
0
0
22
8
2
0
0
Non-
adverse
Effects
0
0
0
0
0
9
9
0
Adverse
Effects
4
4
2
0
0
1
3
4
Frank
Effects
0
0
2
0
0
0
0
0
*Numbers reflect a judgment that whole blood (Haines, 1971) or red blood cell (Wyld et
al., 1992) cholinesterase inhibition of 20% or greater is considered an adverse effect.
This percentage can be debated and is a source of uncertainty for the analysis.
4-37
-------
where:
P = the probability of observing an effect of severity i or less,
s = the severity of the effect,
/ = the severity category 0, 1, 2 or 3,
otj = an unknown intercept parameter associated with severity i,
P = an unknown slope parameter associated with the dose,
D = the logdose of the chemical, aldicarb.
Table 4-4 shows the results of the regression modeling. Using the values in this
table and rounding down, a hypothetical example of a 10% BMD level for use in
developing a HQ (and subsequently using this in a HI calculation) would be roughly
equal to 0.02 mg/kg. (It may be noted that a lower bound on the BMD, a BMDL, would
typically be used in the HQ calculation, but these values are not shown in Dourson et al.
[1997]; thus the BMD is used here for illustration.) Alternatively, if a hypothetical
exposure in a community were equal to 0.01 mg/kg, then the upper bound human risk of
~4% in the table could be used in a response addition calculation for that risk
assessment.
This categorical regression procedure can be expanded beyond a single group of
toxic effects to include other effects whose severity is judged to be of a similar nature
and level (e.g., Table 4-2 could be expanded to include severity judgments for liver and
kidney effects along with cholinesterase inhibition). In addition, duration can be
included as a second dependent variable in the model. Using this procedure, the dose-
response relationship for multiple effects can be modeled and shown as the probability
of toxic effects for a given duration and dose (e.g., the probability of an adverse effect
for a 1-day exposure at 0.1 mg/kg/day), and BMDL estimates can be determined (e.g.,
lower bound on the dose causing a 5% chance of a non-adverse effect). Results of the
categorical regression equation can then be used in response addition and the HI to
present a range of potential health risk for the exposure of interest. In particular, using
Equation 4-1 (from Section 4.2.1) for the HI, the RfD for each chemical can be replaced
by the BMDL for multiple effects divided by an uncertainty factor (e.g., UF = 100) to
account for inter- and intra- species differences. The resulting equation for the multiple
effects HI, for chemicals k= 1, 2,...,n, and exposures Ek, would be:
Hl(effects) =
n
S
(4-6)
4-38
-------
TABLE 4-4
Modeled Probabilities of an Adverse or Frank Effect
Dose (mg/kg)
0.001
0.003
0.01
0.015
0.02
0.025
0.03
0.035
0.04
0.10
Inhibition >20% = Adverse Effect
Mean P(AE or FE)*
0.0014
0.03
0.14
0.44
0.79
0.89
0.95
0.99
Upper 95%CL On P(AE or FE)*
0.00001
0.0007
0.04
0.17
0.36
0.67
0.93
0.97
0.99
1.00
*P(AE or FE) is equal to P(s > 3), i.e., the probability of observing an adverse effect
(severity level 3) or a frank effect (severity level 4), given dose. P(AE or FE) is also
equal to 1 - P(s < 2), i.e., one minus the probability of observing a non-adverse effect or
no effect.
4-39
-------
As an example, suppose chemical A and chemical B are in a mixture with
exposures doses of 8 and 2 mg/kg, respectively. From their individual categorical
regression models of multiple effects, the 10% BMDL's are estimated at 50 and 75
mg/kg, respectively. It is determined that an UF = 100 is appropriate for each chemical.
Therefore, the Hl(effects) = 87(50/100) + 27(75/100) = 16 + 2.7 = -19, which would
indicate potential risk for adverse effects at that environmental exposure.
A probabilistic mixtures risk estimate could also be calculated for multiple effects
using the categorical regression results. Based on Equation 4-2 (and expanding for
more than k = 1 ,..,n chemicals), for ordered severity categories of 1 = no effects, 2 = not
adverse effects, 3 = adverse effects, 4 = frank effects), response addition under
categorical regression for a specific exposure of interest is calculated:
Rm(effects) = 1~Ylpk(severity < 2) (4-7)
k=1
where Pk(severity < 2) is the same as P(s < /) when / = 2, in Equation 4-5 above. These
probabilities can be calculated from the regression modeling results. As an example
using some of the data from Table 4-4, the mean P(s < 2) for a dose of 0.015 mg/kg = 1
- P(s > 3) = 1 - 0.03 = 0.97. Suppose another chemical present in the mixture is
measured at a exposure dose of 0.04 mg/kg and that its categorical regression model
evaluated at that dose yields the P(s < 2) = 0.99. Then, using Equation 4-7 for that two
chemical mixture, Rm(effects) = 1 - (0.97)*(0.99) = 1 - 0.96 = 0.04, the risk estimate for
multiple effects for the mixture exposure of interest.
4.5.2. Interpretation. These two methods for dose-response assessment of multiple
health effects yield very different types of answers. The Hl(effects) is expressed as a
risk indicator and the Rm(effects) is expressed as a probabilistic risk estimate. A group
of chemicals may be screened in as part of a cumulative risk assessment when either
the value of an HI is greater than or equal to some pre-determined level (e.g., 0.5) or a
response addition risk estimate is greater than or equal to an acceptable risk level (e.g.,
1 x 10~6). In either case, when estimates approach or exceed these "cut off" values,
expert judgment of the toxicological significance is used to evaluate the chemicals and
data used in the analysis and to determine the level of concern for the analysis. For a
cumulative risk assessment screening exercise, if either "cut off" value is met or
exceeded, then those chemicals are kept in the cumulative risk assessment. The
factors considered when evaluating dose and response addition in mixture risk
assessments also apply here but only in a rough sense: whether the collection of effects
seem to be toxicologically similar across the set of chemicals or seems to be
4-40
-------
TABLE 4-5
Joint Toxicity: Non-additive Effects of Metal Pairs on Systems/Organs
Using Oral Exposure
Effect of
Metall
on Metals
Arsenic
Cadmium
Chromium
Lead
Not
Additive*
Higher
Lower
Higher
Lower
Higher
Lower
Higher
Lower
Arsenic
Blood
Skin
Kidney
Neurological
Kidney
Blood
Cadmium
Blood
Kidney
Male
reproductive
Male
reproductive
Chromium
Kidney
Lead
Neurological
Blood
Kidney
Neurological
Male
reproductive
Blood
Kidney
* Higher = Effects are greater than expected under additivity
Lower = Effects are less than expected under additivity
Source: ATSDR (2004).
4-41
-------
independent, particularly at the exposure levels under consideration. As described in
the U.S. EPA (2000a) mixture guidance, these formulas give similar results when
component exposures are low.
4.6. EVALUATING INTERACTION EFFECTS
EPA (2000a) defines toxicological interactions as any toxic responses that are
greater than or less than what is observed under an assumption of additivity (e.g., a
departure from dose additivity or response additivity for a group of chemicals). Many
terms are used to represent various kinds of interaction effects (e.g., inhibition,
antagonism, masking). The most common and general of these refer to effects that are
greater than additive (i.e., synergistic) or less than additive (i.e., antagonistic).
The detection of interaction effects varies from toxicological judgment to
statistical determinations. For cumulative risk assessment, interactions information may
be collected from the toxicological and epidemiologic literature and used to inform the
grouping process. EPA has two collections of bibliographic summaries of interaction
studies: the Integral Search System (Arcos et al., 1988) and the MIXTOX database
(Marnicio et al., 1991). ATSDR has also published eleven interaction profiles for
common environmental contaminants (ATSDR, 2006; Pohl et al., 2003). For example,
in Table 4-5, the non-additive interactions are shown for four metals: As, Cd, Cr and Pb
(ATSDR, 2004). As Table 4-5 shows, even when interactions data exist, the situation is
complicated because the direction of interaction can be different for different effects or
for changes in the sequence of exposure. For metals, toxicological interactions are
more troublesome because environmental conditions (e.g., pH) can alter the speciation
and bioavailability of the metals. At a minimum, when evidence of synergistic
interaction is found for two or more chemicals within a group (formed using
Figure 4-7b), those chemicals are included in the cumulative risk assessment. A further
quantitative evaluation may be conducted using the interaction-based HI (see Section
4.6.2 and Chapter 5) or by evaluating interactions using an internal dose HI
(Section 4.3.1).
4.6.1. Toxicology of Interactions. A mixture can consist of chemicals that cause a
unique toxicological expression that was not anticipated from the toxicity of the
individual compounds; the toxicodynamic process of one compound influences that of
another (e.g., one compound causes toxicity and a second compound slows the
process of cellular repair). The toxicity of chemical mixtures is dependent upon the
interactions of mixture components at either toxicokinetic (TK) or toxicodynamic (TD)
4-42
-------
processes, thus, interactions at either level may result in mixtures interactions. TK
processes govern tissue distribution of compounds and include both passive and active
processes. TD processes include the effects or events that are dependent upon the
contact between the toxic chemical species and the biomolecules responsible for the
effect. Interactions at the TK level occur when tissue dosimetry is altered due to gross
tissue alteration or chemicals interact at the same metabolic enzyme.
In addition to separating interactions according to TK or TD, toxicological
interaction among compounds may be direct or indirect. Examples of indirect
interaction include chemicals that may alter the internal dosimetry/metabolism of other
compounds (e.g., enzyme induction, glutathione depletion) and thus exert an indirect
effect on their toxicity. Direct interactions are demonstrated by compounds altering the
same biochemical pathway or cell type or organ/tissue that is directly related to the toxic
effect of the compound. Examples of direct interaction include competition for key
metabolizing enzymes, receptor binding sites and lipid peroxidation leading to
membrane damage and radical formation. Some of these interactions will depend on
the severity of the effect produced. If the effect of the first compound only results in a
slight functional decrement and is recovered quickly or is compensated by the tissue,
then such an effect, whether direct or indirect, may not be sufficient to serve as the
basis for an assumption of interaction. Knowledge that a given effect may be reversible
or compensated for by the cell is coupled with information on the dose-response and
temporal characterization of the reversibility. This applies also to cellular/biochemical
systems which are redundant and may be directly or indirectly related to toxic effects
(e.g., at what point glutathione depletions lead to susceptibility).
It is important to carefully evaluate information on acute toxicities. The
manifestation of acute toxicity (toxicity evident in close temporal proximity to the
exposure) generally requires chemical exposure levels that are greater than those
required to produce delayed effects. Further, doses sufficient to produce acute toxicity
bring a higher likelihood that fundamental biochemistry can be perturbed to produce TK
and/or TD interactions among compounds. Interactions observed with acute toxicity,
however, are generally poor indicators of interaction at lower exposure levels. Tumor
production is a multi-step process, and interactions may be several, ranging from the
classic initiation-promotion type interaction, to adduct formation and inhibition or repair
capacity. For compounds thought to interact in the tumorigenic process, a rich data set
is required to substantiate an interaction. For compounds with a tumorigenic mode of
4-43
-------
action defined to the point that a non-linear, or threshold-like, dose-response
relationship can be defended, it is useful to consider the severity of the underlying effect
(e.g., cytotoxicity and cellular regeneration). For compounds that must be metabolized
to be tumorigenic, TK interactions at the enzyme level are an important aspect to be
evaluated. For additional details on TK and TD interactions, see Appendix C of U.S.
EPA (2000a [Mixtures Guidance]).
4.6.2. A Quantitative Method for Evaluating Interaction Effects. To account for
chemical interactions in a site assessment, EPA recommends applying the HI|NT to
component data (U.S. EPA, 2000a). The main assumption for the HI|NT is that
interactions in a mixture can be adequately represented as departures from dose
addition (Hertzberg et al., 1999). The method follows an obvious approach: begin with
the dose-additive HI (Equation 4-1) and then modify its calculation to reflect the
interaction results, using plausible assumptions to fill in the data gaps. Because
toxicological interactions have been mostly studied with binary mixtures, the HI|NT
includes information only on binary interactions; an assumption is then that higher order
interactions are relatively minor compared to binary interactions. Noting that the first
summation shown below is the additive HI and the second summation shown below is
the modification for interactions, the formula for the HI INT is:
(4"8)
where:
HI|NT= HI modified by binary interactions data,
HQj = hazard quotient for chemical y (unitless, e.g., daily intake/RfD),
fjk = toxic hazard of the kth chemical relative to the total hazard from all
chemicals potentially interacting with chemical y (thus k cannot equal j).
To calculate, the formula is:
J=1
Mjk = interaction magnitude, the influence of chemical k on the toxicity of
chemical/ To calculate, estimate from binary data (see example
calculation below) or use default value = 5
4-44
-------
score for the WOE that chemical k will influence the toxicity of chemical j.
To calculate, the formula is:
JHQj*HQk
gjk = degree to which chemicals /candy are present in equitoxic amounts.
The current WOE classification and scores are given in Table 4-6 (U.S. EPA,
2000a). This scheme does not specifically focus on the types of data available to
support a WOE determination but on the interpretation of the data made by an analyst
or a group of analysts. The binary WOE factor Bjk reflects the strength of evidence that
chemical k will influence the toxicity of chemical j, and that the influence will be relevant
to the risk assessment. In general, the more extrapolation required, the weaker the
evidence is. For example, if the available interaction data were from in vitro studies with
effect measures not directly related to the toxicity of concern, or represented a different
exposure route or duration, then the WOE score would be low. ATSDR has a similar
but more structured scoring rule. The factor does not have to be the same for the
influence of chemicaly on the toxicity of chemical k; i.e., BJk£ BkJ. The WOE
determination begins with a classification of the available information, followed by a
conversion of that classification into a numerical weight.
The term MJk represents the maximum interaction effect that chemicaly can have
on the toxicity of chemical k. As with the WOE score, the interaction magnitude need
not be symmetric; i.e., the magnitude of interactive influence of chemical k on the
toxicity of chemicaly may be different than the corresponding magnitude of chemicaly
on the toxicity of chemical k. When binary mixture toxicological data are available, a
simple calculation can be used to determine a value for My/Jrom Effective Doses (e.g.,
from EDio's that cause a 10% effect in the test animals). EPA uses the proportional
change in effective dose as the interaction magnitude. For example, if the ED10 is
predicted to be 20 mg/kg/day for chemicals y and k together under an assumption of
dose addition, but the observed experimental value fory and k together is equal to 5
mg/kg/day, then the interaction magnitude would equal 20/5 = 4. In the absence of
data, U.S. EPA (2000a) recommends a default value of 5 based on observations of
interactions data from the literature.
This method for modifying the HI is based on commonly discussed principles of
toxicological interactions. The algorithm, however, does not attempt to directly model
toxicological interaction mechanisms. While the interaction magnitude is based on
directly observed toxicity measures, the adjustment for different component doses is
4-45
-------
TABLE 4-6
Default Weighting Factors for the Modified Weight of Evidence
Category
1
II
III
IV
Description
The interaction has been shown to be
relevant to human health effects and the
direction of the interaction is
unequivocal.
The direction of the interaction has been
demonstrated in vivo in an appropriate
animal model, and the relevance to
potential human health effects is likely.
An interaction in a particular direction is
plausible, but the evidence supporting
the interaction and its relevance to
human health effects is weak.
The assumption of additivity has been
demonstrated or must be accepted.
Direction
Greater than
Additive
1.0
0.75
0.50
0.0
Less than
Additive
-1.0
-0.5
0.0
0.0
Source: U.S. EPA(2000a).
4-46
-------
simple and heuristic. Instead of estimating joint toxicity or risk, the interaction HI
method models "concern" for toxicological interactions, which reflects issues of
magnitude as well as likelihood. In this respect, the interaction HI is interpreted the
same way as is the common additive HI; the approach corresponds more closely with
the current use of uncertainty factors in the risk assessment of single chemicals than
with an attempt to biologically model interactions. As more interaction studies are
completed and more interaction mechanisms and modes of interaction are understood,
this method for modifying the HI will be further refined.
4.7. EVALUATING MULTIPLE ROUTE EXPOSURES
A cumulative risk assessment considers exposures to the population from
multiple routes and pathways. Measures or estimates of internal doses may provide an
improved basis both for estimating risks posed by chemical mixtures that occur through
multiple exposure routes. To date, regulatory risk methods have only been published
for simpler and more common approaches that use external exposure levels.
Assessments of multiple route exposures can be complicated because of a lack
of toxicity data for all exposure routes of interest. If data on only one route are
available, then the risk analyst decides if it is appropriate to conduct a route-to-route
extrapolation of the data. Such
extrapolations can be problematic
because of biological differences
among routes in toxic responses or
pharmacokinetic processes. The 2005
cancer guidelines recommend route-to-
route extrapolations only on a case-by-
case basis as supported by available
data. There seems to be general
agreement in the literature that the
most appropriate way to extrapolate
across routes is to employ a PBPK
model. However, both qualitative
assessments and application of simple
quantitative methods of route
extrapolation are used as needed when
data are lacking. Text Box 4-4
describes the uses of route-to-route
EPA Uses of Route to Route Extrapolations
U.S. EPA (2003g) Workshop Report on Inhalation
Risk Assessment (Text Box 4-4)
Office of Solid Waste: only does such extrapolations
when there are findings that indicate it is appropriate.
When it is performed, the approach is similar to that
used to aggregate exposures.
Office of Air Quality Planning and Standards: treats
cancer and non-cancer extrapolations differently. For
cancer, in lieu of an IUR from the hierarchy of sources,
an IUR may be derived from an oral value (using a
rough breathing rate/body weight calculation), with
recognition of added uncertainty. No such rough
extrapolation is done to create RfCs. Because the
Clean Air Act list of hazardous air pollutants is heavily
weighted by respiratory toxicants, such rough non-
cancer route extrapolations are performed because of
the high probability of missing target toxicity.
Office of Pesticide Programs: performs route-to-route
extrapolations with no distinction between cancer and
non-cancer endpoints. Absorption via the inhalation
route (in mg/kg/day) is equal to oral absorption. Air
concentration estimates for human exposure from a
concentration (mg/m3) to an average daily dose
expressed as mg/kg/day so that exposure can be
compared directly to oral NOAEL and LOAEL values.
4-47
-------
extrapolation by several program offices, as presented in a 2003 U.S. EPA workshop
report on inhalation risk assessment (U.S. EPA, 2003h).
4.7.1. Quantitative Approaches to Evaluating Multiple Route Exposures to
Mixtures.
4.7.1.1. Summing Across Routes and PathwaysEPA's Risk Assessment
Guidance for Superfund (1989a) instructs analysts to sum HQs (Equation 4-1) and
cancer risks (Equation 4-3) across exposure routes and exposure pathways, providing
there is evidence of combined exposure pathways to identifiable individuals or groups of
individuals who would consistently face a reasonable maximal exposure. U.S. EPA
(1999b) guidance on preparing Records of Decision for Superfund site assessments
provides further information on this method. (See details of this procedure in
Section 5.5.1.) Although there is no discussion of summing across exposure routes and
pathways in the U.S. EPA (1986b, 2000a) health risk assessment guidance documents
for mixtures, U.S. EPA (1989a, 1999b) establishes this approach as a policy with the
purpose of accounting for any reasonable risk from multiple route and pathway
exposures. U.S. EPA (1999b) provides a template for these calculations in the form of
pre-formatted tables and also shows examples online on its Web site
(http://www.epa.gov/oswer/riskassessment/ragsd/tara.htm). For the purpose of this
report, one recommended approach to account for multiple route exposures is to apply
these procedures to the target organ groups developed in Figure 4-10. Further
discussion of this approach is given in Section 5.5.1 in terms of a cumulative hazard
index (CHI), along with guidance on its interpretation.
4.7.1.2. Summing of Route-Specific Relative Potency FactorsA second
approach is to estimate risks for each group and exposure route using an RPF mixtures
risk assessment approach (U.S. EPA, 2000a) and then sum the risks to yield a total risk
for that group by all routes. The RPF approach is a general methodology for applying
dose addition to mixtures of chemicals that produce toxicity by the same MOA. Text
Box 4-5 shows the mathematical formulas used to develop RPF-based risk estimates,
and Figure 4-12 graphically illustrates the process to be followed in developing an RPF
assessment. To summarize the procedure, doses of mixture components are scaled by
their potency relative to a well-studied component of the chemical mixture (referred to
as the index chemical) using scaling factors called RPFs. The product of each mixture
component's dose and its RPF is considered to be its equivalent dose in units of the
index chemical. These dose equivalents of all the mixture components are summed to
4-48
-------
Concentration
Chemical Al
(Index
Chemical)
xRPFA1 =
(Note:RPFA1=
Concentration
Chemical A2
xRPFAo =
A2
Concentration
Chemical A3
xRPF
A3
Index
Chemical
Concentration
Al
Index
Chemical
Equivalent
Concentration
ofA2
Index
Chemical
Equivalent
Concentration
of A3
Sum index chemical
'equivalent concentration
to estimate total
mixtures exposure in
units of the
index chemical
FIGURE 4-12
Schematic for Relative Potency Factor Approach
Index Chemical
Dose-Response
Mixture
Risk
Equivalent
Concentration
4-49
-------
express the total mixture
dose in terms of an Index
Chemical Equivalent
Dose (ICED).5 The risk
posed by the mixture is
then quantified by
comparing the mixture's
ICED to the dose-
response assessment of
the index chemical. To
implement this approach,
the index chemical must
have an adequate
toxicological dose-
response data set (U.S.
EPA, 2000a). U.S. EPA
(2000a) characterized
the RPF methodology as
a generalized form of the
toxicity equivalence
factor (TEF) methodology that has been used to assess risks. This approach is similar
to the Toxicity Equivalents (TEQ) method used for dioxins (U.S. EPA, 1989b) but
requires a less strict interpretation of the toxicity data. Thus, it is applicable to a larger
group of chemical classes than the TEQ method.
Figure 4-13 illustrates the proposed approach that combines the principles of
dose addition and response addition into one method to assess mixtures risk for
multiple route exposures within a group (e.g., as defined using Figure 4-10). Using two
exposure routes, inhalation and oral, Figure 4-13 illustrates how the approach estimates
risk from exposure to the mixture. Within a target organ group, an index chemical (a
mixture component with high quality dose-response data that acts [or is judged to act]
through the same MOA as the other members of the group for the route of concern) is
selected, and the ICED for the mixture is calculated using the RPF approach (U.S. EPA,
2000a). (Note the text here will only refer to an ICED. However, for clarity in
RPF Formulas for Risk Estimation of a Two Chemical Mixture (Text Box 4-5)
hm!x(d1,d2)=f1(d1 + RPF2*d2)
where:
hmix(di,d2) = mixture hazard or risk from joint exposure to doses d-i of
chemical 1 and d2 of chemical 2 (dose units not specified,
must be consistent for all chemicals)
fi(*) = dose-response function of the index chemical for the
response(s) common to chemical 1 and the other chemicals
RPF2 = potency of chemical 2 relative to that of chemical 1
Let pot, be the potency estimate for chemical /. Then
RPF2 = pot2/pod
For cancer risk, pot: is often given by the slope factor of risk per unit of
dose. Note that if the inverse of the effective dose (e.g., 1/ED10) is used for the
potency, then RPF is the chemical 1 to chemical 2 ratio of the ED values:
RPF2 = ED101/ED102
This mixture hazard formula uses the mixture dose given as the equivalent
dose of the index chemical. Let ICED be the index chemical equivalent dose
based on relative potency estimates (dose units consistent with di and d2).
Then,
ICED = d1 + (RPF2* dz)
and the mixture hazard formula is
hmix(di,d2) = fi(ICED).
Example: With dioxins, the index chemical is 2,3,7,8-TCDD. For the mixture
assessment, the combined doses of all the dioxins are converted into the
equivalent dose of 2,3,7,8-TCDD, and the mixture risk is then determined from
the dose-response data for 2,3,7,8-TCDD.
The ICED has the same mathematical interpretation as the dioxin toxicity equivalents (TEQ). TEQ
refers to the quantification of dioxin concentrations based on the congeners' equivalent 2,3,7,8-TCDD
toxicity (U.S. EPA, 1989b). ICED is applied to mixtures other than dioxins.
4-50
-------
Oral
ICED, Index
Chemical Al =
DoseAl*!
Oral
ICED for
Chemical A2 =
DoseA2*RPFA2/
Oral
ICED for
Chemical A3 =
DoseA3*RPFA3
Dose
Addition for Oral
Inhalation ICEC
for Chemical Al
= ConcAl*RPFAl
Inhalation ICEC
for Chemical A2
= ConcA2*RPFA2
Inhalation ICEC,
Index Chemical A3
= ConcA3*l
Dose \ >
Addition for Inhalation
ICED, Oral
Al Dose
Response
ICED, Oral
I
ICEC, Inhalation
Response Addition for
Group A Mixture Risk
Risk for Oral
Evaluated at
Group A ICED
M Response
A3 Dose
Response
i
CEC, Inhalation
\
Risk for Inhalation
Evaluated at
Group A ICEC
Total Mixture
Risk as Sum of
Risks for Oral
and Inhalation
FIGURE 4-13
Combining Grouped RPF Estimates Across Exposure Routes
(Source: U.S. EPA, 2000e)
4-51
-------
Figure 4-13, the ICED refers to the oral route of exposure, and the ICEC (Index
Chemical Equivalent Concentration] refers to the inhalation route of exposure.) The
ICED is an important concept employed at two levels:
(1) Component ICED - refers to the ICED for an individual chemical
(2) Group ICED - refers to the ICED for all chemicals within the group and route,
formed by summing the component ICEDs.
EPA has proposed the RPF approach as a means for characterizing health risks
associated with mixtures of chemicals that are toxicologically similar (U.S. EPA, 2000a).
To develop an RPF-based risk estimate for a class of chemicals, good toxicological data
are needed for at least one component of the mixture that can be used as the index
chemical. Scientific judgment and analysis of available data are used to assess the
relative toxicity of the other individual components in the mixture. The component
ICEDs are then summed within the group to generate a route-specific ICED. The risk
posed by the group and route of interest can be estimated using the route-specific dose-
response information for the index chemical. For each exposure route, the RPF
approach uses dose-addition to estimate risk for the toxicological outcome common
across the group. An assumption is made that the route-specific risks are independent
of each other (i.e., the toxicity caused by one route does not influence the toxicity
caused by the other route). This condition meets the criteria for applying response
addition; the route-specific risk estimates are added to yield a risk estimate for the
mixture group. Quantitative uncertainty analyses of this approach are complicated by
the general lack of multi-route toxicity studies. It is then important, during the toxicity
assessment, for the risk analyst to identify any studies or dose-response data on the
multi-route mixture exposure that can support (refute) this RPF approach.
4.7.2. Internal Dose Estimates. A third quantitative approach to handling mixtures
assessments for multi-route exposures is to estimate a total internal dose for use in risk
estimation. In 2003, EPA completed a report showing that a multi-route mixtures risk
assessment can be conducted based on internal dose estimates developed in both test
animals and humans for toxicants that do not cause portal of entry effects (Teuschler et
al., 2004; U.S. EPA, 2003b). This approach is mentioned here for completeness but is
resource intensive.
U.S. EPA (2003b) combines exposure modeling results, PBPK modeling results
and the RPF mixtures risk assessment approach. Human internal doses (e.g., blood,
tissue, and organ concentrations) were estimated using PBPK models, accounting for
external exposures from multiple routes (as dictated by the exposure scenario) and
4-52
-------
human pharmacokinetic processes. Hypothetical RPFs were developed for a subset of
chemicals based on test animal data. Although the application of a full PBPK model
was recognized as the preferred approach to estimating rodent internal doses (i.e.,
blood concentrations), for the example data used in the report, administered doses were
assumed to be 100% bioavailable to the rat. The rodent toxic effects were assumed to
be constant between internal and external exposures and were used to evaluate the
human dose-response relationship. The use of internal dose measures (i.e., blood
concentrations in both humans and rodents), both for developing the RPFs based on
rodent data and for indicating human multi-route exposure, provides a consistent basis
for extrapolating across species. However, these approaches are inappropriate for use
with toxicants that elicit responses at points of contact with the body (e.g., skin,
intestinal tract, and nasopharyngeal, bronchial and lung epithelia).
4.8. SUMMARY RECOMMENDATIONS
The toxicity assessment step of the Risk Analysis phase includes the evaluation
of all available and relevant toxicity data, with the goal of simplifying the multiple
chemicals, exposures and effects. The approach presented here focuses on the
identification of common characteristics so that these multiples can be consolidated into
a manageable number of groups. Because the primary risk methods invoke dose
addition or response addition, the grouping processes focus on assumptions of toxic
similarity or toxic independence, respectively. As the chemicals, pathways and effects
are grouped, it is critical to include a discussion of the evidence supporting those key
assumptions. Any decisions to exclude chemicals or exposure pathways from the
cumulative risk assessment may be supported by toxicity arguments that are relevant to
the estimated exposures. When such information is weak or inconclusive, the
chemicals and pathways are retained in the assessment.
4-53
-------
5. CUMULATIVE RISK CHARACTERIZATION
The last phase of cumulative risk assessment, Risk Characterization, assembles
all the information from the Risk Analysis phase and interprets the results in the context
of the problem(s) formulated in the
Elements of Risk Characterization (Text Box 5-1)
Planning and Scoping phase. Text
Results of the integrated analysis
Quality of and confidence in the available data
Uncertainty and sensitivity analyses
Justification of defaults or assumptions
Related research recommendations
Contentious issues and extent of scientific consensus
Effect of alternative assumptions on conclusions and
estimates
Highlights of plausible ranges
Reasonable alternative models
Perspectives through analogy
(U.S. EPA, 2000f)
Box 5-1 lists some important
elements of a Risk Characterization
that are useful to consider in a
cumulative risk assessment. As
described in EPA guidance (U.S.
EPA, 2000f), Risk Characterization
includes two products:
1. Integrative analysis - a
technical presentation of the
predicted risks and uncertainties in an assessment
2. Risk Characterization summary - a condensed version of the results and
uncertainties that emphasizes the recommendations of the analysis, written in a
style that communicates the "bottom lines" to the general public.
The Risk Characterization transparently presents the logic that leads to various
technical decisions in the analysis (e.g., those regarding the inclusion or exclusion of
specific chemical classes or groups of chemicals, specific exposure pathways and the
choice of specific Risk Characterization approaches). The analysis also discusses the
support for analytic assumptions offered by data and professional judgment (e.g.,
assumptions used to group chemicals for use in risk assessment procedures). As
discussed in the Supplementary Mixtures Guidance (U.S. EPA, 2000a), any quantitative
or qualitative risk estimates must be accompanied by the explanation of assumptions
made when estimating the risks and uncertainties associated with the risk estimates.
These obligations of Risk Characterization apply to all risk assessments conducted by
the EPA.
While this discussion focuses on issues to consider in the integrative analysis
product, several aspects of the Risk Characterization summary are also discussed.
Results of the Risk Characterization may impact various non-technical individuals. The
economic and social ramifications of cumulative risk assessments may require that the
Risk Characterization results highlight the important issues and uncertainties and
explore their implications for different audiences. These stakeholders may be
concerned with
5-1
-------
the number of people exposed;
the range of uncertainty around the exposure and health risk estimate;
the critical variables driving the assessment, the existence of data gaps;
the bottom-line conclusion; and
the degree to which the Risk Characterization results support a regulatory
decision.
Section 5.1 presents an overview of Risk Characterization in a cumulative risk
assessment. Section 5.2 describes special Risk Characterization concerns of a
cumulative Risk Analysis including questions that may be used to further guide the
analyst in developing a cumulative Risk Characterization. Section 5.3 presents an
approach for developing an integrative assessment for a cumulative risk assessment
that includes considerations of multiple chemical exposures through multiple exposure
pathways. Section 5.4 discusses potential interaction factors that the analyst may need
to consider and Section 5.5 presents summary recommendations. Finally, section 5.6
presents a hypothetical example using the CHI.
5.1. CHARACTERIZATION OF RISK IN A CUMULATIVE RISK ASSESSMENT
CONTEXT: AN OVERVIEW
Risk Characterizations of cumulative risk assessments evaluate risks posed to
the general population and vulnerable subpopulations. While risk characterizations of
cumulative risk assessments include characterization of single chemical risks, these are
not the focus of this chapter and are not addressed in detail further. As presented in
Chapters 1 and 2, cumulative risk assessments include all steps of the traditional risk
assessment paradigm (i.e., hazard identification, dose-response, exposure assessment
and risk characterization); however, these steps are expanded beyond the elements
pertaining to single chemical assessments to account for the complexities of cumulative
risk (U.S. EPA, 2000a, 2003a). Chapters 3 and 4 of this document describe
approaches for conducting cumulative risk assessments that include analyses of
multiple chemicals, multiple exposure pathways and routes, multiple toxic effects over
distinct time frames and joint exposure response relationships. Following these
evaluations, the pieces are integrated into an overall conclusion about risk, along with
clear descriptions of the analytic limitations and uncertainties (NRC, 1983, 1994).
In the integrative analysis product of the Risk Characterization, the analyst
evaluates the collective information and identifies information gaps, uncertainties at the
interfaces between different process steps and the appropriateness of the potentially
5-2
-------
different levels of analysis across the steps of the risk assessment (uncertainties within
the steps of the process likely have already been identified). For example, combining
dose-response and exposure data that are collected over different durations of time or
at different levels of biological organization necessitates the use of additional
assumptions or additional data gathering activities.
The development of the integrative analysis of a cumulative risk assessment is
typically an iterative process, where information from each process phase is
reconsidered from the perspective of the information generated during the other phases.
Chapters 3 and 4 emphasize the iterative nature of the conduct of a cumulative risk
assessment. The need for this iterative assessment approach continues in the Risk
Characterization. An important iteration is the comparison of analytic results with the
goals set out in the Problem Formulation phase. The description of uncertainties plays
a pivotal role in determining whether these goals have or, perhaps, can been met. If the
results do not sufficiently address the goals, iteration through one or more of the
previous steps might be needed, including the initial Planning and Scoping and Problem
Formulation.
Considerations of multiple chemicals, multiple exposure pathways and routes,
multiple toxic effects over distinct time frames and joint exposure response relationships
complicate the characterization of risks. Text Box 5-2 presents examples of this
complexity for a hypothetical comparative risk decision that considers cumulative risks.
These complications increase the difficulty in understanding the influences of
combinations of underlying assumptions on the analytic results and possibly the
implications of the assessment results.1
5.2. SPECIAL CONCERNS WITH CUMULATIVE RISK CHARACTERIZATION
The EPA guidance documents on Risk Characterization present lists of issues or
questions that may be addressed in the integrative analysis step of Risk
Characterization. Issues important for single chemical Risk Characterization (e.g.,
identifying a single key, supporting toxicity study; addressing only one critical effect; and
deriving a single benchmark risk value with which to judge safety of exposures) may not
be very relevant to the cumulative Risk Characterization. Throughout the analysis, the
analysts make decisions that influence the conclusions of the assessment. Such
decisions may occur during Planning and Scoping, during the iterative exposure and
dose-response analyses and during the integrative analysis in the Risk
1 These complicating factors also make the successful communication of the uncertainties to both the risk
manager and the public in the Risk Characterization summary product difficult.
5-3
-------
Characterization. The
following list of questions
may help to further guide
the analyst in developing a
cumulative Risk
Characterization. It covers
most of the issues raised in
the Risk Characterization
handbook (U.S. EPA,
2000f, Chapter 3).
Overview
Have the goal(s) of
the assessment
been met?
Is the level of the
analysis consistent
throughout?
Are the exposure
scenarios (including
pathways and
routes) considered
appropriate?
Are the type(s) of
exposure data
available, analyzed
and used in the Risk
Characterization
appropriate?
Are the types of
toxicity data
available, analyzed
and used in the Risk
Characterization
appropriate?
Is the choice of
methods for
evaluating risks
posed under the
selected exposure
scenarios
appropriate?
ExampleSite Closure vs. Public Access (Text Box 5-2)
Consider a site with soil contamination due to multiple metals and
organic compounds where the decision alternatives are open public
access or full closure (clay cap and a fence). The risk assessment
evaluates the public access scenario. If the risk assessment outcome
determined that exposures would likely exceed acceptable risk levels,
the risk manager may opt for site closure.
The complications include population-dependent exposure
characteristics. For example, the population near the site may include
adults and children with quite different exposures. In the risk
assessment, children are assumed to be exposed predominantly by
direct contact with soil (dermal absorption and ingestion) and adults
primarily by inhalation of dust. Groundwater, contaminated from
gradual migration through the soil, is a source of drinking water for
some adults and children. This is predicted to result in ingestion route
exposures to relatively low pollutant concentrations.
Complexities to Consider when Evaluating Cumulative Risk
Different proportions of chemicals in inhaled dust compared with
ingested groundwater, leading to different critical effects and
different toxicological interactions
Different toxic sensitivities of adults versus children
Time-varying combined exposure from soil and groundwater that
reflects multiple routes as well as timeframes
Background exposures to pollutants exhibiting similar patterns of
toxicity
Population vulnerabilities.
The integrative analysis may evaluate the relative impact of each of
these complexities on the cumulative risk estimates. Any joint
contributions to risk may be quantified to the extent possible based on
available information.
Risk Characterization Summary
Usual elements of the Risk Characterization (summary of likely
health endpoints, identification of key chemicals)
Based on mixtures risk assessment methods (See Chapter 4),
predictions of adult risk and child risk for multiple chemicals for all
routes combined and over different exposure routes and
timeframes
Quality of the multiroute exposure
Quality of the toxicity information for children and adults
Confidence in summary estimate of cumulative risks
Other descriptions that might be required for this example site include
a comparison of the risk for average exposure vs. high-end exposure
(for adult and for child) and the ranking of the most influential factors
driving the risk estimates (a quantitative sensitivity analysis if
possible).
5-4
-------
To Address Multiple Chemical Exposures, Health Effects and Time Scales
Is there a focus, e.g., an effect caused by a single chemical by one pathway that
dominates the risk? If no single key factor dominates, then what is the best
presentation of the array of possible combinations of factors?
Are the spatial scope and temporal scale of the analysis consistent across
analytic components?
How do composite evaluations compare with multivariate measures?
How much detail and accuracy is lost when combining across effects, such as
with ordinal regression?
How well supported are the number of assumptions and default parameters that
are used and how can that strength of support be reflected in the quantitative
Risk Characterization?
How does the use of surrogates affect the overall uncertainties?
How does relying on an index chemical to represent the group increase the
uncertainties surrounding the contributions of the other chemicals to the
predicted health effects?
Grouping chemicals, pathways and effects structures and simplifies the
assessment. Are there alternative ways of grouping these factors? Are any
factors double-counted by the grouping process?
To Address Interactions
Can the interaction magnitude be estimated for those chemical-pathway
combinations of most importance?
How many interactions cannot be quantified?
Can all identified interactions at least be described for the direction of the
interaction, i.e., do they increase or decrease the risk?
To Address Populations of Concern
How consistent are the risk estimates with those health effects of most concern
to the stakeholders as determined in the Planning and Scoping phase of the
assessment process
If a health effect was the initiating factor or impetus for the cumulative risk
assessment, is that effect adequately addressed in the Risk Characterization?
Are the population vulnerabilities clearly described?
To Address Time Dependencies
What is the likelihood that the mixture composition or exposure pathways will
change over the timeframe being addressed? Can the impact of that change be
quantified in terms of a change in risk?
5-5
-------
How likely is it that the subpopulations of most concern will change location or
their exposure-related behaviors and thus, change their risks over the timeframe
being addressed?
Will any of the alternative remediation options change the mixture composition
(not just the total dose)? Is that change reflected in the way the expected
reduction in risk is calculated?
To Address Consistency of Information
How well do the exposure levels in the dose-response data match the estimated
exposure ranges?
How much extrapolation is required for the risk estimates? How dependent is the
extrapolation on default values?
Are there inconsistencies among the data?
Do some exposure or toxicity units need conversion in order to allow combined
exposure or joint toxicity to be estimated?
How different are the exposure and toxicity measures in terms of level of
understanding, level of accuracy and detail?
How much information is lost when reducing all the measures to the lowest
common level so that grouping and composite analysis can be performed?
To Address Context
How can the Risk Characterization for this site or situation be compared with
Risk Characterizations for other similar sites or situations?
How can multivariate site descriptions and risk evaluations be compared to
determine whether sites are similar to each other?
To Address the Initiating Factor
Elevated Health Effects
Are the health effects that initiated the analysis actually elevated in the
community?
Are populations in the group differentially affected?
Are the effects adequately addressed in the Risk Characterization?
Has the Risk Characterization identified potential cause(s) of the health effects?
Has a causal analysis been undertaken in the context of the cumulative risk
assessment?
Does the Risk Characterization adequately address epidemiologic concepts of
causation (e.g., Hill, 1965; Susser, 1991) or those from evidence-based
toxicology (e.g., Guzelian et al., 2005)?
If the assessment did not identify a specific causal agent or agents, but has
determined that the population is experiencing an increased incidence of an
5-6
-------
adverse health effect(s) (i.e., incidence significantly above normal background),
does the risk characterization adequately describe the exposures to the
contaminants evaluated and the reasons that they are unlikely to be causing the
health effect?
Has the Risk Characterization adequately described the types of studies needed
if the causal agents can not be identified?
Elevated Concentrations
Are the environmental concentrations or biomonitoring data that initiated the
analysis actually elevated in the community?
Are the potential health effects of such elevations adequately addressed in the
Risk Characterization (e.g., Guzelian et al., 2005)?
Is there evidence that the effects anticipated by the toxicology data are actually
being observed in the community?
Has the Risk Characterization identified potential source(s) of the pollutants that
are elevated?
If the Risk Characterization cannot identify a specific source(s) but has
determined that the concentrations are elevated, has the study adequately
described the evaluation of potential sources of such contaminants in the
community? (A description of the evaluation of potential sources that are not
considered to be contributing to the elevated concentrations may aid future
investigations).
If the initiating factor was elevated environmental concentrations, has the Risk
Characterization adequately described the likely exposure pathways for the
population and adequately characterized the risks posed by the elevated
concentrations? Have the spatial and temporal aspects of these analyses been
fully characterized?
If the initiating factor was biomonitoring data, have the likely exposure pathways
and routes been identified?
Has the Risk Characterization adequately characterized the types of risks
associated with the effects and described the types of health studies needed in
the population?
Multiple Sources
Have the quantities of chemicals released from the multiple sources that initiated
the analysis been adequately characterized?
Does evidence show that exposure pathways to vulnerable subpopulations are
complete?
Are there other potential sources of the chemicals in the community? Has the
environmental fate of the released chemicals been adequately evaluated?
5-7
-------
Have the exposure assessment and the toxicity assessment been adequately
conducted?
Has the Risk Characterization adequately integrated the exposure assessment
and dose response assessment?
Has the Risk Characterization characterized the types of risks associated with
the chemicals and described the types of health studies needed in the
population?
It is important to clearly summarize the logic (e.g., quality of the supporting data)
underlying each of the responses to each relevant question to ensure transparency and
clarity of the assessment's conclusions. When possible, for each of the assumptions,
the impact of alternative choices on the resulting risk estimates may be described (e.g.,
a sensitivity analysis) (see Section 5.3). For example, if an exposure pathway is
screened out because adequate data are unlikely to be obtained, then it is useful to
describe the impact of ignoring that pathway in the integrative analysis (i.e., even if the
analyst identifies the direction of potential error; e.g., exclusion of an exposure route is
likely to underestimate the risk slightly because actual exposures to the mixture will then
be slightly higher than those predicted via the routes considered.).
The level of analysis can vary across assessments. The Risk Characterization
discusses whether the analysis was conducted at a screening level or a refined level,
clearly identifying assumptions employed and whether they are consistent with the
stated analytic level. Screening level assessments typically employ many simplifying
assumptions. In a cumulative Risk Analysis, these could include conservative grouping
practices. Chemical grouping into environmental fate groups or health effects groups
could be based on crude criteria in a screening level analysis (e.g., all chemicals
affecting cancer could be included in a single health group in a screening analysis and
the cancer risks associated with such exposures could be estimated using response
addition). Screening level assessments also typically employ steady-state
environmental fate and exposure models; they also may employ deterministic model
variables (e.g., a single high-end exposure factor value, such as all members of the
population consume 2 L of drinking water each day). Refined analyses may employ
fewer assumptions than screening analyses. Chemical groupings may be based on
sophisticated measures (e.g., health effects grouping may be based on MOA analyses).
Dynamic models of environmental fate may be employed in a refined modeling analysis.
Exposure models may include additional temporal or spatial resolution and may be
based on probabilistic analyses. While the level of analysis may be described in the
integrative analysis, these distinctions may be included in explanations of how the
5-8
-------
results of the assessment are interpreted and used in the Risk Characterization
summary.
An important aspect of cumulative risk assessment is the process of identifying
and defining geographic areas, groups of chemicals and exposure scenarios that are
evaluated for further analysis. These decisions about the conduct of the assessment
expand further to take into account appropriate groupings of chemicals using exposure
information (Chapter 3) and judgments regarding similarity of toxic effects and the
potential for interactions, the basis for toxicity groupings (Chapter 4). In the Risk
Characterization the analyst may address broader elements such as appropriateness of
the selected analytic scope, choice of chemicals for analysis, choice of exposure
scenarios, criteria for grouping chemicals, identification of appropriate populations for
analysis and an evaluation that seeks to determine if there are other important factors
not being addressed in the analysis (i.e., looking at the analysis from "outside the box"
to see if the analytic approach really makes sense). At the end of this process, the
analyst identifies the types of effects that might occur, quantifies their likelihood in
different populations and quantifies, where possible, the uncertainties in these
estimates. The risks of any health outcomes that cannot be quantified are to be
described qualitatively, along with suggestions regarding the kinds of information
required for quantitative characterization of the likelihood that such a health effect could
occur. The principles and guidance offered on these matters in the Policy for Risk
Characterization (U.S. EPA, 1995b; see also U.S. EPA, 2000f) and Science and
Judgment in Risk Assessment (NRC, 1994) are applicable to characterizing cumulative
risks.
5.3. A RISK CHARACTERIZATION PROCESS FOR CUMULATIVE RISK
ASSESSMENT
Figure 5-1 presents an approach for characterizing cumulative risks.
Subsequent sections discuss each step in the process.
5.3.1. Populations. In the Problem Formulation phase and the analyses of exposure
(Chapter 3) and toxicity (Chapter 4), the analyst identifies the relevant populations to be
considered in the cumulative assessment. These can include the general population
and vulnerable populations. In the general population assessment, it may be important
to determine whether the exposure assessment accurately depicts exposure factor
variability as well as typical and unique exposure pathways. Following are some
5-9
-------
Populations ConsideredGeneral Population or Subpopulations
Exposure
Groups
Toxicity
> Groups
If
Temporal Analysis
Do relevant multi-chemical
exposures or persistent
effects overlap?
Integrated Groups for Cumulative
Risk Assessment
Extrapolations,
Simplifications
and Omissions
Acceptable?
Conduct Qualitative
Assessment Only or
Revise Analytic Scope
Develop Quantitative Risk Estimates
For General Population and Relevant
Subpopulations
For each Relevant Chemical Group
For each Relevant Pathway and Exposure Route
For each Relevant Timeframe
Conduct Single Chemical Assessment Only,
Discuss All Cumulative Risk Elements Considered,
Describe Data Gaps for Future Research
Document Cumulative Risk Characterization
Including Risk Estimates and
Uncertainty Discussion
Describe Data Gaps for Future Research
Sensitivity and Uncertainty Analysis
Identify Sources of Uncertainty
Develop Integrated Sensitivity Analysis, if
Possible
-Identify Sources of Model Uncertainty
-Identify Sources of Parameter Variability and
Uncertainty
Identify Critical Research Needs
FIGURE 5-1
Schematic of Cumulative Risk Characterization Approach in this Report
5-10
-------
example vulnerable populations that have been identified in cumulative risk scenarios
and published in the literature:
Fetuses - Spontaneous abortions from (possibly) short-term, multiple route
maternal exposures to low doses of multiple drinking water DBFs (Waller et al.,
1998)
Rural resident - Neurological effects from chronic multiple route exposures to
organophosphorous pesticides used in agriculture (U.S. EPA, 2002a)
Subsistence fishing family - Cancer in adults exposed chronically via ingestion of
fish containing PCBs (U.S. EPA, 1996b); neurotoxicities in children exposed in
utero to methylmercury via fish consumption (U.S. EPA, 1997e; U.S. EPA,
2001 b)
Elderly - a potential combination of high exposure and high vulnerability in a
susceptible population that may be vulnerable to a health effect (e.g., cancer)
from chronic, multiple route exposures to high doses of multiple chemicals (U.S.
EPA, 2003a)
Child - asthma from short term inhalation exposures to high levels of particulate
matter in the air (U.S. EPA, 2004e)
The vulnerable subpopulations can be identified in the initial Planning and Scoping
stage of the Problem Formulation phase of the assessment, during the exposure
assessment, the dose-response assessment or Risk Characterization phase.
5.3.2. Temporal Analysis. For each subpopulation and the general population (if
relevant), the analyst may determine whether there are multi-chemical exposures that
occur over a toxicologically relevant timeframe(s), considering both TKs and TDs
(Chapter 3). If no relevant overlaps are evident, then only single chemical
assessment(s) are conducted. If there are relevant overlaps, resulting in exposure
groups, then those groups are evaluated for toxicological similarity (Section 5.3.3).
The temporal relationships among exposures to different chemicals are a critical
consideration in the estimation of cumulative risk. The initial analysis may involve an
examination of the environmental fate of various contaminants over time (Chapter 3).
Using this information for the various exposure media, the exposure analyst can
estimate from the exposure assessment the temporal relationships of each compound
to the other compounds. The exposure and toxicity analysts may then work together to
consider the longevity of each chemical in target organs or tissues and how the
dynamics and kinetics of each compound might influence the mode or mechanism
and/or dose at a target site.
5-11
-------
Some chemicals may bioaccumulate and persist in the blood or in target tissues
for months or even years following exposure (e.g., methylmercury, cadmium). In
contrast, many other environmental pollutants such as PCBs undergo significant
redistribution to non-target tissues such as adipose, effectively decreasing the duration
of exposure at a potential target site (Oberg et al., 2002). As such, characterizing the
"effective" duration of a compound (i.e., time spent above a toxic threshold) in a target
tissue and the manner in which a compound exerts biological effect (toxic mode of
action, MOA) is critical.
The expression of mutagenic/genotoxic effects is presumably independent of the
time-course of tissue exposure. That is, replication of errors in genetic material may
persist long after the toxicant and its original (e.g., a DMA adduct) insult have been
removed from the target tissue (U.S. EPA, 1991b, 2001f; IPCS, 1998). For most other
modes or mechanisms of toxicity, the duration and magnitude (severity) of an adverse
effect may be highly influenced by the dynamics and kinetics of compounds co-located
within the same compartment (e.g., subcellular organelles, cells, whole organs or
tissues). For example, exposure to an inducer of metabolic enzyme systems (e.g.,
cytochrome P450 family) such as chronic ethanol may increase the metabolism and
clearance of environmental pollutants via upregulation of CYP2E1 expression/activity
(Johns et al., 2006). Conversely, exposure to TCDD has been shown to enhance the
accumulation of Cd, and likewise Cd inhibits the biotransformation of TCDD, thus
influencing the bioaccumulation and potential toxicity within target tissues (Regoli et al.,
2005). Thus, important factors in the estimation of cumulative risk involve
characterization of the biological longevity of bioaccumulated compounds at a target
site, the biological longevity of effects, the duration spent above a toxic threshold and
mode or mechanism of action of each. Additionally, the restorative or regenerative
capacity of the tissue(s) affected must be taken into account (e.g., liver versus brain).
5.3.3. Integrative Cumulative Risk Assessment. Based on the temporal analysis of
the exposure and the toxicity information, the analyst forms final chemical groupings for
use in the integrative cumulative Risk Characterization. This step consolidates the
multichemical exposure data and the mixtures dose-response information. The
integrated data might include extrapolations for dose-response elements (animal
species, exposure route or duration, joint toxicity, population susceptibility) and
conversions of measurement units (exposure or dose, toxic effects and assumptions
regarding biological level of organization). Other simplifications (e.g., grouping
chemicals by target organ instead of by more sophisticated knowledge of toxic MOA)
5-12
-------
and other notable omissions (e.g., exclusion of certain exposure pathways, chemicals,
subpopulations, toxic effects, toxicity pathways) may be identified.
For important decisions, expert elicitation may be used to develop final chemical
groups. This practice relies on expert consideration of scientific theories and available
data. The experts then provide judgments in the form of subjective probability
distributions. These judgments can be combined and integrated into the analysis
(DeGroot, 1970; Cooke, 1991); Evans et al. (1994) use this approach for estimating
chemical carcinogenicity in the low-dose region of the dose-response slope.
5.3.4. Evaluation of Extrapolations, Simplifications and Omissions. In this step
the analysts as a group (or an independent group) determines if the extrapolations,
simplifications and omissions employed in the previous step are acceptable. Many of
the extrapolations, simplifications and omissions will result from a lack of knowledge
(e.g., lack of knowledge regarding the toxic MOA; uncertainty regarding exposure
duration), which is sometimes referred to as epistemic uncertainty. The analyst will use
scientific judgment regarding the grouping decisions (e.g., deciding that lack of a
complete exposure pathway would eliminate a certain group of chemicals from the
analysis) and agreements reached during Planning and Scoping (e.g., an agreement to
retain certain relevant chemicals in an assessment regardless of their exposure levels).
It is important to clearly and transparently describe the bases for these judgments,
including the evidence available to support the judgment and the degree of consensus
within the scientific community, so that the underlying logic can be evaluated further. If
the extrapolations, simplifications and omissions are judged to be acceptable, then
quantitative risk estimates are developed. If not, then the analyst describes the
assessment qualitatively, highlighting the limitations in the data; alternatively, the
analytic scope can be revised. In either case the limitations or needed data are
described.
5.3.5. Develop Quantitative Risk Estimates. The analyst next develops quantitative
risk estimates for each subpopulation and the general population (if relevant). This
integrative step in the Risk Characterization includes each relevant chemical group,
each relevant exposure pathway and route and all timeframes analyzed. Detailed
methods are provided in Chapter 4.
5.3.6. Sensitivity and Uncertainty Analysis. Characterization of variability and
uncertainty is integral to all risk assessment steps (Morgan and Henrion, 1990; NRC,
5-13
-------
1994; Cullen and Frey, 1999). Variability refers to population heterogeneity, such as
body weights that vary across individuals. Uncertainty is described as a lack of
knowledge about the correct value for a specific parameter or the correct model. Both
Chapters 3 and 4 discuss uncertainty in the context of exposure assessment and dose-
response assessment, respectively.
In the sensitivity and uncertainty analysis for the Risk Characterization, the
analyst identifies sources of uncertainty in the previous analytic steps and
systematically evaluates their impacts on the final analytic results. This may include in-
depth analyses of the uncertainties in the models and monitoring data used. If models
are used extensively, the analysis could include an examination of model uncertainty
(Cullen and Frey, 1999). Parameter uncertainty and variability are also evaluated at this
stage of the analysis. Section 4 of this chapter provides additional detail on variability
and uncertainty with respect to cumulative exposure assessment and cumulative dose-
response assessment.
5.3.7. Cumulative Risk Characterization. Finally, the analyst develops the
cumulative Risk Characterization for the integrative analysis. This includes a technical
communication of the risk estimates, uncertainty analysis and critical data gaps. Risk
estimates include a description of the expected toxicity for each population of concern, a
quantitative estimate of the risk, and the size of the population likely to be affected. The
Risk Characterization for that population is then a key result of the risk assessment and,
at a minimum, includes the description of risk for the average population exposure,
along with the size of the population. These population groups may reflect those with
high single chemical exposures as well as those with high exposure to interactive
chemical combinations. Subgroups of concern include those that are inherently
sensitive because of biological characteristics and those that are of increased risk
because of the cumulative aspects of risk, namely toxicological interactions. Such
sensitivity might be related to physiologic characteristics or exposure (e.g., lifestyle)
factors that could enhance the synergistic activity of the chemicals. This latter group is
unique to cumulative risk assessment. The risks address those identified in the
Problem Formulation phase and the toxicity groups via all major exposure pathways.
Because the setting includes multiple chemicals with exposure potentially by multiple
routes and time frames, the number of health effects to be addressed can be quite high.
For example, even if one only described risks for the critical toxic effects, ignoring
secondary effects and joint toxicity, there can be different effects for each chemical, by
each route and for each time frame of exposure. Moreover, the potential for several
5-14
-------
sensitive subgroups means that the distribution of effects and severities to consider can
be quite broad.
The cautionary advice most often given for cumulative Risk Characterizations is
to be clear and avoid oversimplification. With sufficient information, each of the
parameter combinations could be assessed separately, resulting in a distribution of risks
that covers the range of combinations of exposure and population subgroup. In many
cases, however, the information required for a complete quantitative Risk
Characterization of these combinations will be unavailable. At the least, the analyst
could provide a recommended risk estimate for the population, such as a central or
median risk estimate for the average individual, along with a risk estimate for the high
end of the population risk distribution. The high-end Risk Characterization describes
the assumed conditions leading to that high risk. Of particular importance is the
plausibility of the co-occurrence of the many factors related to the high-end risk. For
example, the risk associated with a given daily oral exposure might be highest for a
child because of the low body weight. The risk for an exercising adult (all else being
equal) might be highest because of the high daily drinking water intake. For a plausible
high-end risk estimate, the child body weight is combined with the child daily intake and
similarly for the adult; it would be unrealistic to combine the two extremes: a low body
weight (e.g., the 10-kg child) with a high daily oral intake rate (e.g., the exercising adult).
A correlation analysis between intake rates and body weights also may be needed to
clarify the relationship between the two parameters.
The multiplicity of potential health effects in a diverse population raises another
complexity issue: the presentation or evaluation of the combination of different effects.
The traditional approach using a single critical effect avoids this issue so that the
population risk can be attached to one type of toxic endpoint, e.g., reproductive effects.
With cumulative risk assessments, there may be several toxic effects of differing
severity and with different ways to measure or describe them, including some
quantitative and some judgmental. One approach described earlier (Chapter 4) relies
on converting the observed effects into a small set of severity categories so that
different effects can be compared based on their toxic severity. Another approach is to
simplify the effects description by tying the risks to toxicity groups (see Chapter 4 and
Appendix B). In either case, the presentation of results includes a list of all effects
addressed by each risk measure, along with a discussion of the more likely effects.
Because of possible differences in exposure durations and treatability of the effects, it is
useful to include any information on the persistence or reversibility of the most likely
effects.
5-15
-------
Specific population subgroups of main concern might be identified in the
Planning and Scoping stage of the assessment. Some subgroups might be linked to
the initiating factor that led to the cumulative risk assessment. Other subgroups of
concern might be identified during the exposure assessment or the dose-response
analysis. For example, proposed siting of a chemical manufacturing plant might be
nearest to the population subgroup that initially raised the issue, while emissions could
disperse to cause wider-spread exposure. Those subgroups identified in the Planning
and Scoping stage are included in the Risk Characterization. It is important that results
are described in terms of the factors decided on during Problem Formulation to ensure
that the questions of central concern to the stakeholders have been answered.
Several potentially sensitive population subgroups might be identified during the
exposure and toxicity assessment steps. It is good practice to describe the risks to
these subgroups along with estimates of the size of each subgroup, for completeness
as well as improved information for the risk managers. For example, remediation of
organics in groundwater by air stripping might need to be designed to avoid increasing
exposures to potentially sensitive subgroups that might reside downwind of the air
stripper.
5.4. VARIABILITY AND UNCERTAINTY IN EXPOSURE AND DOSE-RESPONSE
This section describes possible sources of variability and uncertainty in
cumulative risk analyses, followed by discussions of variability and uncertainty with
respect to cumulative exposure assessment and cumulative dose-response
assessment.
5.4.1. Usefulness of Variability and Uncertainty Analyses in Cumulative Risk
Assessments. Understanding the uncertainty inherent in a cumulative risk assessment
helps risk managers to understand the possible range of risks posed by the situation
they are evaluating and provides additional insights into risk management opportunities
at the site. Ultimately, uncertainty analyses help the risk manager avoid poor decisions
and improve his or her appreciation for the potential ramifications of each alternative
considered. While no single approach can be used to address all possible sources of
variability and uncertainty in risk assessment, sensitivity analyses provide an
opportunity to evaluate the confidence that can be placed in an assessment and to
identify and prioritize critical research to improve future risk assessments. The NRC
(1994) encourages the development of uncertainty analyses in risk assessments. The
qualitative identification of sources of uncertainty and variability is currently a routine
5-16
-------
component of most human health risk analyses, regardless of their level of
sophistication (e.g., screening level analyses include an identification of sources of
uncertainty and variability).
The EPA recognizes that secondary data are data used for a purpose other than
that for which they were collected. That does not imply that their quality is reduced but
that the appropriateness of their inclusion in the given application be justified.
The quality of experimental data used in a cumulative Risk Analysis can differ
widely and needs to be addressed in a cumulative risk assessment. The data quality
issue is particularly important when the qualitative measures (e.g., response levels or
chemical measurement data collected at the site) or critical qualitative findings (e.g.,
criticality of a certain biochemical event in an MOA pathway or assumptions regarding
the environmental fate of mixture components released to the environment) are
introduced into the cumulative risk assessment.
5.4.2. Exposure Assessment Uncertainty and Variability.
5.4.2.1. Variability in Cumulative Risk Exposure AssessmentsVariability
refers to population heterogeneity, such as tap water consumption rates varying across
individuals and over time (e.g., the same individual may consume more water during hot
arid weather than cold weather leading to seasonal exposure variations). Studies
exhibiting increasing sophistication have been undertaken to increase the
understanding of variability in exposure factors (e.g., drinking water intake rates) in key
subpopulations (e.g., women of reproductive age, pregnant women and children) (U.S.
EPA, 1997c). While tap water consumption can be measured accurately, the true
values will vary across the population and no increase in the level of precision in
measurement techniques will reduce this variability across the population. A second
component of variability involves analyzing correlations among variables (e.g.,
correlations of food intake rates and body weights). Correlation analyses among
exposure factors also continue to improve.
5.4.2.2. Uncertainty in Cumulative Exposure AssessmentsUncertainty or,
more specifically, epistemic uncertainty, is described as a lack of knowledge about a
specific exposure pathway, measure or estimate. Uncertainty is typically more difficult
to quantify than variability.
The exposure scenarios developed for a cumulative risk assessment involve
multiple chemicals and multiple environmental media. In this approach, analyses of
individuals who are co-exposed to chemicals may rely on dynamic fate models. The
5-17
-------
uncertainties in these models and their implications on the assessment results may be
discussed in the Risk Characterization.
The concentrations of these chemicals in various environmental media may be
estimated through direct analytical measurement, predictive modeling or some
combination of the two. Thus, it may be important to examine the sensitivity and
specificity of different analyses used to measure the concentrations of different
chemicals or the same chemicals in different media and, if possible, integrate that
information into the risk assessment or the sensitivity analysis. The quantitative
uncertainty of model predictions for concentrations of chemicals in different media may
also vary. While the scientific understanding of some environmental fate processes
may be well measured and understood, others may be less so. Mathematical models
describing well understood processes may be accepted generally by the scientific
community. For poorly understood processes, there may be competing models. In
such situations, it may be important to analyze model uncertainties.
Exposure assessment uncertainties also include sources of error such as
inaccuracies in the analytical methods for quantifying the level of chemicals in
environmental media (e.g., surface water). There is a true value for such a
concentration in this example, but the available methods may not be sensitive enough
to provide an accurate answer. In the exposure assessment, good scientific practice
would dictate that the treatment of data determined to be below the analytic detection
limit be carefully considered because assumptions regarding the "true" contaminant
concentrations in such samples influence exposure estimates (e.g., Fristachi and Rice,
2007). When combining information on chemical concentrations in the characterization,
clear identification of the limits of the techniques used to estimate these concentrations
is necessary. If the authors of such reports present error bounds, then it is good
practice to state the types of factors included and not included in the analysis.
Ingestion, inhalation and dermal contact rate information may be developed from
different sources. The quality of the sources could vary. The EPA Exposure Factors
Handbook (U.S. EPA, 1997c) recommends specific ingestion rates for foods such as
vegetables and freshwater fish and drinking water; it describes many of the limitations in
the underlying studies. The relevance of the data from the recommended studies to the
populations being evaluated also may be examined. For example, freshwater fish
consumption rates among individuals in certain Native American tribal groups may be
greater than those for the general U.S. population (e.g., Peterson etal., 1995; Toy et al.,
1995); these populations also may consume parts of the fish that are typically not
consumed by the general population.
5-18
-------
In a traditional risk assessment, an exposure is often defined as an event
occurring in a specific place and at a specific time. In cumulative risk assessment, the
focus is on the population of concern so that all relevant exposures are to be included.
The exposures then might encompass a number of events at several locations over
broad and varied time periods. These temporal and spatial aspects of cumulative Risk
Analyses might then require additional consideration as the dose-response data are
integrated in the Risk Characterization.
Finally, the characterization of complex exposures, even to a single chemical,
might include well measured exposures along with those that are conjectural or poorly
understood. For example, concern might exist for consequences of natural disasters
(e.g., lightning induced fires, flooding) or mechanical malfunction (e.g., intermittent
emissions from an aging incinerator), neither of which may have occurred at the site
being assessed. One option is to present the combined exposures and risks
numerically for those aspects that can be quantified and then describe the complete
exposure and risks in qualitative terms, estimating the impact on the risk estimate of the
missing factors. In these situations, the analyst identifies the source of the uncertainty,
the available information to address it and the assumptions invoked in the Risk Analysis
to compensate for the missing information.
5.4.2.3. Methods for Uncertainty AnalysisMethods to quantify uncertainty
are better developed for the field of exposure assessment than the fields of hazard
identification, dose-response assessment and Risk Characterization. Most exposure
assessments quantitatively estimate variability and uncertainty using probabilistic
techniques that make use of Monte Carlo simulations and there is EPA's Guiding
Principles for Monte Carlo Analysis (U.S. EPA, 1997f). Probabilistic exposure models
allow for a robust examination of the factors that lead to exposures and provide a basis
for addressing highly exposed populations. These types of models can examine and
evaluate the impacts of parameter uncertainty and variability. Intake rates and body
weights are highly variable in the population so using a single-point estimate for these
variables instead of probability distributions ignores inherent variability that may
influence exposure estimates. Monte Carlo methods lead to an approximation of a
sampling distribution by statistically incorporating probability distributions of
concentrations in media to which people are exposed and exposure factors, such as
those found in the EPA Exposure Factors Handbook (EFH) (U.S. EPA, 1997c). The
EFH provides a summary of the available statistical data on various factors used in
assessing human exposure. The factors provided include drinking water consumption;
5-19
-------
soil ingestion; inhalation rates; dermal contact factors including skin area and soil
adherence; consumption of fruits and vegetables, fish, meats, dairy products and
homegrown foods; breast milk intake; human activity factors; consumer product use;
and residential characteristics.
Monte Carlo analysis is very useful because it leads to an approximation of an
estimate of the statistic's sampling distribution by treating observed data as if it were the
unknown population and then sampling from this new population. Within a Monte Carlo
simulation, values from each distribution are randomly selected and input to the model;
the output of each iteration is stored in a new distribution that is subsequently analyzed
(U.S. EPA, 1997f).
Because the output of the simulation provides a distribution of results, central
tendency estimates of exposure to the chemicals comprising the mixture as well as the
statistical dispersion around the central tendency estimate (e.g., 5th and 95th percentile
values) can be evaluated. Although beyond the focus of the discussion in this chapter,
other techniques, such as 2-dimmensional Monte Carlo methods, attempt to
quantitatively distinguish between sources of variability and uncertainty (Morgan and
Henrion, 1990; Hoffman and Hammonds, 1994; Simon, 1999). Other methods that do
not randomly sample from the statistical distributions can also be employed, depending
on the goals of the analysis, such as Latin Hypercube sampling, which
disproportionately draws samples from the upper and lower tails of the statistical
distribution (Cullen and Frey, 1999).
5.4.3. Uncertainty and Variability in Dose-Response Assessment for Cumulative
Risk Assessment. Although a variety of quantitative uncertainty methods have been
developed and extensively utilized in exposure assessment, the state of this practice in
dose-response analysis is not as advanced. Most analyses of uncertainty and
variability in dose-response assessment have focused on the qualitative identification of
the sources of uncertainty (e.g., extrapolation from rodent bioassay data to human
dose-response estimates) or the application of default uncertainty factors. Thus,
uncertainty and variability analyses in cumulative risk assessment ultimately may
require both the development of new methodological approaches and also
consideration of additional variables beyond those considered in single chemical dose-
response analysis.
5.4.3.1. Hierarchy of Data Sources for Assessing Exposure-response
Relationships for Chemical MixturesWhen evaluating sources of dose-response
5-20
-------
data for chemical mixtures, EPA's mixture guidance documents (U.S. EPA, 1986a,
2000a) suggest the following order of preference for data:
1) The mixture of interest
2) A sufficiently similar mixture
3) Mixture components
Dose-response data from human studies is preferred to animal bioassay data; animal
toxicology data is preferred to ex vivo, in vitro and in silica data.
Often, epidemiology data (i.e., occupational data) and toxicological data are
available only for commercial mixtures of environmental contaminants. If the
environmental mixture is similar to the commercial mixture, then the commercial mixture
dose-response data can be utilized in the assessment. If, in the environment, the
mixture components are differentially transported, partitioned, transformed, degraded
or bioaccumulated, then the use of the commercial mixture dose-response data can be
highly uncertain, because the environmental mixture to which people may be exposed is
likely to differ from the commercial mixture. Consequently, the toxicity of the
toxicologically tested mixture may differ from the environmental mixture and such data
may not be very useful in the dose-response analysis. The analyst may judge whether
the environmental mixture is sufficiently similar to the tested; these analyses are
typically based on ratios and concentrations of components. If the analyst judges the
mixtures to be sufficiently similar then the dose-response data from the tested mixture is
used. If the mixtures are judged to not be sufficiently similar, then mixture component
data are used (Chapter 4; see also U.S. EPA, 1986a, 2000a).
Epidemiologic data is preferred over animal toxicological data, because, following
exposure, there may be differences in absorption, distribution, metabolism (retention)
and elimination across species. Also, the toxicity (toxicodynamics) of mixtures can vary
across species. Even if there are epidemiologic data, there are possible sources of
uncertainty to consider when using epidemiologic data for cumulative risk. These
include (1) study design issues such as potential confounding and other biases,
inadequate sample size and follow-up, (2) the choice of the dataset, (3) specification of
the dose-response model, (4) estimation of exposure and dose and (5) unrecognized
variability in susceptibility (Stayner et al., 1999). Further, it is important to note that
exposures to the same environmental mixtures can vary substantially across time and
place due to differential partitioning, etc. (see previous discussion of environmental fate
and transport factors and sufficient similarity; also see discussion of sufficient similarity
in U.S. EPA, 2000a).
5-21
-------
In vivo whole mixture data (e.g., the testing of concentrated whole DBP mixtures
in rodent bioassays conducted by Simmons et al. [2002]) are preferred over ex vivo, in
vitro, and in silica data. In vivo data account for absorption, distribution, metabolism
and elimination of the mixture. The other listed sources of toxicity data either do not
account for these factors or only partially account for them (e.g., addition of the S9
fraction to Salmonella reverse mutation assays). These other sources of toxicity data
may not be useful for identifying secondary effects of chemical mixtures (i.e., effects
associated with doses that are higher than those needed to elicit the primary or critical
effect).
The uncertainties associated with simple mixture component methods such as
the HI, RPFs and response addition are generally considered to be larger than those
associated with dose-response data based on whole mixtures. For example, analyses
that rely on such methods may not assess the toxicity associated with all components of
the mixture.
The HI approaches are based on a defined RfD or a surrogate for secondary
effects (see Text Box 4-2 on Target Organ Toxicity Doses in Chapter 4). The RfD is
derived by the application of uncertainty factors, which may each be thought of as
values from a distribution of values (when the true value is unknown) ranging from 1-10
(Swartout et al., 1998).2 In addition, the true value of the NOAEL or LOAEL is subject
to the details of experimental design of the toxicity study, not the least of which is the
selection of dose spacing. Thus, it is important to recognize the uncertainty in the UFs
themselves and in the resulting RfDs. Additional attention may be given to whether the
critical effect was determined in animals or was detected in humansRfD values
resulting from the latter case are more certain than RfD values derived from toxicity
characterized in research animals.
Additivity methods such as dose addition and response addition (U.S. EPA,
2000a) are relatively simple mathematical models depicting responses to mixtures
comprised of chemicals sharing a common toxic MOA or independent MOA while
affecting the same target tissue or toxic endpoint, respectively. These component
models typically are associated with greater uncertainty than dose-response
assessments developed when the toxicity of the whole mixture is evaluated.
It may be important to discuss the uncertainties associated with grouping mixture
components into common MOA subgroups or in assuming that they independently
affect the same target tissue and to describe the implications of other toxicity groupings.
2 The EPA has recognized the potential overlap in the areas of uncertainty by establishing a cap of 3000
when four areas of uncertainty are required to derive a reference value (U.S. EPA, 2002e).
5-22
-------
When determining groups of chemicals (as shown in Figure 4-7b), the evaluation of
component data includes steps that require consideration of target organ-specific data.
Toxicity databases, such as the EPA IRIS database, may provide toxicological
information only on a single critical effect (i.e., that effect occurring at the lowest
exposure level). Additional data such as those in other EPA documents, ATSDR
toxicological profiles and interaction profiles, or those obtained from primary literature
searches may be necessary to identify additional effects and target organs. Whether
adequate dose-response data are available affects the grouping of chemicals and the
potential for estimating the joint toxicity of the chemical combinations. When
information on secondary effects is inadequate, it may be important to address the
impact of this uncertainty, particularly regarding joint toxicity that may be
underestimated for those secondary effects.
Many biological systems comprise diverse biochemical processes and multiple
organs or tissues. A single chemical insult to a specific part of the overall pathway (e.g.,
endocrine function) may result in a toxicity seemingly unrelated to the chemico-
biological interaction. Multiple chemicals may have different mechanisms of action yet
impact the same system. These may result in upregulation or downregualtion of
individual steps of the system. The end result is that some effects may compensate for
other effects and that some effects may interact in a synergistic or additive manner. If
there are credible data that support synergistic or antagonistic interactions, then these
may be included in the cumulative risk assessment. However, in the absence of such
data, the assumption of a common or independent MOA used to support the choice of
dose addition or response addition, respectively, is identified as an area of assumption
and uncertainty.
5.4.3.2. Choice of Dose MetricChemical components of a mixture may
contain parent compounds and environmental transformation and degradation products.
Thus, chemical metabolism and its effect on toxicity may be important to consider.
When toxicity arises due to the formation of a bioactive metabolite, exposures to the
metabolite are better justified as dose metrics than exposure to the parent chemical. To
the extent possible, cumulative risk assessments identify when toxicity is due to a
chemical metabolite.
Mixtures dose response data also may be based on internal dose measures
(e.g., measures of chemicals in blood or in the target tissue). Such measures may
address chemical absorption, distribution, metabolism and elimination. Such measures
may be preferred to potential dose measures. When toxicity results from the formation
5-23
-------
of a metabolite, internal dose measures of the metabolite are better justified as dose
metrics than potential doses of the parent chemical.
5.4.3.3. Analysis of Uncertainties Associated with Dose-response Data
Both model-independent and model-dependent approaches using toxicological and/or
epidemiologic data are used in risk assessments. Major issues with choosing and
implementing different approaches to dose-response estimation include knowledge of
MOA and biological relationships at low dose. Due to limitations in experimental or
epidemiologic data, dose-response relationships are often extrapolated to the low dose
region from measured responses at high doses. Exposure misclassification can also
distort dose-response relationships in epidemiological studies, in particular for low
doses or exposures. Therefore, toxicological data may be preferred for dose-response
assessments for certain exposures. Although common practice in risk assessment,
extrapolation from the observed region of the dose-response curve to low dose levels
introduces uncertainty into the assessment.
Most dose-response models output central tendency estimates and confidence
limits based on variability from the dose-response function, but they provide little or no
characterization of uncertainty. Uncertainty can be introduced into a dose-response
estimate and subsequent risk values through model structure and parameter selection.
Uncertainties in quantitative data are, perhaps, the most worrisome in that the
cumulative risk assessment may be based on a less than certain measure of effect or
exposure (in the context that HQs are defined as exposures divided by acceptable level
of exposures). Both concerns over data quality and uncertainty in quantitative data may
be based on several factors including proximity of the measured event to the
toxicological event of interest, the likelihood of measurement (detection) error,
erroneous selection of a given dose-response model, biases (reported or not) in
experimental design as well as incongruence between the study purpose and the risk
assessment application. These technical potential defects and uncertainties can
adversely impact confidence in the cumulative risk assessment outcome through
imparting a higher level of uncertainty in quantitative data, owing to data quality issues
including issues of data collection and data application. These issues may be
highlighted when combining data from multiple single-chemical studies to develop a
cumulative risk assessment.
Commonly, uncertainty in toxicological data is addressed using uncertainty
factors. Uncertainty factors are used to compensate for deficits in knowledge
concerning the accuracy of toxicity data and the difficulty in estimating the health effects
5-24
-------
in a different species and/or in different exposure conditions. Other methods, such as
PBPK modeling using Monte Carlo simulation are available for evaluating human
variability in the dose-response assessment.
5.4.3.4. Consideration of Multiple EffectsCumulative risk assessments also
address multiple effects that result from exposures to one or more chemicals. Multiple
effects of a chemical mixture include the primary (critical effect) and secondary effects
(effects that result from higher exposures than the exposures producing primary effect.
Because the critical organ is already affected, further changes or alterations in the
toxicity in that organ are seldom studied or documented. Secondary effects may arise
due to several reasons:
The metabolism of the compound could become saturated with increasing dose
and higher doses may result in circulating levels of toxic parent chemical or
higher circulating levels of a toxic metabolite that may saturate clearance
mechanisms
Damage to the primary organ
A higher degree of resistance or repair capacity in the secondary organs such
that a higher level of exposure is required to produce these effects
Secondary tissues may differ in the inherent biochemistry and cellular
organization from the primary organ such that a different MOA is active in the
secondary tissues, and the exposure concentrations necessary to drive that MOA
are higher than the exposure necessary to stimulate the MOA in the primary
organ
In general, the latter is the assumed reason for secondary effects in cumulative risk
assessment, but the former three circumstances may be considered. These differ from
the latter in that the latter may be more directly employed as the basis for the
assumption of independence between the critical effect and secondary effects.
5.4.3.5. Duration of EffectsIn developing a cumulative risk assessment, it is
assumed that the period during which effects will become manifest is similar for all
contaminants. Data may be available to address whether certain chemicals differ in
their periods of latency for observable health effects. When different latency periods
(and durations of toxic effects) exist for components of a chemical mixture, then the
likelihood of a simultaneously-expressed toxicity is reduced. When possible, it may be
useful to address the quantification of the latency period and duration of toxic effects
and uncertainty in these estimates.
5-25
-------
5.4.4. Variability and Uncertainty Summary. In summary, quantifying the sources of
variability and uncertainty in the integrative analysis can be quite complicated. The
types of issues typically evaluated in quantitative uncertainty analyses include model
uncertainty, parameter uncertainty and uncertainty in assumptions that are developed
because of missing information. Identification of the sources of uncertainty in an
exposure assessment can increase the level of confidence in results and may further
help to determine the type of research needed to reduce it in future assessments.
Addressing uncertainty and variability increases the clarity and transparency of
cumulative risk assessments and may be quantitatively estimated to the extent possible.
When only qualitative characterizations are provided, their basis is described along with
suggestions for ways to improve and quantify those characterizations.
As has been discussed in several previous EPA risk assessment guidance
reports, a critical part of the uncertainty analysis concerns the possible impact of
missing information. For example, if the risk assessment produces a CHI<1, the
indication of safety may be false due to an information gap. The CHI calculation may be
evaluated and quantified where possible to estimate the likely change if the missing,
critical information were obtained. One example approach treats the possible impact on
a mixture risk estimate from unidentified chemicals in drinking water (U.S. EPA, 2003b).
Chemicals and exposure pathways that are not quantitatively included in the risk
assessment are placed in a watch list, so that when sufficient information becomes
available, their contribution to the cumulative risk can be assessed.
5.5. EXAMPLE EVALUATIONS OF QUANTITATIVE APPROACHES TO
CUMULATIVE RISK CHARACTERIZATION
Much of the process of cumulative risk assessment involves information sharing,
planning discussions and qualitative or judgment based decisions. The goal of this
information sharing is to develop accurate estimates of risks. Because cumulative risk
assessment includes many factors, some of which vary over time, the ideal risk
calculations would utilize supporting measurements and studies that may not exist. For
example, Section 4.7.1 presents a modified RPF approach for exposure to mixtures by
multiple pathways. The RPF approach requires information demonstrating that the
chemicals included in the calculation have similar toxicological MOAs. Such information
is not always available on all chemicals of concern. To illustrate the Risk
Characterization issues involved with quantitative risk assessment, some, but not all, of
the quantitative approaches presented in Chapter 4 are examined here in terms of
feasibility and impact on the risk assessment.
5-26
-------
5.5.1. Example Cumulative Risk Characterization: Cumulative Hazard Index. As
an alternative to the RPF approach of Chapter 4, the integration of multiple chemical
exposures along multiple pathways can be quantitatively represented in a simple
fashion by the CHI. The common dose-additive HI combines multi-chemical exposures
by summing the component exposure levels after each has been scaled by division by
that chemical's RfD (for ingestion) or RfC (for inhalation). (See Section 4.2.2.1 for a
complete description of the dose-additive HI.) The CHI as discussed here expands on
the uses and interpretation of the CHI used by Superfund in site evaluations. The
Superfund guidance first recommends calculating each chemical's exposure for each
completed pathway and then converting each into a pathway-specific, or more properly,
a route-specific HQ in the usual way. EPA's Risk Assessment Guidance for Superfund
(1989a) instructs analysts to sum HQs (Equation 5-1) across exposure routes and
exposure pathways, providing there is evidence of combined exposure pathways to
identifiable individuals or groups of individuals that would consistently face a reasonable
maximal exposure. For each chemical, the pathway HQs are summed to give the Risk
Characterization reflecting that chemical's total exposure to the individual or population
and expressed as a total HQ across exposure routes with those routes explicitly stated.
The CHI is then the sum of these totals across chemicals.
5.5.1.1. Calculation Steps The CHI calculation that follows is based on the
Superfund guidance (U.S. EPA, 1999b).
This equation solves for pathway-specific HQ for chemical/
where:
k = one of the pathways
Ejk = exposure for that pathway and
RVjk = the risk-based toxicity value for pathway k, such as the RfD for the water
pathway or the RfC for the air pathway.
This equation solves for total HQ for chemical j across m pathways:
(5-2)
k=1
The CHI across pathways and chemicals is then the sum across chemicals of the
total HQs:
(5-3)
5-27
-------
where n is the number of chemicals in the assessment. Not all chemicals need to be
present in a given pathway, and a given chemical need not be present in all pathways.
This latter condition means that in Equation 5-2, some terms might = 0.
5.5.1.2. InterpretationThe numerical value of CHI is an index of concern in
the same vein as the common dose-additive HI used for mixture Risk Characterization.
The numerical value is not interpreted as a risk number. For example, although a
higher CHI value indicates more concern for possible health effects, CHI = 8 does not
necessarily indicate a site hazard that is 4 times worse than if CHI = 2. The purpose of
the CHI is to express or indicate the degree of concern over possible toxic effects from
onsite exposure.
As with the mixture HI, the value of 1 could be used as the decision point for
determining whether further assessment or remedial action is warranted. When CHI>1,
the quality and nature of the CHI is generally examined. The analyst may re-evaluate
the exposure assessment to determine if more details are available, such as information
suggesting co-exposure by multiple pathways and review the dose-response
assessment, particularly the assumptions of similarity and no interaction (see Section
5.2.1.3), along with the other assumptions described in the EPA mixture guidance (U.S.
EPA, 2000a). The interpretation of CHI is not different from that of any HI calculation.
CHI is a risk indicator. If CHI>1, then the analyst examines the individual HQs in the
calculation to see if any one chemical, one exposure route or small group of chemicals
is driving the risk indicator.
When CHI<1, the indication is that no significant hazard exists by the chemicals
and pathways addressed. The key assumption to be checked is "no interaction." If any
indication of synergy exists from the supporting toxicity studies, then the analyst
evaluates the pathways involving those interacting chemicals in more detail. The
second check to be made is of the uncertainties, in particular the missing information
(see Section 5.3 for more details and suggestions).
Because the CHI involves simple sums, the summation can proceed in either
order:
Either sum across pathways for each chemical and then across chemicals
(Superfund's approach, given above in Equations 5-2 and 5-3)
Sum across chemicals to get a pathway specific HI and then sum Hi's across
pathways
The first sequence of summing gives an index of total risk per chemical and thus
identifies which chemicals are posing the highest hazard or risk. That approach might
5-28
-------
be useful in predicting the toxic effects that are most likely or of highest severity, keying
on the critical effects of those chemicals.
The second sequence gives an index of total risk per pathway, which might
assist in determining the preferred remediation approach. This approach might suggest
focusing on treating or mitigating the high-risk pathway without paying much attention to
the specific contaminants in that pathway. The best approach might be to perform both
intermediate calculations and present both the highest risk chemicals and highest risk
pathways to the decision makers. Previous experience by EPA in risk assessments of
Superfund waste sites indicates that in many cases risks will be dominated by one or
two chemicals and by one or two exposure pathways. These two intermediate
calculations will then help explain the extent of that dominance and provide support for
further simplification or reduction in the scope of the cumulative risk assessment.
This calculation is analogous to the cumulative risk approach used by the EPA
Office of Pesticide Programs (OPP). Although OPP uses margins of exposure (MOEs)
instead of HQs, once they are scaled by an uncertainty factor for species differences,
the total MOEs become nearly identical to the inverse of the total HQ. The primary
difference is in use of uncertainty factors. OPP considers whether there are
deficiencies in the database that apply to the chemicals as a group. The concern is tied
to the FQPA legislation that requires an additional safety factor when children's health is
an issue. If evidence indicates that another critical effect is produced by an identified
mechanism of toxicity at a dose significantly lower than the dose used in the risk
approach, then an additional database uncertainty factor is applied to the mixture
assessment to be protective for the young. OPP notes the importance of only applying
an uncertainty factor for database uncertainties once, i.e., either to a specific individual
chemical or as a group factor (U.S. EPA, 2002d).
5.5.1.3. Assumptions with Cumulative Hazard IndexThe Risk
Characterization step addresses the assumptions in the CHI determination and the
likely conditions under which the approach would be reasonable and those under which
it would be inappropriate. Similar to the use of the mixture HI (U.S. EPA, 2000a), the
CHI is useful for a screening level risk assessment because it is simple to determine
once the exposures have been estimated. The simple summation carries with it two
assumptions:
There are no interactions across exposure pathways or chemicals in terms of
toxicity
5-29
-------
There are no interactions across chemicals in terms of fate and transport or in
terms of single or multi-route uptake by the exposed individual
The main weakness then seems to be this assumption of no interactions. By
drawing analogies to mixture risk procedures, one can define conditions under which
this exposure additivity, i.e., the lack of interaction, is plausible. The chemical
properties under which the HI for mixtures risk is plausible all relate to concepts of
functional or structural similarity. The assumptions for the CHI are
each of the chemicals incorporated into the CHI are toxicologically similar for all
the pathways included in its pathway HQ calculation and have no significant
portal of entry effects (route-specific primary toxicity). Similarity here can be
indicated by the same toxic MOA, same primary target organs or similar general
type of toxic effect (e.g., cancer, reproductive toxicity). (For further discussion of
toxic similarity, see Section 4.4.) This property supports the combining of
exposures across pathways because for a given chemical, the same main toxic
effects occur for all pathways;
the chemicals grouped for a given pathway are toxicologically similar for that
pathway according to the requirements for dose addition. This property supports
the combining of chemicals for a given pathway, i.e., the pathway HI; and
perhaps most unique to cumulative risk assessment, the chemicals do not affect
each other's fate and transport, regardless of pathway.
Text Box 5-3 shows an example illustration. Text Box 5-3b briefly illustrates
calculations using dose and response addition.
ExampleSite Safety (Text Box 5-3)
Consider the case where a cumulative risk assessment goal is to determine with high confidence
whether a site is safe prior to initiating any site clean-up activities. One risk description could include
an overly conservative (health protective) estimate, perhaps based on the high-end exposure estimates
for each of the possible routes. If the risks predicted by this conservative approach are considered by
the risk manager of the site to be within acceptable levels, then any refined risk estimate is likely to be
lower, indicating high confidence of no health concern. For Risk Characterization described by the
CHI, then if CHI<1, this screening level conclusion is there is no health concern.
This approach is similar to the screening calculation of an HI that includes all chemicals,
temporarily ignoring the requirement of same target organ: if the mixture's screening assessment gives
Hl<1, even when including all target organs, then there is a conclusion of no health concern because
an improved and more appropriate HI restricted to a specific target organ would be even lower (U.S.
EPA, 2001 c). If CHI>1, then additional evaluation, perhaps a more refined analysis, is recommended.
Because the CHI is a conservative overestimate of the HI, a value exceeding the acceptable levels
does not imply the expectation of toxic effects but only that a more detailed risk assessment is needed.
For screening analyses, conservative CHI criteria that are <1 also can be employed by the risk
manager.
5-30
-------
5.5.2. Categorical Regression Calculations for Multiple Effects and Pathways.
One complication of cumulative risk, recognized in the EPA's Framework (U.S. EPA,
2003a), concerns the risk estimation and communication of multiple toxic effects. The
inclusion in the risk assessment of multiple stressors, pathways, exposure timeframes
and subpopulations increases the likelihood of multiple effects of concern. One
approach is to subdivide the Risk Characterization so that each division of the
document addresses only one of the likely toxic effects. This approach provides the
opportunity to include details that may be difficult to incorporate into a single
comprehensive risk characterization. An alternative is to address the multiple effects
directly in a single composite measure as described in Chapter 4.
5.5.2.1. CalculationsTwo formulas are given in Chapter 4 for describing
multiple effects for k = 1,...,n chemicals (see Section 4.5.1). These are restated here,
one based on the HI (Equation 5-4) and one based on response addition (Equation 5-5):
HI (effects) =
k=l
(5-4)
BMDLk/
and
Rm (effects) = ^Pk (severity > 2) (5-5)
k=1
or more accurately as
Rm(effects) = 1-flpk(severity < 2) (5-6)
k=1
where:
BMDL = Benchmark dose lower bound
UF = Uncertainty factor.
In both formulas, the underlying dose-response data, which include all effects of
concern, are first converted into dose-severity data by assigning each effect to a
severity category, where categories 3 and 4 represent toxic or lethal effects. (Greater
detail about these concepts can be found in Chapter 4 and in Appendix C.)
Equation 5-6, Rm(effects), is the probabilistic risk of any adverse effect for the
mixture. It is the general form of Equation 4-2, response addition for only two
chemicals. As with the common response addition for mixtures, Equations 5-5 and 5-6
become essentially identical for low risks (e.g., P/<<0.01). The representation in
5-31
-------
Equation 5-6 might be easier to follow because its factors are the results of categorical
regression as given in Equation 4-3.
Equation 5-4, Hl(effects), represents the HI for multiple effects from exposure to
the mixture. In Equation 5-4, the benchmark dose lower bound (BMDL) is derived from
categorical regression on the dose-severity data, and represents the dose associated
with a fixed low probability of toxicity, e.g., P(severity>2) = 0.10. The BMDL is scaled to
human terms by the uncertainty factor so that the denominator is similar to the RfD and
the formula corresponds to the standard mixture HI formula. The Hl(effects) calculated
in Equation 5-4 could be used in the CHI calculation of the pathway HI and would then
avoid the assumption of toxic similarity of the chemicals in that pathway. Because all
effects are included, the pathway HI and the resulting CHI would also reflect all effects
in the underlying dose-response data.
In Equation 5-5, the first step is to convert the doses in the supporting toxicity
data into human equivalent doses. That converted set of dose-response data is then
modeled using categorical regression as described above (and in Section 4.5.1). The
resulting regression formula is then used with the actual exposure estimates to generate
probabilities or risks of toxic effects (i.e., severity>2). The risk for the mixture is then
given by the sum of these chemical-specific risks. The mixture risk is not attached to
any particular toxic effect, as is the common single chemical benchmark risk, but
instead reflects all toxic effects in the underlying dose-response data and is then the risk
or probability of any toxicity. The interpretation of the risk addition approach is
straightforward for a mixture of chemicals in one pathway or environmental medium,
i.e., examining the assumption of independent toxic action among the chemicals. For
this regression on overall severity, this assumption might be described as the toxicity of
one chemical having no effect on the toxicity of another chemical in the mixture, which
is more plausible if the component doses are all low. The combined mixture risk is then
an estimate of the probability of toxicity (any effect) from one or more of the chemicals.
The extension to cumulative risk in terms of a combination across pathways is not as
clear.
For both evaluations of multiple effects, Rm(effects) and Hl(effects), the effects
observed in the animal studies are converted to severity categories, and then the model
produces the probability of observing a certain severity of effect, given dose. When
applied to a cumulative risk problem, the result does not correlate to any particular
health effect in the population but only provides evidence of whether or not there are
mixture risks of concern, taking into account that the chemicals cause multiple health
effects. If these metrics are large enough to raise concerns for a certain exposure
5-32
-------
scenario, then additional investigations may evaluate the population for health impacts,
or a decision could be made to begin some type of remediation.
5.5.3. Assumptions with Multi-route Formulas for Multiple Effects. The calculation
formulas for hazard or risk, for multiple effects by multiple routes, are similar to those
used for simple mixtures, but the assumptions are less clear and more difficult to
evaluate. For Equation 5-4, the use of an HI implies the assumption of similar toxicity
across the chemicals. The regression on all effects makes the interpretation more
complex. Because the BMDL indicates a specific risk of toxicity, the HI represents an
increasing concern as more chemicals approach or exceed their benchmark risk level.
The combining of multiple lower confidence bounds on the benchmark dose has not
been sufficiently investigated to allow a probabilistic interpretation in terms of a
confidence bound on the HI calculated in Equation 5-4.
Both of these approaches for addressing mixture risk for multiple effects are new
and have not been implemented in actual site assessments. One aspect related to
screening level assessments is the decision to base probabilistic risks on severity>2,
which means overt toxic effects. If a more conservative approach to the screening
assessment is indicated, then the calculations could be based instead on severity>1,
which would include effects that are not necessarily adverse. Further exploration of the
numerical properties of these approaches and scientific assumptions with respect to
transport and toxicity are encouraged.
5.5.4. Combination of Exposures of Different Time Frames. Risk estimates for
different time frames are developed by the analyst for the combined dose-duration
influence on toxicity. With complex aggregate exposures, the overlapping of exposures
that have quite different time courses is possible. An example is a low continuous
exposure (e.g., ambient air and drinking water) combined with intermittent exposure to
industrial pulse emissions, perhaps once a week at moderate to high levels. For acute
exposure to many chemicals, peak tissue concentration seems most appropriate as a
predictor of toxicity, i.e., accumulated dose or simple time-weighted averaging may be
inappropriate measures of toxicity (Boyes et al., 2000). For longer exposure periods,
simple cumulative dose (Haber's rule) often is inappropriate although a modified form
does seem acceptable as a dose-duration metric. The combining of joint exposures
over differing time frames uses the exposure metric appropriate to each exposure
period.
5-33
-------
The EPA and various scientists have published guidance, issue reports and
research results on the impact of exposure duration on toxicity but, to date, these only
consider single exposures for a fixed duration (Miller et al., 2000; Strickland and Guth,
2002; U.S. EPA, 1998d, 1999d, 2000c, 2004f; Zwart and Woutersen, 1988). The
complications with cumulative risk assessment include (1) the potential overlap of
exposures of different durations, (2) the persistence of an internal dose exhibited by a
long half-life in the body (e.g., half-life of methyl mercury is estimated to be 45-70 days
[U.S. EPA, 2007]) and (3) the persistence of an effect (e.g., the classical toxicodynamic
concept of cancer initiation vs. cancer promotion). The combination exposures are
evaluated jointly, as described in Chapters 3 and 4. When exposure duration is short,
less than a few days, then the analyst may undertake the following steps:
Estimating the combined exposure during the short exposure period, based on
the combination of the short and longer exposures. For example, a brief
exposure to a hepatic toxicant might be combined with a longer-term exposure to
another hepatic toxicant by summing their exposure levels or internal dose
concentrations, yielding a higher exposure level for the short duration. The
analyst may also estimate the persistence of the short-term physiologic changes
at the cellular or tissue level (see Section 5.2.4 below).
Developing a Risk Characterization specific to this short exposure period,
focusing on those significant effects that do not persist beyond the short
exposure period.
Determining whether any impacts from the short exposure (e.g., changes in cell
populations, upregulated enzymes) are likely to persist well into the longer
exposure period. Those effects are incorporated into the description of likely
toxicity for the longer period. The persistent effects might be increased by the
longer exposure and might influence other effects caused by the longer
exposure.
Clearly, additional research on risks posed by exposures to multiple chemicals
over different time frames would be useful.
5.6. OUTCOMES FROM CUMULATIVE RISK CHARACTERIZATION
The outcome of the cumulative Risk Characterization may provide a useful
integration of the data needed by the risk manager to make decisions regarding a
cumulative risk initiating factor. Results of the analysis may aid the risk manager in
deciding the extent of potential health risks from population exposures and whether
remedial action is necessary. A cumulative Risk Characterization may include sensitive
information such as the number of people exposed, risk estimates for health endpoints
of concern to the community, uncertainties regarding the exposure and health risk
5-34
-------
estimates and bottom-line conclusions in support of a regulatory decision. Thus, results
of the Risk Characterization are communicated clearly, with important issues and
uncertainties highlighted. Finally, it may be useful to articulate the identification of data
gaps, chemicals placed on a watch list and research needs that may improve the Risk
Characterization.
5.6.1. Interpretation of Results in the Context of Interaction Factors. The Risk
Characterization may be used to decide from among several risk management
response alternatives, from recommended changes in individual lifestyles of the
affected population to official governmental action. These responses often will involve
changing one or more factors in the scenario. For example, a remedial action could
include moderate reduction of all exposures or substantial reduction of some key
exposures. Because the cumulative risk assessment considers interactions (e.g., in
transport and toxicity), those same interactions will affect the post-remediation risk
assessment. Any remedial decisions will be enhanced if the key interactions are
identified and discussed in the Risk Characterization. Summaries that include a
quantification of toxicological interactions using the Interaction-Based HI (Section 4.6.2)
can show numerically how the remediation may have affected the risks by calculating
this metric before and after remediation. Other summaries that may include only a
qualitative indication of the direction of potential interactions might still be useful for
setting priorities or changing the degree of conservatism used in the assessment.
Table 5-1 illustrates whether joint toxicity is greater than, less than or equal to dose
addition for oral exposures to binary combinations of Cd, Pb, As and Cr (ATSDR, 2004).
This information can be used to qualitatively modify the interpretation of an additive (HI)
calculation containing these chemicals.
Background exposures, those exposures that are not necessarily site-related or
source-related, also can contribute to interactions. In many risk assessments, site
contamination is often assessed as an incremental exposure, and thus incremental risk,
i.e., the risk from the site exposure that exceeds background. Where possible, in
cumulative risk assessments background exposures are included in the exposure
estimate and in evaluations of interactions. Inclusion of background sources of
exposure may improve the characterization of population risk. These also can facilitate
comparisons of various sources of exposure (e.g., exposures from a group of sources
vs. background) in the analysis.
5-35
-------
TABLE 5-1
Joint Toxicity: Summary of Pairwise Toxic Interactions by Organ/System*
Metal
Interactions
Higher than
additive
Additive
Lower than
additive
Blood
As+Cd
As+Pb
Cd+Pb
Kidney
As+Cd
As+Cd
As+Cr
As+Pb
Cd+Pb
Neurological
As+Pb
Cd+Pb
Male
Reproductive
Cd+Pb
As+Cd
Skin
Cr+As
Cardiovascular
As+Cr
Cd+Pb
*AII exposures are oral. This table summarizes information in Table 4-2.
Source: ATSDR (2004).
5-36
-------
5.6.2. Interpretation of Results in Context of Problem Formulation. Results
highlight those risk estimates that address the issues identified in the Problem
Formulation phase according to the consensus details of the Planning and Scoping
phase (see the Planning and Scoping documents referred to in Chapter 1 for details.)
The risk assessment may contribute useful information to the risk management
decisions. In particular, the uncertainties may be linked to the stakeholder concerns
and interpreted in the context of the risk management options as well as the risk
estimates themselves. If the results do not seem to be compatible with the scope or are
not sufficiently accurate or detailed to be useful to the risk management decisions, then
the Planning and Scoping stage and Problem Formulation phase may be revisited. For
example, if the primary concern is risks caused by contamination at the site, then a
comparison may be needed with risks from exposures to background or off-site
contamination.
5.6.3. Interpretation of Results in Context of the Initiating Factor. Different initiating
factors may require different approaches to the cumulative risk assessment and thus
result in different outcomes and interpretations. For example, epidemiologic
approaches may be the best choice when the initiating factor for a cumulative risk
assessment is a population illness identified by a disease cluster (Section 2.5). When
the initiating factor is a particular source, environmental concentration or biomonitoring
result, the investigation may also require analyses using epidemiologic methods that
can be linked with the examination of chemicals and their sources. Thus, results will be
relevant to an existing population with known exposures to environmental chemicals
and observable health conditions. The observable health conditions can then be linked
with chemical exposures and the risks to subpopulations can be calculated (e.g., using
relative risk measures such as odds ratios). Vulnerability factors (e.g., age, smoking,
health status) are generally accounted for in epidemiologic analyses and can be
discussed in the Risk Characterization.
The approaches on multiple route exposures to multiple chemicals, discussed in
Chapters 3 and 4, may be used for any of the initiating factors. In certain cases,
estimates of potential impacts on human health may be made for a hypothetical
population (e.g., a population that may be exposed due to future land uses of a
contaminated site). This could be done even when the initiating factor is a current
population illness. Such analyses may include expected or anticipated exposures to
environmental chemicals and the potential for effects that are not yet exhibited but may
occur in later years (e.g., cancers that are expressed only after a long latency period).
5-37
-------
Thus, a chemical- or source-based analysis may be necessary and a link between
disease endpoints and chemical exposures articulated, along with an estimate of
potential population risks. The analyst may perform a WOE assessment to support
whether exposures to multiple chemicals are occurring and if significant health effects
may be anticipated in the population of interest. The population characteristics may
need to be articulated, including consideration of vulnerability factors.
5.7. SUMMARY
In summary, this chapter has stressed the importance of the Risk
Characterization phase of cumulative risk assessment and has endeavored to consider
issues in the context of evaluating multiple chemicals, exposures and effects, including
interaction effects, with respect to the population characteristics. Issues regarding
uncertainty, variability and sensitivity analysis have been discussed and a schematic
(Figure 5-1) has been presented for conducting a cumulative Risk Characterization. An
integrative technical analysis of the predicted risks is typically produced, as well as a
summary of the results and uncertainties of the Risk Analysis. Risk Characterization
results may be used by risk managers in the final Decision-Making stage of a
cumulative risk assessment; thus the Planning and Scoping process, data sources,
analytical techniques, logic used to make various technical decisions and uncertainty
analysis must be scientifically sound and presented in a transparent manner.
5-38
-------
6. REFERENCES
ACS (American Chemical Society). 2003. Long-Range Research Initiative. Available
at http://www.uslri.org/.
Alavanja, M.C.R., D. Sandier, S. McMaster et al. 1996. The agricultural health study.
Environ. Health Perspect. 104(4):362-369.
Alavanja, M.C.R., N.L. Sprince and E. Oliver. 2001. Nested case-control analysis of
high pesticide exposure events from the Agricultural Health Study. Am. J. Ind. Med.
39(6):557-563.
Alavanja, M.C.R., C. Samanic, M. Dosemeci etal. 2003. Use of agricultural pesticides
and prostate cancer risk in the Agricultural Health Study Cohort. Am. J. Epidemiol.
157(9):800-814.
Amdur, M.O., J. Doull and C.D. Klaassen. 1993. Casarett and Doull's Toxicology, 4th
ed. McGraw-Hill, Inc., New York, NY.
ARB (Air Resources Board). 1997. Chlorine. Available at
http://www.scorecard.org/chemical-profiles/html/chlorine.html. Also see NPI, 2005 (as
cited in Scorecard, accessed 2006).
Arcos, J.C., Y.T. Woo and D.Y. Lai. 1988. Database on binary combination effects of
chemical carcinogens. Environ. Carcinogen. Revs. J. Environ. Sci. Health (Part C).
6(1): 1-164.
Aschengrau, A., S. Rogers and D. Ozonoff. 2003. Perchloroethylene-contaminated
drinking water and the risk of breast cancer: additional results from Cape Cod,
Massachusetts, USA. Environ. Health Perspect. 111(2): 167-173.
ATSDR (Agency for Toxic Substances and Disease Registry). 1994a. Toxicological
Profile for Chlordane. U.S. Department of Health and Human Services, Public Health
Service. May. Available at http://www.atsdr.cdc.gov/toxprofiles/tp31.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 1994b. Toxicological
Profile for 1,1-Dichloroethene. U.S. Department of Health and Human Services, Public
Health Service. May. Available at http://www.atsdr.cdc.gov/toxprofiles/tp39.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 1996. Toxicological
Profile for cis-, trans-1,2-Dichloroethene (update). U.S. Department of Health and
Human Services, Public Health Service. August. Available at
http://www. atsdr. cdc. gov/toxprof i Ies/tp87. htm I.
ATSDR (Agency for Toxic Substances and Disease Registry). 1997a. Toxicological
Profile for Chloroform (update). U.S. Department of Health and Human Services, Public
Health Service. September. Available at http://www.atsdr.cdc.gov/toxprofiles/tp6.html.
6-1
-------
ATSDR (Agency for Toxic Substances and Disease Registry). 1997b. Toxicological
Profile for Tetrachloroethylene (PERC). U.S. Department of Health and Human
Services, Public Health Service. September. Available at
http://www.atsdr.cdc.gov/toxprofiles/tp18.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 1997c. Toxicological
Profile for Trichloroethylene (TCE). U.S. Department of Health and Human Services,
Public Health Service. September. Available at
http://www.atsdr.cdc.gov/toxprofiles/tp19.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 1997d. Toxicological
Profile for Vinyl Chloride (update). U.S. Department of Health and Human Services,
Public Health Service. September. Available at
http://www.atsdr.cdc.gov/toxprofiles/tp20.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 1999a. Toxicological
Profile for Lead. U.S. Department of Health and Human Services, Public Health
Service. July. Available at http://www.atsdr.cdc.gov/toxprofiles/tp13.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 1999b. Toxicological
Profile for Cadmium (Update). U.S. Department of Health and Human Services, Public
Health Service. July. Available at http://www.atsdr.cdc.gov/toxprofiles/tp5.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 1999c. Toxicological
Profile for Mercury. U.S. Department of Health and Human Services, Public Health
Service. March. Available at http://www.atsdr.cdc.gov/toxprofiles/tp46.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 1999d. Toxicological
Profile for Uranium. U.S. Department of Health and Human Services, Public Health
Service. September. Available at http://www.atsdr.cdc.gov/toxprofiles/tp150.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 2000a. Toxicological
Profile for Arsenic (Update). U.S. Department of Health and Human Services, Public
Health Service. September. Available at http://www.atsdr.cdc.gov/toxprofiles/tp2.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 2000b. Toxicological
Profile for Chromium (Update). U.S. Department of Health and Human Services, Public
Health Service. September. Available at http://www.atsdr.cdc.gov/toxprofiles/tp7.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 2000c. Toxicological
Profile for Polychlorinated Biphenyls (PCBs) (update). U.S. Department of Health and
Human Services, Public Health Service. November. Available at
http://www.atsdr.cdc.gov/toxprofiles/tp17.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 2000d. Case studies in
environmental medicine: arsenic toxicity: exposure pathways. October. Available at
http://www.atsdr.cdc.gov/HEC/CSEM/arsenic/exposure pathwavs.html.
6-2
-------
ATSDR (Agency for Toxic Substances and Disease Registry). 2001. Toxicological
Profile for 1,2-Dichloroethane. Update. U.S. Department of Health and Human
Services, Public Health Service, Atlanta, GA. Available at
http://www.atsdr.cdc.gov/toxprofiles/tp38.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 2002a. Toxicological
Profile for Aldrin and Dieldrin. U.S. Department of Health and Human Services, Public
Health Service. September. Available at http://www.atsdr.cdc.gov/toxprofiles/tp1.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 2002b. Toxicological
Profile for Beryllium. U.S. Department of Health and Human Services, Public Health
Service. September. Available at http://www.atsdr.cdc.gov/toxprofiles/tp4.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 2002c. Guidance
Manual for the Assessment of Joint Toxic Action of Chemical Mixtures. U.S.
Department of Health and Human Services, Public Health Service, Atlanta, GA.
ATSDR (Agency for Toxic Substances and Disease Registry). 2003a. Draft
Toxicological Profile for Carbon Tetrachloride. Agency for Toxic Substances and
Disease Registry, Atlanta, GA. September. Available at
http://www.atsdr.cdc.gov/toxprofiles/tp30.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 2003b. Draft
Toxicological Profile for Nickel. U.S. Department of Health and Human Services, Public
Health Service,. September. Available at
http://www.atsdr.cdc.gov/toxprofiles/tp15.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 2003c. Draft
Toxicological Profile for Zinc. U.S. Department of Health and Human Services, Public
Health Service. September. Available at http://www.atsdr.cdc.gov/toxprofiles/tp60.html.
ATSDR (Agency for Toxic Substances and Disease Registry). 2004. Interaction Profile
for Arsenic, Cadmium, Chromium and Lead. Agency for Toxic Substances and Disease
Registry, Atlanta, GA. Available at
http://www.atsdr.cdc.gov/interactionprofiles/IP-metals1/ip04.pdf.
ATSDR (Agency for Toxic Substances and Disease Registry). 2006. Interaction
Profiles for Toxic Substances. Agency for Toxic Substances and Disease Registry,
Atlanta, GA. Available at http://www.atsdr.cdc.gov/interactionprofiles/.
Baghurst, P.A., A.J. McMichael, N.R. Wigg et al. 1992. Environmental exposure to
lead and children's intelligence at the age of seven years, The Port Pirie Cohort Study.
New Eng. J. Med. 327(18): 1279-1284.
Birnbaum, L.S. 1995. Developmental effects of dioxins and other endocrine disrupting
chemicals. Neurotoxicology. 16(4):748.
6-3
-------
Black, M.R., D.M. Medeiros, E. Brunett et al. 1988. Zinc supplements and serum lipids
in young adult white males. Am. J. Clin. Nutr. 47:970-975 (as cited in ATSDR, 2003c).
Bolz, R.E. and G.L. Tuve. 1973. CRC Handbook of Tables for Applied Engineering
Science. CRC Press, Cleveland, OH.
Bosch, H.M., A.B. Rosefield, R. Huston, H.R. Shipman and F.L. Woodward. 1950.
Methemoglobinemia and Minnesota well supplies. J. Am. Water Works Assoc.
42:161-170 (as cited in U.S. EPA, 2006e).
Boyes, W.K., P.J. Bushnell, K.M. Crofton, M. Evans and J.E. Simmons. 2000.
Neurotoxic and pharmacokinetic responses to trichloroethylene as a function of
exposure scenario. Environ. Health Perspect. 108(Suppl. 2):317-322.
Brown, R. 1999. Personal communication from R. Brown to M. MacDonell, Argonne
National Laboratory, Argonne, IL (for the Hanford groundwater/vadose zone integration
project report). November, (as cited in U.S. DOE, 1999).
Bruckner, J.V., W.F. MacKenzie, W. Muralidhara et al. 1986. Oral toxicity of carbon
tetrachloride: Acute, subacute and subchronic studies in rats. Fundam. Appl. Toxicol.
6:16-34 (as cited in ATSDR, 2003a).
CCDE (City of Chicago Department of the Environment). 2003. Polynuclear Aromatic
Hydrocarbon Background Study: Illinois. Prepared by Tetra Tech EM Inc. under
contract to CCDE. February 24.
CDC (Centers for Disease Control and Prevention). 1984. Organophosphate
insecticide poisoning among siblingsMississippi. Morb. Mortal. Wkly. Rep.
33(42):592-594.
CDC (Centers for Disease Control and Prevention). 2006. Third National Report on
Human Exposure to Environmental Chemicals. July 2005. Revised February, 2006.
Centers for Disease Control and Prevention, Atlanta, GA. NCEH Pub. No. 05-0570.
Available at http://www.cdc.gov/exposurereport/.
Cebrian, M.E., A. Albores, M. Aquilar and E. Blakely. 1983. Chronic arsenic poisoning
in the north of Mexico. Human Toxicol. 2:121-133 (as cited in U.S. EPA, 2006e).
Chen, J.J., Y.J. Chen, G.E. Rice, K. Hamernik, A. Protzel and R.L. Kodell. 2001. Using
dose addition to estimate cumulative risks from exposures to multiple chemicals. Reg.
Toxicol. Pharmacol. 34(1):35-41.
Clement, T.P, M.J. Truex and P. Lee. 2002. A case study for demonstrating the
application of US EPA's monitored natural attenuation screening protocol at a
hazardous waste site. J. Contam. Hydrol. 59(1-2): 133-162.
CNN. 2004. Poisoning diagnosis'rock solid.' CNN World, Dec 12. Available at
http://www.cnn.com/2004/WORLD/europe/12/12/vushchenko/index.html.
6-4
-------
Cooke, R.M. 1991. Experts in Uncertainty: Opinion and Subjective Probability in
Science. Oxford University Press, New York. 321 pp.
Cox, C., T.W. Clarkson, D.O. Marsh et al. 1989. Dose-response analysis of infants
prenatally exposed to methyl mercury: An application of a single compartment model to
single-strand hair analysis. Environ. Res. 49(2):318-332 (as cited in ATSDR, 1999c).
CPDP (Carcinogenic Potency Database Project). 2004. Summary of Carcinogenic
Potency Database by Target Organ (Table 1). Funded by the National Toxicology
Program, National Institute of Environmental Health Sciences, U.S. Department of
Energy through Lawrence Berkeley National Laboratory, and the University of California
at Berkeley. March 23. Available at
http://potencv.berkelev.edu/pdfs/NCINTPPathology.pdf.
Crump, K.S., T. Kjellstrom, A.M. Shipp, A. Silvers and A. Stewart. 1998. Influence of
prenatal mercury exposure upon scholastic and psychological test performance:
Benchmark analysis of a New Zealand cohort. Risk Anal. 18(6):701 -713.
Cullen, A.C. and H.C. Frey. 1999. Probabilistic Techniques in Exposure Assessment:
A Handbook for Dealing with Variability and Uncertainty in Models and Inputs. Plenum
Press, New York, NY.
Cutler, J.J., G.S. Parker, S. Rosen, B. Prenney, R. Healey and G.G. Caldwell. 1986.
Childhood leukemia in Woburn, Massachusetts. Public Health Rep. 101(2):201-205.
Dawson, B.V., P.O. Johnson, S.J. Goldberg et al. 1993. Cardiac teratogenesis of
halogenated hydrocarbon contaminated drinking water. J. Am. Coll. Cardiol.
21:1466-1472 (as cited in ATSDR, 1997c).
DeGroot, M. 1970. Optimal Statistical Decision. McGraw-Hill, New York, NY. 489pp.
De Oliveira, F.S., M.R. Viana, A.R. Antoniolli et al. 2001. Differential effects of lead and
zinc on inhibitory avoidance learning in mice. Braz. J. Med. Biol. Res. 34:117-120 (as
cited in ATSDR, 2003c).
DeRosa, C.T., J.F. Stara and P.R. Durkin. 1985. Ranking of chemicals based upon
chronic toxicity data. Toxicol. Ind. Health. 1(4): 177-192.
Dourson, M.L., S.P. Felterand D. Robinson. 1996. Evolution of science-based
uncertainty factors in noncancer risk assessment. Regul. Toxicol. Pharmacol.
24:108-120.
Dourson, M.L., L.K. Teuschler, W.M. Stiteler and P.R. Durkin. 1997. Categorical
regression of toxicity data: A case study using aldicarb. Reg. Toxicol. Pharmacol.
25:121-129.
DTSC (Department of Toxic Substances Control). 2003. Johnson and Ettinger (1991)
Model forTSC Intrusion into Buildings. Version 3.0-Modification 1. July.
6-5
-------
Durant, J.L., J. Chen, H.F. Hemond and W.G. Thilly. 1995. Elevated incidence of
childhood leukemia in Woburn, Massachusetts: NIEHS Superfund basic research
program searches for causes. Environ. Health Perspect. 103(Suppl. 6):93-98.
E-Doc (Electronic Doctor) Index of Medical Terminology, (c) E-Doc 1998-99. Available
at http://www.edoc.co.za/.
Edmond, C., J.E. Michalek, L.S. Birnbaum and M.J. DeVito. 2005. Comparison of the
use of a physiologically based pharmacokinetic model and a classical pharmacokinetic
model for dioxin exposure assessments. Environ. Health Perspect.
113(12): 1666-1668. Available at
http://www.ehponline.org/members/2005/8016/8016.html.
EFSA (European Food Safety Authority). 2006. EFSA Colloquium 7 - Cumulative Risk
Assessment of pesticides to human health: The way forward. Available at
http://www.efsa.europa.eu/en/science/colloquium series/colloquium 7.html.
EHTPT (Environmental Health Tracking Project Team). 2000. America's
Environmental Health Gap: Why the Country Needs a Nationwide Health Tracking
Network. Technical Report. Prepared by EHTPT, Johns Hopkins School of Hygiene
and Public Health, Department of Health Policy and Management. Sponsored by the
Pew Environmental Health Commission. September. Available at
http://www.pewtrusts.com/pdf/hhs enviro health gap technical.pdf.
Elbetieha, A. and M.H. AI-Hamood. 1997. Long-term exposure of male and female
mice to trivalent and hexavalent chromium compounds: Effect on fertility. Toxicology.
116:39-47 (as cited in ATSDR, 2000b).
Elder, L, K. Poirier, M. Dourson et al. 2002. Mathematical modeling and quantitative
methods. Food Chem. Toxicol. 40:283-326.
Engel, L.S., H. Checkoway, M.C. Keifer et al. 2001. Parkinsonism and occupational
exposure to pesticides. Occup. Environ. Med. 58:582-589.
Evans, J.S., J.D. Graham, G.M. Gray and R.L. Sielken Jr. 1994. A distributional
approach to characterizing low-dose cancer risk. Risk Anal. 14:25-34.
Farland, W. and M.L. Dourson. 1992. Noncancer Health Endpoints: Approaches to
Quantitative Risk Assessment. In: Comparative Environmental Risk Assessment,
R. Cothern, Ed. Lewis Publishers Inc., Boca Raton, LA. December.
Feldman, R.G. 1992. Manganese as possible ecoetiologic factor in Parkinson's
disease. Ann. New York Acad. Sci. 648:266-267.
Feron, V.J., F.R. Cassee and J.P. Groten. 1998. Toxicology of chemical mixtures:
European perspective. Environ. Health Perspect. 106(Suppl. 6): 1281-1289.
6-6
-------
Ferrario, J.B., C.J. Byrne and D.H. Cleverly. 2000. 2,3,7,8-Dibenzo-p-dioxins in mined
clay products from the United States: Evidence for possible natural origin. Environ. Sci.
Technol. 34:4524-4532.
Fischer, P.W.F., A. Giroux and A.R. L'Abbe. 1984. Effect of zinc supplementation on
copper status in adult man. Am. J. Clin. Nutr. 40:743-746 (as cited in ATSDR, 2003c).
Foy, H.M., S. Tarmapai, P. Eamchan et al. 1992. Chronic arsenic poisoning from well
water in a mining area in Thailand. Asia Pac. J. Pub. Health. 6(3): 150-152 (as cited in
ATSDR, 2000a).
Freundt, K.J. and H.A. Ibrahim. 1990. Growth of rats during a subchronic intake of the
heavy metals Pb, Cd, Zn, Mn, Cu, Hg, and Be. Pol. J. Occup. Med. 3:227-232 (as cited
in ATSDR, 2002a).
Fristachi, A and G. Rice. 2007. Estimating the relative contribution of drinking water to
the daily ingestion of N-nitrosodimethylamine (NDMA). J. Water Health. (In Press).
Gierthy, J.F., K.F. Arcaro and M. Floyd. 1997. Assessment of PCB estrogenicity in a
human breast cancer cell line. Chemosphere. 34(5-7): 1495-1505.
Gilman, A.P., D.C. Villeneuve, V.E. Secours et al. 1998a. Uranyl nitrate: 28-day and
91 -day toxicity studies in the Sprague-Dawley rat. Toxicol. Sci. 41 (1): 117-128 (as cited
in ATSDR, 1999d).
Gilman, A.P., D.C. Villeneuve, V.E. Secours et al. 1998b. Uranyl nitrate: 91-day toxicity
studies in the New Zealand white rabbit. Toxicol. Sci. 41(1): 129-137 (as cited in
ATSDR, 1999d).
Gold, L.S., N.B. Manley, T.H. Slone and J.M. Ward. 2001. Compendium of chemical
carcinogens by target organ: Results of chronic bioassays in rats, mice, hamsters, dogs,
and monkeys. Toxicol. Pathol. 29(6):639-652. Available at
http://potencv.berkeley.edu/text/ToxicolPathol.pdf.
Gorell, J.M., C.C. Johnson, B.A. Rybicki et al. 1999. Occupational exposure to
manganese, copper, lead, iron, mercury and zinc and the risk of Parkinson's disease.
Neurotoxicology. 20(2-3):239-248.
Grandjean, P., P. Weihe, R. White et al. 1997. Cognitive deficit in 7-year old children
with prenatal exposure to methylmercury. Neurotoxicol. Teratol. 19(6):417-428.
Gunderson, V.M., K.S. Grant-Webster, T.M. Burbacher et al. 1988. Visual recognition
memory deficits in methylmercury-exposed Macaca fascicularis infants. Neurotoxicol.
Teratol. 10(4):373-379 (as cited in ATSDR, 1999c).
Guth, D. J. 1996. Acute exposure response assessment for 1,1,1 -trichloroethane using
stratified ordinal regression. Presented at the 89th Annual Meeting of the Air and Waste
Management Association, Nashville, TN. June 23-28.
6-7
-------
Guth, D.J., A.M. Jarabek, L. Wymer and R. Hertzberg. 1991. Evaluation of risk
assessment methods for short-term inhalation exposure. Presented at the 84th Annual
Meeting of the Air and Waste Management Association, Vancouver, British Columbia.
June 16-21.
Guth, D.J., R.J. Carroll, D.G. Simpson et al. 1997. Categorical regression analysis of
acute exposure to tetrachloroethylene. Risk Anal. 17(3):321 -332.
Guzelian, P.S., M.S. Victoroff, N.C. Halmes, R.C. Janes and C.P. Guzelian. 2005.
Evidence-based toxicology: a comprehensive framework for causation. Human Exp.
Toxicol. 24:161-201.
Haber, L.T., J.S. Dollarhide, A. Maierand M.L. Dourson. 2001. Noncancer risk
assessment: principles and practice in environmental and occupational settings. In:
Patty's Toxicology, 5th ed, E. Bingham, B. Cohrssen and C.H. Powell, Ed. Wiley and
Sons, Inc., New York, NY. p. 169-232.
Haddad, S., R. Tardif, C. Viau and K. Krishnan. 1999. A modeling approach to account
for pharmacokinetic interactions in the calculation of biological hazard index for
chemical mixtures. Toxicol. Lett. 108:303-308.
Haddad, S., R. Tardif, C. Viau and K. Krishnan. 2001. A modeling approach to account
for toxicokinetic interactions in the calculation of biological hazard index for chemical
mixtures. Toxicol. Sci. 63:125-131.
Haines, R.G. 1971. Ingestion of aldicarb by human volunteers: A controlled study of
the effect of aldicarb on man. In: EPA Pesticide Petition No. 1F1008. Unpublished
report.
Hertzberg, R.C. 1989. Extrapolation and scaling of animal data to humans: Fitting a
model to categorical response data with application to species extrapolation of toxicity.
Health Phys. 57(Suppl. 1):405-409.
Hertzberg, R.C. and M. Miller. 1985. A statistical model for species extrapolating using
categorical response data. Toxicol. Ind. Health. 1(4):43-63.
Hertzberg, R.C. and L. Wymer. 1991. Modeling the severity of toxic effects. Air &
Waste Management Association. In: Proceedings papers from the 84th Annual Meeting
and Exhibition, June 16-21, 1991, British Columbia. Air and Waste Management
Association.
Hertzberg, R.C., G.E. Rice and L.K. Teuschler. 1999. Methods for health risk
assessment of combustion mixtures. In: Hazardous Waste Incineration: Evaluating the
Human Health and Environmental Risks, S. Roberts, C. Teaf and J. Bean, Ed. CRC
Press, Boca Raton, FL. p. 105-148.
Hileman, B. 2001. The environment and Parkinson's. Chem. Eng. News. 79(38).
Available at http://www.mindfullv.org/Health/Parkinsons-And-Environment.htm.
6-8
-------
Hoffman, P.O. and J.S. Hammonds. 1994. Propagation of uncertainty in risk
assessments: The need to distinguish between uncertainty due to lack of knowledge
and uncertainty due to variability. Risk Anal. 14(5):707-712.
Hoppin, J.A., D.M. Umbach, S.J. London, M.C.R. Alavanja and D.P. Sandier. 2002.
Chemical predictors of wheeze among farmer pesticide applicators in the Agricultural
Health Study. Am. J. Respir. Grit. Care Med. 165(5):683-689.
Howard, P.H., Ed. 1989. Handbook of Environmental Fate and Exposure Data for
Organic Chemicals. Lewis Publishers Inc., Boca Raton, PL.
Hricko, A. 1994. Rings of controversy around benzene. Environ. Health Perspect.
102(3):276-281.
HSDB (Hazardous Substances Data Bank). 1991. Chlorine (CASRN 7782-50-5).
National Library of Medicine. Available at http://www.toxnet.nlm.nih.gov/.
HSDB (Hazardous Substances Data Bank). 2006. U.S. National Library of Medicine.
Bethesda, MD. Available at http://toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen7HSDB.
IPCS (International Programme on Chemical Safety). 1998. Environmental Health
Criteria 202. Selected non-heterocyclic polycyclic aromatic hydrocarbons. International
Programme on Chemical Safety.
IPCS (International Programme on Chemical Safety). 2006. HomePage. World Health
Organization. Available at http://www.who.int/ipcs/en/.
IRRST (Institut de reserche Robert-Sauve en sante et en securite du travail). 2003.
Mixtures of Substances in Workplaces: A Utility Program for Evaluating the Toxic Risk.
Available at http://www.irsst.gc.ca/en/ outil 100024.html.
Johns, D.O., W.E. Daniell, D.D. Shen, D.A. Kalman, R.L. Dills and M.S. Morgan. 2006.
Ethanol-induced increase in the metabolic clearance of 1,1,1-trichloroethane in human
volunteers. Toxicol. Sci. 92:61-70.
Johnson, P.C. and R.A. Ettinger. 1991. Heuristic model for predicting the intrusion rate
of contaminant vapors into buildings. Environ. Sci. Technol. 25:1445-1452.
Jones, N.L., T.P. Clement and C.M. Hansen. 2006. A three-dimensional analytical tool
for modeling reactive transport. Ground Water. 44(4):613-617.
Kedderis, G.L. 1997. Extrapolation of in vitro enzyme induction data to humans in vivo.
Chem. Biol. Interact. 107:109-121.
Khan, AT., A. Atkinson, T.C. Graham et al. 2001. Effects of low levels of zinc on
reproductive performance of rats. Environ. Sci. (Tokyo) 8:367-381 (as cited in ATSDR,
2003c).
6-9
-------
Khera, K.S. and S.A. Tabacova. 1973. Effects of methylmercuric chloride on the
progeny of mice and rats treated before or during gestation. Food Cosmet. Toxicol.
11:245-254 (referenced in ATSDR, 1999c).
Kjellstrom, T., P. Kennedy, S. Wallis et al. 1986. Physical and mental development of
children with prenatal exposure to mercury from fish. Stage 1: Preliminary test at age 4.
Natl. Swed. Environ. Protec. Bd., Rpt. 3080. (Solna, Sweden)
Kjellstrom, T., P. Kennedy, S. Wallis et al. 1989. Physical and mental development of
children with prenatal exposure to mercury from fish. Stage 2: Interviews and
psychological tests at age 6. Natl. Swed. Environ. Prot. Bd., Rpt. 3642. (Solna,
Sweden)
Kopp, S.J., T. Glonek, H.M. Perry Jr. et al. 1982. Cardiovascular actions of cadmium at
environmental exposure levels. Science. 217:837-839 (as cited in ATSDR, 1999b).
Kynast, G. and E. Saling. 1986. Effect of oral zinc application during pregnancy.
Gynecol. Obstet. Invest. 21:117-123 (as cited in ATSDR, 2003c).
Lagakos, S.W., B.J. Wessen, M. Aelen. 1986. An analysis of contaminated well water
and health effects in Woburn, Massachusetts. J. Am. Stat. Assoc. 81(395):583-596.
Liao, K.H., I.D. Dobrev, J.E. Dennison et al. 2002. Application of biologically-based
computer modeling to simple or complex mixtures. Environ. Health Perspect.
110(Suppl. 6):957-963.
Lipscomb, J.C. 2003. How differences in enzyme expression can translate into
pharmacokinetic variance and susceptibility to risk. J. Child. Health. 1:189-202.
Lipscomb, J.C. 2004. Evaluating the relationship between variance in enzyme
expression and toxicant concentration in health risk assessment. Human Ecol. Risk
Assess. 10:39-55.
Mackay, D., W.Y. Shiu, K-C Ma and S.C. Lee. 2006. Physical-Chemical Properties and
Environmental Fate for Organic Chemicals. CRC Press, Boca Raton, FL.
MADEP (Massachusetts Department of Environmental Protection). 2002. Technical
Update: Background Levels of Polycyclic Aromatic Hydrocarbons and Metals in Soil.
Available at http://www.mass.gov/dep/cleanup/laws/orspub03.htm.
Marnicio, R.J., P.J. Hakkinen, S.D. Lutkenhoff, R.C. Hertzberg and P.O. Moskowitz.
1991. Risk analysis software and data bases: Review of Riskware '90 Conference and
Exhibition. Risk Anal. 11:545-560.
Martin, S.A., Jr, D.P. Sandier, S.D. Harlow, D.L. Shore, A.S. Rowland and M.C.R.
Alavanja. 2002. Pesticide use and pesticide-related symptoms among black farmers in
the Agricultural Health Study. Am. J. Ind. Med. 41(3):202-209.
6-10
-------
McGuire, T.M., C.J. Newell, B.B. Looney et al. 2004. Historical analysis of monitored
natural attenuation: A survey of 191 chlorinated solvent sites and 45 solvent plumes.
Remediation. 99:112. Available at
http://www.gsi-net.com/Publications/McGuire HistoricalMNA pa.pdf.
Mehendale, H.M. 1995. Toxicodynamics of low level toxicant interactions of biological
significance: inhibition of tissue repair. Toxicology. 105:251-266.
Miller, F.J., P.M. Schlosser and D.B. Janszen. 2000. Haber's Rule: A special case in a
family of curves relating concentration and duration of exposure to a fixed level of
response for a given endpoint. Toxicology. 149:20-34.
Morgan, M.G. and M. Henrion. 1990. Uncertainty: A Guide for Dealing with Uncertainty
in Quantitative Risk and Policy Analysis. Cambridge University Press, Cambridge, UK.
Morgareidge, K., G.E. Cox and M.A. Gallo. 1976. Chronic feeding studies with
beryllium in dogs. Food and Drug Research Laboratories, Inc. Submitted to the
Aluminum Company of America, Alcan Research and Development, Ltd.,
Kawecki-Berylco Industries, Inc., and Brush-Wellman, Inc. (as cited in ATSDR, 2002c).
Mumtaz, M.M., K.A. Poirier and J.T. Coleman. 1997. Risk assessment for chemical
mixtures: Fine-tuning the hazard index approach. J. Clean Technol. Environ. Toxicol.
Occup. Med. 6(2): 189-204.
Myers, G.J., P.W. Davidson, C.F. Shamlaye et al. 1997. Effects of prenatal
methylmercury exposure from a high fish diet on developmental milestones in the
Seychelles child development study. Neurotoxicology. 18(3):819-289 (as cited in
ATSDR, 1999c).
Nadeenko, V.G., V. Lenchenko, S.B. Genkina and T.A. Arkhipenko. 1978. The
influence of tungsten, molibdenum, copper and arsenic on the intrauterine development
of the fetus. TR-79-0353. Farmakologiya i Toksikologiya. 41:620-623 (as cited in
U.S. EPA, 2006e).
Nakashima, M. S. Wu, T. Shigenko et al. 2005. A Field Application of HRC© Bio-
Barrier and Subsurface Microorganism Research. Proceedings of the Fourth
International Conference on Remediation of Chlorinated and Recalcitrant Compounds
(Monterey, CA, May 2004). Battelle Press.
Naranjo, E., F. Hellweger, L.H. Wilson and P. Anid. 2000. Mapping risk from mining
activities: A case study of Oruro, Bolivia. Proceedings of the Twentieth Annual ESRI
User Conference, June 26-30, San Diego, CA. Available at
http://gis.esri.com/librarv/userconf/procOO/professional/papers/PAP480/p480.htm.
NCHS (National Health and Nutrition Examination Survey). 2002. NHANES 1999-2000
Laboratory Components. U.S. Department of Health and Human Services, Centers for
Disease Control and Prevention, National Center for Health Statistics, Hyattsville, MD.
Available at http://www.cdc.gov/nchs/data/nhanes/blood.pdf.
6-11
-------
Neiger, R.D. and G.D. Osweiler. 1989. Effect of subacute low level dietary sodium
arsenite on dogs. Fundam. Appl. Toxicol. 13:439-451 (as cited in ATSDR, 2000a).
Nickel Institute. 1999. Nickel Allergic Contact Dermatitis. Available at
http://www.nidi.org/index.cfm/ci id/99.htm.
Nogawa, K., R. Honda, T. Kido et al. 1989. A dose-response analysis of cadmium in
the general environment with special reference to total cadmium intake limit. Environ.
Res. 48:7-16 (as cited in ATSDR, 1999b).
NoMiracle (Novel Methods for Integrated Risk Assessment of Cumulative Stressors in
Europe). 2006. HomePage. Available at http://nomiracle.jrc.it/default.aspx.
Norris, G., S.N. YoungPong, J.Q. Koenig, T.V. Larson, L. Sheppard and J.W. Stout.
1999. An association between fine particles and asthma emergency department visits
for children in Seattle. Environ. Health Perspect. 107(6):489-493.
NPI (National Pollutant Inventory). 2005. Chlorine Fact Sheet. Australian Government,
Department of the Environment and Heritage, Canberra, Australia (Oct.). Available at
http://www.npi.gov.au/database/substance-info/profiles/20.html.
NRC (National Research Council). 1983. Risk Assessment in the Federal Government:
Managing the Process. Committee on the Institutional Means for Assessments of Risk
to Public Health, Commission on Life Sciences. National Academy Press, Washington,
DC.
NRC (National Research Council). 1994. Science and Judgment in Risk Assessment.
Committee on Risk Assessment of Hazardous Air Pollutants, Board on Environmental
Sciences and Technology, Commission on Life Sciences. National Academy Press,
Washington, DC.
NRC (National Research Council). 2003a. Exposure of the American Population to
Radioactive Fallout from Nuclear Weapons Tests: A Review of the CDC-NCI Draft
Report on a Feasibility Study of the Health Consequences to the American Population
from Nuclear Weapons Tests Conducted by the United States and Other Nations.
Board on Radiation Effects Research. National Academy Press, Washington, DC.
NRC (National Research Council). 2003b. A Review of the Dose Reconstruction
Program of Defense Threat Reduction Agency (J. Till, Chair). Board on Radiation
Effects Research. National Academy Press, Washington, DC.
NRC (National Research Council). 2004. Review of the Army's Technical Guides on
Assessing and Managing Chemical Hazards to Deployed Personnel. A report by the
Subcommittee on the Toxicological Risks to Deployed Military Personnel, Committee on
Toxicology. National Academy Press, Washington, DC.
6-12
-------
NRC (National Research Council) 2007. Models in Environmental Regulatory Decision
Making. Committee on Models in the Regulatory Decision Process. National Academy
Press, Washington, DC. 286 pp.
NTP (National Toxicology Program). 1996. Final Report on the Reproductive Toxicity
of Potassium Dichromate (Hexavalent) (CAS No. 7778-50-9) Administered in Diet to SD
Rats. NTIS No. PB97-125355. National Institute of Environmental Health Sciences,
Research Triangle Park, NC (as cited in ATSDR, 2000b).
NTP (National Toxicology Program). 2002. Report on Carcinogens, 10th ed.
U.S. Department of Health and Human Services, Public Health Services, National
Institute of Environmental Health Sciences, Washington, DC. December. Available at
http://ehp.niehs.nih.gov/roc/toc10.html.
NYSDOH (New York State Department of Health). 2003. Protecting Our Children from
Lead: The Success of New York's Efforts to Prevent Childhood Lead Poisoning.
January 21. Available at http://www.health.state.nv.us/nvsdoh/lead/childlead.pdf.
Oberg, M., A. Sjodin, H. Casabona, I. Nordgren, E. Klasson-Wehler and H. Hakansson.
2002. Tissue distribution and half-lives of individual polychlorinated biphenyls and
serum levels of 4-hydroxy-2,3,3',4',5-pentachlorobiphenyl in the rat. Toxicol. Sci.
70(2):171-182.
O'Connor, G.T. and D.R. Gold. 1999. Cockroach allergy and asthma in a 30-year-old
man. Environ. Health Perspect. 107(3):243-247. March. Available at
http://ehp.niehs.nih.gov/members/1999/107p243-247oconnor/oconnor-full.html.
OMB (Office of Management and Budget). 2006. Proposed Risk Assessment Bulletin.
Washington, DC. Available at
http://www.whitehouse.gov/omb/inforeg/proposed risk assessment bulletin 010906.pd
f.
Parker, G.S. and S.L. Rosen. 1981. Woburn: Cancer Incidence and Environmental
Hazards, 1969-1978. Massachusetts Department of Public Health. Available at
http://www.sph.umich.edu/geomed/grabber/1981ca.pdf.
Paulu, C., A. Aschengrau and D. Ozonoff. 2002. Exploring associations between
residential location and breast cancer in a case-control study. Environ. Health
Perspect. 110(5):471-478.
Perera, P.P., V. Rauh, W-Y. Tsai et al. 2003. Effects of transplacental exposure to
environmental pollutants on birth outcomes in a multi-ethnic population. Environ. Health
Perspect. 111:201-205.
Perera, F., D. Tang, Y.-H. Tu et al. 2004. Biomarkers in maternal and newborn blood
indicate heightened fetal susceptibility to procarcinogenic DNA Damage. Environ.
Health Perspect. 112:1133-1136.
6-13
-------
Perry, H.M. Jr., M.W. Erlanger, T.O. Gustafsson et al. 1989. Reversal of cadmium-
induced hypertension by D-myo-inositol-1,2,6-trisphosphate. J. Toxicol. Environ. Health
28:151-159 (as cited in ATSDR, 1999b).
Peterson, D.E., M.S. Kanarek, M.A. Kuykendall et al. 1995. Fish consumption patterns
and blood mercury levels in Wisconsin Chippewa Indians. Arch. Environ. Health.
49(1):53-58.
Pohl, H.R. and H.G. Abadin. 1995. Utilizing uncertainty factors in minimal risk levels
derivation. Regul. Toxicol. Pharmacol. 22:180-188.
Pohl, H.R., N. Roney, S. Wilbur, H. Hansen and C.T. DeRosa. 2003. Six interaction
profiles for simple mixtures. Chemosphere. 53:183-197.
Public Health Service. 1981. Cancer in Woburn, Massachusetts. EPI-80-37-2.
Centers for Disease Control and Prevention, Atlanta, GA.
RAIS (Risk Assessment Information System). 1991. Toxicity Summary for Cadmium.
Updated 8/29/97, accessed September 2003. Available at
http://risk.lsd.ornl.gov/index.shtml.
RAIS (Risk Assessment Information System). 1995. Toxicity Summary for Nitrate.
Accessed September 2003. Available at http://risk.lsd.ornl.gov/index.shtml.
Rao, V.R., K. Levy and M. Lustik. 1993. Logistic regression of inhalation toxicities of
perchloroethylene - Application in noncancer risk assessment. Reg. Toxicol.
Pharmacol. 18:233-247.
Regoli, F., M. Nigro, M. Benedetti et al. 2005. Interactions between metabolism of
trace metals and xenobiotic agonists of the aryl hydrocarbon receptor in the antarctic
fish Trematomus bernacchii: environmental perspectives. Environ. Toxicol. Chem.
24(6): 1475-1482.
Richardson, J.P. 2004. Monitoring, education and partnerships through the Georgia
Southeast and Coastal Region Training Center. Poster presented at the 2004 National
Monitoring Conference, Chattanooga, TN, May 17-20. Available at
http://water.usgs.gov/wicp/acwi/monitoring/conference/2004/.
Santucci, B, R. Manna F and C. Cannistraci et al. 1994. Serum and urine
concentrations in nickel-sensitive patients after prolonged oral administration. Contact
Dermatitis. 30:97-101 (as cited in ATSDR, 2003b).
Schroeder, H.A. and M. Mitchener. 1975. Life-term studies in rats: Effects of
aluminum, barium, beryllium,and tungsten. J. Nutr. 105:421-427 (as cited in ATSDR,
2002c).
6-14
-------
Schroeder, H.A., J.J. Balassa and W.H. Vinton Jr. 1965. Chromium, cadmium and lead
in rats: Effects on lifespan, tumors and tissue levels. J. Nutr. 86:51-66 (as cited in
ATSDR, 2000b).
Sette, W.F. 1992. Memorandum: Joint OPPT/OW/ORD Review of 1992 Aldicarb
Human Study. U.S. Environmental Protection Agency, Washington, DC. Sept. 4, 1992.
Shiwen, C., Y. Lin, H. Zhineng et al. 1990. Cadmium exposure and health effects
among residents in an irrigation area with ore dressing wastewater. Sci. Total Environ.
90:67-73 (as cited in ATSDR, 1999b).
Shuval, H.I. and N. Gruener. 1972. Epidemiological and toxicological aspects of
nitrates and nitrites in the environment. Am. J. Public Health. 62(8): 1045-1052 (as
cited in RAIS, 1995).
Simmons, J.E., S.D. Richardson, T.F. Speth et al. 2002. Development of a research
strategy for integrated technology-based toxicological and chemical evaluation of
complex mixtures of drinking water disinfection byproducts. Environ. Health Perspect.
110(6): 1013-1024.
Simon, C., H. Manzke, H. Kay and G. Mrowetz. 1964. Occurrence, pathogenesis, and
possible prophylaxis of nitrite induced methemoglobinemia. Zeitschr. Kinderheilk.
91:124-138 (German) (as cited in U.S. EPA, 2006e).
Simon, T.W. 1999. Two dimensional Monte Carlo simulation and beyond: A
comparison of several probabilistic risk assessment methods applied to a Superfund
site. Human Ecol. Risk Assess. 5(4):823-843.
Soni, M.G., S.K. Ramaiah, H. Mumtaz, H.M. Clewell and H. Mehendale. 1999.
Toxicant-inflicted injury and stimulated tissue repair are opposing toxicodynamic forces
in predictive toxicology. Regul. Toxicol. Pharmacol. 29:165-174.
Stange, A.W., D.E. Hilmas, F.J. Furman and T.R. Gatliffe. 2001. Beryllium
sensitization and chronic beryllium disease at a former nuclear weapons facility. Appl.
Occup. Environ. Hyg. 16(3):405-417.
Stayner, L, A.J. Bailer, R. Smith, S. Gilbert, F. Rice and E. Kuempel. 1999. Sources of
uncertainty in dose-response modeling of epidemiological data for cancer risk
assessment. Ann. NY Acad. Sci. 895:212-222.
Stephenson, J. 2000. Exposure to home pesticides linked to Parkinson's disease. J.
Am. Med. Assoc. 283(23):3055-3056.
Strickland, J.A. and D.J. Guth. 2002. Quantitative exposure-response assessment
approaches to evaluate acute inhalation toxicity of phosgene. Human Ecol. Risk
Assess. 8(3):511-536.
Susser, M. 1991. Philosophy in epidemiology. Theor. Med. 12(3):271-273.
6-15
-------
Suter, G.W. 1999. Developing conceptual models for complex ecological risk
assessments. Hum. Ecol. Risk Assess. 5:375-396.
Suter, G.W., T. Vermeire, W.R. Munns Jr. and J. Sekizawa. 2003. Framework for the
integration of health and ecological risk assessment. Hum. Ecol. Risk Assess.
9:281-301.
Swartout, J.C., P.S. Price, M.L. Dourson, H.L. Carlson-Lynch and R.E. Keenan. 1998.
A probabilistic framework for the reference dose (probabilistic RfD). Risk Anal.
18(3):271-282.
TCEQ (Texas Commission on Environmental Quality). 1999. Texas-Specific
Background Concentrations, Texas Risk Reduction Program (TRRP) Rule, Figure:
30 TAG Section 350.51 (m), September 2. Available at
http://www.tceq.state.tx.us/assets/public/remediation/trrp/350revisions.doc.
TCEQ (Texas Commission on Environmental Quality). 2002. Risk Levels, Hazard
Indices, and Cumulative Adjustment, Texas Natural Resource Conservation
Commission (TNRCC) Regulatory Guidance, Remediation Division, RG-366/TRRP-18,
August. Available at http://www.tceq.state.tx.us/.
TCEQ (Texas Commission on Environmental Quality). 2003. Texas Risk Reduction
Program (TRRP) Rule Protective Concentration Level (PCL) Tables, Chemical/Physical
Properties. March. Available at http://www.tnrcc.state.tx.us/permittinq/trrp.htm.
Telles, N.C. 1981. Cancer Mortality in Woburn-A Three-Decade Study (1949-1978).
Massachusetts Department of Public Health.
Teuschler, L.K., M.L. Dourson, W.M. Stiteler, P. McClure and H. Tully. 1999. Health
risk above the reference dose for multiple chemicals. Reg. Toxicol. Pharmacol.
30:S19-S26.
Teuschler, L.K., G.E. Rice, C.R. Wilkes, J.C. Lipscomb and F.W. Power. 2004. A
feasibility study of cumulative risk assessment methods for drinking water disinfection
by-product mixtures. J. Toxicol. Environ. Health A. 67:755-777.
The New Lexicon: Webster's Dictionary of the English Language. 1989. Lexicon
Publications, Inc., New York, NY.
The On-line Medical Dictionary (c) Academic Medical Publishing & CancerWEB
1997-98. Distributed by CancerWEB under license from Academic Medical Publishing.
Accessed July-September 2001. Available at
http://www.betterhealth.vic.qov.au/bhcv2/bhcsite.nsf/paqes/bhc medicaldictionarv?open
document.
Toy, K.A., G.D. Gawne-Mittelstaedt, N.L. Pollisarand S. Liao. 1995. A Fish
Consumption Survey of the Tulalip and Squaxin Island Tribes of Puget Sound. Report
to Tulalip Tribes, Department of the Environment. Seattle, WA.
6-16
-------
TOXNET (Toxicology Data Network). 2005. Tetrachloroethylene. National Library of
Medicine. Available at http://toxnet.nlm.nih.gov/ (accessed 2006).
Tucker, A.N., V.M. Sanders, D.W. Barnes et al. 1982. Toxicology of trichloroethylene
in the mouse. Toxicol. Appl. Pharmacol. 62:351-357 (as cited in ASTDR, 1997c).
U.S. Consumer Product Safety Commission. 2005. Interim Enforcement Policy Lead
Levels. Washington, DC. Available at
http://www.cdc.gov/nceh/lead/ACCLPP/Mar%202005/Docs/03%20-
%20lnterim%20Policv%20Lead%20Levels-Kristina%20Hatlelid.pdf.
U.S. DOE (Department of Energy). 1999. Risk/Impact Technical Report for the
Hanford Groundwater/Vadose Zone Integration Project. Prepared by Argonne National
Laboratory for U.S. Department of Energy Center for Risk Excellence, Argonne, IL.
January. DOE/CH/CRE-7-1999.
U.S. DOE (Department of Energy). 2005. Natural Attenuation Monitor. Prepared by
SRNL for U.S. Department of Energy, Savannah River Site, GA. Issue 3, March. (This
work at the Savannah River Site is also highlighted at
http://www.epa.gov/ada/highlights/ian2006 highlights.html.)
U.S. EPA. 1980. Guidelines and Methodology Used in the Preparation of Health Effect
Assessment Chapters of the Consent Decree Water Criteria Documents. Fed. Reg.
45(231 ):79347-79357.
U.S. EPA. 1981. Ambient Water Quality Criteria Document: Chlorine. U.S.
Environmental Protection Agency, Washington, DC. EPA 450/3-78-005. (as cited by
HSDB, 1991; accessed 2006).
U.S. EPA. 1985. Guideline for Determination of Good Engineering Practice Stack
Height (Technical Support Document for the Stack Height Regulations) - Revised. U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards,
Research Triangle Park, NC. June. EPA/450/4-80/023R.
U.S. EPA. 1986a. Guidelines for the Health Risk Assessment of Chemical Mixtures.
U.S. Environmental Protection Agency, Office of Research and Development,
Washington, DC. September. EPA/630/R-98/002.
U.S. EPA. 1986b. Guidelines for Carcinogen Risk Assessment. Fed. Reg.
51(185):33992-34003.
U.S. EPA. 1987. The Risk Assessment Guidelines of 1986. U.S. Environmental
Protection Agency, Office of Health and Environmental Assessment, Washington, DC.
EPA/600/8-87/045.
6-17
-------
U.S. EPA. 1989a. Risk Assessment Guidance for Superfund: Volume 1, Human Health
Evaluation Manual (Part A). U.S. Environmental Protection Agency, Office of
Emergency and Remedial Response, Washington, DC. EPA/540/1-89/002. (Also see
Parts B-D.)
U.S. EPA. 1989b. Interim Procedures for Estimating Risks Associated with Exposures
to Mixtures of Chlorinated Dibenzo-p-dioxins and -dibenzofurans (CDDs and CDFs) and
1989 Update. U.S. Environmental Protection Agency, Risk Assessment Forum,
Washington, DC. EPA/625/3-89/016.
U.S. EPA. 1991 a. Guidelines for Developmental Toxicity Risk Assessment. Federal
Register. 56(234):63798-63826.
U.S. EPA. 1991b. Drinking Water Criteria Document for PAH. Prepared by the Office
of Health and Environmental Assessment, Environmental Criteria and Assessment
Office, Cincinnati, OH for the Office of Water Regulations and Standards, Washington,
DC.
U.S. EPA. 1992a. Guidelines for Exposure Assessment. U.S. Environmental
Protection Agency, Risk Assessment Forum, Washington, DC. EPA/600/Z-92/001.
U.S. EPA. 1992b. Screening Procedures for Estimating the Air Quality Impact of
Stationary Sources, Revised. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. October. EPA/454/R-92/019.
U.S. EPA. 1994a. Revised Interim Soil Lead Guidance for CERCLA Sites and RCRA
Corrective Action Facilities. U.S. Environmental Protection Agency, Office of Solid
Waste and Emergency Response, Washington, DC. OSWER Directive #9355.4-12.
U.S. EPA. 1994b. Chemical Summary for Chlorine. Office of Pollution Prevention and
Toxics, Washington, DC. EPA/749/F-94/010a, Available at
http://www.epa.goV/chemfact/s chlori.txt.
U.S. EPA. 1995a. Profile of the Metal Mining Industry. U.S. Environmental Protection
Agency, Office of Compliance, Office of Enforcement and Compliance Assurance,
Washington, DC. September. EPA/310/R-95/008. Available at
http://www.epa.gov/compliance/resources/publications/assistance/sectors/notebooks/rn
etminsnpt1.pdf.
U.S. EPA. 1995b. Policy for Risk Characterization. Memorandum from Agency
Administrator Carol M. Browner, Washington, DC. March 21.
U.S. EPA. 1995c (et sequelae). Compilation of Air Pollutant Emission Factors. U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards.
EPA AP-42. Available at http://www.epa.gov/ttn/chief/ap42/.
6-18
-------
U.S. EPA. 1996a. Soil Screening Guidance, Technical Background Document. U.S.
Environmental Protection Agency, Office of Solid Waste and Emergency Response,
Washington, DC. May. EPA/540/R-95/128. Available at
http://www.epa.gov/superfund/resources/soil/introtbd.htm.
U.S. EPA. 1996b. PCBs: Cancer Dose-Response Assessment and Application to
Environmental Mixtures. U.S. Environmental Protection Agency, Office of Research
and Development, National Center for Environmental Assessment, Washington Office,
Washington, DC. EPA/600/P-96/001F.
U.S. EPA. 1997a. Guidance on Cumulative Risk Assessment, Part 1. Planning and
Scoping. U.S. Environmental Protection Agency, Science Policy Council, Washington,
DC. Attachment to memo dated July 3, 1997 from the Administrator, Carol Browner,
and Deputy Administrator, Fred Hansen, titled "Cumulative Risk Assessment Guidance-
Phase I Planning and Scoping." Available at
http://www.epa.gov/OSA/spc/2cumrisk.htm.
U.S. EPA. 1997b. Chemical and Radiation Leukemogenesis in Humans and Rodents
and the Value of Rodent Models for Assessing Risks of Lymphohematopoietic Cancers.
U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment, Washington, DC. May. EPA/600/R-97/090.
Available at http://www.epa.gov/ncea/pdfs/lympho.pdf.
U.S. EPA. 1997c. Exposure Factors Handbook-Volumes I, II, and III (General
Factors, Food Ingestion Factors, and Activity Factors). U.S. Environmental Protection
Agency, Office of Research and Development, National Center for Environmental
Assessment, Washington, DC. August. EPA/600/P-95/002Fa. Available at
http://www.epa.gov/ncea/pdfs/efh/front.pdf.
U.S. EPA. 1997d. Research on Risk Assessment Issues with Commercial Mixtures
Using Toxaphene as a Case Study. U.S. Environmental Protection Agency, National
Center for Environmental Assessment, Cincinnati, OH.
U.S. EPA. 1997e. Mercury Study Report to Congress. U.S. Environmental Protection
Agency, Office of Research and Development, Office of Air Quality, Planning &
Standards, Washington, DC. EPA/452/R-97/003.
U.S. EPA. 1997f. Guiding Principles for Monte Carlo Analysis. U.S. Environmental
Protection Agency, Risk Assessment Forum, Washington, DC. EPA/630/R-97/001.
U.S. EPA. 1998a. Methodology for Assessing Health Risks Associated with Multiple
Pathways of Exposure to Combustor Emissions. U.S. Environmental Protection
Agency, Office of Research and Development, National Center for Environmental
Assessment, Cincinnati, OH. December. EPA/600/R-98/137.
U.S. EPA. 1998b. Guidelines for Neurotoxicity Risk Assessment. Federal Register.
63(93): 26926-26954. EPA/630/R-95/001 F.
6-19
-------
U.S. EPA. 1998c. Guidelines for Ecological Risk Assessment. Federal Register.
63(93): 26846-26924. EPA/630/R-95/002F.
U.S. EPA. 1998d. C x T: Historical perspectives, current issues, and approaches. In:
Summary of the U.S. EPA Workshop on the Relationship Between Exposure Duration
and Toxicity. U.S. Environmental Protection Agency, National Center for Environmental
Assessment, Washington, DC. September. EPA/600/R-99/081.
U.S. EPA. 1998e. Handbook for Air Toxics Emission Inventory Development,
Volume I: Stationary Sources. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Office of Air and Radiation, Washington, DC.
EPA/454/B-98/002.
U.S. EPA. 1999a. Guidance for Identifying Pesticide Chemicals and Other Substances
that Have a Common Mechanism of Toxicity. U.S. Environmental Protection Agency,
Office of Pesticide Programs, Washington, DC.
U.S. EPA. 1999b. A Guide to Preparing Superfund Proposed Plans, Records of
Decision, and Other Remedy Selection Decision Documents: 6.0: Writing the Record of
Decision. U.S. Environmental Protection Agency, Office of Solid Waste and Emergency
Response, Washington, DC. EPA/540/R-98/031.
U.S. EPA. 1999c. Sociodemographic Data Used for Identifying Potentially Highly
Exposed Populations. U.S. Environmental Protection Agency, National Center for
Environmental Assessment, Washington, DC. July. EPA/600/R-99/060. Summary
information (not the report) is available at
http://oaspub.epa.gov/eims/eimscomm.getfile7p download id=428679.
U.S. EPA. 1999d. Reregistration Eligibility Decision Facts for Chlorine Gas. U.S.
Environmental Protection Agency, Office of Prevention, Pesticides, and Toxic
Substances, Washington, DC. February. EPA/738/F-99/001. Available at
http://www.epa.gov/oppsrrd1/REDs/factsheets/4022fact.pdf.
U.S. EPA. 1999e. Guidance for Performing Aggregate Exposure and Risk
Assessments. U.S. Environmental Protection Agency, Office of Pesticide Programs,
Washington, DC. October. Available at
http://www.pestlaw.eom/x/guide/1999/EPA-19991029A.html.
U.S. EPA. 1999f. Screening Level Ecological Risk Assessment Protocol for Hazardous
Waste Combustion Facilities. Peer Review Draft. U.S. Environmental Protection
Agency, Office of Solid Waste and Emergency Response, Washington, DC. EPA/R6-
098/002A. November. Available at
http://www. epa. gov/epaoswer/hazwaste/com bust/ecorisk. htm.
U.S. EPA. 1999g. Frequently Asked Questions (FAQs) on the Adult Lead Model.
Technical Review Workgroup for Lead Guidance Document. U.S. Environmental
Protection Agency, Washington, DC. April.
6-20
-------
U.S. EPA. 1999h. Handbook for Criteria Pollutant Inventory Development: A
Beginner's Guide for Point and Area Sources. U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Office of Air and Radiation, Washington,
DC. September. EPA/454/R-99/037.
U.S. EPA. 1999i. Risk Assessment Guidance for Superfund (Volume 3, Part A:
Process for Conducting Probabilistic Risk Assessment). Draft. U.S. Environmental
Protection Agency, Office of Solid Waste and Emergency Response, Washington, DC.
December. Available at http://www.epa.gov/oswer/riskassessment/rags3adt/.
U.S. EPA. 2000a. Supplementary Guidance for Conducting Health Risk Assessment of
Chemical Mixtures. U.S. Environmental Protection Agency, Risk Assessment Forum,
Washington, DC. EPA/630/R-00/002. Available at
http://www.epa.gov/ncea/raf/pdfs/chem mix/chem mix 08 2001.pdf.
U.S. EPA. 2000b. Community Risk-Based Air Screening: A Case Study in Baltimore,
MD. Baltimore Community Environmental Partnership, Air Committee Technical
Report. U.S. Environmental Protection Agency, Office of Pollution Prevention and
Toxics, Washington, DC. April. EPA/744/R-00/005.
U.S. EPA. 2000c. CATREG Software Documentation. Office of Research and
Development, Washington, DC. EPA/600/R-98/053F.
U.S. EPA. 2000d. CATREG Software User Manual. Office of Research and
Development, Washington, DC. EPA/600/R-98/052F.
U.S. EPA. 2000e. Conducting a Risk Assessment of Mixtures of Disinfection By-
Products (DBPs) for Drinking Water Treatment Systems. U.S. Environmental Protection
Agency, Office of Research and Development, National Center for Environmental
Assessment, Cincinnati, OH. EPA/600/R-03/040.
U.S. EPA. 2000f. Science Policy Council Handbook: Risk Characterization. U.S.
Environmental Protection Agency, Science Policy Council, Washington, DC.
EPA/1 OO/B-00/002.
U.S. EPA. 2000g. Guidance for the Data Quality Objectives Process (QA/G-4). U.S.
Environmental Protection Agency, Washington, DC. Available at
http://www.epa.gov/guality/gs-docs/g4-final.pdf.
U.S. EPA. 2000h. Guidance for Data Quality Assessment: Practical Methods for Data
Analysis. U.S. Environmental Protection Agency, Office of Environmental Information,
Washington, DC. July. EPA/600/R-96/084. Available at
http://www.epa.gov/region10/www/offices/oea/epagag9.pdf.
U.S. EPA. 2000i. Trichloroethylene: Hazard Summary. Technology Transfer Network
Air Toxics Website, Created in April 1992; Revised in January 2000. Office of Air and
Radiation, Washington, DC. Available at
http://www.epa.gov/ttn/uatw/hlthef/tri-ethy.html.
6-21
-------
U.S. EPA. 2001 a. General Principles for Performing Aggregate Exposure and Risk
Assessments. U.S. Environmental Protection Agency, Office of Pesticide Programs,
Washington, DC. Fax-On-Demand. Fax no. (202) 401-0527. Item no. 6043.
U.S. EPA. 2001 b. Methylmercury Reference Dose. Integrated Risk Information
System. Available at http://www.epa.gov/iris/subst/0073.htm.
U.S. EPA. 2001 c. Risk Assessment Guidance for Superfund. Vol. I. Human Health
Evaluation Manual (Part D), Standardized Planning, Reporting, and Review of
Superfund Risk Assessments. U.S. Environmental Protection Agency, Office of Solid
Waste and Emergency Response, Washington, DC.
U.S. EPA. 2001 d. Guidance for Characterizing Background Chemicals in Soil at
Superfund Sites. External Review Draft. U.S. Environmental Protection Agency,
Washington, DC. June. EPA/540/R-01/003.
U.S. EPA. 2001 e. Risk Assessment Guidance for Superfund. Vol. 3, Part A: Process
for Conducting Probabilistic Risk Assessment (RAGS 3A). U.S. Environmental
Protection Agency, Office of Emergency and Remedial Response, Washington, DC.
EPA/540/R-02/002.
U.S. EPA. 2001f. Workshop on Approaches to Polycyclic Aromatic Hydrocarbon (PAH)
Health Assessment. Discussion document. U.S. Environmental Protection Agency,
Office of Research and Development, National Center for Environmental Assessment,
Washington, DC. NCEA-3-1105.
U.S. EPA. 2002a. Organophosphate Pesticides: Revised Cumulative Risk
Assessment. U.S. Environmental Protection Agency, Office of Pesticide Programs,
Washington, DC. Available at http://www.epa.gov/pesticides/cumulative/rra-op/.
U.S. EPA. 2002b. Region/ORD Workshop on Cumulative Risk Assessment.
November 4-8, 2002, Dallas, TX. Office of Science Policy, Washington, DC. Available
at http://www.epa.gov/osp/regions/cmrskrpt.pdf.
U.S. EPA. 2002c. Guidance on Cumulative Risk Assessment of Pesticide Chemicals
That Have a Common Mechanism of Toxicity. U.S. Environmental Protection Agency,
Office of Pesticide Programs, Washington, DC. Available at
http://www.epa.gov/oppfead1/trac/science/cumulative guidance.pdf.
U.S. EPA. 2002d. Ground Water and Drinking Water Technical Fact Sheet on
1,1-Dichloroethylene. U.S. Environmental Protection Agency, Office of Ground Water
and Drinking Water, Washington, DC. November. Available at
http://www.epa.goV/OGWDW/dwh/t-voc/11 -dichl.html.
6-22
-------
U.S. EPA. 2002e. A Review of the Reference Dose and Reference Concentration
Processes. U.S. Environmental Protection Agency, Risk Assessment Forum,
Washington, DC. EPA/630/P-02/002F. Available at
http://epa.gov/iriswebp/iris/RFD FINALM l.pdf.
U.S. EPA. 2002f. Lessons Learned on Planning and Scoping of Environmental Risk
Assessment. Memorandum from Science Policy Council. January. Available at
http://www.epa.gov/osp/spc/llmemo.htm.
U.S. EPA. 2002g. Region 9 Preliminary Remediation Goals Table 2002 Update.
Technical Memorandum (from Stanford Smucker, Regional Toxicologist, to PRG Table
Users). U.S. Environmental Protection Agency, Washington, DC. October. Available
at http://www.epa.gov/region09/waste/sfund/prg/files/02userguide.pdf.
U.S. EPA. 2002h. Guidance for Evaluating the Vapor Intrusion to Indoor Air Pathway
from Groundwater and Soils (subsurface vapor intrusion guidance). Draft. Federal
Register. 67(230):71169-71172. November 29. Available at
http://www.epa.gov/correctiveaction/eis/vapor.htm.
U.S. EPA. 2002i. Child-Specific Exposure Factors Handbook. Interim Report. U.S.
Environmental Protection Agency, National Center for Environmental Assessment,
Washington, DC. EPA/600/P-00/02b. Available at
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=55145.
U.S. EPA. 2002J. Guidance on Environmental Data Verification and Data Validation.
Office of Environmental Information, Washington, DC. EPA/240/R 02/004. Available at
http://www.epa.gov/qualitv/qs-docs/g8-final.pdf.
U.S. EPA. 2002k. Groundwater and Ecosystems Restoration Research. Biochlor
Version 2.2. March 2002. Available at
http://www.epa.qov/ada/csmos/models/biochlor.html (last updated March 2006).
U.S. EPA. 2003a. Framework for Cumulative Risk Assessment. U.S. Environmental
Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Washington, DC. EPA/600/P-02/001 F. Available at
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=54944.
U.S. EPA. 2003b. The Feasibility of Performing Cumulative Risk Assessments for
Mixtures of Disinfection By-Products in Drinking Water. U.S. Environmental Protection
Agency, Office of Research and Development, National Center for Environmental
Assessment, Cincinnati, OH. EPA/600/R-03/051.
U.S. EPA. 2003c. Exposure and Human Health Reassessment of
2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds. National
Academy Sciences (NAS) Review Draft. U.S. Environmental Protection Agency,
Exposure Assessment and Risk Characterization Group, Washington, DC.
EPA/600/P-00/001Cb.
6-23
-------
U.S. EPA. 2003d. Guideline on Air Quality Models, Appendix W of CFR Part 51, April.
Available at http://www.arb.ca.gov/toxics/harp/docs/40CFR APPW.pdf.
U.S. EPA. 2003e. Considerations in Risk Communication: A Digest of Risk
Communication as a Risk Management Tool. U.S. Environmental Protection Agency,
National Risk Management Research Laboratory, Cincinnati, OH. March.
EPA/625/R-02/004. Available at
http://www.epa.gov/ORD/NRMRL/Pubs/625r02004/625r02004.pdf.
U.S. EPA. 2003f. Developing Relative Potency Factors for Pesticide Mixtures:
Biostatistical Analyses of Joint Dose-Response. U.S. Environmental Protection Agency,
Office of Research and Development, National Center for Environmental Assessment,
Cincinnati, OH. EPA/600/R-03/052. Available at
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=66273.
U.S. EPA. 2003g. Guidance for Developing Ecological Soil Screening Levels. Revised
February 2005. U.S. Environmental Protection Agency, Office of Solid Waste and
Emergency Response, Washington, DC. OSWER Directive 9285.7-55. Available at
http://www.epa.gov/ecotox/ecossl/pdf/ecossl guidance chapters.pdf.
U.S. EPA. 2003h. Region/ORD Workshop on Inhalation Risk Assessment: A
Superfund Focus: Summary Report. U.S. Environmental Protection Agency,
Washington, DC. September 9-12, 2003. Available at
http://intranet.epa.gov/ospintra/scienceportal/htm/complete.htmffinhale.
U.S. EPA. 2003i. Region 3 Risk-Based Concentrations (RBC) Tables. Technical
Background Document (from Jennifer Hubbard, Regional Toxicologist, to RBC Table
Users). U.S. Environmental Protection Agency, Washington, DC. October. Available
at http://www.epa.gov/reg3hwmd/risk/human/info/cover.pdf.
U.S. EPA. 2003J. User's Guide for Evaluating Subsurface Vapor Intrusion into
Buildings. Draft. Prepared by Environmental Quality Management under Contract
#68-W-01-058 to U.S. Environmental Protection Agency, Office of Emergency and
Remedial Response, Washington DC. June 19. Available at
http://www.epa.gov/superfund/programs/risk/airmodel/guide.pdf.
U.S. EPA. 2003k. Draft Guidance on the Development, Evaluation, and Application of
Regulatory Environmental Models. Council on Regulatory Models, Office of Science
Policy, Office of Research and Development, Washington, DC.
U.S. EPA. 2004a. Risk Assessment Guidance for Superfund Volume I: Human Health
Evaluation Manual (Part E, Supplemental Guidance for Dermal Risk Assessment).
U.S. Environmental Protection Agency, Office of Solid Waste and Emergency
Response, Washington, DC. EPA/540/R/99/005. Available at
http://www.epa.gov/oswer/riskassessment/ragse/.
6-24
-------
U.S. EPA. 2004b. Framework for Inorganic Metals Risk Assessment. U.S.
Environmental Protection Agency, Office of Research and Development, Risk
Assessment Forum, Washington, DC. EPA/630/P-04/068B.
U.S. EPA. 2004c. Air Screening Assessment for Cook County Illinois and Lake County
Indiana. Prepared by Argonne National Laboratory, Argonne, IL, in support of the
U.S. EPA Region V Cumulative Risk Initiative, for U.S. Environmental Protection
Agency, Office of Pollution Prevention and Toxics and Region V. (In press.)
U.S. EPA. 2004d. Human Exposure Measurements: National Human Exposure
Assessment Survey (NHEXAS). Office of Research and Development, National
Exposure Research Laboratory. Accessed March 2004. Available at
http://www.epa.gov/heasd/edrb/nhexas.htm.
U.S. EPA. 2004e. Air Quality Criteria for Particulate Matter. U.S. Environmental
Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Research Triangle Park, NC. EPA/600/P-99/002aF.
Available at http://cfpub.epa.gov/ncea/cfm/partmatt.cfm.
U.S. EPA. 2004f. Health-based Short-term Advisory Levels: Pilot Guide. National
Homeland Security Research Center, Cincinnati, OH.
U.S. EPA. 2004g. Benchmark Dose Software. U.S. Environmental Protection Agency,
Washington, DC. Accessed February 18. Available at
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=20167.
U.S. EPA. 2004h. Supplemental Guidance for Dermal Risk Assessment, Part E of Risk
Assessment Guidance for Superfund, Human Health Evaluation Manual (Volume I).
Memorandum from M.B. Cook (Director, Office of Superfund Remediation and
Technology Innovation, Washington, DC) to Superfund National Policy Managers and
Regional Toxics Integration Coordinators, Regions 1-10. OSWER 9285.7 02 EP (Aug.
16). Available at
http://www.epa.gov/oswer/riskassessment/ragse/pdf/part e impl 2004 final.pdf.
U.S. EPA. 2005a. Wells G & H Fact Sheet. U.S. Environmental Protection Agency,
Region 1, Boston, MA. Available at
http://www.epa.gov/NE/superfund/sites/wellsgh/factsh.html.
U.S. EPA. 2005b. Human Health Risk Assessment Protocol for Hazardous Waste
Combustion Facilities, Final. U.S. Environmental Protection Agency, Office of Solid
Waste and Emergency Response (5305W), Washington, DC. EPA/520/R-05/006.
Available at http://www.epa.gov/epaoswer/hazwaste/combust/risk.htm.
U.S. EPA. 2005c. Human Health Medium-Specific Screening Levels. U.S.
Environmental Protection Agency, Region 6, Dallas, TX. November. Available at
http://www.epa.gov/earth1r6/6pd/rcra c/pd-n/r6screenbackground.pdf.
6-25
-------
U.S. EPA. 2005d. Guidelines for Carcinogen Risk Assessment. U.S. Environmental
Protection Agency, Risk Assessment Forum, Washington, DC. EPA/630/P-03/001B.
U.S. EPA. 2005e. Supplemental Guidance for Assessing Susceptibility from Early-Life
Exposure to Carcinogens. U.S. Environmental Protection Agency, Risk Assessment
Forum, Washington, DC. EPA/630/R-03/003F.
U.S. EPA. 2005f. All-Ages Lead Model (AALM) Version 1.05 (External Review Draft).
U.S. Environmental Protection Agency, Washington, DC. EPA/600/C-05/013. Available
at http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=139314.
U.S. EPA. 2005g. Technical Support Document for the Final Clean Air Mercury Rule:
Air Quality Modeling. U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Research Triangle Park, NC.
U.S. EPA. 2006a. Office of Water: Protocol for Developing Sediment Total Maximum
Daily Load (TMDL): Glossary. Terminology Reference System. U.S. Environmental
Protection Agency, Washington, DC. Available at
http://iaspub.epa.gov/trs/trs proc qry.alphabet?p term nm=V&p reg auth id=1&p dat
a id=11570&p version=1.
U.S. EPA. 2006b. Technical Factsheet on: Chlordane. U.S. Environmental Protection
Agency, Office of Water, Office of Ground Water and Drinking Water, Washington, DC.
Available at http://www.epa.gov/safewater/dwh/t-soc/chlordan.html (last updated Feb.
28, 2006).
U.S. EPA. 2006c. Technical Factsheet on 1,1-Dichloroethylene. U.S. Environmental
Protection Agency, Office of Water, Office of Ground Water and Drinking Water,
Washington, DC. Available at http://www.epa.goV/safewater/dwh/c-voc/11 -dichl.html
(last updated Feb. 28, 2006).
U.S. EPA. 2007. Integrated Risk Information System (IRIS). U.S. Environmental
Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Washington, DC. Available at www.epa.gov/iris.
Verma, A.K., G.T. Bryan and C.A. Reznikoff. 1985. Tumor promoter 12-0-
tetradecanoylphorbol-13-acetate receptors in normal human transitional epithelial cells.
Carcinogenesis. 6(3):427-432.
Vogel, T.M. and P.L. McCarty. 1985. Biotransformation of tetrachloroethylene to
trichloroethylene, dichloroethylene, vinyl chloride, and carbon dioxide under
methanogenic conditions. Appl. Environ. Microbiol. 49(5): 1080-1083.
Waller, K., S.H. Swan, G. DeLorenze and B. Hopkins. 1998. Trihalomethanes in
drinking water and spontaneous abortion. Epidemiology. 9(2):134-140.
6-26
-------
Walton, G. 1951. Survey of literature relating to infant methemoglobinemia due to
nitrate-contaminated water. Am. J. Public Health. 41:986-996 (as cited in U.S. EPA,
2006e).
Weischer, C.H., W. Kordel and D. Hochrainer. 1980. Effects of NiCI2 and NiO in Wistar
rats after oral uptake and inhalation exposure, respectively. Zent Bakteriol. Mikrobiol.
Hyg. (B) 171:336-351 (as cited in ATSDR, 2003b).
WHO (World Health Organization). 1996. Chlorine in Drinking-water: Background
Document for Development of WHO Guidelines for Drinking-water Quality. World
Health Organization, Geneva. Available at
http://www.who.int/water sanitation health/dwq/chlorine.pdf.
Wright, J.M. and J. Keller-Byrne. 2005. Environmental determinants of Parkinson's
Disease. Arch. Environ. Occup. Health. 60(1):32-38.
Wyld, P.J., C.E. Watson, W.S. Nimmo and N. Watson. 1992. A Safety and Tolerability
Study of Aldicarb at Various Dose Levels in Healthy Male and Female Volunteers.
Study submitted to U.S. EPA by Rhone-Poulenc Company (as summarized by Sette,
1992).
Yadrick, M.K., M.A. Kenney and E.A. Winterfeldt. 1989. Iron, copper, and zinc status:
Response to supplementation with zinc or zinc and iron in adult females. Am. J. Clin.
Nutr. 49:145-150 (as cited in U.S. EPA, 2006e).
Zhang, J. and X. Li. 1987. Chromium pollution of soil and water in Jinzhou. J. Chinese
Prev. Med. 21:262-264 (as cited in ATSDR, 2000b).
Zwart, A. and R.A. Woutersen. 1988. Acute inhalation toxicity of chlorine in rats and
mice: Time-concentration-mortality relationships and effects on respiration. J. Haz. Mat.
19:195-208.
6-27
-------
7. GLOSSARY
Important terms used in this document and in cumulative health risk
assessments are defined below. Definitions have been extended to include the
implications for cumulative risk assessment. Many general risk terms are not included
because standard definitions are readily available elsewhere. In particular, EPA and
ATSDR have developed extensive glossaries of risk assessment terms (available at
http://www.epa.gov/iris/gloss8.htm, http://oaspub.epa.gov/trs/ and
http://www.atsdr.cdc.gov/glossarv.html).
Absorbed dose. The concentration of a chemical inside the body, upon being taken in
through an absorption barrier, e.g., skin absorption, ingestion (see dose).
Acute toxicity. Adverse effect expressed within a short time (generally from minutes to
a day) following exposure to an agent (here, chemical). Most experimental acute
toxicity studies involve response to a single, large dose of an agent, although
occasionally to multiple exposures given within a short time period. EPA defines acute
exposure to be 24 hours or less.
Additivity. Concept that cumulative or joint risk can be represented by adding the
component information, commonly used for chemical doses or their toxic responses.
Additivity is the default assumption for evaluating health effects of multiple chemicals.
Specifically, an additive formula for the toxicity of multiple chemicals is some function of
a linear combination of the component exposures or toxic responses (such as a
weighted sum). Exposure can be represented by the external exposure level or the
internal dose, and toxic response can be represented by the frequency or probability of
toxicity or the measure of toxic effect. (The terms exposure and effect must be explicitly
defined for additivity to be meaningful for a given combination of chemicals.)
Agent. An environmental chemical that could cause harm to human health. (More
broadly interpreted, this term can include biological stressors such as anthrax and
physical stressors such as noise and heat as well as stressors causing impacts other
than toxicity. This document focuses on chemicals and human health effects.)
Aggregate exposure. The combined exposure of a receptor (individual or population)
to a single chemical. The chemical can originate from multiple sources and be present
in multiple media, and exposures can occur by different routes and over different time
periods. Under current EPA definitions, aggregate exposure does not translate to
cumulative risk because it addresses only one chemical; however, combining aggregate
exposures by addressing two or more chemicals would constitute a cumulative risk
assessment.
Antagonism. The process by which two or more chemicals together exert an effect
that is lower than would be predicted by simple addition, which is usually defined as
adding the doses or responses of the individual chemicals. For example, copper has
7-1
-------
been shown to protect against cadmium poisoning. Thus, depending on their levels
(compared with those at which this sparing effect is observed), ingesting both could
reduce the combined toxic response predicted from summing the individual responses.
Additivity must be clearly defined (e.g., dose or response addition) to appropriately
assess whether antagonism exists, and care must be taken to understand the dose-
response relationships. For example, if dose addition were applied when in fact the
chemicals were toxicologically independent (meaning response addition should be
applied), then the result would be lower than expected and could be misinterpreted as
antagonism.
Bioactivation. Process by which a chemical or its metabolite is biochemically
converted to a reactive intermediate. For example, chloroform is converted in the body
to the reactive intermediate phosgene (which was historically used as a chemical
weapon). In a mixture, one chemical can trigger the toxic effects of another by affecting
its bioactivation.
Biomolecule. Any molecule synthesized by an organism, e.g., an enzyme or other
protein.
Chemical antagonism. The process by which two or more chemicals undergo a
chemical reaction to produce a different chemical, which has a lower toxic effect than
that predicted from adding the toxic responses of the original chemicals; this toxic effect
might also qualitatively differ from those of the original chemicals (see antagonism).
Chemical exposure class. A group of chemicals that are physically and chemically
similar, primarily in chemical structure and potential for environmental transformation
and transport (as directly linked to potential exposure). For example, chlorinated
ethanes are considered a chemical exposure class because they are generated by the
same commercial process and have similar fate and transport characteristics so are
often found together in the environment.
Chemical mixture. Two or more chemicals that coexist (e.g., whether at a generating
source, dispersed in the environment, or inside a person) and could contribute to
combined toxicity; their actual identities or origins might or might not be known.
Examples include: (1) Aroclor 1254 (a commercial combination of PCB congeners) in
soil and (2) benzene and ethanol together in the body due to workplace exposures to
benzene followed by drinking beer at home. In parallel with the common risk
assessment term for single chemicals, this can also be referred to as the "mixture of
concern" (see whole mixture and complex mixture).
Chemical synergism. The process by which two or more chemicals undergo a
chemical reaction to produce a different chemical, which has a greater toxic effect than
that predicted from adding the toxic responses of the original chemicals; this toxic effect
might also qualitatively differ from those of the original chemicals (see synergism).
7-2
-------
Chemical toxicity class. A group of chemicals that are toxicologically similar, primarily
due to similarities in chemical structure and biologic activity. Such a group with similar
toxicities could also be a chemical exposure class, e.g., if they were produced by the
same commercial process and frequently coexist in the environment. Where the
composition of such a group is well controlled (e.g., by a standard generating process),
the mixture could be evaluated as a single chemical. Examples include dioxins,
coplanar (dioxin-like) polychlorinated biphenyls (PCBs), and ketones; these similar
groups of compounds can also interact toxicologically with chemicals outside their class.
Complex interaction. The interaction produced by three or more chemicals acting
together that cannot be described according to other interaction definitions. (For two
chemicals, see pair-wise interaction.)
Complex mixture. A mixture containing so many chemicals that any estimate of its
toxicity based on the toxicities of its components is too uncertain to be useful. The
chemical composition of this type of mixture could vary over time or with different
generating conditions. The various components of complex mixtures can be produced
as commercial products or they can be generated simultaneously as byproducts of a
process (e.g., diesel exhaust emissions), or they can coexist because of disposal
practices. To assess risks for complex mixtures, exposure and toxicity data for the
complete mixture are preferred (see whole mixture method).
Component(s). Single chemicals that make up a mixture. These could be further
classified by the type of toxicity they cause. For example, the individual toxicities of
dichloroethylene and acetone ingested together could be separately assessed, as well
as their potential for toxicologic interaction.
Component-based method. An approach for evaluating a mixture using exposure and
dose-response information for the individual chemicals in that mixture. This approach is
useful for comparing mixtures that contain the same chemicals but in differing
concentrations and proportions to determine whether they are similar mixtures. (See
whole mixture method for comparison.)
Contact. The connection between a receptor (person) and a chemical (e.g., in soil,
water, or air). Contact can be continuous (constant) or intermittent (e.g., only occurring
at discrete times during a day or season).
Critical effect. The toxic effect characterized by the lowest observed adverse effect
level (LOAEL), which represents the lowest dose at which any adverse effect is
observed regardless of its nature (e.g., severity) and serves as the basis of the toxicity
values used to assess noncancer effects (see reference dose, reference concentration,
and toxicity value).
Cumulative risk. The combined risk to a receptor (individual or population) from
exposures to multiple agents (here, chemicals) that can come from many sources and
exist in different media, and to which multiple exposures can be incurred over time to
7-3
-------
produce multiple effects. (Health risks are the focus of this document.) More than one
chemical must be involved for the risk to be considered cumulative.
Detoxify. Diminish or remove the toxicologic effect of a chemical, e.g., by metabolic or
chemical reaction with another (sometimes referred to as detoxicate).
Dose. The amount of a chemical that enters into the body (from being administered,
taken, or absorbed), usually expressed as milligrams of substance per kilogram of body
weight. If the exposure surface crossed is an absorption barrier, the dose is an
absorbed dose/uptake dose; otherwise it is an intake dose. The dose represents the
amount available for interaction, e.g., with other chemicals, metabolic processes or
biologically significant receptors.
Dose addition. The process by which the doses of individual chemicals in a mixture
are summed to represent an overall mixture dose. This approach assumes that the
chemicals are toxicologically similar, with each behaving as a concentration or dilution
of an index chemical in that mixture (effectively as a senior or junior clone). The mixture
dose is estimated by summing equivalent doses of the individual chemicals, which are
determined by scaling the toxic potency of each to that of the index chemical (see index
chemical and hazard index).
Effect. The health endpoint resulting from the chemical exposure(s), which can be
estimated or observed (such as increased liver enzyme levels, cardiac arrhythmia, or
cancer). Human health effects are typically estimated from effects observed in animal
toxicity studies, with various adjustment factors applied as appropriate.
Endpoint. An observable or measurable biological event; this can be an observed
effect or a chemical concentration (e.g., of a metabolite in a target tissue) used as an
index of an exposure.
Exposure. The contact between a chemical and the outer boundary of an organism,
quantified as the amount available at the exchange boundaries (e.g., skin, lungs, or
gut). This contact can be intermittent or continuous. The total amount of exposure is
determined by multiplying the exposure time, frequency and duration.
Exposure duration. The total length of time over which an exposure occurs, given in
years for chronic exposures. Unless time-weighted averaging can be justified, repeated
exposures should consider duration to be the time period from start to end of the
exposure. For example, if an individual contacts a chemical 10 minutes a day for
350 days a year over 8 years, the exposure duration is 8 years.
Exposure frequency. How often a receptor is exposed to a chemical over a year, for
chronic exposures. For example, if an individual contacts a chemical 10 minutes a day
for 350 days a year over 8 years, the exposure frequency is 350 days/year.
7-4
-------
Exposure pathway. The physical course a chemical takes from its source to a
receptor. If an exposure is occurring the exposure pathway is considered complete.
The elements of a complete pathway are (1) a chemical source (e.g., waste lagoon) and
mechanism of release (e.g., volatilization or leaching); (2) contaminant fate (such as
physical or chemical changes) and transport through the environment (e.g., air, water,
and soil); (3) an exposure point, or the location where the receptor comes in contact
with either the source itself or a medium carrying the chemical; and (4) an exposure
route.
Exposure route. The way a chemical gets inside an individual who comes in contact
with it, e.g., inhalation, ingestion, or dermal absorption.
Exposure time. How long a receptor is in intermittent or continuous contact with a
chemical over a day. For example, if an individual is in contact 10 minutes a day for
350 days a year over 8 years, the exposure time is 10 minutes/day.
Extrapolation. The process by which information is inferred to fill a gap in existing
data. Commonly used to estimate the response at a low dose, often well below the
range of the experimental data, or equitoxic doses across species. The better
approaches use biologically based mathematical models.
Hazard identification. The process of determining whether exposure to a given
chemical or mixture could cause harm (adverse health effects). It can also involve
qualitatively indicating the nature of the likely health effects.
Index chemical. The one chemical in a mixture against which the toxicities of the other
chemicals are normalized so equivalent doses can be calculated and summed to
represent the total dose of the mixture. Two key criteria are used to select an index
chemical: first, good toxicity data should exist (with a clearly defined dose-response
relationship), and second, it should represent the whole group well. To illustrate,
2,3,7,8-TCDD is the index chemical for dioxins because it has the best toxicity data and
is considered a good representative of this group of compounds; the concentrations of
the other dioxins are multiplied by their individual potencies relative to this isomer, then
summed as "2,3,7,8-TCDD equivalents" to arrive at the dose for the dioxin mixture.
Induction. The initiation or elicitation of a certain response, which can be beneficial or
adverse. The response can be evaluated across a wide scale, from the genetic and
cellular level to the tissue and whole-organism level. For example, at the genetic level
the activity of a regulatory protein can induce increased expression of a certain gene,
while at the molecular level the binding of a chemical to a biomolecule can induce an
enzyme to increase its reaction rate or initiate a series of biochemical reactions that can
ultimately result in an adverse health effect (such as kidney hyperplasia).
Inhibition. The process by which a chemical that is not itself toxic acts on another
chemical that is toxic and makes that chemical less toxic. (More broadly, this term
means the limitation or prevention of a certain response, which could be beneficial or
7-5
-------
adverse. For example, if the response is cell growth, one toxic chemical might inhibit
the growth of certain cells needed for a system to function properly, while another might
inhibit cell proliferation that would otherwise lead to tumor formation [e.g., a
chemotherapeutic agent]. For mixtures, this term is often used to describe beneficial
inhibition as indicated above.)
Initiating factor. A condition involving more than one chemical that catalyzes a
cumulative risk study, such as (1) multiple sources/releases, (2) measured or inferred
chemical concentrations, or (3) illness in a given population.
Interaction. Generally, the influence or action of one chemical on the behavior or effect
of another, which can be mutual or reciprocal. In the environment, interactions among
chemicals can alter their physicochemical forms and transport characteristics (e.g.,
increasing or decreasing mobility and bioavailability). Within the body, one chemical
can interact with another (or others) to cause toxicity, increase or decrease a response,
or completely change the response expected from the individual chemicals acting alone.
Both pharmacokinetics and pharmacodynamics could be altered by the interactions of
chemicals that can target different organs or organ functions and can result from
simultaneous or sequential exposures (so long as they are present at the same time
within the body, e.g., due to pharmacokinetic overlap). The EPA has defined toxic
interactions as being less or more than additive.
Interindividual variability. Differences among individuals within the same species,
e.g., differential susceptibility of humans to a given heath effect from exposure to a
given hazard, which can result from metabolic or other pharmacokinetic differences. To
illustrate for a physical hazard (ultraviolet radiation), one person might sunburn after
spending an hour outside, while another might not burn for several more hours, i.e., until
the exposure is much greater. Similar variability exists for exposures to chemicals and
within other species (see intraspecies variability).
Internal dose. The dose of a chemical inside the body. Depending on the nature of
the data, this can be expressed as (1) the total absorbed dose of the original chemical
(also referred to as the parent compound), (2) the concentration of the parent
compound in target tissues, (3) the total amount of the toxicologically active metabolite,
or (4) the concentration of the toxicologically active chemical species in the target
tissues.
Interspecies variability. Differences between different species (e.g., between rats and
mice, or between rats and humans). A factor of 10 is often applied to account for these
differences in deriving a standard toxicity value to estimate human health effects from
animal studies, as indicated by the appropriate scientific data.
Intraspecies variability. Differences within a single species (e.g., among rats or
among mice, but not between rats and mice). A factor of 10 is often applied to account
for these differences in deriving a standard toxicity value to estimate human health
effects as indicated by the appropriate scientific data (see interindividual variability).
7-6
-------
Joint toxicity. The toxic outcome resulting from the interaction of a set of two or more
chemicals. This outcome can be lower than, equal to, or greater than that predicted by
adding the doses or responses of the component chemicals acting alone.
No observed interaction. The negative outcome of a study of two or more chemicals,
which indicates that they do not interact at the levels studied, to alter either behavior or
effect. For example, considering toxic interactions, if two chemicals were administered
together or coexist within the body due to pharmacokinetic overlap (when exposure
timing differs), and if the effect produced does not differ from that expected by the two
chemicals acting alone (which could also be no effect), then no interaction would be
observed. (Note: this term was used to categorize study outcomes for EPA's Mixtox
data base.)
Parent compound. The original form of a chemical prior to its transformation in the
environment (e.g., by photolysis or microbial degradation) or its transformation within
the body (e.g., by metabolism).
Pharmacodynamics (PD). The study of the biochemical and physiological effects of
drugs and their mechanisms of action, or what they do to the body (see toxicodynamics
for the parallel study of toxic chemicals).
Pharmacokinetics (PK). The study of the absorption, distribution, metabolism, and
excretion of a drug in and from the body (see toxicodynamics for the parallel study of
toxic chemicals).
Physiologically based pharmacokinetic (PBPK) model. A mathematical model that
estimates the dose to a target tissue or organ by taking into account the rates of
absorption into the body, distribution among organs and systems, metabolism, and
elimination. It typically takes the form of compartments that represent organs and
tissues, linked by flow (e.g., blood) exchanges, with associated weights, volumes, flow
rates and fractions, partition coefficients, and metabolic constants based on
physiological studies. These mechanistic PBPK models translate exposure to tissue
concentrations, characterizing tissue dosimetry for different species, doses and route
extrapolations. (Although PBPK models can offer insights into metabolic interactions for
mixtures, integrating multiple contaminants greatly increases the amount of data
needed for parameter estimates.)
Potentiation. The process by which a chemical that is not itself toxic acts on another
chemical that is toxic and makes that chemical more toxic. (More broadly, this term
means the enhancement of a certain response, which could be beneficial or adverse.
For mixtures, this term is often used to describe an enhanced adverse response, as
indicated above.)
Receptor. The individual or population group actually or potentially exposed to a
chemical (receptors can be real or hypothetical). For contaminated sites, various
receptors are typically hypothesized to evaluate potential risks under likely future uses,
7-7
-------
to help guide risk management decisions. In cases where real people might be
incurring exposures (e.g., including cleanup workers), these should clearly be assessed.
Reference concentration (RfC). An estimate (with uncertainty spanning perhaps an
order of magnitude) of a continuous inhalation exposure to the human population
(including sensitive subgroups) that is likely to be without an appreciable risk of
deleterious effects during a lifetime. It can be derived from a NOAEL, LOAEL or
benchmark concentration, with uncertainty factors generally applied to reflect limitations
of the data used. Generally used in U.S. EPA's noncancer health assessments.
Reference dose (RfD). An estimate (with uncertainty spanning perhaps an order of
magnitude) of a daily oral exposure to the human population (including sensitive
subgroups) that is likely to be without an appreciable risk of deleterious effects during a
lifetime. It can be derived from a NOAEL, LOAEL, or benchmark dose, with uncertainty
factors generally applied to reflect limitations of the data used. Generally this is used in
U.S. EPA's noncancer health assessments.
Reference value (RfV). An estimate of an exposure for a given duration to the human
population (including susceptible subgroups) that is likely to be without an appreciable
risk of adverse health effects over a lifetime. It is derived from a BMDL, a NOAEL, a
LOAEL, or another suitable point of departure, with uncertainty/variability factors applied
to reflect limitations of the data used. [Durations include acute, short-term, subchronic,
and chronic and are defined individually in this glossary.] [Reference value is a term
proposed in the report A Review of the Reference Dose and Reference Concentration
Processes (U.S. EPA, 2002e), and is a generic term not specific to a given route of
exposure. U.S. EPA develops numerical toxicity values for the RfD and RfC only; no
numerical toxicity values are developed for the RfV.]
Response addition. The process by which the toxic response of each chemical in a
mixture is summed to represent an overall mixture response. This approach assumes
the chemicals are toxicologically independent, and the toxic response can be defined as
a rate, incidence, risk, or probability of effect. For mixtures, the response equals the
conditional sum of the toxic responses for individual chemicals as defined by the
formula for the sum of independent event probabilities. For two-chemical mixtures, this
means the incremental toxic effect from exposure to the first chemical is the same
whether the second chemical is present or not. (Response addition underlies the
standard process for estimating combined cancer risks by summing the cancer risks of
individual chemicals.)
Risk. The probability (for carcinogens) or potential (for noncarcinogens) that adverse
health effects to result from chemical exposures (see cumulative risk). (More broadly,
this term also covers other types of risks and other stressors, but the focus of this
document is the potential for harm to human health from exposures to multiple
chemicals.)
7-8
-------
Similar components. Single chemicals that cause or are expected to cause the same
type biologic activity based on toxicity studies or chemical structure (e.g., as analogues,
reflecting the structure-activity relationship). In addition to similar characteristics in
terms of physiological processes and toxicity within the body, these chemicals would
also be considered to have similar fate and transport characteristics in the environment.
Evidence of toxic similarity can include (1) similarly shaped dose-response curves,
(2) parallel log-probit or logit dose-response curves for quantal (presence-absence) data
on the number of animals (or people) exhibiting a specific response, and (3) the same
mechanism of action or toxic endpoint. Trichloroethylene and tetrachloroethylene are
examples of similar components.
Similar mixtures. Mixtures of similar chemicals although they might differ slightly from
one another (e.g., same chemicals but in slightly different proportions or the same
chemicals in nearly the same proportions but missing a few or have a few new ones).
Similar mixtures cause or are expected to cause the same type of biologic activity, and
they would act by the same modes of action or affect the same toxic endpoints. In
addition to similar characteristics in terms of physiological processes and toxicity within
the body, these chemicals would also be considered to have similar fate and transport
characteristics in the environment. Varying grades of gasoline (e.g., from regular to
super-premium) are examples of similar mixtures.
Simple mixture. A set of chemicals that is small enough for each individual chemical to
be identified, so the toxicity of the mixture can be characterized by combining the
toxicities and considering the interactions of the component chemicals. For example,
acetone, methylene chloride, and ethanol present together in water to which someone
could be exposed would comprise a simple mixture.
Slope factor. An upper bound, approximating a 95% confidence limit, on the increased
cancer risk from a lifetime exposure to an agent. This estimate, usually expressed in
units of proportion (of a population) affected per mg/kg-day, is generally reserved for
use in the low-dose region of the dose-response relationship, that is, for exposures
corresponding to risks less than 1 in 100.
Source. The location of the environmental chemical(s) being assessed (e.g., an
incinerator stack or waste lagoon), from which it is released and can subsequently be
transported through the environment.
Stressor. A chemical that could cause harm. More broadly, this term also covers
biological agents such as anthrax and physical agents such as noise and heat. The
umbrella definition provided in the Framework for Cumulative Risk (U.S. EPA, 2003a)
extends to any physical, chemical or biological agent that can induce an adverse
response, e.g., a chemical, noise, loss of habitat, or lack of food or water.
Substrate. The substance to which another material attaches or upon which it acts, for
example an environmental chemical or biomolecule upon which an enzyme acts. This
can be a chemical that binds to the active site of an enzyme or other protein in the body.
7-9
-------
Synergism. The process by which two or more chemicals together exert an effect that
is greater than would be predicted by simple addition, which is usually defined as
adding the doses or responses of individual components. For example, depending on
their levels (compared with those at which the toxic interaction is observed), inhaling
both carbon tetrachloride and acetone could produce a more toxic liver response than
would be predicted from summing the individual responses. Additivity must be clearly
defined (e.g., dose or response addition) to appropriately assess whether synergism
exists; care must be taken to understand the dose-response relationships. For
example, if response addition were applied when in fact the chemicals were dose-
additive, then the result would be higher than expected and could be misinterpreted as
synergism.
Target Organ. The biological organ adversely affected by a given chemical or mixture.
Toxicity value. The standard value used to translate chemical exposures (doses) to
estimates of cancer risks or the potential for noncarcinogenic effects. The cancer or
noncancer toxicity value is specific to the chemical (or mixture), route of exposure, and
duration over which the exposure occurs. These values are typically derived from
animal studies, with adjustment factors applied to develop estimates for humans. For
the cancer endpoint the toxicity value is termed the slope factor, and for noncarcinogens
it is termed the reference concentration (RfC) for inhalation exposure and reference
dose (RfD) for oral exposure.
Toxicodynamics (TD). The sequence of events at the cellular and molecular levels
leading to a toxic response following exposure to a chemical. This involves the
processes underlying the effect severity, reversibility, recovery, and adaptive response.
(See the general term pharmacodynamics, which was developed for drug studies.
Although the TD term is often used in risk assessments of environmental chemicals,
pharmacodynamics could be a more appropriate term for certain chemicals, e.g.,
essential metals, depending on the exposure levels.)
Toxicokinetics (TK). The characterization and quantification of the time course of
absorption, distribution, and metabolism (or biotransformation) in the body and
elimination (or excretion) from the body of a chemical taken in. (See the general term
pharmacokinetics, which was developed for drug studies. Although the TK term is often
used in risk assessments of environmental chemicals, pharmacokinetics could be a
more appropriate term for certain chemicals, e.g., essential metals, depending on the
exposure levels.)
Toxicologic interaction class. A group of chemicals that are toxicologically similar in
terms of the direction of toxicologic interaction (synergism, antagonism or additivity).
For any given interacting chemical, when paired with other members of this group the
direction of the interaction would be the same. This group can be defined as a
toxicologic interaction class only for specific toxic endpoints. Ketones and selenium
compounds are examples of interaction classes.
7-10
-------
Unable to assess. The effect of the chemical (mixture) cannot be classified, for
example due to lack of proper control groups; lack of statistical significance; or poor,
inconsistent, or inconclusive data in the available toxicity studies.
Uncertainty factor (UF). An adjustment factor applied to experimental data in deriving
toxicity values used to estimate health risks and the potential for noncancer effects.
These factors are applied to account for (1) variation in susceptibility among members
of the human population; (2) uncertainty in extrapolating animal data to humans;
(3) uncertainty in extrapolating from data obtained in a study with less-than-lifetime
exposure; (4) uncertainty in extrapolating from a lowest-observed-adverse-effect level
(LOAEL) instead of a no-observed-adverse-effect level (NOAEL); and (5) uncertainty
associated with extrapolation when the database is incomplete (which might be
addressed by a modifying factor).
Whole mixture. A mixture that is evaluated in its entirety, usually with exposure levels
for the entire mixture unadjusted for any differences among the toxic potencies of its
component chemicals. Some whole mixtures can be defined and are reproducible, e.g.,
where the process that created them is well understood. Other whole mixtures are
defined by groups of structurally similar chemicals that often co-occur. Examples
include total chromium and compounds and total petroleum (hydrocarbons). This term
is often applied to highly complex mixtures with components that cannot be fully
identified or reproducibly measured. Diesel exhaust, gasoline and toxaphene are
specific examples.
Whole mixture method. An approach in which the whole mixture is treated as a single
entity, similar to the way single chemicals are assessed, and thus requires dose-
response information for the whole mixture. This approach is used for complex
mixtures; and it is best applied to mixtures with a composition that is constant over the
entire exposure period. It differs from the component-based method because the
toxicity information inherently reflects unidentified chemicals in the mixture as well as
any interactions that might be occurring among the chemicals. (See the component-
based method for comparison.)
7-11
-------
APPENDIX A
CUMULATIVE RISK TOOLBOX
This appendix identifies resources that can be used to address various elements
of cumulative risk assessments for specific situations and contaminated sites. Several
have been applied at sites being addressed by the EPA and U.S. DOE. Many of these
resources are also useful for other types of cumulative risk analyses, and tools from
EPA studies for several regulatory programs are also included here. In addition to
resources provided by the EPA, included herein are many documents from other
Federal Agencies and organizations. They are included to provide sources of
information to the reader that may be useful in conducting a cumulative risk
assessment, however, their inclusion does not necessarily imply their review or
endorsement by the EPA.
Many federal, state, academic and professional organizations have developed
general risk assessment guidelines and tools for a variety of situations. While some
resources clearly consider multiple exposures to multiple chemicals, such as the
standard Risk Assessment Guidance for Superfund (U.S. EPA, 1989a), relatively few
are described as explicitly assessing cumulative risks by specifically addressing
groupings or joint toxicity, or by being population focused. The main body of this report
includes discussions of how more recent cumulative risk approaches can enhance the
traditional risk assessment approach. The toolbox of information resources presented
in this appendix includes many tools developed for general risk assessments that can
also be used or adapted for population specific cumulative risk assessments, or whose
underlying approaches offer insights for these assessments. This toolbox is not
intended to be comprehensive; the aim is simply to highlight those resources that could
be useful for cumulative health risk assessments. This appendix focuses on chronic
exposures, but some resources related to acute or subchronic exposures (such as
those developed for health and safety in the workplace) are also included.
Resources that support planning, scoping, and problem formulation, including
stakeholder involvement, are identified in Section A.1. Those that support evaluations
of contaminant fate and transport and exposure, which range from summary data on
physicochemical constants to specific transport and exposure models, are highlighted in
Section A.2. Resources that support the toxicity evaluation are offered in Section A.3,
and those that support the characterization of risk and uncertainty and presentation of
results are highlighted in Section A.4. Several resources cover more than one of these
topics; where this is the case, they are generally listed within their main area of
A-1
-------
emphasis. The information reproduced here is believed accurate as of the publication
date. The intent is to post these resources on the EPA's Web site and update them
regularly.
A.1. RESOURCES FOR PLANNING, SCOPING AND PROBLEM FORMULATION
Topics addressed during iterative planning, scoping and problem formulation
include the purpose and scope of the assessment (which involves considering multiple
chemicals, exposures, effects and population groups), the products needed, the data to
be collected and synthesized, the general assessment approach and stakeholder
involvement. Cumulative risk assessments are complex because of the very large
number of potential combinations of chemicals and interactions inherent to
environmental settings.
During this initial and iterative phase of a cumulative assessment, a main focus is
on which chemicals present are most likely to interact and what the nature of those
interactions might be. The internet has emerged as a very valuable tool for stakeholder
involvement. It can be used to easily provide information about the project and
associated scientific issues for a wide audience, which can be browsed on-line or
downloaded at the user's convenience. It can also be used to notify interested parties
of upcoming meetings or the availability of specific reports for the site. Project websites
and e-mails can also be used to effectively solicit and receive stakeholder inputs about
the project. Limited-access web sites can be also used to share and evaluate draft
information as it is developed.
The usefulness of internet-based approaches for stakeholder involvement is
described further below, and examples of specific tools are included in Table A-1. (Note
that most resources presented in this toolbox are available through the internet.)
1. Low cost to involve many stakeholders. Although fixed costs to build a website
can be somewhat high, the marginal cost to involve additional stakeholders is
nearly zero, so the internet can be cost-effective for projects with extensive
stakeholder participation. For example, a document can be posted on a website
very cheaply; in contrast, mailing would require postage, printing, and paper
costs with marginal costs that do not diminish significantly with additional users
(essentially free via the internet method). Receiving stakeholder inputs through
the web or e-mail can also save costs compared with paper-based approaches.
2. Wide geographical reach. Using a website and e-mail allows ready access to
information and opportunity for participation regardless of stakeholder location, in
contrast to traditional methods that typically focus on people nearby. This is
particularly important when travel to public meetings is restricted (e.g., due to
cost, schedule, or physical disabilities). This broad accessibility can increase
participation because additional people become aware of the project (e.g.,
A-2
-------
through web searches). The use of e-mail can also be effective because
information can be delivered to a broad set of stakeholders at their desktops.
3. Availability. Information posted to a public website is available 24 hours a day, 7
days a week, and can be accessed at times convenient to the user - which can
also increase participation. (People without computers could access the internet
from libraries or other such facilities during regular hours.) Likewise, e-mails can
be opened at the user's convenience.
4. Extent of information. Large amounts of data and other information can be
provided via the internet, much more than would be reasonable by other means
(meetings and paper). Further, this information can be reviewed at whatever
level of detail and pace the user prefers.
5. Immediacy. Information can be made available essentially immediately via the
internet. This can be especially useful for situations that might arise when the
level of concern is high (e.g., when wildfires or accidents cause acute releases).
6. Data interactivity. Websites can integrate the capabilities of many different
databases, geographic information systems (GISs), graphing, and other tools so
stakeholders can play with data and information in ways that would not be
possible under traditional methods (e.g., with hard copies). This can include
"clicking" on specific locations to identify multiple chemicals present there, or
searching to find all locations with a specific combination of chemicals (e.g.,
which could be known to interact).
7. Flexibility. Information shared vie the web or e-mail can be made available in
different types of electronic formats, which can facilitate use by multiple parties.
Also, websites and e-mail communications can be readily adapted to
accommodate new types of information as it is developed.
Selected resources that can be used to support planning, scoping, and problem
formulation for cumulative risk assessments, including stakeholder involvement, are
briefly described below. Selected information is also summarized in Table A-1 at the
end of this section.
Framework for Cumulative Risk Assessment (U.S. EPA). The Framework
document released in spring 2003 identifies an umbrella structure for cumulative
risk assessments, identifies key issues and defines common terms. It
summarizes basic elements of the cumulative risk assessment process and
presents a flexible structure for conducting cumulative risk assessments. Neither
a procedural guide nor a regulatory requirement, this framework is expected to
evolve over time. The document does not present protocols to address specific
risk issues; rather it provides good information about important aspects of
cumulative risk (U.S. EPA, 2003a). A main foundation of this document, the
report is available at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=54944.
Planning and Scoping for Cumulative Risk Assessment (U.S. EPA).
Guidance was published in 1997 by the EPA Office of Science Policy, Science
Policy Council, which reflects the EPA's policy statement for planning and
A-3
-------
scoping for cumulative risk assessments (U.S. EPA, 1997a). This guidance
presents ideas for broad-based approaches, including consideration of multiple
endpoints, sources, pathways and routes of exposure; community-based
decision making; flexibility in achieving goals; case-specific responses; a focus
on all environmental media; and holistic reduction of risk. This report is available
at http://www.epa.gov/OSA/spc/2cumrisk.htm. Lessons learned from cumulative
risk case studies are captured in a companion technical memorandum (report)
(U.S. EPA, 2002f), available at http://www.epa.gov/osp/spc/llmemo.htm.
Environmental Justice Geographic Assessment Tool (U.S. EPA) and Similar
Ranking/Prioritization Tools. Designed jointly by the EPA Office of
Environmental Information and Office of Environmental Justice, this tool is a CIS-
based module to support front-end scoping of cumulative assessments. It
combines environmental, socioeconomic and health indicators in statistical
tables, and it was initially developed to evaluate potential issues related to
environmental justice. Where a community-based approach is applied, this tool
can be helpful in identifying the risk problems to be assessed. (Although
presented here within the planning/problem formulation stage, this can also be a
used to support risk characterization.)
Site Conceptual Exposure Model (SCEM) Builder (DOE). The SCEM Builder
was developed by the DOE Office of Environmental Policy and Guidance in 1997
to support planning, scoping, and problem formulation for risk assessments at
contaminated sites, by providing a tool to build SCEMs. An SCEM is a visual
representation of scenarios that organizes information about sources of
contamination, release mechanisms, exposure pathways and receptors for a site
and can be used to address data gaps. These conceptual models are often used
to develop data quality objectives (DQOs) and prioritize field sampling activities,
in order to help reduce uncertainty associated with risk characterization. Using
this tool, analysts can build SCEMs for a given site and modify variables to refine
the model, e.g., to reflect stakeholder inputs. This tool can also be used to
develop SCEMs for various "what-if" scenarios to help bound data uncertainties.
It is available at http://tis.eh.doe.gov/oepa/programs/scem.cfm.
Stakeholder Involvement (U.S. EPA, DOE). Several resources exist that
document the procedures and approaches implemented to support stakeholder
involvement activities in risk assessment projects. These range from national
policy guidance documents to site-specific reports that chronicle the approaches
taken by individual projects to solicit input from stakeholders and incorporate
their concerns and ideas into the analysis plan. Guidance from the EPA
Superfund and Environmental Justice programs (captured in Table A-1)
encourages community involvement and can be useful for cumulative risk
assessments at contaminated sites.
A number of stakeholder involvement examples exist that can offer insights for
cumulative risk assessment projects. Many are available for contaminated DOE
sites, where citizen advisory boards have been established to provide input
during planning and scoping and as assessments progress. The mission or
charter language prepared by these advisory boards can offer clues for other
A-4
-------
projects. Such language typically includes general "rules of engagement"
(including respect for diverse opinions) as well as specific roles and
responsibilities (notably with regard to providing advice and recommendations
instead of making management decisions for the project).
For example, a citizen's advisory board (CAB) was created to facilitate public
outreach for the DOE Savannah River Site. That CAB consists of 25 individuals
from South Carolina and Georgia chosen by an independent panel of citizens
from approximately 250 applicants that reflect the cultural diversity of the local
population. The CAB has considered itself a major component of the risk
assessment/management team for the site and maintains a website
(www.srs.gov/general/outreach/srs-cab) that offers ideas that can be useful for
similar programs at other sites.
A stakeholder advisory board has also been established at the DOE Hanford site
in Washington. Information on the Hanford Advisory Board (HAB) is available at
http://www.hanford.gov/public/boards/hab/. This Board created a calendar for
public involvement that lists upcoming meetings and other events at which input
from affected parties and stakeholders is encouraged. Nearly a decade ago, an
advisory group that included many stakeholders and a technical expert team
from the project considered an approach for a comprehensive impact
assessment for the Columbia River that flows next to the site; that effort is no
longer underway as defined at that time, but related information can be found on
the internet (e.g., see http://www.hanford.gov/docs/rl-96-16/). The DOE
management at Hanford has also put together a comment response tracking
system, as have other sites, to coordinate the issues identified by stakeholders
during the iterative planning and scoping phase and throughout the assessment
process (which at this site will last for decades), and to track follow-ups.
A stakeholder involvement program is underway for an ongoing sitewide
cumulative risk assessment and risk reduction project at the DOE Los Alamos
National Laboratory (LANL) in New Mexico. This approach has been developed
and is being implemented by the independent Risk Assessment Corporation
(RAC) team is under the Risk Analysis, Communication, Evaluation, and
Reduction (RACER) project. The primary objectives of this project are to
develop:
1. A process for extensive stakeholder involvement in the risk assessment and
decision-making processes for LANL
2. A methodology to estimate contemporary (current) human health risks and
ecological impacts from LANL using available data on chemicals and
radionuclides measured in environmental media
3. A methodology to implement a comprehensive risk-informed decision
analysis framework, including a prospective risk and ecological impact
assessment and other quantitative and qualitative criteria, to guide long-term
management of risks and ecological impacts at LANL
A-5
-------
4. A consistent approach for efficiently compiling, using and updating data to
support the risk assessment and decision-making processes
Guidelines developed by RAC for involving stakeholders in this project are
included on the project website at
http://www.racteam.com/LANLRisk/Reports/Guidelines%20for%20lnvolvement%
2010-30.pdf. The RACER project is also involving local schools in science
projects, including to provide input to exposure scenarios. This input is also
being solicited in one-on-one meetings with others at various locations in the
community (businesses and homes).
A much earlier scientific educational partnership was established more than a
decade ago at the Weldon Spring site. Information about that Partners in
Education program can be found at http://web.em.doe.gov/wssrap/pie.html.
Every community will have its own priorities and levels of interest. More
examples are given in Table A-1.
Data Quality Objectives and Assessment (U.S. EPA). The EPA has
developed a series of documents that provide guidelines to help ensure that the
data collected are appropriate for their intended use (see Table A-1). These
documents outline a systematic planning process for developing performance
criteria for the collection, evaluation and use of environmental data. This process
can be used to focus communication among interested parties and to form the
basis for selecting decision points for a risk assessment project. The overall
approach is called the DQO process, and it is detailed in Guidance for the Data
Quality Objectives Process (U.S. EPA, 2000g). The seven-step planning
approach to develop sampling designs for data collection is iterative and applies
to all scientific studies, but it is particularly useful for addressing problems that
have two clear alternatives. The final outcome of the DQO process is a design
for collecting data (including the number of samples, location of samples and
collection method) that acknowledges the limits on the data collection and the
probabilities of making decision errors. Guidance can be found at
http://www.epa.gov/quality/qs-docs/g4-final.pdf.
The EPA has also developed Data Quality Assessment (DQA) guidance
(U.S. EPA, 2000h) that describes procedures to help ensure that data used in
risk assessments are appropriate for their intended use with respect to quality,
quantity and type. Also provided are statistical and analytical tools that can be
used to review DQOs and sampling designs, review preliminary data, select
statistical tests to summarize and analyze data, verify the assumptions of the
statistical test, and perform appropriate calculations.
A-6
-------
TABLE A-1
Selected Resources for Planning, Scoping and Problem Formulation
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Resources for Overall Planning, Scoping and Problem Formulation
Framework for Cumulative Risk
Assessment (U.S. EPA)
http://cfpub.epa.gov/ncea/cfm/recordispl
ay.cfm?deid=54944
Provides a flexible framework for cumulative risk
assessments; identifies the basic elements of the process,
describes a number of technical and coordination issues
and defines terms.
Defines general structure and components
of cumulative risk assessments; serves as
the foundation for this report.
Guidance on Cumulative Risk
Assessment - Part 1, Planning and
Scoping (U.S. EPA)
http://www.epa.gov/OSA/spc/2cumrisk.
htm
This guidance directs each office of the EPA 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 in
all cases for which relevant data are available. It
describes general approaches and concepts for planning
and scoping for cumulative risk assessments.
Identifies four key steps for planning and
scoping: determine overall purpose and risk
management objectives for assessment;
determine the scope, problem statement,
participants and resources; determine the
risk dimensions and technical elements that
may be evaluated; and formulate a
technical approach including a conceptual
model and an analysis plan for conducting
the assessment.
Lessons Learned on the Planning and
Scoping of Environmental Risk
Assessments (U.S. EPA)
http://www.epa.gov/osa/spc/pdfs/handb
ook.pdf
Provides early feedback to EPA scientists and managers
regarding EPA's experiences with planning and scoping
as the first step in conducting environmental
assessments. It is intended to reinforce the importance of
formal planning and dialogue prior to conducting complex
cumulative assessments and to provide case studies
"lessons learned" for anyone involved in planning an
assessment.
Provides information and feedback from the
Part 1 planning guidance that offer insights
for designing and conducting cumulative
risk assessments.
A-7
-------
TABLE A-1 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Environmental Justice Geographic
Assessment Tool (U.S. EPA)
http://www.epa.gov/enviro/ej/
CIS-based module designed for front-end scoping of
cumulative assessments. Combines environmental,
socioeconomic and health indicators in statistical tables.
Initially developed to evaluate potential environmental
justice (EJ) issues.
Allows interactive mapping and review of
regulated facilities, environmental
monitoring sites, bodies of water, land use,
community demographics and
streets/schools/hospitals. Can be adapted
or linked as a module to assess cumulative
risks for various communities (i.e., not
limited to EJ issues).
SCEM Builder Model (DOE)
http://tis.eh.doe.gov/oepa/programs/sce
m.cfm
Graphics tool designed to develop a site conceptual
exposure model for a contaminated site.
General graphics tool that can be used to
set up a conceptual model for the site, to
guide stakeholder inputs for a cumulative
risk assessment.
Risk Screening Environmental
Indicators (RSEI) (U.S. EPA)
http://www.epa.gov/opptintr/rsei/
Screening tool that compares toxic chemicals released to
the environment from industrial sources. Offers way to
examine rankings and trends and set priorities for further
action.
Allows data to be sorted by chemical,
media, and geographic area. Preliminary
analyses can identify situations of relatively
higher concern during scoping.
Resources for Stakeholder Involvement
Community Air Screening How To
Manual (U.S. EPA)
http://www.epa.gov/oppt/cahp/howto.ht
ml
Explains how to form a partnership, clarify goals, develop
a detailed local source inventory, use a risk-based
process to identify priorities and develop options for risk
reduction. Developed by the EPA's Office of Pollution
Prevention and Toxics based on the Baltimore, MD,
approach. (Expected to be published in spring 2004.)
Presents and explains a step-by-step
process a community can follow to: form a
partnership to access technical expertise,
identify and inventory local sources of air
pollutants, review these sources to identify
known hazards that might pose a health risk
to the community and set priorities and
develop a plan for making improvements.
Covers only the air pathway.
A-8
-------
TABLE A-1 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Superfund Community Involvement
Handbook, Appendix on Community
Involvement Requirements (U.S. EPA)
http://www.epa.gov/superfund/action/co
mmunity/index.htm
Superfund guidance on suggested community
involvement structure, communications and approach.
For contaminated Superfund sites, the lead agency
informs public of the availability of technical assistance
grants (TAG). TAG is a grant program that provides funds
for citizen groups to hire independent technical advisors to
help them understand/comment on technical decisions re:
Superfund cleanup actions.
Developed for the EPA's Superfund
program, the information about community
involvement, including forming community
advisory groups (CAGs), is useful for
cumulative risk assessments at
contaminated sites.
Hanford Site, Hanford Advisory Board
(HAB), Public Involvement Resources
and Calendar (DOE site)
http://www.hanford.gov/orp/?page=5&p
arent=1
http://www.hanford.gov/public/calendar/
The HAB was set up to provide recommendations and
advice to DOE, EPA and the Washington Department of
Ecology on a number of issues related to cleanup of the
Hanford site.
The HAB has developed mission language,
a meeting schedule/calendar and other
information that can serve as examples for
other projects.
Los Alamos National Laboratory (LANL)
Risk Analysis, Communication,
Evaluation, and Reduction (RACER)
project (DOE site)
http://www.racteam.com/LANLRisk/Rep
orts/Guidelines%20for%20lnvolvement
%2010-30.pdf
The RACER project is founded on extensive stakeholder
involvement. Established by the RAC team, this project is
developing an open process for assessing cumulative
risks at LANL and for creating a decision analysis
framework for risk reduction, as well as an integrated
database (containing data from multiple collecting
organizations) to support data evaluations and trend
analyses, site risk assessments and the overall decision-
making process for environmental management at LANL.
Stakeholder participation is actively sought, both open
progress meetings and one-on-one meetings are held (in
various settings), and the internet (project website and
e-mail) is also used to announce upcoming activities and
the availability of draft documents for stakeholder
comment, and to solicit inputs.
Insights for cumulative assessments can be
found in: RAC guidelines for stakeholder
involvement, open survey questions, plans
for soliciting (in various venues) and
summarizing inputs to guide the
assessment and suggestions for pursuing
grants for ongoing stakeholder involvement
(aimed to be administered through an
independent group), as well as other plans
and products that can be found on the
project website.
A-9
-------
TABLE A-1 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Community Air Screening How To
Manual (U.S. EPA)
http://www.epa.gov/oppt/cahp/howto.ht
ml
Explains how to form a partnership, clarify goals, develop
a detailed local source inventory, use a risk-based
process to identify priorities and develop options for risk
reduction. Developed by the EPA's Office of Pollution
Prevention and Toxics based on the Baltimore, MD,
approach. (Expected to be published in spring 2004.)
Presents and explains a step-by-step
process a community can follow to: form a
partnership to access technical expertise,
identify and inventory local sources of air
pollutants, review these sources to identify
known hazards that might pose a health risk
to the community and set priorities and
develop a plan for making improvements.
Covers only the air pathway.
Superfund Community Involvement
Handbook, Appendix on Community
Involvement Requirements (U.S. EPA)
http://www.epa.gov/superfund/action/co
mmunitv/index.htm
Superfund guidance on suggested community
involvement structure, communications, and approach.
For contaminated Superfund sites, the lead agency
informs public of the availability of technical assistance
grants (TAG). TAG is a grant program that provides funds
for citizen groups to hire independent technical advisors to
help them understand/comment on technical decisions re:
Superfund cleanup actions.
Developed for the EPA's Superfund
program, the information about community
involvement, including forming community
advisory groups (CAGs), is useful for
cumulative risk assessments at
contaminated sites.
Hanford Site, Hanford Advisory Board
(HAB), Public Involvement Resources
and Calendar (DOE site)
http://www.hanford.gov/orp/?page=5&p
arent=1
http://www.hanford.gov/public/calendar/
The HAB was set up to provide recommendations and
advice to DOE, EPA and the Washington Department of
Ecology on a number of issues related to cleanup of the
Hanford site.
The HAB has developed mission language,
a meeting schedule/calendar and other
information that can serve as examples for
other projects.
A-10
-------
TABLE A-1 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Los Alamos National Laboratory (LANL)
Risk Analysis, Communication,
Evaluation, and Reduction (RACER)
project (DOE site)
http://www.racteam.com/LANLRisk/Rep
orts/Guidelines%20for%20lnvolvement
%2010-30.pdf
The RACER project is founded on extensive stakeholder
involvement. Established by the RAC team, this project is
developing an open process for assessing cumulative
risks at LANL and for creating a decision analysis
framework for risk reduction, as well as an integrated
database (containing data from multiple collecting
organizations) to support data evaluations and trend
analyses, site risk assessments and the overall decision-
making process for environmental management at LANL.
Stakeholder participation is actively sought, both open
progress meetings and one-on-one meetings are held (in
various settings), and the internet (project website and
e-mail) is also used to announce upcoming activities and
the availability of draft documents for stakeholder
comment, and to solicit inputs.
Insights for cumulative assessments can be
found in: RAC guidelines for stakeholder
involvement, open survey questions, plans
for soliciting (in various venues) and
summarizing inputs to guide the
assessment and suggestions for pursuing
grants for ongoing stakeholder involvement
(aimed to be administered through an
independent group), as well as other plans
and products that can be found on the
project website.
Savannah River Site Citizen's Advisory
Board (CAB) (DOE site)
http://www.srs.gov/general/outreach/srs
-cab
The CAB provides advice and recommendations DOE,
EPA, and the South Carolina Department of Health and
Environmental Control on environmental remediation,
waste management and related issues. Meetings and
public comment sessions are held regularly and are open
to the public.
Recommendations and information on
workshops published on this website can
offer insights for similar projects.
Multnomah County Protocol for
Assessing Community Excellence in
Environmental Health (PACE-EH)
http://www.pace-eh.org
Pilot assessments performed in five neighborhoods of
Portland, Oregon, resulted from a community health
assessment team's efforts to prioritize environmental
health concerns.
Multipathway issues identified that can offer
insights for other studies include poor
indoor air quality (including mold and
mildew), exposure to lead-based paint and
unsafe grounds.
A-11
-------
TABLE A-1 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Onondaga Lake Partnership (OLP)
website http://www.onlakepartners.org
Aim is to promote cooperation among government
agencies and others involved in managing environmental
issues of Onondaga Lake and the Onondaga Lake
watershed in Syracuse, New York. The website presents
information about pollutants, health risks, cleanup
projects, and opportunities for public involvement in this
complex cleanup project for a heavily polluted lake in a
major metropolitan area, with high level of public concern.
Similar to previous example, illustrates how
a variety of scientific information,
documents, program management
information, presentations, video clips,
image gallery and an e-mail announcement
list can be shared for cumulative risk
assessment projects.
Depleted Uranium Hexafluoride
Management Information Network
(DOE project)
http://www.ead.anl.gov/uranium
Presents information for the DOE inventory of depleted
uranium hexafluoride (DUF6). Includes basic scientific
information on uranium, depleted uranium and DUF6; the
DOE program for managing the DUF6 inventory; research
and development for beneficial uses of DU, and public
involvement opportunities. Environmental impact
statements (EISs) and other reports are included.
(Several hundred thousand visitors since 1997.) Used
comment response management system (CRMS), web-
enabled software which expedites responses to
government and public comments about this and other
EISs.
Similar to previous example, illustrates how
various reports, presentations, video clips,
image gallery and an e-mail announcement
list can be shared for a cumulative risk
assessment project.
Resources for Guiding Data Quality
Guidance for the Data Quality
Objectives Process (QA/G-4) (U.S.
EPA) http://www.epa.gov/gualitv/gs-
docs/g4-final.pdf
Guidance on the data quality objectives (DQO) process, a
systematic planning process for environmental data
collection. Designed to help analysts ensure that data are
collected for a specific purpose. Includes determination of
chemicals to evaluate or test for, media and locations of
concern, and detection limits.
Developed for the recommended planning
process when environmental data are used
to select between two opposing conditions,
this general guidance is useful for
cumulative assessments. Focus is placed
on the cumulative risk questions to be
answered, while maintaining awareness of
appropriate statistical techniques that
should be considered to produce defensible
results.
A-12
-------
TABLE A-1 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Decision Error Feasibility Trials (DEFT)
Software (QA/G-4D) (U.S. EPA)
http://www.epa.gov/qualitv/qs-docs/g4d-
final.pdf
Computer-based software for determining the feasibility of
data quality objectives defined using the DQO process.
Enables statistical sample size planning and can be used
to estimate costs associated with obtaining a specific
precision in environmental data (such as how many
samples are required to determine whether environmental
concentrations are above or below background or risk-
based concentrations).
General analytical guidance can be applied
to multiple media and multiple
contaminants. This tool calculates the
appropriate number of environmental
samples required to statistically answer
whether soil or water concentrations are
above or below a risk-based level, which
could be adapted to grouped chemicals.
Guidance on Choosing a Sampling
Design for Environmental Data
Collection (QA/G-5S) (U.S. EPA)
http://www.epa.gov/gualitv/gs-docs/g5s-
final.pdf
Guidance on applying standard statistical sampling
designs (such as simple random sampling) and more
advanced sampling designs (such as ranked set
sampling, adaptive cluster sampling) to environmental
applications.
Can be useful to identify co-located
contaminants to support grouping for a
cumulative risk assessment at a
contaminated site or situation.
Guidance for Quality Assurance Project
Plans for Modeling (QA/G-5M) (U.S.
EPA) http://www.epa.gov/guality/gs-
docs/g5m-final.pdf
General guidance for developing quality assurance project
plans (QAPPs) for modeling projects.
Can be useful to cumulative risk
assessments, particularly where air or
groundwater models are needed to
extrapolate small data sets to the site or
community level.
Guidance on Environmental Data
Verification and Data Validation (QA/G-
8) (U.S. EPA)
http://www.epa.gov/guality/gs-docs/g8-
final.pdf
Guidance to help organizations verify and validate data.
Applying this to laboratory analytical data allows analysts
to understand uncertainties associated with concentration
measurements (which impact assessment results).
Useful for determining appropriate data for
the chemicals to be evaluated in a
cumulative risk assessment; important to
results, especially when using conservative
screening approaches.
A-13
-------
TABLE A-1 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Guidance for Data Quality Assessment:
(DQA): Practical Methods for Data
Analysis (QA/G-9) (U.S. EPA)
http://www.epa.gov/qualitv/qs-docs/q9-
final.pdf
Describes procedures and methodologies for ensuring
sound data are used in the risk assessment. Provides
tools that can be used to review DQOs and sampling
design, review preliminary data, select statistical tests to
summarize and analyze data, verify the assumptions of
the statistical test and perform calculations.
These tools can indicate differences in the
statistical robustness that might affect data
combinations for chemical
groupings/selection of representative
concentrations. For instance, if certain data
were collected according to DQOs
established with DEFT (see earlier entry)
while other data were collected under a
different program that required fewer
samples, then care must be taken when
combining those data.
A-14
-------
A.2. RESOURCES FOR ENVIRONMENTAL FATE AND TRANSPORT ANALYSES
Several tools that can be used to evaluate environmental fate and transport of
chemicals to support cumulative health risk assessments are highlighted below.
Selected information is summarized in Table A-2 at the end of this section.
ChemFinder Database (Private, via U.S. EPA). The ChemFinder database is
an online, EPA-linked search engine that provides access to information on the
chemical, physical, product and biological properties of a large number of
chemicals. Developed by Cambridgesoft, this tool can be searched by common
name, brand name, Chemical Abstract Service (CAS) number, chemical formula,
or other designations, including chemical structure. ChemFinder searches
chemical information from a large pool of websites worldwide, including
government and multilateral agencies, universities and private institutions. The
ChemFinder search engine is available for free use via the EPA Office of
Pesticide Programs at http://www.epa.gov/oppfead1/pmreg/pits/index.html and
can also be found at http://chemfinder.cambridgesoft.com/.
Risk Assessment Protocols for Hazardous Waste Combustion Facilities
(U.S. EPA). In 1998, EPA Region 6 identified the need for a guidance document
that consolidated information presented in earlier EPA documents and in reports
from state environmental agencies, to provide an integrated set of procedures for
conducting site-specific combustion risk assessments addressing multiple
sources and exposure scenarios. Two documents were prepared: the Human
Health Risk Assessment Protocol for Hazardous Waste Combustion Facilities
(HHRAP; U.S. EPA, 2005b), and the Screening Level Ecological Risk
Assessment Protocol for Hazardous Waste Combustion Facilities (SLERAP;
U.S. EPA, 1999f). The objectives of these documents were to (1) apply the best
available methods for evaluating risk to human health and the environment from
operations of hazardous waste combustion units and (2) develop repeatable and
documented methods for consistency and equity in permitting decisions.
In addition to providing methodologies for evaluating multi-media, multi-pathway
risks, Volume II of the guidance contains information and data on the chemical,
physical, and environmental properties of many chemicals that can be used to
model environmental fate and transport and exposure. These data can also be
used to predict what chemicals are likely to behave similarly in the environment,
to support groupings for cumulative risk assessments.
Soil Screening Guidance (U.S. EPA). The EPA has developed an extensive
set of environmental and physical constants and parameters that can be used to
model the fate and transport of chemicals in soil and to develop risk-based soil
screening levels (SSLs) to protect human health (U.S. EPA, 1996a).
The primary goal is to provide simple screening information and a method for
developing site-specific screening levels, so it also serves as a tool to support
exposure-based screening. The guidance includes both detailed models and
generic SSLs, which can be used to quickly (and conservatively) assess what
areas or pathways might not warrant a detailed assessment. Developed for use
A-15
-------
at National Priorities List sites, the concepts can be extended to other sites and
situations. The guidance also includes tables of chemical-specific constants,
such as the organic carbon partition coefficient (Koc), the soil-water partition
coefficient (Kd), and water and air diffusivity constants (Di,w and Di,a), to support
the evaluation of fate and transport.
Background Determinations (U.S. EPA, Others). Concentrations that
appropriately represent "background" levels (naturally occurring or ambient) are
location-specific and help provide context for the fate and transport of site
chemicals. The EPA has prepared extensive guidance on various approaches
for characterizing background, as well as protocols for determining whether a
contaminated site's concentrations are statistically above background. For
example, see Guidance for Characterizing Background Chemicals in Soil at
Superfund Sites (U.S. EPA, 2001 d).
Data on background concentrations of inorganics (notably in soil) can be found in
several sources, and these data can provide an initial general context for site- or
community-specific risk analyses. The information sources include toxicological
profiles developed by the EPA for Toxic Substances and Disease Registry
(ATSDR) and reports from the U.S. Geological Survey and universities. EPA
sources include the Ecological Soil Screening Level Guidance (U.S. EPA,
2004e), which gives 50 state-specific ranges, and regional guidelines, and
"typical" values provided as technical background to risk-based screening levels
(U.S. EPA, 2002g, 2003i). The EPA Region 6 includes background
concentrations in its Human Health Medium-Specific Screening Levels document
(U.S. EPA, 2005c), and the associated database contains screening values and
the physical and chemical parameters that were used to derive those values.
Background data can also be found in state-specific documents, such as the
Texas Risk Reduction Program Guidelines (TCEQ, 1999), which include
background concentrations for the state. The Massachusetts Department of
Environmental Protection (MADEP) has published state-specific background
levels of PAHs and metals in soil
(http://www.tceq.state.tx.us/assets/public/remediation/trrp/350revisions.doc)
(MADEP, 2002). City or other location-specific resources can also be found (as
described in Chapter 3 of this report), such as the City of Chicago Department of
Environment values for "background" PAHs (CCDE, 2003), which have been
adopted by Illinois EPA as indicative of PAH concentrations in Chicago soil (see
http://www.epa.state.il.us/land/site-remediation/urban-area-pah-study.pdf).
Vapor Intrusion (U.S. EPA, Others). Vapor intrusion can be an important
pathway when volatile organic chemicals in subsurface media (soil, groundwater,
and non-aqueous phase liquids) could migrate to air inside a building. Risks
from this pathway are often combined with other exposure pathways for indoor
air (e.g., inhalation of volatiles during showering) to quantify aggregate risks for
single chemicals (e.g., benzene) and cumulative risks for a group of chemicals
(e.g., chlorinated solvents).
A-16
-------
This pathway has been evaluated using a model based on the allometric
equation given in Johnson and Ettinger (1991). That model is a one-dimensional
spreadsheet that estimates convective and diffusive transport of chemical vapors
to indoor air from sources near a building's perimeter. The model ignores
attenuating factors (e.g., biological degradation) and assumes an infinite source
over the exposure duration of the receptor (e.g., 25 years for a commercial or
industrial worker). A detailed description of the vapor intrusion model is provided
in draft Guidance for Evaluating the Vapor Intrusion to Indoor Air Pathway from
Groundwater and Soils (U.S. EPA, 2002h) and the draft User's Guide for
Evaluating Subsurface Vapor Intrusion into Buildings (U.S. EPA, 2003J).
Separate versions of the spreadsheet model are available for evaluating potential
source concentrations (e.g., soil gas or groundwater data).
Both screening-level and advanced versions of the models are available for
each. The screening-level version limits user inputs to the most sensitive
parameters and allows the user to define only a single soil stratum above the
source. The advanced version allows users to enter additional site-specific data
for soil and building parameters and incorporates up to three soil strata for which
soil properties can be varied. In February 2003, the EPA released Version 3.0 of
the vapor intrusion model, which contained updated toxicity values and other
physical/chemical parameters. This model and associated guide are still
undergoing review. Certain state agencies (e.g., California) have modified that
model to include state-sanctioned toxicity values or other model parameters
(DTSC, 2003). Other organizations are also developing approaches (including
other federal agencies).
While the Johnson and Ettinger model is most widely recognized for vapor
intrusion, several states have adopted simple equations based on this
methodology to evaluate the indoor air pathway on a screening level. For
example, the Risk Evaluation/Corrective Action Program (RECAP) of the
Louisiana Department of Environmental Quality (LDEQ) has developed a set of
publicly available spreadsheets that contain equations and chemical-specific
information that can be used to predict conservative concentrations of VOCs in
indoor air for industrial and nonindustrial buildings constructed over groundwater
plumes. Chemical concentration values for multiple chemicals calculated by the
models could be combined to evaluate cumulative exposure.
Fate and Transport/Risk Assessments (U.S. EPA, Others). For risk
assessments at contaminated sites, urban environments and other situations
potentially impacted by multiple sources or sources distant from the population of
concern, it is often necessary to simulate the behavior of multiple chemicals in
different environmental media. Hundreds of computer models have been
developed to model various aspects of horizontal and vertical contaminant fate
and transport in the environment. Some are very general and conceptual, while
others are quite specific to certain media characteristics and applications. The
use and applicability of individual models varies widely depending on the project
objectives and specificity required, so it is important for the model chosen to be
appropriate for the given site setting. For example, the Center for Subsurface
A-17
-------
Modeling Support (CSMoS) within the EPA's Office of Research and
Development (ORD) (located in Ada, Oklahoma) maintains an online database of
public groundwater and vadose zone fate and transport models. This database
is accessible at http://www.epa.gov/ada/csmos.html.
Other tools that support characterization and modeling of the movement and
behavior of chemicals in the environment include the EPA Soil Screening
Guidance (described above), as well as environmental data compiled by many
organizations for specific regions and conditions. Data of interest typically
include soil type (e.g., sand, loam, clay); drainage characteristics, hydraulic
conductivity, depth to groundwater, water quality parameters, organic carbon
content and various other constants and coefficients.
Environmental data are also available through databases maintained by the U.S.
Geological Survey, state natural resources departments, colleges and
universities, U.S. Department of Agriculture (USDA) Natural Resources
Conservation Service (NRCS) field offices (offices in most county seats), USDA
soil surveys (available for most counties at NRCS offices and local libraries),
scientific textbooks and journals, internet resources and professional
organizations. Other organizations have also developed groundwater models
that can be used for cumulative risk assessments (not available through the EPA
website), as indicated in Table A-2.
A-18
-------
TABLE A-2
Selected Resources for Evaluating Fate and Transport
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Soil Screening Guidance
(U.S. EPA)
http://www.epa.gov/superfun
d/resources/soil/introtbd.htm
Provides tools for developing screening levels for, and
conducting, risk assessments involving soil and
groundwater. Useful input parameters and technical
background for environmental models.
Standard constants, coefficients and soil data
that can be useful to cumulative risk
assessments.
SESOIL (SEasonal SOIL
compartment model)
In the public domain,
although updated versions
are available from
RockWare, Inc.
http://www.rockware.com/
SESOIL is a one-dimensional vertical transport screening-
level model for the unsaturated (vadose) zone that can be
used to simulate the fate of contaminants in soil to support
site-specific cleanup objectives. Simulates natural
attenuation based on diffusion, adsorption, volatilization,
biodegradation, cation exchange and hydrolysis. The
model can evaluate one chemical at a time; does not
predict interactions in environmental media.
Results can indicate how far a contaminant
plume will migrate; predicted concentrations can
be compared to media-specific standards and
can be used to estimate single-chemical risks
based on standard default exposure parameters,
locations, and times. The location- and time-
specific predictions for single chemicals can be
overlain to support grouping decisions for a
cumulative assessment.
AT123D
(Analytical Transient 1-, 2-
and 3-Dimensional
simulation of waste transport
in the aquifer system)
http://www.scisoftware.com/
Generalized three-dimensional groundwater transport and
fate model. Transport and fate processes simulated
include advection, dispersion, adsorption and biological
decay. The model can evaluate one chemical at a time;
does not predict interactions in environmental media.
As above.
Summers model
http://www.seview.com/
Screening level leachate program that estimates
groundwater concentrations based on mixing. Simulates
dilution of soil in groundwater. The model can evaluate one
chemical at a time; does not predict interactions in
environmental media.
As above.
A-19
-------
TABLE A-2 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Draft guidance and user's
guide for evaluating vapor
intrusion into buildings (U.S.
EPA);
LDEQ spreadsheets to
screen vapor intrusion
pathway
Provides a model to estimate convective and diffusive
transport of chemical vapors to indoor air. Could offer
insights where indoor air exposures are a concern.
(Currently under review.) LDEQ provides set of equations
that enable screening of the vapor intrusion pathway.
Model output can be used to support cumulative
risk assessments, as concentrations of multiple
chemicals can be evaluated simultaneously.
(The following models are available for download from the CSMoS website, http://www.epa.gov/ada/csmos/models.html.)
2DFATMIC and 3DFATMIC
Simulates subsurface flow, transport, and fate of
contaminants that are undergoing chemical and/or
biological transformations. Applicable to transient
conditions in both saturated and unsaturated zones.
model can evaluate one chemical at a time; does not
predict interactions in environmental media.
The
Results can indicate how far a contaminant plume
will migrate; predicted concentrations can be
compared to media-specific standards and can be
used to estimate single-chemical risks based on
standard default exposure parameters, locations
and times. The location- and time-specific
predictions for single chemicals can be overlain to
support grouping decisions for a cumulative
assessment.
BIOCHLOR
Screening model that simulates remediation by natural
attenuation of dissolved solvents at sites with chlorinated
solvents. Can be used to simulate solute transport without
decay and solute transport with biodegradation modeled as
a sequential first-order process within one or two different
reaction zones. The model can evaluate one chemical at a
time; does not predict interactions in environmental media.
As above.
A-20
-------
TABLE A-2 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
BIOPLUME II and
BIOPLUME III
Two-dimensional contaminant transport under the influence
of oxygen-limited biodegradation (BIOPLUME II) and under
the influence of oxygen, nitrate, iron, sulfate and
methanogenic biodegradation (BIOPLUME III). Models
advection, dispersion, sorption, biodegradation (aerobic
and anaerobic) and reaeration (BIOPLUME II) and through
instantaneous, first order, zero order or Monod kinetics
(BIOPLUME III). BIOPLUME III was developed primarily for
the modeling of natural attenuation of organic contaminants
in groundwater; it is particularly useful at petroleum-
contaminated sites. The model can evaluate one chemical
at a time; does not predict interactions in environmental
media.
As above.
BIOSCREEN
Screening-level groundwater transport model that simulates
natural attenuation of dissolved-phase hydrocarbons.
Based on the Domenico analytical contaminant transport
model and can simulate natural attenuation based on
advection, dispersion, adsorption and biological decay.
Estimates plume migration to evaluate risk at specific
locations and times. The model can evaluate one chemical
at a time; does not predict interactions in environmental
media.
As above.
A-21
-------
TABLE A-2 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
(The following models are available for download from the CSMoS website, http://www.eDa.gov/ada/csmos/models.html, except as indicated.)
CHEMFLO
Simulates one-dimensional water and chemical movement
in the vadose zone. Models advection, dispersion, first-
order decay and linear sorption. The model can evaluate
one chemical at a time; it does not predict interactions in
environmental media.
Results can indicate how far a contaminant plume
will migrate; predicted concentrations can be
compared to media-specific standards and can be
used to estimate single-chemical risks based on
standard default exposure parameters, locations
and times. The location- and time-specific
predictions for single chemicals can be overlain to
support grouping decisions for a cumulative
assessment.
GEOEAS
Enables geostatistical analysis of spatially correlated data.
Can perform basic statistics, scatter plots/linear and
nonlinear estimation (kriging). The model can evaluate one
chemical at a time; it does not predict interactions in
environmental media.
As above.
GEOPACK
Enables geostatistical analysis of spatially correlated data.
Can perform basic statistics, variography, linear and
nonlinear estimation (kriging). The model can evaluate one
chemical at a time; it does not predict interactions in
environmental media.
As above.
A-22
-------
TABLE A-2 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
HSSM
Can simulate: light non-aqueous phase liquid (LNAPL) flow
and transport of a chemical constituent of the LNAPL from
the ground surface to the water table; radial spreading of
the LNAPL phase at the water table; and dissolution and
aquifer transport of the chemical. One-dimensional in the
vadose zone, radial in the capillary fringe, two-dimensional
vertically averaged analytical solution of the advection-
dispersion equation in the saturated zone. The model can
evaluate one chemical at a time; it does not predict
interactions in environmental media.
As above.
Visual MODFLOW (available
for a fee from the developer)
and MODFLOW
(U.S. Geological Survey),
many iterations/updates;
most recent is MODFLOW-
2000
One of the most accessible and widely used models
available. Numerically solves the three-dimensional
ground-water flow equation for a porous medium by using a
finite-difference method. Visual MODFLOW output is
graphic, including two- and three-dimensional maps;
designed to model flow, can evaluate one chemical at a
time (information input by user); it does not predict
interactions in environmental media.
As above.
A-23
-------
TABLE A-2 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
(The first three models below are available for download from the CSMoS website, http://www. epa.gov/ada/csmos/models. html.)
PESTAN
Vadose zone modeling of the transport of organic
pesticides. Models advection, dispersion, first-order decay
and linear sorption. The model can evaluate one chemical
at a time; it does not predict interactions in environmental
media.
Results can indicate how far a contaminant plume
will migrate; predicted concentrations can be
compared to media-specific standards and can be
used to estimate single-chemical risks based on
standard default exposure parameters, locations,
and times. The location- and time-specific
predictions for single chemicals can be overlain to
support grouping decisions for a cumulative
assessment.
Soil Transport and Fate
(STF) Database
Database providing information concerning the behavior of
organic and a few inorganic chemicals in the soil
environment. Focus is on one chemical at a time;
interactions not addressed.
General-use tool can be used to evaluate
environmental contaminants for cumulative risk
assessments.
UTCHEM
Three-dimensional model that simulates non-aqueous
phase liquid (NAPL) movement in the subsurface. Can
address: multiple phases; dissolution and/or mobilization
by non-dilute remedial fluids; chemical and microbiological
transformations; and changes in fluid properties as a site is
remediated.
General-use tool can be used to evaluate
environmental contaminants for cumulative risk
assessments. Interesting for cumulative risk
because NAPL is commonly a complex mixture
itself and can be present in multiple phases, which
are assessed by the model.
MT3D (links to MODFLOW)
http://www.ess.co.at/ECOSI
M/MANUAL/mt3d.html
Three-dimensional transport model for simulating
advection, dispersion and chemical reactions in
groundwater systems; assumes first-order decay. Can
address one chemical at a time.
Chemical reaction can be addressed with a loss
term (information on chemical must be input by
user) but degradation product not tracked.
Heavily dependent on extensive characterization
of site setting (can be hard to get sufficient data
for all parameters needed).
A-24
-------
TABLE A-2 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
SWIFTIII (private)
Three-dimensional flow (transient and steady state) and
solute transport (advection, dispersion, sorption and decay)
in fractured porous media; uses finite difference method;
addresses chemical reactions with second-order decay;
also models radionuclides.
Similar to above, but can address more than one
chemical: parent plus degradation product(s)
(chain of two). (As above, user must input
information about each chemical.)
MULKOM codes, including
TMVOC (and predecessor
T2VOC)
(DOE/Lawrence Berkeley
Laboratory, http://www-
esd.lbl.gov/TOUGH2
Three-dimensional, three-phase flow of water, air, and
volatile organic compounds in saturated and unsaturated
zone to support remediation (e.g., soil vapor extraction).
TMVOC can address more than one volatile organic (e.g.,
to model a spill of fuel hydrocarbons or solvents).
Similar to above, but can address a mixture of
volatile organic compounds. Like the others
models, depends heavily on extensive site setting
characterization (hard to get data needed for all
parameters, for results to be meaningful).
A-25
-------
A.3. RESOURCES FOR EXPOSURE ANALYSES
Many exposure models are well suited to assessing multiple exposures to
multiple chemicals at contaminated sites and other multimedia situations, although this
is generally performed by combining predictions for individual chemicals. Tools range
from relatively straightforward screening models to comprehensive multimedia, multiple-
pathway exposure models, as summarized below and in Table A-3 at the end of this
section. Certain models presented here also support other portions of the risk
assessment process. For example, the model for subsurface vapor migration soil
(Johnson and Ettinger, 1991) is commonly considered an environmental fate and
transport tool, but it can also serve as a multimedia exposure assessment resource
because it considers both soil and groundwater inputs to predict concentrations in
indoor air. Several supporting documents are also available that provide exposure
factors, their bases, and receptor parameters that are used in various exposure models.
Exposure Factors (U.S. EPA). Risk assessments rely on exposure models to
represent various environmental and receptor-specific factors that can affect
exposures to chemicals. For example, exposure factors cover exposure
duration, time involved in certain activities, body weight and surface area, intake
rates (e.g., inhalation, ingestion of food, soil, water), and many others parameters
needed to estimate representative risks. The EPA has summarized extensive
data in a set of exposure factor handbooks based on many studies, which
consider statistical and relative contributions of many potential sources of human
exposures to chemicals in air, drinking water, vapor, food, and soil. These
handbooks include:
Exposure Factors Handbook, Volume I - General Factors (U.S. EPA, 1997c),
seewww.epa.gov/ncea/pdfs/efh/front.pdf.
Exposure Factors Handbook, Volume II - Food Ingestion Factors (U.S. EPA,
1997c), see www.epa.gov/ncea/pdfs/efh/front.pdf.
Exposure Factors Handbook, Volume III - Activity Factors (U.S. EPA, 1997c),
seewww.epa.gov/ncea/pdfs/efh/front.pdf.
Child-Specific Exposure Factors Handbook (Interim Report) (U.S. EPA,
2002i), see http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=55145.
Sociodemographic Data Used for Identifying Potentially Highly Exposed
Populations (U.S. EPA, 1999c), see
http://oaspub.epa.gov/eims/eimscomm.getfile7p download id=428679.
Fact Finder CD-ROM searches data from the Exposure Factors Handbook
and Sociodemographic Data Used for Identifying Potentially Highly Exposed
Populations (referenced above), see
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=23650.
A-26
-------
3MRA Model (U.S. EPA). The 3MRA model is a multimedia, multipathway,
multireceptor exposure and risk assessment model being developed by the EPA to
assess releases from land-based waste management units. After simulating
releases from disposal units, modules model fate and transport through the
environment, estimate exposure to receptors and calculates distributions of risks to
receptors. This screening-level model is intended to be applied on a site-specific
basis to generate risk-based standards (considering exit levels, e.g., to exit from
specific regulations). Risks are assessed at individual sites to provide input to a
representation a national distribution of risks. The national distribution of risks is the
basis for determining waste stream constituent concentrations that meet regulatory
criteria established to be protective of human health and ecological receptors (as
determined by EPA policy). To establish national regulatory limits, site-based risk
results are combined to evaluate national risk (i.e., to determine the percentage of
nationwide receptors that are protected at various levels). For example, from this
information a limit might be established to ensure protection of 95% of all receptors
within 2 miles of a waste management unit at all sites across the nation. The 3MRA
methodology uses a Monte Carlo scheme to quantify uncertainty (e.g., from natural
variability or based on selection of representative sites). The resulting national
criteria would represent threshold waste concentrations not considered hazardous
(and not requiring Subtitle C disposal). The model is available at
http://www.epa.qov/ceampubl/mmedia/3mra/.
Exposure and Fate Assessment Screening (E-FAST) Tool (U.S. EPA). This
computer-based model can provide screening-level estimates of general
population, consumer, and environmental exposures to concentrations of
chemicals released to air, surface water, landfills and from consumer products.
Potential inhalation, dermal and ingestion doses resulting from these releases
are estimated. Modeled concentrations and doses are designed to reasonably
overestimate exposures for use in screening-level assessments. The model is
available from http://www.epa.gov/opptintr/exposure/docs/efast.htm.
Lead Exposure (U.S. EPA). The traditional reference dose approach used to
estimate health risks does not apply to lead because most human health effects
data are based on blood lead concentrations rather than external dose. Blood
lead concentration is an integrated measure of internal dose, reflecting total
exposure from all sources (e.g., both site-related and background sources for
Superfund sites) (ATSDR, 1999a). Both the EPA and the California EPA
Department of Toxic Substances Control (CalEPA DTSC) have developed
models to estimate blood lead concentrations from exposures to lead from
various media, including soil, water, air and food. The EPA tool for evaluating
lead risks (the All Ages Lead Model) (U.S. EPA, 2005f) predicts lead
concentrations in body tissue and organs for a hypothetical individual based on a
simulated lifetime of lead exposure, and then extrapolates to a population of
similarly exposed individuals.
The EPA has also developed a set of models for evaluating lead exposures and
risks for non-residential adults. The models and supporting literature,
methodologies and technical information for these analyses are available at
A-27
-------
http://www.epa.gov/superfund/programs/lead/products.htm. Documents on the
website include descriptions of how bioavailability and uptake factors for the adult
lead model were determined. Examples of useful support documents also
available from the EPA include Revised Interim Soil Lead Guidance for CERCLA
Sites and RCRA Corrective Action Facilities (U.S. EPA, 1994a) and Frequently
Asked Questions on the Adult Lead Model (U.S. EPA, 1999g).
The National Human Exposure Assessment Survey (NHEXAS) (U.S. EPA).
NHEXAS was developed by the EPA's Office of Research and Development
(ORD) in the early 1990s to provide critical information about multipathway,
multimedia population exposure distribution to chemical classes. The first phase
consisted of three pilot studies with the objectives of: evaluating the feasibility of
NHEXAS concepts, methods and approaches for the conduct of future
population-based exposure studies; evaluating the utility of NHEXAS data for
improved risk assessment and management decisions; testing the hypothesis
that the distributions of exposure given by modeling and extant data do not differ
from the measurement-based distributions of exposure; defining the distribution
of multipathway human exposures for a relatively large geographic area; and
stimulating exposure research and forging strong working relationships between
government and nongovernment scientists. The NHEXAS web site is located at
http://www.epa.gov/nerl/research/nhexas/nhexas.htm. NHEXAS data are
available in the Human Exposure Database System (HEDS) at
http://www.epa.gov/heds/.
Hotspots Analysis and Reporting Program (HARP) Tool (California Air
Resources Board, CARB). The State of California's Air Toxics "Hot Spots"
program requires stationary air emission sources within the state to report the
types and quantities of certain substances routinely release into the air. The
recent HARP software package is designed to create and manage facility
emissions inventory databases; prioritize facilities; model atmospheric dispersion
of chemicals from one or multiple facilities using EPA models ISCST3 and BPIP;
calculate cancer and noncancer (acute and chronic) health impacts using
guidance developed by CalEPA (in 2003); use point estimates or data
distributions of exposures to calculate inhalation and multipathway risks; perform
stochastic health risk analyses; calculate potential health effects for individual
receptors, population exposures, cumulative impacts for one or multiple facilities
and one or multiple pollutants, and potential health effects using ground-level
concentrations; and present results as tables and isopleth maps. The results can
be printed, added to word processing documents, or input to a Geographic
Information Systems (CIS) program. The HARP model can be downloaded from
http://www.arb.ca. gov/toxics/harp/downloads.htm#2.
Dietary Exposure Potential Model (DEPM) (U.S. EPA). The DEPM estimates
dietary exposure to multiple chemicals based on data from several national,
government-sponsored food intake surveys and chemical residue monitoring
programs. The DEPM includes recipes developed specifically for exposure
analyses that link consumption survey data for prepared foods to the chemical
residue information, which is normally reported for raw food ingredients, to
A-28
-------
estimate daily dietary exposure. Consumption in the model is based on 11 food
groups containing approximately 800 exposure core food types, established from
over 6500 common food items. The summary databases are aggregated in a
way that allows the analyst to select appropriate demographic factors, such
age/sex groups, geographical regions, ethnic groups and economic status. The
model also includes modules for evaluating chemical exposures from residues,
soil, and tap water. The model is available from the EPA's National Exposure
Research Laboratory (NERL) at http://www.epa.gov/nerlcwww/depm.htm.
Health Registries (Centers for Disease Control and Prevention, CDC;
Others). Several organizations maintain databases that contain information on
the frequencies and types of diseases and other health-related information, such
as on cancer, asthma, and birth defects, and blood lead levels. This information
can be evaluated in concert with modeled or measured chemical exposure data
to correlate potential influences of multiple exposures and to calibrate risk
models. For example, the CDC maintains a national registry of cancer cases,
including cancer type and target tissue, as well as demographic and location
information.
Many states have established cancer and other disease registries to monitor
trends over time; determine patterns in various populations; guide planning and
evaluation of control programs; help set priorities for allocating health resources;
advance clinical, epidemiologic, and health services research; and provide
information for a national database of cancer incidence. The National Cancer
Registry is searchable online http://www.cdc.gov/cancer/natlcancerdata.htm.
The CDC website also contains links to various state registries. Other resources
that can be useful for identifying populations at potential risk include the
U.S. Census Bureau (http://www.census.gov/), state and local government health
departments and other health organizations. An additional useful resource is the
report Sociodemographic Data Used for Identifying Potentially Highly Exposed
Populations (U.S. EPA, 1999c).
National Occupational Research Agenda (NORA) (National Institute for
Occupational Safety and Health, NIOSH). Within NIOSH, NORA has identified
a number of research areas for mixed occupational exposures, with an aim to
protect individuals in the workplace from exposures to multiple chemicals. The
mixed exposures team website
(http://www2a.cdc.gov/nora/noratopictemp.asp?rscharea=me) provides links to
current and past studies, as well as information on how to join a listserv group to
discuss topics related to mixed exposures. Scientific knowledge developed
through this effort can offer insights for assessing combined the effects of
chemicals at contaminated sites, occupational settings and other scenarios
involving multiple chemicals.
Tool for the Reduction and Assessment of Chemical and Other Impacts
(TRACI) (U.S. EPA). TRACI is an impact assessment tool for assessing multiple
chemical impact and resource-use categories to analyze various study designs.
Impacts that can be modeled include: ozone depletion; global warming;
acidification; eutrophication; photochemical smog; cancer risk and noncancer
A-29
-------
health effects; human health criteria; ecotoxicity; fossil fuel use; land use; and
water use. The program includes quantitative data on human carcinogenicity
and noncarcinogenicity (based on human toxicity potentials), acidification, smog
formation and eutrophication. The model uses a probabilistic approach to
determine spatial scale(s) for other impact categories such as acidification, smog
formation, eutrophication and land use. Information is available at
http://www.epa.gov/ordntrnt/ORD/NRMRL/pubs/600r02052/600r02052.htm .
Technology Transfer Network, TTN (U.S. EPA). This is an on-line information
resource for tools to support air pathway analyses. The TTN maintains a
Clearinghouse for Inventories and Emission Factors (CHIEF) website
(http://www.epa.gov/ttn/chief/) that contains links to many of the relevant
documents on methods and data for constructing emissions inventories available
for download, including the Handbook for Criteria Pollutant Inventory
Development: A Beginner's Guide for Point and Area Sources (U.S. EPA,
1999h); Handbook for Air Toxics Emission Inventory Development, Volume I:
Stationary Sources (U.S. EPA, 1998e); and Compilation of Air Pollutant Emission
Factors (U.S. EPA, 1995cet seq.). The EPA also maintains a Support Center for
Regulatory Air Models (SCRAM) website (http://www.epa.gov/ttn/scram/), which
provides information on codes described in the Guideline on Air Quality Models
(U.S. EPA, 2003d) and includes downloadable models and guidance.
Information from TTN is included in the discussion of the air pathway in
Section 4.4 of this report.
A-30
-------
TABLE A-3
Selected Resources for Evaluating Exposure
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Exposure Factors
Guidance (U.S. EPA)
general:
http://www.epa.gov/ncea/
pdfs/efh/front.pdf
child:
http://cfpub.epa.gov/ncea/
cfm/recordisplay.cfm?deid
=55145
Provides extensive values and underlying bases for many
factors that affect exposures. Examples include exposure
duration, frequency, surface area, inhalation rates per activity
level and age/gender, as well as ingestion rates, including for
incidental soil ingestion and by food type, based on age and
gender. Because children are often more heavily exposed to
environmental toxicants than adults, the EPA also published
the Child-Specific Exposure Factors Handbook is to provide a
summary of the available and up-to-date statistical data on
various factors assessing children exposures.
Excellent compendium of values for
exposure parameters that can be
reviewed to determine those most
appropriate for a given site/setting (for
both adults and children). Can be
used to assess multiple pathways and
activities/intake rates associated with
multiple chemicals.
Sociodemographic Data
Used for Identifying
Potentially Highly
Exposed Populations
(U.S. EPA)
http://cfpub.epa.gov/ncea/
cfm/recordisplay.cfm?deid
=22562
Fact Finder searches and returns data from the
Sociodemographic Data Used for Identifying Potentially Highly
Exposed Populations document. These data assist analysts in
identifying and enumerating potentially highly exposed
populations. Due to unique social and demographic
characteristics, various segments of the population may
experience exposures different from those of the general
population, which in many cases could be higher. It is helpful
for risk or exposure analysts evaluating a diverse population to
first identify and then characterize certain groups within the
general population who could be at risk for greater
contaminant exposures (and related effects).
This document presents data relating
to factors which potentially impact an
individual or group's exposure to
environmental contaminants based on
various activity patterns, different
microenvironments, and other Socio-
demographic data such as age,
gender, race and economic status.
Populations potentially more exposed
to multiple chemicals of concern,
relative to the general population, is
also addressed in this database.
A-31
-------
TABLE A-3 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
3MRA (U.S. EPA)
http://www.epa.gov/ceam
publ/mmedia/3mra/index.
htm
(CEAM)
Developed for screening-level assessment of potential human
and ecological health risks from chronic exposures to
chemicals released from land-based waste management units
containing listed waste streams. Site-based and intended for
national-scale application to generate risk-based standards
(e.g., levels to exit from hazardous waste regulation),
evaluates human and ecological receptors and captures
uncertainty and variability in risk estimates. (Ecological
exposure and risk focuses on population effects related to key
species within habitats found in the proximity of sites.)
Can quantify exposure via multiple
pathways after a simulated release.
Human receptors include adult/child
residents, home gardeners, beef and
dairy farmers, and recreational fishers.
Pathways include inhalation of
outdoor air and indoor air during
showering, ingestion of drinking water
and ingestion of farming products and
fish.
E-FAST (U.S. EPA)
http://www.epa.gov/opptin
tr/exposure/docs/efast. ht
m
Provides screening-level estimates for general population,
consumer, and environmental exposures to concentrations of
chemicals released to air, surface water, landfills and from
consumer products. Modeled estimates of concentrations and
doses are designed to reasonably overestimate exposures, for
use in screening-level assessments.
Default exposure parameters are
available, but site-specific values are
recommended to be used. Can
predict exposure concentrations for
comparison to media-specific
standards.
All Ages Lead Model
(U.S. EPA):
http://cfpub.epa.gov/ncea/
cfm/recordisplay.cfm?deid
=139314
Predicts lead concentrations in body tissue and organs for a
hypothetical individual based on a simulated lifetime of lead
exposure, and then extrapolates to a population of similarly
exposed individuals.
Useful for evaluating the impact of
possible sources of lead in a specific
human setting where there is a
concern for potential or real exposures
to lead. The results can be correlated
with risks from other contaminants, if
interactions with lead are known to
occur.
A-32
-------
TABLE A-3 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
CALTOX Model
(CalEPA)
Spreadsheet-based model that relates the concentration of a
chemical in soil to the risk of an adverse health effect for a
person living or working on or near a site. Determines
chemical concentration in the exposure media of breathing
zone air, drinking water, food and soil that people inhale,
ingest and contact dermally, and uses the standard equations
found in EPA's RAGS (U.S. EPA, 1989a) to estimate exposure
and risk.
Can be used to assess multiple
exposures; has tended to be more for
research than practical applications.
Defaults are available but site-specific
values are recommended. Can
predict exposure concentrations that
can be compared to media-specific
standards and used to estimate
single-chemical risks.
Dietary Exposure
Potential Model (DPEM)
(U.S. EPA)
http://www.epa.gov/nerlcw
ww/depm.htm
The DEPM estimates dietary exposures to multiple chemicals
based on data from several national, government-sponsored
food intake surveys and chemical residue monitoring
programs.
Can be used to assess exposures to
multiple chemicals by ingestion of
food and tap water, including as
potential context for ambient
exposures in the area of a site.
Disease registries
(multiple organizations,
including CDC:)
http://www.cdc.gov/cancer
/natlcancerdata. htm
A number of databases exist for cancer and other health-
related information, such as asthma and birth defects.
Data could be used to indicate key
community health concerns or for
exploratory investigation of certain
diseases that might increase the
vulnerability of certain people exposed
to chemicals from a contaminated site.
However, the links to diseases from
environmental exposures or directly to
environmental pollutants as a causal
or contributing factor is not usually
clear.
A-33
-------
TABLE A-3 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Tool for the Reduction
and Assessment of
Chemical and Other
Impacts (TRACI) (U.S.
EPA)
http://www.epa.gov/ordntr
nt/ORD/NRMRL/pubs/600
r02052/600r02052.htm
TRACI is an impact assessment tool for evaluating multiple
chemical impact and resource-use categories so various study
designs can be analyzed.
Can be used to model and compare
exposures to multiple chemicals and
health risks associated with different
projects. For example, can
graphically analyze the reduction in
risk projected from one
implementation design versus
another.
NORA Mixed Exposures
Team (NIOSH)
http://www2a.cdc.gov/nor
a/noratopictemp.asp?rsch
area=me
Provides technical and support information on projects
involving mixed exposures in the workplace. Research
reflected on the website could provide insights for cumulative
risk assessment projects.
Information resource for mixtures in
the workplace; can offer insights for
cumulative assessments at
contaminated sites.
A-34
-------
A.4. RESOU RCES FOR TOXICITY ANALYSES
Resources that can be used to support toxicity analyses for cumulative risk
assessments are highlighted below and summarized in Table A-4. Topics include:
(1) development of toxicity factors, including for whole mixtures; (2) identification of
toxicity criteria for similar or surrogate compounds or mixtures to represent a mixture or
its components and (3) joint toxicity of the components of a mixture.
Integrated Risk Information System, IRIS (U.S. EPA). The IRIS database is a
key source of information on chronic toxicity, including standard toxicity values
(reference doses and concentrations), cancer slope factors and corresponding
risk-based concentrations. These values have undergone a thorough review
process including EPA internal review, expert scientific external review and
public review. They represent expert EPA consensus, and they are widely used
within the United States and by other countries. Toxicity values and target tissue
information included in IRIS summaries can be used in a cumulative risk
assessment to identify chemicals that primarily or secondarily affect similar target
tissues or systems. Chemical interactions other than addition are not quantifiable
using toxicity criteria from IRIS; however, information in the accompanying study
summaries can be used to qualitatively assess the nature and magnitude of
certain interactions, and the primary literature can be further pursued for
additional information. Toxicity criteria are presented in a way that supports
addition (the default approach) to estimate risks and the potential noncancer
effects of chemicals. This information is available at http://www.epa.qov/iris/.
Toxicological Profiles and Interaction Profiles (ATSDR). The ATSDR, within
the U.S. Centers for Disease Control and Prevention (CDC), has developed
toxicological profiles for many individual chemicals that summarize information
about sources and uses as well as key data from the scientific literature
regarding toxicity and behavior and levels in the environment. These profiles can
be valuable for cumulative risk assessments because they describe in detail the
effects of the given chemical, as well as its primary environmental and metabolic
transformation products, on specific target organs and biological functions. In
addition, where possible, the toxicological profiles discuss known interactions of
the topic chemical with other chemicals. These profiles are available at
http://www.atsdr.cdc.gov/toxpro2.html.
The ATSDR has also developed a mixtures program and has drafted a guidance
manual that presents an assessment approach, and perhaps more importantly
has drafted nine interaction profiles for seven specific chemical combinations and
two general mixtures. The specific chemical combinations are: (1) arsenic,
cadmium, chromium, and lead; (2) benzene, toluene, ethylbenzene, and xylene;
(3) lead, manganese, zinc, and copper; (4) cyanide, fluoride, nitrate, and
uranium; (5) cesium, cobalt, PCBs, strontium, and trichloroethylene;
(6) 1,1,1-trichloroethane, 1,1-dichloroethane, trichloroethylene, and
tetrachloroethylene and (7) arsenic, hydrazine, jet fuels, strontium-90, and
trichloroethylene. These interaction profiles evaluate data on the toxicology of
the whole mixture where available, and where not available data are evaluated
A-35
-------
for the joint toxicity of chemicals in the mixture (often as pairs). These drafts are
available at http://www.atsdr.cdc.gov/iphome.html.
Supplementary Guidance for Conducting Health Risk Assessment of
Chemical Mixtures (U.S. EPA). This guidance published in summer 2000
updates the EPA's 1986 guidelines for chemical mixtures (U.S. EPA, 2000a). It
describes approaches that depend on the type, nature, and quality of available
data. The report includes equations, definitions, discussions of toxicologic
interactions and pharmacokinetic models and approaches for assessing whole
mixtures, surrogate mixtures and individual mixture components. The whole-
mixture discussion includes the whole-mixture reference dose (RfD) and
concentration (RfC) and slope factors; comparative potency; and environmental
transformations. The component discussion includes the hazard index (HI);
interaction-based HI; relative potency factors (RPF); and response addition.
Toxicity criteria are presented for several common product mixtures, such as
polychlorinated biphenyls (PCBs). This guidance is available at
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=20533.
Database for Airborne Workplace Chemicals (Institut de Recherche Robert-
Sauve en Sante et en Securite du Travail, IRSST). This health and safety
research institute in Quebec, Canada, has developed a database that covers a
large number of chemicals commonly found in the workplace, and also found at
many contaminated sites. This database contains information on occupational
standards, chemical-specific health effects, target organs (and chemical-specific
groupings), toxicokinetics, effect levels and mode of action where available. The
database also includes a calculation tool that allows up to 10 chemicals to be
assessed at a time, comparing the concentration of interest to the occupational
standard (many are similar to ours) to produce a sum of ratios, using an additivity
default (IRRST, 2003).
Relative Potency Factors for Pesticide Mixtures, Biostatistical Analyses of
Joint Dose Response (U.S. EPA). In response to requirements of the Food
Quality Protection Act of 1996, the EPA recently published a technical report that
presents research and methodologies for developing relative potency factors by
which cumulative risks from exposures to mixtures such as organophosphate
pesticides, dioxins, and PCBs can be assessed (U.S. EPA, 2003f). The
document presents three scenarios for which biostatistical methods for toxicity
assessment can be accomplished, including use of dose addition in simple cases
where common modes of toxicity are present, integration of dose and response
addition for cases where toxicities are independent and joint dose-response
modeling for cases where the mode of action is uncertain. The report, published
by NCEA in coordination with OPP, is available at
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=66273.
Cumulative Risk of Pesticides with Common Toxic Mechanism (U.S. EPA).
In response to the Food Quality Protection Act, the EPA Office of Pesticide
Programs (OPP) recently released an assessment of the risks associated with
cumulative exposures to various formulations of organophosphate (OP)
pesticides (U.S. EPA, 2002a). This report updated the preliminary assessment
A-36
-------
released a year earlier. For this assessment, the EPA evaluated potential
exposures to 30 OPs, including via food, drinking water and residential uses and
applied methodologies to account for variability in exposures based on age,
seasonal, and geographic factors. The cumulative risk assessment report is
available at http://www.epa.gov/pesticides/cumulative/rra-op/.
Dose Addition for Cumulative Risks from Exposures to Multiple Chemicals
(U.S. EPA). As part of the response to the Food Quality Protection Act of 1996,
which requires consideration of cumulative risk from exposures to multiple
chemicals that have a common mechanism of toxicity, NCEA published a paper
describing three dose addition-based techniques that can be used to estimate
cumulative risk (Chen et al., 2001). The three methods include the hazard index
(HI), point-of-departure index (PODI) and toxicity equivalence factor (TEF), all of
which are based on estimates of a point of departure (as the effective dose for a
10 percent response, or ED10) and reference doses of individual chemicals. A
formal statistical procedure is also proposed to estimate cumulative risk by fitting
the dose-response model of the mixture under dose addition and estimating
relative potency between two chemicals from that model.
Long-Range Research Initiative, LRI (American Chemistry Council, ACC).
Through its LRI program, the ACC sponsors scientific research aimed at better
understanding the potential impacts of chemicals on human health and the
environment, including wildlife (ACS, 2003). Cumulative risk is a priority
research area within the LRI program, and studies are ongoing. Reports and
papers prepared from this research can provide insights for cumulative risk
assessments at contaminated sites. Research topics include improved methods
for understanding toxicodynamics, applications of physiologically-based
pharmacokinetic (PBPK) models to predict target tissue dose and response, and
exposure assessment of mixtures. The LRI holds a conference each year at
which ongoing and completed research is presented. The summary report of the
recent annual conference, with abstracts of research projects presented, can be
found at http://www.uslri.com/.
Chemical Mixtures Toxicology Studies (Netherlands, TNO). International
research is currently underway to improve the understanding of potential risks of
chemical mixtures with different modes of action. For example, a team led by Dr.
John Groten of the TNO Nutrition and Food Research Institute of the Netherlands
is researching the use of mechanistic models to describe interactions between
mixture components expected to act by different modes of action. In an ongoing
pilot study (funded by ACC/LRI), the TNO team is using PBPK models to assess
possible toxicokinetic interactions between compounds in an applied mixture,
and comparing them to empirical dose-response modeling of observed
pathological changes in liver, blood and kidney. The aim is to apply the method
developed to other chemical mixtures. Other studies have developed and
applied statistical experiments combining multivariate data analysis and modeling
in in vitro and in vivo studies on various chemical mixtures such as petroleum
hydrocarbons, aldehydes, food contaminants, industrial solvents and mycotoxins
(Feronetal., 1998).
A-37
-------
Scientific Studies on Toxicology/Mixtures (National Institute of Environmental
Health Sciences, NIEHS). Research areas of the NIEHS, within the National
Institutes of Health (NIH), U.S. Department of Health and Human Services (DHHS),
include toxicology, mixtures and environmental health. The Institute sponsors the
National Toxicology Program (NTP), which coordinates toxicological testing
programs; strengthens the science base in toxicology; develops and validates
improved testing methods; and provides information about potentially toxic
chemicals to health regulatory and research agencies, scientific and medical
communities, and the public. Fact sheets and reports on chemicals and related
risks, and data and findings from NTP-related studies are available at
http://www.niehs.nih.gov/. This website also links to other research projects and
programs within the organization and summaries of past and ongoing studies that
can provide insights for cumulative risk assessments at contaminated sites. A
search engine on the website can be used to identify research and tools for specific
applications, including those related to cumulative risk. NIEHS also publishes
Environmental Health Perspectives, a monthly journal that often summarizes
research papers relevant to chemical mixtures, and some issues and supplements
have been entirely dedicated to mixtures. Also, NIH maintains the National Library
of Medicine Toxic Substances Data Bank and other valuable databases and
biomedical links.
Toxic Substances Research Initiative, TSRI (Health Canada). The Canadian
environmental health department (Health Canada) has developed a program called
the Toxic Substances Research Initiative (TSRI). The primary focus of this initiative
is assessment of cumulative effects to human and ecological receptors. To date,
TSRI has spent more $7 million to fund 23 research projects in this priority research
area. Resulting technical reports and other publications are available at
http://www.hc-sc.gc.ca/ahc-asc/media/nr-cp/2000/2000 69bk2 e.html. One
example research study is the evaluation of the pharmacokinetics and cumulative
health effects of mixtures of disinfection byproducts, led by Dr. Kannan Krishnan of
the University of Montreal.
Toxicity Values for Diesel Particulate Matter (DPM) Mixture (California EPA).
Risks of whole mixtures are evaluated using toxicity criteria developed for that
mixture where data are available. In 1998, the CalEPA Office of Environmental
Health Hazard Assessment (OEHHA) completed a 10-year human health
assessment of the mix of chemicals in diesel exhaust. From the results the
California Air Resources Board (GARB) identified diesel particulate matter (DPM)
exhaust as a toxic air contaminant (TAG) that poses a threat to human health.
This exhaust results from combustion of diesel fuel in internal combustion
engines. Its composition varies based on engine type, operating conditions, fuel
composition, lubricating oil and whether an emission control system is present.
The DPM exhaust is a complex mixture of thousands of fine particles, commonly
known as soot; this contains 47 compounds classified by the EPA as hazardous
air pollutants and by GARB as TACs. These compounds include many known or
suspected carcinogens, such as benzene, arsenic, formaldehyde and nickel.
A-38
-------
The GARB evaluation exhaust takes
into account its individual
components; chemicals commonly
found in diesel exhaust are shown in
Text Box A-1.
The report prepared from the GARB
assessment Proposed Identification of
Diesel Exhaust as a Toxic Air
Contaminant was formally reviewed
and approved by a scientific review
panel. The panel deemed data from
human epidemiological studies of
occupationally exposed populations to
be applicable for quantitative risk
assessment. After considering the
results of the meta-analysis of human
studies, as well as the detailed
analysis of railroad workers, the panel
developed a unit risk estimate
expressed in terms of diesel
particulates, which was then used to
derive an inhalation slope factor of
1.1 (mg/kg-day)"1. This type of
approach might offer useful insights
not only for assessments involving
diesel exhaust but also for
assessments at sites with other
chemical mixtures.
Toxicity/Risk Technical Resource
(U.S. EPA National Center for
Environmental Assessment,
NCEA). As a major research center
within the EPA Office of Research and
Development (ORD), NCEA serves as
the EPA's national resource for
human health and ecological risk
assessment. The Center conducts
risk assessments as well as research
to improve the state-of-the-science,
and also provides guidance and
technical support to analysts. This
organization manages and is
responsible for updating the content of
the IRIS database (U.S. EPA, 2007). Analysts can contact NCEA for help when
toxicity values are not available in IRIS. Information available online at
http://cfpub.epa.gov/ncea/ can offer useful insights for cumulative risk
Toxic Air Contaminants in Diesel Exhaust*
(Text Box A-1)
Acetaldehyde
Acrolein
Aluminum
Ammonia
Aniline
Antimony compounds
Arsenic
Barium
Benzene
Beryllium compounds
Biphenyl
Bis [2-ethylhexyl]phthalate
Bromine
1,3-Butadiene
Cadmium
Chlorinated dioxins
Chlorine
Chlorobenzene
Chromium
Cobalt compounds
Copper
Cresol
Cyanide compounds
Dibenzofuran
Dibutylphthalate
Ethyl benzene
Formaldehyde
Hexane
Lead compounds
Manganese compounds
Mercury compounds
Methanol
Methyl ethyl ketone
Naphthalene
Nickel compounds
4-Nitrobiphenyl
Phenol
Phosphorus
Polycyclic aromatic hydrocarbons
Propionaldehyde
Selenium compounds
Silver
Styrene
Sulfuric acid
Toluene
Xylene isomers and mixtures
Zinc
These have either been identified in diesel exhaust
or are presumed to be in the exhaust based on
observed chemical reactions and/or their presence
in the fuel or oil. Additional information at
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm7deid
=29060.
A-39
-------
assessments. Ongoing research is being conducted by NCEA in the
development of PBPK models for use in risk assessments, the evaluation of
different risk assessment approaches, the modified hazard index approach for
chemical mixtures assessments and the significance of indirect exposure
pathways and quantitative models of variability for assessing uncertainty.
Statistical/Computer Tools in Development (Universities, Research
Institutes). Statistically based methods and computer tools that can model
interactions and effects associated with multiple chemicals are being developed.
A main area of study involves applying physiologically based pharmacokinetic/
pharmacodynamic (PB-PK/PD) models to chemical mixtures. Many researchers
are working in this area (e.g., M. Anderson, K. Krishnan, and R. Yang), and
advances continue to be made. An example of a computer-based approach for
predicting toxicological interactions of chemical mixtures is reaction network
modeling, which has been to model complex chemical processes in petroleum
engineering. For this effort, reaction network modeling incorporates various
statistical methods (including Monte Carlo-type analysis) to predict chemical
reaction rates, products, and outcomes. A molecular-based model (BioMOL) is
in development, which uses this reaction network modeling approach to predict
effects of chemicals in complex biological systems (Liao et al., 2002).
BMDS (U.S. EPA). This software was developed by the EPA to perform fitting of
mathematical models to toxicological dose-response data for a particular toxic
effect (U.S. EPA, 1995c). The user evaluates the results to select a benchmark
dose (BMD) that is associated with a predetermined benchmark response
(BMR), such as a10% increase in the incidence of a particular lesion or a 10%
decrease in body weight gain. A goal of the BMD approach is to define a starting
point of departure for the computation of a reference value (RfD or RfC) or slope
factor that is more independent of study design than the traditional method that
uses a single experimental dose, such as the no-observed-adverse-effect level
(NOAEL). The hazard index uses RfDs or RfCs in a dose addition formula to
scale the exposure levels in a mixture, producing an indicator of the extent of
concern for toxicity. The BMD values used with dose addition could allow
estimation of a BMD for the mixture, allowing the mixture dose to be interpreted
in terms of the risk of a particular effect.
CatReg (U.S. EPA). This categorical regression tool was developed by the EPA
to conduct meta-analyses of toxicological data, i.e., to analyze data or results
from multiple studies including to assess different severity levels. The tool is a
customized software package that runs under S-PLUS (MathSoft, Inc.), and a
free version written in R is under development. Additional context is offered as
follows (from U.S. EPA, 2000c): "Meta-analysis becomes valuable when
individual experiments are too narrow to address broad concerns. For example,
in acute inhalation risk assessment, it is important to investigate the combined
effects of concentration and duration of exposure but few published experiments
vary both the concentration and the duration of exposure. By combining
information from multiple studies, the contribution of both concentration and
duration to toxicity can be estimated. Moreover, the combined analysis allows
A-40
-------
the analyst to investigate variation among experiments, an important benchmark
for the level of model uncertainty." For cumulative health risk assessments,
CatReg can be applied to evaluate grouped chemicals considering multiple
effects and multiple routes. Therefore, this tool can also be used to support
toxicity values.
Risk-Based Screening Levels (U.S. EPA). Risk-based screening criteria have
been developed for environmental media (including soil, drinking water, and air)
by several organizations. For example, EPA Regions 3, 6, and 9 have
developed risk-based concentrations (RBCs), medium-specific screening levels
(MSSLs), and preliminary remediation goals (PRGs), respectively. These
screening values are based on very conservative default assumptions for
exposure and environmental parameters and incorporate toxicity values for
cancer and non-cancer effects from IRIS and other EPA documents (e.g., the old
Health Effects Assessment Summary Tables (HEAST), which have not been
updated since 1997). EPA's Office of Air Quality Planning and Standards has
compiled long and short term inhalation and oral toxicity values from several data
sources (e.g., values from ATSDR, IRIS and California EPA) that can be found at
http://www.epa.gov/ttn/atw/toxsource/summary.html. Information for the MSSLs
is presented in technical guidance (U.S. EPA, 2005c) and can be found at
http://www.epa.gov/earth1r6/6pd/rcra c/pd-n/r6screenbackground.pdf. The
PRGs developed from the guidance (U.S. EPA, 2002g) can be found at
http://www.epa.gov/region09/waste/sfund/prg/files/02userguide.pdf. The RBCs
are described in a technical memorandum (U.S. EPA, 2003i) and can be found at
http://www.epa.gov/reg3hwmd/risk/human/info/cover.htm. These screening
criteria can be used to narrow the focus of the assessment to those chemicals of
potential concern likely to contribute the most to overall risks associated with the
site. However, the screening values do not reflect site-specific exposure routes
and are of limited usefulness for site-specific cumulative risk assessments
because they do not consider relevant setting and exposure information.
A-41
-------
TABLE A-4
Selected Resources for Evaluating Joint Toxicity
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Integrated Risk Information
System (IRIS) (U.S. EPA)
http://www.epa.gov/iris
An electronic database containing information on human
health effects that may result from exposure to various
chemicals in the environment. Describes toxic effects, dose
concentrations and reference inhalation dose concentrations
for oral and inhalation exposures of over 500 chemicals.
Good resource for identifying individual toxicological effects
for an extensive list of chemicals. Combined with specific
exposure information, the data in IRIS can be used for
characterization of the health risks of a given chemical in a
given situation and provide toxic effects of a particular
chemical within a chemical mixture.
Toxicity values and target organ information
included in IRIS summaries can be used in
cumulative risk assessments to identify
chemicals that primarily or secondarily affect
similar target tissues or systems. Chemical
interactions other than addition are not
quantifiable using these toxicity criteria;
however, the nature(s) and magnitudes of some
interactions could be predicted. Toxicity criteria
are calibrated such that health effects and
cancer risks can be readily summed where
effects are assumed to be additive.
Technical resource
(U.S. EPA)
http://www.epa.gov/ncea
NCEA is a technical resource for many topics relevant to
cumulative assessments. These EPA scientists provide
guidance and support to analysts across a broad scope of
assessment issues, including cumulative health risk.
Serves as a source of single chemical and
chemical mixture toxicity assessments and risk
assessment methods development.
Interaction profiles (draft)
(ATSDR)
http://www.atsdr.cdc.gov/ipho
me.html
These interaction profiles summarize available toxicity data
for mixtures and assesses joint toxicity. Drafts exist for nine
combinations (see accompanying text). Information
includes critical effect levels and directions of interactions
with confidence indicators by organ/system, and also
includes representative chemicals.
Useful for assessing cumulative risks when
exposures involve chemicals covered in the
profiles. Good resource for finding specific
toxicity data organized by organ/system to
determine at what levels joint toxicity could be
exerted among chemical sets without having to
search in the primary literature. Some
secondary effects information is included.
A-42
-------
TABLE A-4 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Supplementary Guidance for
Conducting Health Risk
Assessment of Chemical
Mixtures (U.S. EPA)
http://cfpub.epa.gov/ncea/cfm
/record isplay.cfm?deid=2053
3
This guidance presents approaches for assessing risks of
mixtures, as dictated by the nature and quality of available
data (e.g., for mixtures, surrogate mixtures or individual
mixture components). Provides formulas, definitions and
discussions of toxic interactions and pharmacokinetic
models. (Does not address exposures, just toxicity.)
Presents more detailed information on
considerations and calculational approaches for
assessing mixtures, going beyond the
summaries included in Chapters 5 and 6 of this
report.
TOXNET, other databases
(NIH)
http://toxnet.nlm.nih.gov/cgi-
bin/sis/htmlgen?HSDB
NIH sponsors many databases for toxicology and
environmental health, including TOXNET and Haz-Map
(hazardous chemicals and occupational disease), and
MEDLINE links to biomedical journals.
Useful source of single-chemical information,
will also reflect emerging data relevant to
cumulative risks as they are developed.
Chemical database
(IRSST)
http://www.irsst.gc.ca/fr/ outil
100015.html
Database for airborne chemicals in the workplace that
includes the Canadian occupational standards (many are
the same as U.S. standards) and identifies target organs,
effect levels from toxicity studies, and, where available,
mode of action information; includes a sum-of-ratios tool to
assess airborne chemicals compared to standards, for up to
10 at a time. (The database is in French; it is currently
being translated to English.)
Good source of useful inhalation toxicity
information for a large number of chemicals.
The tool can be used to organize chemicals by
target organ/effect and levels can be ratioed to
a reference level (occupational standard), with
an option for calculating a sum of ratios for
10 chemicals at a time (assumes additivity) for a
combined estimate.
Revised Cumulative Risk
Assessment of Pesticides
That Have a Common
Mechanism of Toxicity (U.S.
EPA)
http://www.epa.gov/pesticides
/cumulative/rra-op
Identifies methods, review toxicities, develop relative
potency factors and present risks associated with
cumulative exposures to organophosphate pesticides.
Document reviewed toxicity, product and exposure data for
30 organophosphate and presented detailed findings on
cumulative risks.
One of the first comprehensive risk
assessments addressing cumulative risk; offers
good insights for multipathway assessments.
A-43
-------
TABLE A-4 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Studies within Long-Range
Research Initiative (LRI)
(ACC)
http://www.uslri.org/
Industry-funded scientific program includes a cumulative risk
focus area. Ongoing research in this area is addressing
assessment methods and toxicity studies for mixtures.
Research results could offer insights for
cumulative risk assessments at contaminated
sites.
BMDS (U.S. EPA)
http://cfpub.epa.gov/ncea/cfm
/record isplay.cfm?deid=2016
7
BMDS is designed to fit mathematical models to dose-
response data so that the results allow selection of a
benchmark dose (BMD) that is associated with a
predetermined benchmark response (BMR), such as a 10%
increase in the incidence of a particular lesion or a 10%
decrease in body weight gain. General guidance is
available. Technical guidance document for BMDS is
available online (external review draft). Periodic revision.
BMD values used with dose addition could allow
estimation of a BMD for the mixture. For toxicity
endpoints usually described by virtually safe
levels (RfDs and RfCs), this approach would
provide a risk-based dose associated with risk
of a particular effect.
CatReg (U.S. EPA)
http://cfpub.epa.gov/ncea/cfm
/record isplay.cfm?deid=1816
2
Categorical regression model developed for meta-analysis
of toxicology data. Still in development, this could be useful
for evaluating different types of data in evaluating potential
cumulative health risks.
CatReg can be used to evaluate multiple effects
within a chemical grouping (e.g., as grouped by
target organ or system) and can also be used
as a tool to support the health effect estimate
(e.g., hazard index) from multiple-route
exposures.
Risk-based screening levels
(see text, can be found
through:
http://www.epa.gov/region09/
waste/sfund/prg/,
http://www.epa.gov/reg3hwm
d/risk/eco, and
http://epa.gov/earth1 r6/6pd/rc
ra c/pd-n/screen.htm
Screening criteria for environmental media (soil, drinking
water, and air) based on specified risk levels, based on
conservative assumptions and extant toxicity values (some
are outdated); developed by various EPA regions, offices,
and other organizations. For example, EPA Regions 3, 6,
and 9 have developed risk-based concentrations (RBCs),
medium-specific screening levels (MSSLs), and preliminary
remediation goals (PRGs), respectively.
Not designed for cumulative risk assessment,
because they are chemical-specific and not
based on specific pathways or target organs.
However, they could be useful for narrowing the
assessment focus (e.g., during data evaluation)
to those chemicals most likely to contribute to
overall risks at a site.
A-44
-------
A.5. RESOURCES TO CHARACTERIZE RISK AND UNCERTAINTY AND PRESENT
RESULTS
Many assumptions are made when assessing human health risks of multiple
chemicals from environmental exposures. Thus, it is important for the risk results and
associated uncertainties to be well characterized and clearly presented so this
information can be appropriately interpreted to guide sound decisions. This can involve
graphical illustrations of statistical and spatial information, as highlighted below.
Selected tools to support this final phase of the cumulative risk assessment are
summarized in Table A-5.
Spatial Analysis and Decision Assistance (SADA) (U.S. EPA and
U.S. Nuclear Regulatory Commission, NRC). The NRC joined the EPA to
support a very useful integrated software package to support human and
ecological cumulative risk assessments, working with the University of
Tennessee. The human health module of this tool includes the equations from
the standard Superfund guidance (U.S. EPA, 1989a) and contains flexible land
use scenarios and exposure pathways. These can be combined as indicated to
represent overall exposure for the representative receptors evaluated. The input
data for these pathways can be tailored to reflect site-specific conditions;
interactions are not considered. This tool emphasizes the spatial distribution of
contaminant data, and modules cover visualization, geospatial analysis,
statistical analysis, sampling design and decision analysis. Outputs can be
tabular or graphical, and can be used to identify where risk results exceeds a
target value. Many SADA capabilities are also covered by the Fully Integrated
Environmental Location Decision Support (FIELDS) system, which is coordinated
through EPA Region 5 and accessible from ArcView. The SADA tool is available
at http://www.tiem.utk.edu/~sada/.
Probabilistic Resources (U.S. EPA, Others). Risk assessments commonly
present human health risks as single-point estimates (e.g., 1 x 10~5), following the
EPA's basic risk assessment guidance for contaminated sites (U.S. EPA, 1989a).
Such estimates provide little information about the underlying uncertainty or
variability. The uncertainty typically spans at least an order of magnitude and
often much more. Monte Carlo simulation offers one way of considering
uncertainty and variability, as it relies on multiple descriptors using statistical
techniques to calculate a quantity repeatedly with inputs selected randomly from
a reasonable population of values (U.S. EPA, 1999i). Results approximate a full
range of reasonably possible outcomes and are typically plotted as graphs (e.g.,
frequency distributions) or tabulated. However, this approach has several
limitations, which affect its acceptance as a preferred assessment method.
Limitations include the following: difficulty in distinguishing between variability
and uncertainty; use of exposure parameters developed from short-term studies
for long-term exposure; and sensitivity of the tails of the distributions, which can
be of greatest interest, to input distributions. Nevertheless, Monte Carlo
A-45
-------
simulation approaches offer one way to represent uncertainty and variability in
the risk results.
RESRAD (DOE Argonne National Laboratory). The original RESidual
RADioactivity code was designed to evaluate radiological risks and develop
radiological cleanup levels. It can cover 14 combined exposure pathways and is
used by DOE for radioactively contaminated sites and by NRC for dose
evaluations to support decommissioning and waste disposal requests.
Subsequent additions to the family of codes include RESRAD-CHEM (which
calculates risks and hazard indices across 9 exposure pathways and includes a
database of chemical properties, transfer factors, and toxicity values for about
150 chemicals), RESRAD-BASELINE (which covers both radionuclides and
chemicals and uses measured concentrations as input), and RESRAD-OFFSITE
(with includes a two-dimensional dispersion groundwater model and the
CAP-88PC air dispersion model). Outputs can be tabular and graphic, and the
code includes a Monte Carlo module for probabilistic analyses. The code
incorporates transformation over time for radioactive decay, but like many others
it does not address environmental transformation of chemicals or interactions.
Regional Air Modeling Initiative (RAIMI) (U.S. EPA). The Regional Air
Modeling Initiative (RAIMI) approach developed by EPA Region 6 is CIS-based
and looks at multiple sources across the EPA programs. This tool was
developed by Region 6 and uses multiple emissions data sources to assess
community-level inhalation impact by evaluating an unlimited number of
stationary and mobile air toxics sources. It utilizes both air and risk modeling
components. RAIMI also supports source attribution analyses, so individual
sources can be for targeted reductions rather than simply revealing areas of
concern. Initial findings indicate that a small number of sources may be
responsible for the majority of impact. Such models aim to become useful
beyond Region 6, as the EPA moves to risk-based approaches across all
programs. In the RAIMI approach, cumulative information does not necessarily
take into account the effect of complex mixtures, as additivity is assumed. At a
July 2003 meeting of the Advisory Board, several potential applications of this
tool were identified, including using the RAIMI dataset in conjunction with the
cumulative risk framework; predicting future risk, or the impact of past regulation;
or integrating data sources. The tool is already being used to identify useful
databases and emissions inventories. The model has been submitted to the
EPA's Council for Regulatory Environmental Modeling (CREM) for validation.
The tool currently focuses on one medium (air) so it would need to link with other
modules to address other sources of risk (such as from community drinking water
or food residues) for a full cumulative assessment. Information is available at
http://www.epa.qov/earth1 r6/6pd/rcra c/raimi/raimi.htm.
Cumulative Risk Index Analysis (U.S. EPA). The Cumulative Risk Index
Analysis (CRIA) System is a multi-purpose environmental assessment tool based
on CIS technology from EPA Region 6. This CIS-based screening system uses
data from major government databases and inputs from technical and regulatory
professionals to mathematically transform information relevant to cumulative risk
A-46
-------
to visual forms such as GIS maps and tables. The system has been used to
assess and display human health, ecological, socio-economic and regulatory risk
information. The framework developed for implementing CRIA is available from
the EPA website at http://www.epa.gov/osp/presentations/cumrisk/carnev.pdf.
Region 6 has conducted over 6500 cumulative risk assessments in
environmental justice communities using its Comparative Cumulative Risk
System.
Other GIS Tools (Private). Several government agencies and private
companies have developed GIS programs to simultaneously assess exposures
of multiple chemicals by a single receptor. For example, ESRI, Inc., has
developed the screening-level risk assessment module RISKMOD for its ArcView
platform; this tool calculates cumulative risks from multiple contaminants. For
carcinogens, risk is calculated for each exposure pathway by summing the
individual lifetime excess cancer risks for each chemical associated with that
pathway. For noncarcinogens, the hazard quotients for each exposure pathway
can be summed to produce a hazard index for that pathway (Naranjo et al.,
2000). A case study illustrating how RISKMOD was applied to assess risks for a
Bolivian mine site is available at
http://gis.esri.com/librarv/userconf/procOO/professional/papers/PAP480/p480.htm.
Cumulative Adjustment of Protective Concentration Levels (PCLs) (Texas,
TCEQ). PCLs are a set of toxicity-based screening criteria developed by TCEQ
for use in risk assessments of sites in the state. Whereas the individual PCLs
were derived for evaluation of risks from individual chemicals, the TCEQ has
developed an equation for downward adjustment of the PCLs for use when
evaluating risks where at least 10 carcinogenic or noncarcinogenic chemicals of
concern (COC) are present for a specific exposure pathway. The adjustments
result in reduced PCLs for individual chemicals based on the ratio of the
measured concentration of each COC to its PCL. If the sum of these ratios
exceeds a predetermined value (here, 10), adjusted PCL values may be
necessary for some COCs to ensure that state risk reduction rule mandates are
met (i.e., cumulative cancer risks for multiple carcinogenic COCs cannot exceed
1 x 10~4, and the hazard index for multiple noncarcinogenic COCs cannot exceed
10). The COCs to be adjusted are determined based on a decision process
outlined in the Cumulative Adjustment guidance document (TCEQ, 2002). The
adjustment process is a simplistic budgeting exercise in which the analysts are
able to choose the PCLs to be lowered and the magnitude of the reduction. The
guidance document is available at http://www.tceg.state.tx.us/.
Framework for Risk Analysis in Multimedia Environmental Systems
(FRAMES) (U.S. EPA). The EPA has developed an integrated software system
with support from Pacific Northwest National Laboratory, to conduct screening-
level assessments of health and ecological risks for hazardous waste
identification rule (HWIR) chemicals from land-based waste management units.
A-47
-------
TABLE A-5
Selected Resources for Characterizing Risk and Uncertainty and Presenting Results
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
SADA (Spatial Analysis and Decision
Assistance) (DOE, NRC, UT)
http://www.tiem.utk.edu/~sada/
Integrated set of software with flexible land use
scenarios and exposure pathways to assess health
risks. The tool emphasizes spatial distribution of
contaminant data; modules cover visualization,
geospatial analysis, statistical analysis, sampling
design and decision analysis. Outputs can be tabular
or graphical. (Also covers ecological risks, aims to
support integrated decisions.)
Useful for cumulative risk assessments;
can combine pathways to assess overall
exposures and summed risks/hazard
indices for receptors of interest. Input
data can reflect site-specific conditions;
interactions are not considered.
RESRAD (RESidual RADioactivity) (DOE-
ANL)
http://www.ead.anl.gov/resrad
(family of codes, including RESRAD-
CHEM and BASELINE for chemicals)
The original code was designed to guide radiological
cleanup criteria for contaminated sites and assess
doses and risks from residual radionuclides. Sister
codes cover chemical contaminants to support a
combined evaluation of risks and hazard indices at
sites with radionuclides and chemicals. Includes a
screening groundwater model, links to an air
dispersion model and includes a probabilistic module.
Outputs are graphics and tables.
Useful for cumulative assessments at
radioactively and chemically
contaminated sites; can assess
sensitivity, covers natural radioactive
decay (but not environmental
transformation) to address changes over
time; produces risk and hazard indices
summed across multiple contaminants
and pathways; does not address
interactions.
Monte Carlo Analysis-Based Resources
(U.S. EPA, others)
Statistical methods for addressing uncertainty and
variability in estimating health risks by developing
multiple descriptors to calculate a quantity repeatedly
with randomly selected scenarios for each
calculation. Most useful for single-point risk
estimates; can be a useful as a presentation tool
because graphics show range of scenarios and
outputs.
Combining approximations for multiple
sources of potential risk (e.g.,
environmental and lifestyle risk) can be
complicated. Could be used to evaluate
cumulative risks by combining results for
individual exposures that consider
variability and uncertainty.
A-48
-------
TABLE A-5 cont.
Resource and Access
Purpose and Scope
Cumulative Risk Remarks
Regional Air Impact Modeling Initiative
(RAIMI) (U.S. EPA)
http://cfpub2.epa.gov/crem/crem report.cf
m?deid=74913
Risk-based prioritization tool developed by Region 6
to support regional risk-based prioritization at a
community-level resolution, from exposures to
multiple airborne contaminants from multiple sources
via multiple exposure pathways. Designed to support
cross-program analyses. Includes Risk-MAP, to
estimate health risks from exposures to chemical
emissions over large areas.
Assesses multiple contaminants and
multiple sources for EPA programs, for air
contaminants. Designed to consider
source-specific and contaminant-specific
contributions to cumulative exposures
associated with the air pathway.
Cumulative Risk Index Analysis (CRIA)
(U.S. EPA)
http://www.epa.gov/osp/presentations/cu
mrisk/carney.pdf
Analyze and present cumulative risks spatially and
statistically using a CIS-based tool designed by EPA
Region 6. Useful for projects where quality toxicity,
geographical and exposure data exist. Useful for
cumulative impacts analysis in National
Environmental Policy Act (NEPA) projects, including
ecological stressors and sources of pollutants
impacting humans.
Designed specifically for spatial
presentation of cumulative risks. Can
compare human health and ecological
risks. 90 environmental criteria are in
use, with 45 used to identify multimedia
inspection targets. Also considers
cultural resource concerns and sensitive
subpopulations.
Environmental Load Profile (U.S. EPA)
http://www.epa.gOv/region02/community/e
i/guidelines.htm#step4
Compares indicators of well-being with statewide-
derived benchmarks. A screening-level tool
developed by EPA Region 2, as a companion to the
Environmental Justice Demographic Screening Tool.
Similar to RAIMI and CRIA above but
considers only Toxics Release Inventory
(TRI) emissions, air toxics and facility
density, in screening mode. A more
detailed investigation fora community's
burden should be conducted at the local
level.
A-49
-------
APPENDIX B
TOXICITY INFORMATION TO SUPPORT GROUPINGS
This appendix illustrates how toxicity data can be organized to support screening
and grouping for cumulative risk assessments. Information presented here is expected
to show how such toxicity data can be used in conjunction with the toxicity
considerations presented in Chapter 4. More detailed chemical-specific information
sources are also available (e.g., resources listed in Appendix A). Note that these data
on toxicity values are constantly being updated as assessments are revised and
created. Users should always check with IRIS and other reliable data sources for
current toxicological qualitative evaluations and reference values. Also, note that the
tables in this Appendix B focus on noncancer endpoints and may not reflect
carcinogenicity or other endpoints of concern for a given chemical.
B.1. EXAMPLE TOXICITY MATRICES FOR SELECTED CHEMICALS
The primary toxicological effects for a set of example chemicals often
encountered at a contaminated site are summarized in this appendix to illustrate how
this information can be used to support grouping for an evaluation of joint toxicity and
potential interactions. These chemicals were selected for study to support a site-
specific integrated risk evaluation (at the U.S. Department of Energy's Hanford site).
This primary toxicity information can be used to help group the chemicals by common
target organ or system, by common mode of action or by potential for interaction
considering common metabolites or metabolic pathways. Primary effects for oral
exposures are provided in Table B-1, and those for inhalation exposures are
summarized in Table B-2. The toxicity values presented in Tables B-1 and B-2 are from
the EPA's IRIS database, current to 2007. The reference doses and lowest secondary
toxicological effect levels for these study chemicals are compared in Table B-3.
To simplify the presentation of information, the tables are presented together
after the references for this appendix. A glossary of toxicity terms to support the
grouping of chemicals by effects is presented following these tables.
B.2. SUPPORTING INFORMATION ON TOXICOLOGICAL CONCEPTS
Information used to derive the primary toxicity valuesoral reference doses
(RfDs) and inhalation reference concentrations (RfCs)are provided in Section B.2.1.
These primary data are also compared to the data describing effects that are
B-1
-------
considered secondary (occurring at higher doses than the primary or critical effect) in
Section B.2.2.
B.2.1. Derivation of Primary Toxicity Factors. As described in EPA's document
A Review of the Reference Dose and Reference Concentration Processes (U.S. EPA,
2002e), the critical effect used in dose-response assessments is currently associated
with the lowest no-observed-adverse-effect level (NOAEL), and various uncertainty
factors are applied to the dose at this critical-effect level to derive the RfD or RfC. An
experimental exposure level is selected from the critical-effect study that represents the
highest level tested in which no adverse effect was demonstrated. This NOAEL is the
key data point obtained from the study of the dose-response relationship and has
traditionally served as the primary basis for evaluating potential human health risks.
This approach is based on the assumption that if the critical toxic effect is prevented,
then all toxic effects are prevented. A chemical can elicit more than one toxic effect,
even in one test animal, or in tests of the same or different duration (acute, subchronic
and chronic exposure studies). In general, NOAELs for these effects will differ. In
addition, this approach assumes that the sequence of various health effects with
increasing exposure for a particular chemical is maintained across species (U.S. EPA,
2002e).
A more recent approach used to derive RfDs and RfCs is the benchmark dose
(BMD) method. Use of the NOAEL in determining RfDs and RfCs has long been
recognized as having limitations in that it (1) is limited to one of the doses in the study
and is dependent on study design; (2) does not account for variability in the estimate of
the dose-response; (3) does not account for the slope of the dose-response curve; and
(4) cannot be applied when there is no NOAEL, except through application of an
uncertainty factor (U.S. EPA, 2004g). A goal of the BMD approach is to define a
starting point-of-departure for the computation of a reference value (RfD or RfC) or
slope factor that is more independent of study design. Use of BMD methods involves
fitting mathematical models to dose-response data and using the different results to
select a BMD that is associated with a predetermined benchmark response, such as a
10% increase in the incidence of a particular lesion or a 10% decrease in body weight
gain, which would be termed the BMDio (U.S. EPA, 2004g). Note that for the study
chemicals, the primary RfD for beryllium and the primary RfC for chromium VI
(particulates) are both based on this newer BMD approach, as opposed to the standard
NOAEL/LOAEL approach used to derive toxicity data for the other chemicals.
B-2
-------
B.2.2. Comparison of Primary and Lowest Secondary Effects. The primary and
lowest secondary effects and respective concentrations (i.e., RfDs and
LOAELs/NOAELs) are given for each chemical for the oral pathway in Table B-3. The
secondary effects data were selected as the lowest doses from the entire set of studies
discussed in the sections on subchronic and chronic levels of significant exposure in the
toxicological profiles prepared by the Agency for Toxic Substances and Disease
Registry (ATSDRJ. Human and animal studies were evaluated separately.
As shown in this table, the lowest doses yielding secondary effects are higher
than the respective RfDs for all the study chemicals. This is to be expected because
RfDs are set to be protective of the lowest adverse effects, or critical effects. For all but
three chemicals, the RfDs are lower than both the lowest NOAEL and LOAEL values for
secondary effects from human and animal studies.
The three chemicals where RfDs could overlap NOAELs are trivalent chromium,
nickel, and zinc. For trivalent chromium, nickel, and zinc, some of the lowest NOAEL
values for secondary effects are below the RfD, but none of the LOAEL values for
secondary effects are below the RfD. The RfD for trivalent chromium is 1.5 mg/kg-day,
while the lowest animal NOAEL is a lower value of 0.46 mg/kg-day. However, the
lowest animal LOAEL (5 mg/kg-day) is above the RfD. The RfD for nickel is
0.02 mg/kg-day and the lowest human NOAEL is also 0.02 mg/kg-day. No human
LOAEL was reported for nickel, but the lowest animal NOAEL (0.97 mg/kg-day) is
above the RfD. The RfD for zinc is 0.3 mg/kg-day, while the lowest human NOAEL is
0.06 mg/kg-day, a lower value. However, the lowest human (0.71 mg/kg-day) and
animal LOAELs (0.5 mg/kg-day) are both higher than the RfD. These overlaps can be
viewed as indications of the quantitative uncertainties when using LOAELs and
NOAELs.
All secondary adverse effects identified in the collection of human and animal
studies reported in the ATSDR toxicological profiles for the 15 study chemicals occur at
concentrations above the RfDs (all LOAELs were above the RfDs). Thus, although
some actual LOAELs for secondary effects may be lower than the LOAEL for the
primary effect (as discussed in Section B.2.3), the series of uncertainty factors applied
during the RfD derivation process ensured that the RfD based on a critical effect is at
least below other available LOAELs. The levels resulting in secondary effects would not
typically be seen on contaminated sites, as the lowest LOAELs for secondary effects
are generally several orders of magnitude higher than the RfDs. This fact is a
testament to the necessity for uncertainty factors during RfD development, given the
findings noted in Section B.2. Because hazard indices estimated for contaminated sites
B-3
-------
are often less than 10, these effects would not generally be expected to occur, except in
cases of high concentrations (e.g., following a major release, for which acute or short-
term exposure levels would be relevant rather than chronic values), multiple routes of
exposure or where interactions occur. Thus, although effect-specific RfDs can be
derived for data-rich chemicals, which would yield useful information for a cumulative
risk assessment involving chemical mixtures, such an approach might not be needed.
Obviously, obtaining secondary effects data for less-studied compounds would be more
difficult but would give a fuller picture of the array of toxic effects exerted by each
chemical. Another example of what a secondary effect analysis might find is discussed
below.
B.2.3. Secondary Effects Findings: Case Study Chemicals. Although the
discussion above notes that the RfDs based on primary effects appear protective of all
effects for the example chemicals studied, it should be noted that the RfD or RfC is
protective partly because of the use of uncertainty factors. Except for a few cases
where no or minimal UFs are used (e.g., when chronic human toxicity data are
available), part of the magnitude of UFs is to account for equitoxic dose extrapolation or
scaling, and part is to be protective in the face of quantitative uncertainty. Thus,
uncertainty factors serve multiple purposes. Some secondary effects might occur at
concentrations lower than the primary NOAEL or LOAEL, but because of study
difficulties might have not been selected as the critical study. Consequently, one
purpose of the UFs not often recognized is to provide some assurance that the RfD or
RfC is protective of secondary effects.
The secondary effects summary for the study chemicals discussed below is
abstracted from the ATSDR toxicological profiles and includes some examples of
LOAELs for secondary effects that are lower than the primary effect LOAEL. These are
the types of secondary effects that should be prioritized in a cumulative health
assessment, as they would be the first to be manifested upon cumulative source or
cumulative pathway exposure in addition to the primary effects. This is not a
comprehensive review of all LOAELs for the study chemicals where a LOAEL is below
the primary effect LOAEL, but rather a cross-section of considerations. Highlights are
as follows:
A human oral arsenic study found nervous system effects including
fatigue, headaches, dizziness, insomnia, and numbness at a secondary
effect LOAEL of 5 x 10"3 mg/kg-day (below the primary effect LOAEL of
1.4 x 10~2 mg/kg-day). Dermal effects of oral exposure have been
documented at LOAELs below the LOAEL from the key study for the same
B-4
-------
dermal primary effect in at least three studies. Two recent studies found
cardiovascular effects at a LOAEL below the dermal-based primary effect
LOAEL; increased cerebrovascular disease and cerebral infarction were
indicated at a LOAEL of 2 x 10~3 mg/kg-day in a 1997 study. Palpitations,
chest discomfort and cyanosis of the extremities were indicated in a 1994
study that also documented dermal effects at 5 x 10~3 mg/kg-day.
Increased serum bilirubin has also been observed at a lower LOAEL than
the primary effect; however, the biological significance of this endpoint
alone may be questionable.
A human inhalation study of beryllium found increased T-cell activity and
chronic beryllium disease at a reported LOAEL of 5.2 x 10~4 mg/m3 (below
the primary effect LOAEL of 5.5 x 10"4 mg/m3). Although this is
mathematically slightly lower than the study selected as the critical study
in the IRIS file derivation of the RfC, the difference is not significant, as the
primary effect basis for the RfC was also a human (more recent 1996)
occupational study of chronic beryllium disease.
Mercury has been reported in at least six developmental studies and
seven neurological studies to result in adverse effects below the primary
effect-based LOAEL of 0.633 mg/kg-day. Four studies found impacts to
the kidneys at LOAELs below the primary effect-based LOAEL as well.
For nickel, 15 studies found effects below the primary effect-based LOAEL
of 50 mg/kg-day. A handful of the studies also found effects below the
NOAEL of 5 mg/kg-day. Specifically, 1993, 1999 and 2000 studies
(captured in the 2003 update to the ATSDR toxicological profile) indicate
reproductive impacts in animals below the primary NOAEL.
Uranium studies found secondary effects at LOAELs below that which the
oral RfD was based. Specifically, endocrine effects and cellular hepatic
and kidney changes were observed in one study. Other minor renal
effects were also noted at lower LOAELs than that used to develop the
oral RfD.
Cancer data are also given in the ATSDR toxicological profiles. For example,
human lung cancer and skin cancer due to arsenic exposure were also reported at
LOAELs below the noncancer primary effect LOAEL; however, cancer risks are typically
evaluated separately from the noncancer hazards so this would be accounted for in a
cancer risk assessment.
Thus, the full body of available literature and resulting toxicity factors, NOAELs
and LOAELs need to be considered and evaluated when performing a cumulative risk
assessment to ensure that the risk assessment takes into account all possible
significant effects and their respective effect levels. While the primary RfDs and RfCs
are considered protective and are often based on the effect seen at the lowest chemical
concentration or dose, the secondary effects discussed above should be prioritized and
B-5
-------
considered in a cumulative health assessment, as they would be the first to be
manifested upon cumulative source or cumulative pathway exposure in addition to the
primary effects.
B.3. GLOSSARY OF TOXICOLOGICAL EFFECTS
Abdominal pain See Pain. Indicates effect is seen in the abdominal region.
Abnormality Unusual function or irregularity.
Abnormal electromyographic findings See Abnormality. In this effect, measurements
indicating that the electrical voltage generated by body muscles is irregular.
Abnormal nerve conduction See Abnormality. Indicates the effect is manifested in
nerve conduction.
Abortion The premature expulsion from the uterus of the products of conception of
the embryo or of a nonviable fetus. Natural abortions are typically called miscarriages.
Aborted or stillborn fetuses See Abortion, Stillbirth.
Absorption alterations See Alterations. Indicates effect is seen in gastrointestinal
tract absorption.
Acinar cell necrosis and metaplasia in pancreas See Necrosis and Metaplasia.
Indicates effects are seen in the acinar cells of the pancreas.
Adenocarcinoma A form of cancer that involves cells from the lining of the walls of
many different organs of the body.
Adenoma A benign epithelial tumor in which the cells form recognizable glandular
structures or in which the cells are clearly derived from glandular epithelium.
Adhesions Fibrous bands or structures by which parts abnormally adhere.
Adnexal changes Alterations in appendages. For example, in gynecology the
adnexa are the appendages of the uterus, namely the ovaries, Fallopian tubes and
ligaments that hold the uterus in place.
Albuminuria The presence of protein in the urine, principally albumin, generally
indicating disease.
Alkaline phosphatase An enzyme that catalyses the cleavage of inorganic phosphate
non-specifically from a wide variety of phosphate esters and having a high (greater than
8) pH optimum.
B-6
-------
Alopecia Baldness, absence of the hair from skin areas where it normally is present.
ALT activity changes Changes in a liver enzyme that plays a role in protein
metabolism; see also AST. Elevated serum levels of ALT are a sign of liver damage
from disease or drugs. Synonym: serum glutamic pyruvic transaminase.
Alterations Changes, such as increase or decrease.
Altered sperm chromatin structure See Alterations. Indicates effect seen in the
chromatin structure of sperm.
Alveolar proteinosis A very rare disease in which a phospholipid is widely distributed
in cells and accumulates in the alveolar spaces in the lung. In some cases the
underlying cause is unknown. In others it may relate to an infection or an immune
system dysfunction. The net effect is a progressive interference in the ability of the lung
(alveoli) to exchange oxygen and carbon dioxide. Symptoms include cough, weight
loss, fatigue, shortness of breath and nail abnormalities (clubbing).
Anemia Too few red blood cells in the bloodstream, resulting in insufficient oxygen
supply to tissues and organs.
Anisokaryosis Cells or cell nuclei that vary considerably in size.
Anorexia The uncontrolled lack or loss of the appetite for food.
Arterial insufficiency Failure of arteries to function adequately, resulting in insufficient
oxygen supply to cells, tissues or organs.
Arterial [oxygen] tension The pressure of the blood within an artery, the arterial
pressure. Also called the intra-arterial pressure.
Arterial thickening Increase in the thickness of the arterial walls, resulting in impaired
function and restricted flow.
Arterial thickening in pancreas See Arterial thickening. Indicates effect is seen in the
pancreas.
Arterial thickening in stomach and intestines See Arterial thickening. Indicates effect
is seen in the stomach and intestines.
Ascites An effusion and accumulation of serous fluid in the abdominal cavity.
Synonyms: abdominal dropsy, peritoneal dropsy, hydroperitonia, hydrops abdominis.
AST activity changes Changes in a liver enzyme that plays a role in protein
metabolism; see also ALT. Elevated serum levels of AST are a sign of liver damage
from disease or drugs. Synonym: serum glutamic oxaloacetic transaminase.
Astroglial hypertrophy See Astrogliosis.
B-7
-------
Astrogliosis Hypertrophy of the astroglia, usually in response to injury. Astroglia
(astrocytes) are the largest and most numerous neuroglial cells in the brain and spinal
cord. They regulate the extracellular ionic and chemical environment, and "reactive
astrocytes" (along with microglia) respond to injury.
Ataxia Failure of muscular coordination, irregularity of muscular action.
Atelectasis A term used to describe partial or complete collapse of the lung, usually
due to an obstruction of a bronchus (with mucus plug, infection or cancer). Symptoms
of atelectasis include low-grade fever, dry cough, chest pains and mild shortness of
breath.
Atrophy A wasting away, a diminution in the size of a cell, tissue, organ or part.
Autoimmune glomerulonephritis A condition in which an individual's immune system
starts reacting against his or her own tissues, causing diseases such as
glomerulonephritis (inflammation of the cluster of blood vessels at the beginning of the
kidney tubule where unconcentrated urine is formed by filtration of the blood).
Autonomic dysfunction See Dysfunction. Indicates effect is seen in the autonomic
nervous system (Neurons that are not under conscious control, comprising two
antagonistic components, the sympathetic and parasympathetic nervous systems. The
autonomic nervous system regulates key functions including the activity of the cardiac
(heart) muscle, smooth muscles (e.g., of the gut), and glands. The autonomic nervous
system has two divisions: 1. The sympathetic nervous system that accelerates the heart
rate, constricts blood vessels, and raises blood pressure. 2. The parasym pathetic
nervous system slows the heart rate, increases intestinal and gland activity, and relaxes
sphincter muscles.
Azotemia A higher than normal blood level of urea or other nitrogen containing
compounds in the blood. The hallmark test is the serum BUN (blood urea nitrogen)
level. Usually caused by the inability of the kidney to excrete these compounds.
Basal cell carcinoma See Carcinoma. Indicates effects is seen in the relatively
undifferentiated cells in an epithelial sheet that give rise to more specialized cells act as
stem cells.
Behavioral changes See Alterations. Indicates effect is seen on normal or usual
behavior.
Bile duct enlargement/proliferation See Enlargement, Proliferation. Indicates effect is
seen in bile ducts.
Blackfoot disease Syndrome characterized by a progressive loss of circulation in the
hands and feet, leading ultimately to necrosis and gangrene.
Blastogenesis Multiplication or increase by gemmation or budding.
B-8
-------
Blastogenic activity See Blastogenesis.
Bleeding in the gut See Hemorrhage. Indicates effect is seen in the gut.
Blood phosphate A salt of phosphoric acid present in blood or blood serum, the clear
liquid that separates from blood on clotting.
Body weight alterations See Alterations. Indicates effect is manifested as a change
in body weight. See also Weight gain, Weight loss.
Body weight gain See Weight gain. Indicates effect is for whole body weight.
Body weight loss See Weight loss. Indicates effect is for whole body weight.
Bone accretion The growing together of bones.
Bone marrow retention alterations See Retention alterations. Indicates effect is
manifested in the bone marrow.
Brain cell degeneration See Degeneration. Indicates effect is manifested in brain
cells.
Brain, reduced number of myelinated fibers Fewer neural connections within the
brain.
Bronchiectasis Persistent and progressive dilation of bronchi or bronchioles as a
consequence of inflammatory disease (lung infections), obstruction (tumor) or
congenital abnormality (for example cystic fibrosis). Symptoms include fetid breath and
paroxysmal (spastic) coughing, with the expectoration of mucopurulent matter. It may
affect the bronchioles uniformly (cylindric bronchiectasis) or occur in irregular pockets
(sacculated bronchiectasis) or the dilated bronchi may have terminal bulbous
enlargements (fusiform bronchiectasis).
Bronchitis Inflammation of one or more bronchi, usually secondary to infection.
Bronchopneumonia/bronchiopneumonia Inflammation of the lungs that usually
begins in the terminal bronchioles. These become clogged with a mucopurulent
exudate forming consolidated patches in adjacent lobules. The disease is frequently
secondary in character, following infections of the upper respiratory tract, specific
infectious fevers and debilitating diseases. In infants and debilitated persons of any age
it may occur as a primary affection. Synonyms: bronchial pneumonia, bronchoalveolitis,
bronchopneumonitis, lobular pneumonia.
Carcinoma A malignant new growth that arises from epithelium, found in skin or,
more commonly, the lining of body organs, for example: breast, prostate, lung, stomach
or bowel. Carcinomas tend to infiltrate into adjacent tissue and spread (metastasize) to
distant organs, for example: to bone, liver, lung or the brain.
B-9
-------
Cardiac inotropy See Inotropy. Indicates effect is seen in the cardiac muscles.
Casts (in urine) White blood cell casts indicate pyelonephritis, but they are not
always present in the urine.
Cell-mediated cytotoxicity See Cytotoxicity. Indicates cells convey effect.
Cell-mediated immune response Immune response that involves effector T
lymphocytes and not the production of humoral antibody. Responsible for delayed
hypersensitivity and in defense against viral infection and intracellular protozoan
parasites.
Cellular degeneration/changes See Degeneration. Indicates effect is seen within
cells.
Central lobe necrosis See Necrosis. Indicates effect is seen in the central lobe of the
liver.
Centrilobular necrosis See Central lobe necrosis.
Cerebral infarction Infarction of brain tissue.
Cerebrovascular disease A general term which encompasses a variety of diseases
which affect (via the occlusive effects of atherosclerosis) the arteries which supply the
brain.
Chronic conjunctivitis See Conjunctivitis.
Cirrhosis Liver disease characterized pathologically by loss of the normal
microscopic lobular architecture, with fibrosis and nodular regeneration. The term is
sometimes used to refer to chronic interstitial inflammation of any organ.
Cloudy swelling in kidneys See Inflammation. Indicates effect is seen in kidneys.
Confusion Disturbed orientation in regard to time, place or person, sometimes
accompanied by disordered consciousness.
Congenital malformations Abnormal formation of a structure evident at birth
Conjunctivitis Inflammation of the conjunctiva, generally consisting of conjunctival
hyperemia associated with a discharge.
Contractility Capacity for becoming short in response to a suitable stimulus.
Cough A rapid expulsion of air from the lungs typically in order to clear the lung
airways of fluids, mucus or material.
Cramps See Pain. Indicates effect is seen in abdomen.
B-10
-------
Cyanosis A bluish discoloration, applied especially to such discoloration of skin and
mucous membranes due to excessive concentration of reduced hemoglobin in the
blood.
Cysts Any closed cavity or sac that is lined by epithelium often contains liquid or
semi-solid material.
Cytomegaly A condition or disease characterized by abnormally enlarged cells.
Cytotoxicity The quality of being poisonous, or toxic, to individual cells.
Damage See Injury.
Death See Survival.
Decline in conditioned responses Reduced frequency of learned behaviors in
response to triggering stimulus.
Decrease in Hb and H values Lowered hemoglobin content, resulting in reduced
oxygen carrying capacity and possible anoxia. Hemoglobin is the Four subunit globular
oxygen carrying protein of vertebrates and some invertebrates. There are two alpha
and two beta chains (very similar to myoglobin) in adult humans, the heme moiety (an
iron-containing substituted porphyrin) is firmly held in a nonpolar crevice in each peptide
chain.
Decreased alkaline phosphatase See alkaline phosphatase.
Decreased arterial tension See arterial tension. Reduction in the pressure of blood
within an artery.
Decreased avoidance response Reduction in learned ability to respond to a cue that
is instrumental in avoiding a noxious experience.
Decreased blood or serum phosphate levels See blood phosphate and serum
phosphate.
Decreased cardiac contractility See contractility. Indicates effect is seen in the
cardiac muscles.
Decreased caudal ossification See Ossification. Indicates effect is seen at a position
more toward the cauda or tail of an organism.
Decreased corpuscular volume See Anemia. Indicates reduced volume of red blood
cells.
Decreased DMA in brain areas Reduction in genetic material in the brain.
Decreased fetal body weight See Weight Loss. Indicates decrease is in the fetus.
B-11
-------
Decreased immunoglobulins Reduction in the specific protein substances that are
produced by plasma cells to aid in fighting infection. Some immunoglobulins (gamma
globulin) take part in various immune responses of the body to bacteria or foreign
substances (allergens, tumor or transplanted tissue). Examples include IgG, IgM, IgA,
IgD and IgE.
Decreased macrophage activity Reduction in the function of macrophages, which are
relatively long lived phagocytic cell of mammalian tissues, derived from blood monocyte.
Macrophages from different sites have distinctly different properties. Macrophages play
an important role in killing of some bacteria, protozoa and tumor cells, release
substances that stimulate other cells of the immune system and are involved in antigen
presentation.
Decreased pulmonary bactericidal activity Reduction in the body's defense
mechanisms to kill bacteria in the lungs.
Decreased response rate for learned behaviors Increased time to respond to
triggering stimuli. See also Decline in Conditioned Responses.
Decreased tactile-kinesthetic function Reduction of the tactile the sense of touch or
pressure by which muscular motion, weight, position are perceived.
Decreased T-cell activity See T-cell.
Decreased sperm count Decrease in the number of sperm in the ejaculate (when
given as the number of sperm per milliliter it is more accurately known as the sperm
concentration or sperm density).
Decreased survival See Survival.
Decreased vasoreactivity Reduction in the blood vessels' ability to change caliber in
response to stimulus, thus affecting blood flow.
Degeneration Reduced size or function of a cell, tissue, organ or part.
Dehydration Excessive loss of body water.
Delayed ossification Indicates a delay in the formation of bone or of a bony
substance, the conversion of fibrous tissue or of cartilage into bone or a bony
substance. See also Reduced Ossification.
Demyelination See Myelin degeneration.
Depigmentation See Pigmentation changes. The removal or loss of pigment,
especially melanin.
Depression A lowering or decrease of functional activity. Also a mental state of
depressed mood characterized by feelings of sadness, despair and discouragement.
B-12
-------
Depression ranges from normal feelings of the blues through dysthymia to major
depression.
Dermal effects Effects on the skin.
Dermatitis Inflammation of the skin.
Desquamation of tubular cells The shedding or exfoliation of epithelial elements of
the renal tubules.
Diabetes mellitus Relative or absolute lack of insulin leading to uncontrolled
carbohydrate metabolism. In juvenile onset diabetes (that may be an autoimmune
response to pancreatic cells) the insulin deficiency tends to be almost total, whereas in
adult onset diabetes there seems to be no immunological component but an association
with obesity.
Diarrhea A morbidly frequent and profuse discharge of loose or fluid evacuations
from the intestines, without tenesmus; a purging or looseness of the bowels; a flux.
Diffuse erythematous and scaly rash Redness and scaling of the skin produced by
congestion of the capillaries, which may result from a variety of causes.
Diffuse palmar or plantar hyperkeratosis See Hyperkeratosis. Indicates effect is
seen on palms of hands and soles of feet and is widespread in nature.
Diffuse pigmentation See Pigmentation. Indicates pigmentation is widespread.
Dilation Expanded in internal diameter.
Disorientation See Confusion.
Distribution alterations Changes in distribution.
Diuresis Increased excretion of urine. Can be due to metabolic conditions such as
diabetes, where the increased glucose level in the blood causes water to be lost in the
urine. Can also be produced specifically by diuretic drugs that increase sodium and
water loss from the kidney.
DOPAC (Dopachrome oxidoreductase) Decarboxylates and converts dopachrome to
5,6-dihydroxyindole.
Dysfunction Failure to function normally.
Dyspepsia Difficult or painful digestion, indigestion.
Edema The presence of abnormally large amounts of fluid in the intercellular tissue
spaces of the body, usually applied to demonstrable accumulation of excessive fluid in
the subcutaneous tissues. Edema may be localized, due to venous or lymphatic
obstruction or to increased vascular permeability or it may be systemic due to heart
B-13
-------
failure or renal disease. Collections of edemous fluid are designated according to the
site, for example ascites (peritoneal cavity), hydrothorax (pleural cavity) and
hydropericardium (pericardial sac). Massive generalized edema is called anasarca.
Embryolethality See Abortion, Stillbirth.
Emaciation Excessive leanness; a wasted condition of the body.
Emesis Vomiting, an act of vomiting. Also used as a word termination, as in
hematemesis.
Emphysema A pathological accumulation of air in tissues or organs, applied
especially to such a condition of the lungs.
Encephaloceles Hernia of the brain; infarction of brain tissue.
Enhanced inflammatory response Increased sensitivity to tissue injury causing an
inflammatory response, which is a part of innate immunity. Inflammation occurs when
tissues are injured by viruses, bacteria, trauma, chemicals, heat, cold or any other
harmful stimulus. Chemicals including bradykinin, histamine, serotonin and others are
released by specialized cells. These chemicals attract tissue macrophages and white
blood cells to localize in an area to engulf (phagocytize) and destroy foreign
substances. A byproduct of this activity is the formation of pus, which is a combination
of white blood cells, bacteria and foreign debris.
Enlarged nuclei Increase in size of the cellular nucleus.
Enlarged nuclei of tubular cells See Enlarged nuclei. Indicates cells affected are
kidney tubular cells.
Enlargement Increased size. See also Weight gain.
Enzyme activity stimulation See Increased enzyme activity.
Enzyme inhibition - Arrest or restraint of a enzyme process(es).
Eosinophilia The formation and accumulation of an abnormally large number of
eosinophils in the blood.
Epitaxis (epitasis) The period of violence in a fever or disease; paroxysm.
Epithelial degeneration See Degeneration. Indicates effect is manifested in the
epithelium.
Epithelial degradation See Epithelial degeneration.
Eroded luminal epithelium in the stomach See Degeneration. Indicates effect is seen
in the luminal epithelium of the stomach.
B-14
-------
Erythroid hyperplasia of bone marrow See Hyperplasia. Indicates effect is seen in
erythrocytes of the bone marrow.
Exencephaly See Terata. Condition in which the brain is located outside of the skull.
This condition is usually found in embryos as an early stage of anencephaly. As an
exencephalic pregnancy progresses, the neural tissue gradually degenerates. It is
unusual to find an infant carried to term with this condition because the defect is
incompatible with survival.
Excretion reduction A decline in production of waste products. See also Abnormal
Retention. May include reduced urinary output.
Eye defects in fetus See Terata. Indicates malformation of the fetal eye.
Fatigue Weakness.
Fatty changes See Fatty infiltration.
Fatty infiltration Accumulation of fatty acids as triglycerides in the liver. Focal fatty
infiltration may mimic neoplastic or other low-density parenchymal lesions, including
abscesses and hemangiomas. Fatty liver has also been associated with diabetes,
obesity, use of corticosteroids and other drugs (including chemotherapy), Cushing's
disease, total parenteral nutrition, starvation, hyperlipidemia, pregnancy, cystic fibrosis,
Reye's syndrome, malignancy, jejunoileal bypass and other causes.
Fertility The capacity to conceive or induce conception and thus generate offspring.
Fetotoxicity Toxicity manifested in the fetus.
Fibrosis The formation of fibrous tissue, fibroid or fibrous degeneration.
Focal necrosis See Necrosis. Indicates effect is seen in localized area.
Folliculitis Inflammation of a follicle or follicles, used ordinarily in reference to hair
follicles, but sometimes in relation to follicles of other kinds.
Functional denervation Reduced capacity of existing neurons resulting in effective
disfunction at the neural termination.
Functional impairment Reduction of normal function in a cell, organ, tissue or part.
Gangrene Death of tissue, usually in considerable mass and generally associated
with loss of vascular (nutritive) supply and followed by bacterial invasion and
putrefaction.
Gasping The act of opening the mouth convulsively to catch the breath; a labored
respiration; a painful catching of the breath.
B-15
-------
Gastrointestinal hemorrhage See Hemorrhage. Indicates effect is seen in the
gastrointestinal tract.
Gastrointestinal irritation See Irritation. Indicates effect is seen in the gastrointestinal
tract.
Genitourinary defects See Terata. Indicates malformation occurring in the
(urogenital) genital and urinary organs.
Glucosuria A condition in which glucose is discharged in the urine; diabetes mellitus.
Glycogen level changes Alterations in levels of the branched polymer of D glucose,
which serves as the major short-term storage polymer of animal cells and is particularly
abundant in the liver and to a lesser extent in muscle.
Granule cell loss Reduction in number of granule cells, a type of neuron, in the
cerebellum.
Granuloma Chronic inflammatory lesion characterized by large numbers of cells of
various types (macrophages, lymphocytes, fibroblasts, giant cells), some degrading and
some repairing the tissues.
Granulomata See Granuloma.
Gross gastrointestinal lesions See Lesions. Indicates widespread effect is seen in
the gastrointestinal tract.
Gross physical abnormalities See Terata. Indicates fetal malformations are
significant and relate to the basic components of the body. See also Skeletal
Malformations, Increases in Skeletal Variations.
Headache See Pain. Indicates effect is seen in the head or sinuses.
Heart abnormalities in fetus See Terata. Indicates malformations affecting the heart.
Heart disease Common condition where vessels (arteries) that carry blood to the
heart muscle become narrowed with fatty deposits. The heart then cannot get the
oxygen and other nutrients it needs. A complete blockage of one of these vessels may
result in a heart attack.
Hematemesis The vomiting of blood.
Hemolysis Disruption of the integrity of the red cell membrane causing release of
hemoglobin.
Hemoperitoneum Intraabdominal bleeding, accompanied by abdominal pain. The
liver or spleen may increase in size. If the bleeding is severe enough, the blood
pressure and hematocrit may fall.
B-16
-------
Hemorrhage Bleeding. The escape of blood from the vessels. Small hemorrhages
are classified according to size as petechiae (very small), purpura (up to 1 cm) and
ecchymoses (larger). The massive accumulation of blood within a tissue is called a
hematoma.
Hemosiderin deposits Deposits of a mammalian iron storage protein (related to
ferritin but less abundant).
Hemosiderin deposits in hepatic macrophages See Hemosiderin deposits. Indicates
effect is seen in liver macrophages, which are relatively long-lived phagocytic cells of
mammalian tissues, derived from blood monocytes.
Hemosiderin deposits in liver See Hemosiderin deposits. Indicates effect is seen in
liver.
Hemosiderin deposits in kidney See Hemosiderin deposits. Indicates effect is seen
in kidney.
Hepatoma Carcinoma derived from liver cells. Also known as hepatocarcinoma or
hepatocellular carcinoma.
Hepatomegaly Enlargement of the liver.
Hepatotoxicity Toxicity manifested in the liver.
Histopathological changes Microscopic changes in diseased tissues.
Histopathological changes in heart tissue See Histopathological changes. Indicates
effect is manifested in heart tissue.
Histopathological changes in lungs See Histopathological changes. Indicates effect
is manifested in lung tissue.
Humoral immune response Immune responses mediated by antibodies.
Hypalgesia Decreased pain response.
Hyperemia An excess amount of blood in an organ. Active hyperemia is increased
blood supply to an organ, usually for physiologic reasons (exercise). Passive
hyperemia is engorgement of an organ with venous blood, usually the result of
inadequate circulation (heart failure).
Hyperkeratosis Hypertrophy of the corneous layer of the skin, or any of various
conditions marked by hyperkeratosis.
Hyperkeratosis of foot See Hyperkeratosis. Indicates effect is seen in the feet.
Hyperpigmentation Darkening of the skin. See also Pigmentation.
B-17
-------
Hyperplasia The abnormal multiplication or increase in the number of normal cells in
normal arrangement in a tissue.
Hypertension Persistently high arterial blood pressure. Hypertension may have no
known cause (essential or idiopathic hypertension) or be associated with other primary
diseases (secondary hypertension). This condition is considered a risk factor for the
development of heart disease, peripheral vascular disease, stroke and kidney disease.
Hypertrophy The enlargement or overgrowth of an organ or part due to an increase
in size of its constituent cells.
Hypertrophy of pancreas islet cells See Hypertrophy. Indicates effect is seen on the
cells of the Islets of Langerhans (or islet cells) within the pancreas.
Hypoplasia The incomplete development or underdevelopment of an organ or tissue.
Hypopigmentation A condition caused by a deficiency in melanin formation or a loss
of pre-existing melanin or melanocytes. It can be complete or partial and may result
from trauma, inflammation and certain infections.
Hypothermia A low body temperature, as that due to exposure in cold weather or a
state of low temperature of the body induced as a means of decreasing metabolism of
tissues and thereby the need for oxygen, as used in various surgical procedures,
especially on the heart or in an excised organ being preserved for transplantation.
Impaired lymphocytic/leukocytic function See impairment. Indicates effect is seen in
the normal function of lymphocytes and leukocytes.
Impaired peripheral vision Reduction in visual capacity, particularly in the periphery
of the normal field of vision.
Impaired liver mitochondrial respiration See Impairment. Indicates effect is seen in
the respiration of the liver mitochondria.
Impaired renal mitochondrial respiration See Impairment. Indicates effect is seen in
the respiration of the kidney mitochondria.
Impairment Reduction in normal function.
Increased cerebral infarction Infarction (an area of tissue death due to a local lack of
oxygen) of brain tissue.
Increased cerebrovascular disease Increase in any of a variety of diseases which
affect (via the occlusive effects of atherosclerosis) the arteries which supply the brain.
May lead to stroke.
Increased DOPAC concentration See DOPAC, increased enzyme activity, and
increased enzyme levels.
B-18
-------
Increased enzyme activity Metabolic increase via stimulation of enzyme systems.
Increased enzyme levels See Increased enzyme activity. Higher measurable
circulating or tissue enzymes.
Increased glycogen see Glycogen level changes.
Increased heart weight See Organ weight gain. Indicates effect is manifested in the
heart tissue.
Increased kidney weight See Organ weight gain. Indicates effect is manifested in the
kidney tissue.
Increased leukocyte count An abnormal accumulation of white blood cells.
Increased liver weight See Organ weight gain. Indicates effect is manifested in the
liver tissue.
Increased lung weight See Organ weight gain. Indicates effect is manifested in the
lung tissue.
Increased MCH See MCH, increased enzyme activity, and increased enzyme levels.
Increased resorptions The loss of substance through physiologic or pathologic
means, such as loss of dentin and cementum of a tooth or of the alveolar process of the
mandible or maxilla. In a reproductive context, implies embryos are not carried to term
but are instead absorbed into the uterine wall. See also Fertility, Reduced Birth Rate,
and Reduced Litter Size, as increased resorptions are related to pregnancy outcome.
Increased response to sheep red blood cells Heightened sensitivity to immune
challenge.
Increased serum enzyme levels See Increased enzyme levels. Indicates effect is
manifested in circulating serum enzymes.
Increased SCOT See SCOT, increased enzyme activity and increased enzyme
levels.
Increased skeletal variations See Terata. See also Gross physical abnormalities.
Increased stillbirth See Stillbirth.
Increased urea See urea. Indicates a higher than normal excretion of urea in urine.
Increased vasopasticity Enhanced constriction of blood vessels.
B-19
-------
Inflammation A localized protective response elicited by injury or destruction of
tissues, which serves to destroy, dilute or wall off (sequester) both the injurious agent
and the injured tissue. Histologically, it involves a complex series of events, including
dilatation of arterioles, capillaries and venules, with increased permeability and blood
flow, exudation of fluids, including plasma proteins and leukocytic migration into the
inflammatory focus.
Infiltration The diffusion or accumulation in a tissue or cells of substances not normal
to it or in amounts of the normal. Also, the material so accumulated. See Macrophage
infiltration.
Inotropy Muscular contractions.
Interstitial bronchiole pneumonia See Bronchiopneumonia. Indicates effect is seen in
the interspaces of the lung tissue.
Interstitial lung disease A heterogeneous group of noninfectious, nonmalignant
disorders of the lower respiratory tract, affecting primarily the alveolar wall structures but
also often involving the small airways and blood vessels of the lung parenchyma.
"Interstitial" refers to the fact that the interstitium of the alveolar walls is thickened,
usually by fibrosis. This group of diseases is usually inflammatory.
Intraepidermal carcinoma See Carcinoma. Indicates effect is seen within the
epidermis.
Intromission Insertion; introduction.
Initial body weight loss See Weight loss.
Injury Result of assault by an external force, organic or physiologic dysfunction, or a
pathogen.
Intestinal hyperemia See Hyperemia. Congestion of the blood in the intestines.
Irritation of the eyes See Irritation. Indicates effect is seen in the eye.
Irritation Local inflammation of cutaneous or mucosal surfaces.
Ischemic heart disease Disease of the heart characterized by a low oxygen state
usually due to obstruction of the arterial blood supply or inadequate blood flow leading
to hypoxia in the tissue.
Karyomegaly The condition of a cells nucleus being abnormally enlarged (i.e., for
reasons other than it being polyploid).
B-20
-------
Keratosis A skin lesion that is abnormally sensitive to the effects of ultraviolet light
(sunlight). Thought to be a precancerous skin lesion that is more common in the fair-
skinned or elderly individual. Usually a discrete slightly raised, red or pink lesion
located on a sun-exposed surface. Texture may appear as rough, gritty or scaly.
Labored breathing See Gasping.
Lesions Any pathological or traumatic discontinuity of tissue or loss of function of a
part.
Lassitude Weakness, exhaustion.
Leukocytosis A term used to describe an abnormal elevation on the white blood cell
count. Elevated counts can be seen in cases of inflammation and infection.
Leukoderma An acquired disorder that selectively destroys (or that results in the
selective disappearance) of some or all melanocytes residing in the interfollicular
epidermis and occasionally in the follicle as well. The mechanism(s) by which the
melanocytes are lost (or by which melanocytes are made to disappear) may be multiple
but are not yet identified unequivocally.
Leukopenia Abnormal decrease in the number of white blood cells.
Lethal Dose 50 The amount, or dosage, of a toxin necessary to kill 50% of the
experimental subjects.
Leydig cell tumor The most common nongerminal tumor of the testis, derived from
the leydig cells. It is rarely malignant. This tumor appears among 1-3% of testicular
tumors and although they may be seen in children, the median age of appearance is 60
years. They are sometimes seen in women as ovarian tumors. Clinically, symptoms
are usually related to the endocrine abnormalities induced by this tumor.
Lipid peroxidation Peroxidase-catalyzed oxidation of lipids using hydrogen peroxide
as an electron acceptor.
Loss of circulation Reduced oxygen supply to cells, organs, or parts.
Loss of dexterity Decrease in readiness and grace in physical activity; decrease in
skill and ease in using the hands.
Lung irritation See Irritation. Indicates effect is manifested in the lung.
Lymphoma Malignant tumor of lymphoblasts derived from B lymphocytes.
Lysosomal inclusions Accumulations of the undigested substrate within cells caused
by an enzyme deficiency.
B-21
-------
MCH (Mch4 proteaseAn) An enzyme. An aspartate-specific cysteine protease
containing two fadd-like domains.
Macrocytic anemia See Anemia. Indicates the effect is caused by enlarged red
blood cells.
Macrophage infiltration See Infiltration. Indicates effect is an accumulation of
macrophages.
Melanoderma Abnormal blackness of skin.
Melanosis A disorder caused by a disturbance in melanin pigmentation; melanism.
Melena Bloody or dark black or tarry bowel movements.
Memory loss Disturbances in registering an impression, in the retention of an
acquired impression or in the recall of an impression.
Mental sluggishness Delayed reactions or fatigue arising in consequence of mental
effort.
Metabolism alterations See Alterations. Indicates the effect is manifested in
metabolic processes; may reflect and increase or decrease in metabolism.
Metaplasia The change in the type of adult cells in a tissue to a form that is not
formal for that tissue.
Methemoglobinemia The presence of methemoglobin in the blood, resulting in
cyanosis. A small amount of methemoglobin is present in the blood normally, but injury
or toxic agents convert a larger proportion of hemoglobin into methemoglobin, which
does not function reversibly as an oxygen carrier.
Microgranuloma See Granuloma. Indicates the effect is small, little.
Mineralization Production of bone minerals from collagen, important in the
progressive growth and development of normally calcifying bone, cartilage, tendon,
dentin and cementum among vertebrate tissues. Collagen represents the principal
organic component in such tissues and it strictly mediates the nucleation, growth, and
development of the mineral, a calcium phosphate salt (apatite). The interaction
between collagen and mineral leads to a composite tissue having improved strength
and biomechanical properties different from those of either component separately
considered. Conversely, changes in collagen content, assembly or aggregation could
have profound effects on mineralization and subsequently on the nature of tissue
integrity and mechanical behavior.
Miscarriage See Abortion.
B-22
-------
Mitochondria! Respiration Impairment See Impairment. Indicates reduction in the
energy produced in the mitochondria, which are specialized membrane structures within
a cell that provide energy for a cell by the addition of substances acted upon by
enzymes
Mortality See Survival.
Motility Ability of the spermatozoa to move by flagellate swimming.
Muscular hypertrophy see Hypertrophy.
Myelin degeneration See Degeneration. Indicates the effect is seen in the material
making up the myelin sheath of nerve axons.
Narcosis State of unconsciousness.
Nausea An unpleasant sensation, vaguely referred to the epigastrium and abdomen
and often culminating in vomiting. See Also Dyspepsia, Emesis, Vomiting.
Necrosis Death of a tissue.
Nephrosis A type of nephritis that is characterized by low serum albumin, large
amount of protein in the urine and swelling (edema). Swelling, weight gain, high blood
pressure and anorexia are key features. Nephrotic syndrome can be seen with a
number of illness that cause damage to the kidney glomerulus. Examples include
diabetes, hereditary disorders, lupus, multiple myeloma, amyloidosis,
glomerulonephritis, minimal change disease and membranous glomerulonephritis.
Nephrotoxicity Toxicity to the kidney.
Nerve conduction Neural transport of an electronic impulse.
Neuropathy A general term denoting functional disturbances and/or pathological
changes in the peripheral nervous system. If the involvement is in one nerve it is called
mononeuropathy, in several nerves, mononeuropathy multiplex, if diffuse and bilateral,
polyneuropathy. The etiology may be known for example arsenical neuropathy, diabetic
neuropathy, ischemic neuropathy, traumatic neuropathy) or unknown. Encephalopathy
and myelopathy are corresponding terms relating to involvement of the brain and spinal
cord, respectively. The term is also used to designate noninflammatory lesions in the
peripheral nervous system, in contrast to inflammatory lesions (neuritis).
Neonatal survival See Perinatal mortality.
Nonspecific brain injury See Injury. Indicates effect is seen in the brain, but specific
etiology or precise effect is unknown.
Nonspecific hepatotoxicity See Hepatotoxicity.
B-23
-------
Not specified Not otherwise specified. No additional information is immediately
available.
Numbness Lacking sensation.
Oliguria Secretion of a diminished amount of urine in relation to the fluid intake.
Ossification The formation of bone or of a bony substance, the conversion of fibrous
tissue or of cartilage into bone or a bony substance.
Organ Weight Gain Increase in the mass of an organ. May indicate injury to the
organ or increase in organ function in response to a stimulus.
Ossification the formation of bone or of a bony substance, the conversion of fibrous
tissue or of cartilage into bone or a bony substance. See Delayed ossification; Reduced
ossification.
Osteomalacia A condition marked by softening of the bones (due to impaired
mineralization, with excess accumulation of osteoid), with pain, tenderness, muscular
weakness, anorexia and loss of weight, resulting from deficiency of vitamin D and
calcium.
Osteoporosis A reduction in the amount of bone mass, leading to fractures after
minimal trauma.
Pain Sensation of discomfort, distress, or agony.
Pale skin Skin lacking freshness or ruddiness; a sickly whiteness; lack of color or
luster; wanness.
Palmar and plantar keratosis See Keratosis. Indicates effect is seen on palms of
hands and soles of feet.
Palpitations Irregular and violent heartbeats.
Pancreatitis Acute or chronic inflammation of the pancreas, which may be
asymptomatic or symptomatic and which is due to autodigestion of a pancreatic tissue
by its own enzymes.
Paresthesia Paralysis.
Perforation A hole made through a part or substance.
Periocular edema See Edema. Indicates effect is seen around the eyes.
Perinatal mortality Mortality occurring in the period shortly before and after birth, (in
humans defined as beginning with completion of the twentieth to twenty eighth week of
gestation and ending 7-28 days after birth); see also Stillbirth, Abortion, Mortality.
B-24
-------
Peripheral nervous system impairment See Impairment. Indicates effect is seen in
the nerves of the PNS, which connect the central nervous system (CMS) with sensory
organs, other organs, muscles, blood vessels and glands.
Peripheral Pertaining to or situated at or near the periphery, situated away from a
center or central structure.
Persistent extensive hyperkeratosis See Hyperkeratosis. Indicates condition is
widespread and difficult to treat.
Pharyngitis Inflammation of the pharynx.
Pheochromocytoma A tumor of the adrenal gland, which produces catecholamines
(noradrenaline and adrenaline). Although the tumor is usually benign it produces
hypertension, pounding headaches, tachycardia, palpitations, apprehension, facial
flushing, nausea and vomiting.
Pigmentation Coloration, especially abnormally increased coloration, by melanin.
Pigmentation changes Increase or decrease in pigment, especially melanin.
Pigmentation in hepatic macrophages See Pigmentation. Indicates effect is seen in
the liver macrophages, which are relatively long-lived phagocytic cells derived from
blood monocytes.
Pneumonia Inflammation of the lungs with consolidation.
Pneumonitis Inflammation of the lung secondary to viral or bacterial infection.
Portal hypertension Any increase in the portal vein (in the liver) pressure due to
anatomic or functional obstruction (for example alcoholic cirrhosis) to blood flow in the
portal venous system. Indicators of portal hypertension are: esophageal varices,
hemorrhoids, enlarged veins on the anterior abdominal wall (caput Medusae) and
ascites.
Possible vascular complications See Vascular complications.
Production Creation of a product.
Proliferation Increase in numbers; the reproduction or multiplication of similar forms,
especially of cells and morbid cysts.
Prostration Absolute exhaustion.
Proteinuria Too much protein in the urine. This may be a sign of kidney damage.
Pulmonary vasculitis See Vasculitis. Indicates effect is seen in the respiratory tract.
Rales Abnormal breathing sounds heard through a stethoscope.
B-25
-------
Raynaud's disease Paroxysmal (i.e., occurring in spasms or seizures) bilateral
cyanosis of the digits due to arterial or arteriolar contraction.
RBC functional impairment See Impairment. Indicates failure of the red blood cells to
function, primarily resulting in poor oxygen distribution.
Reduced birth rate Fewer live births than expected. See also Stillbirth, Increased
resorptions, Abortion, and Reduced fertility.
Reduced growth rate Failure to gain weight normally. See also Weight gain, Weight
loss.
Reduced clavicle Also called the collar bone, it articulates with the shoulder on one
end (at the acromion process of the scapula) and the sternum (breast bone) on the
other.
Reduced fertility See Fertility. Failure to conceive normally.
Reduced fine motor performance See Impairment. Indicates effect is noted in fine
motor skills.
Reduced glycogen Reduction in the polysaccharide occurring especially in the liver
and muscle, where it is stored as a sugar-supply reserve, capable of complete
conversion to glucose when needed. See also Glycogen level changes.
Reduced heart rate Depressed heart rate.
Reduced litter size See Reduced birth rate.
Reduced lung function See Impairment. Indicates effect is seen on pulmonary
function.
Reduced nerve conduction See Impairment. Indicates effect is seen in nerve
conduction.
Reduced ossification Indicates a reduction in the formation of bone or of a bony
substance, the conversion of fibrous tissue or of cartilage into bone or a bony
substance. See also Delayed Ossification.
Reduced short-term memory See Memory Loss. Indicates effect is manifested in
short-term retention.
Reduced sperm motility See Motility. See also Fertility. Indicates effect is seen in
sperm.
Reduced sperm production See Production. See also Fertility. Indicates effect is
seen in sperm.
B-26
-------
Reduced urinary output Lower volume (whether due to excretion reduction or
concentration of wastes) of urine production. See also Excretion Reduction.
Respiratory tract inflammation See Inflammation. Indicates effect is seen in the
respiratory tract.
Resorption The loss of substance through physiologic or pathologic means.
Respiratory tract injury See Injury. Indicates effect is seen in the respiratory tract.
Retention alterations Changes in the persistent keeping within the body of matters
normally excreted; thus, decreased excretion is also increased retention. See also
Excretion Reduction.
Reticulin sclerosis See Sclerosis. Indicates effect is seen in the reticulin, the
constituent protein of reticulin fibers found in extracellular matrix.
Rhinitis Inflammation of the mucous membrane of the nose.
Rhinorrhea The free discharge of a thin nasal mucus.
Rickets A condition caused by deficiency of vitamin D, especially in infancy and
childhood, with disturbance of normal ossification. The disease is marked by bending
and distortion of the bones under muscular action, by the formation of nodular
enlargements on the ends and sides of the bones, by delayed closure of the fontanelles,
pain in the muscles and sweating of the head.
Scaling Dry patches of skin resembling fish scales. See also Dermatitis.
Scaling of skin See Scaling.
Sciatic and optic nerve injury See Injury. Indicates effect is seen in the sciatic (hip
region) and optic (eye) nerves.
Sclerosis An induration or hardening, especially hardening of a part from
inflammation and in diseases of the interstitial substance. The term is used chiefly for
such a hardening of the nervous system due to hyperplasia of the connective tissue or
to designate hardening of the blood vessels.
Seizures Attacks of cerebral origin consisting of sudden and transitory abnormal
phenomena of a motor, sensory, autonomic or psychic nature resulting from transient
dysfunction of the brain.
Serum phosphate See blood phosphate.
SCOT An enzyme produced by the liver. Elevated levels of SCOT in the blood
indicate a liver problem.
B-27
-------
Skeletal defects See Terata. Indicates skeletal malformation, may be a considered a
(see also) Gross Physical Abnormality.
Skin inflammation See Dermatitis.
Sleep disorders Disturbances of usual sleep patterns or behaviors.
Spasm of digital arteries A sudden but transitory constriction of the arteries of the
digits (e.g., one of the terminal divisions of a limb appendage, such as a finger or toe).
Squamous cell carcinoma See Carcinoma. Indicates effect is seen in the flat thin
cells found in the outer layer of the skin.
Stillbirth Delivery of a dead fetus. See also Abortion.
Stomach adhesions See Adhesions. Indicates effect is seen in the stomach.
Survival Living or continuing living. Decreased survival is increased mortality,
increased death rate.
Swelling of the eyes See Edema. Indicates effect is seen in or near the eyes.
T-cell A class of lymphocytes, so called because they are derived from the thymus
and have been through thymic processing. Involved primarily in controlling cell-
mediated immune reactions and in the control of B-cell development. The T-cells
coordinate the immune system by secreting lymphokine hormones.
Terata Malformation in an embryo; birth defect.
Testicular degeneration or atrophy See Degeneration, Atrophy. Indicates effect is
seen in the testicles.
Thin and dilated coronary arteries See Thinning, Dilation. Indicates effect is seen in
coronary arteries.
Thinning Reduced thickness, as of vessel walls.
Thrombosis The formation, development or presence of a thrombus.
Tingling of hands and feet Detection of a feeling in extremities indicated.
Tonsilitis Inflammation of the tonsil.
Toxic nephrosis Toxicity or destruction observed in kidney cells. See also
Nephrotoxicity.
Tremors An involuntary trembling or quivering.
Trembling See Tremors.
B-28
-------
Tubular degeneration See Degeneration. Indicates effect is seen in kidney tubules.
Ulcer A local defect or excavation, of the surface of an organ or tissue, which is
produced by the sloughing of inflammatory necrotic tissue.
Ulceration See Ulcer. The formation or development of an ulcer.
Ulcerative cecitis Inflammation of the cecum, a blind pouch-like commencement of
the colon in the right lower quadrant of the abdomen at the end of the small intestine.
The appendix is a diverticulum that extends off the cecum.
Urea The final nitrogenous excretion product of many organisms.
Vacuolization Formation into, or multiplication of, vacuoles.
Vacuolization of fasciculata cells in adrenal cortex See Vacuolization. Indicates
effect is seen on the fasciculata cells in adrenal cortex, the outer portion of the fatty
acids that inhibit inflammation in allergic responses.
Vacuolization of pancreas islet cells See Vacuolization. Indicates effect is seen on
the cells of the Islets of Langerhans (or islet cells) within the pancreas.
Vascular complications Complications pertaining to blood vessels or indicative of a
copious blood supply.
Vasculitis Inflammation of a vessel.
Vesiculation The state of containing vesicles, or the process by which vesicles are
formed. A vesicle is a closed membrane shell, derived from membranes either by a
physiological process (budding) or mechanically by sonication.
Viability The quality or state of being viable; specifically, the capacity of living after
birth.
Vibration sensation Detection of a feeling of oscillation.
Vomiting See Emesis. See also Nausea, Dyspepsia.
Wart formation Formation of a benign tumor of basal cell of skin, the result of the
infection of a single cell with wart virus (Papilloma virus). Virus is undetectable in basal
layer, but proliferates in keratinizing cells of outer layers.
Weight gain Increase in body mass.
Weight loss Decrease in body mass.
B-29
-------
Note: These definitions have been adapted from the following sources:
The On-line Medical Dictionary (c) Academic Medical Publishing & CancerWEB
1997-98. Available at
http://www.betterhealth.vic.gov.au/bhcv2/bhcsite.nsf/pages/bhc medicaldictionarv?open
document. Accessed July-September 2001. Distributed by CancerWEB under license
from Academic Medical Publishing.
The New Lexicon: Webster's Dictionary of the English Language. 1989 edition.
Lexicon Publications, Inc., New York, NY.
ATSDR (Agency for Toxic Substances and Disease Registry). 2000a. Toxicological
Profile for Arsenic (Update). September.
E-Doc (Electronic Doctor) Index of Medical Terminology, (c) E-Doc 1998-99. Available
at http://www.edoc.co.za/.
B-30
-------
TABLE B-1
Example Noncancer Data Table: Primary Effects from Oral Exposures3
Chemical
Arsenic (inorganic)
(As)
Beryllium (Be)
Bromodichloro-
methane (BDCM)
Cadmium (Cd)
Carbon tetrachloride
(CCI4)
Chromium III
(insoluble salts) (Crlll)
Chromium VI
(Cr VI)
Dichloroacetic Acid
(DCA)
Mercury (based on
mercuric chloride) (Hg)
Primary
System/Organ
Affected
Skin, cardiovascular
system
Gastrointestinal
system
Kidney,
Developing fetus
Kidney
Liver
Liver, spleen
No observed effect
Reproductive
system, Developing
fetus, Liver, Brain
Kidney
Primary Noncancer Effect
Hyperpigmentation, keratosis,
possible vascular complications
Small intestinal lesions
Renal cytomegaly
Proteinuria
Lesions (mild centrilobular
vacuolization, increased serum
sorbitol dehydrogenase activity)
Decreased organ weights
No observed effect
Lesions in the testes, cerebrum,
cerebellum, liver
Autoimmune glomerulonephritis
Primary Effect
LOAEL
(mg/kg-day)
0.014
Not established
(benchmark dose is
0.46)b
17.9
Not established
(NOAEL is 0.005
[water], 0.01 [food])
7.1
Not established
(NOAEL is 1468)
Not established
(NOAEL is 2.5)
12.5
0.317
Oral RfD
(mg/kg-day)
0.0003
0.002
0.02
0.0005 (water)
0.001 (food)
0.0007
1.5
0.003
0.004
0.0003
Oral RfD Combined
Uncertainty Factor
3
300
1000
10
1000
900
1000
3000
1000
B-31
-------
TABLE B-1 cont.
Chemical
Nickel (soluble salts)
(Ni)
Nitrate (NO3)
Nitrite (NO2)
Polychlorinated
Biphenyls (PCBs)
(Arochlor1016)
Trichloroethylene3
(TCE)
Uranium (soluble
salts) (U)
Zinc (Zn)
Primary
System/Organ
Affected
Kidney, liver, spleen
Blood
Blood
Reproductive
system, Brain
Liver, kidney, and
developing fetus
Kidney
Blood
Primary Noncancer Effect
Decreased body and organ
weights
Methemoglobinemia
Methemoglobinemia
Reduced birth weights
Disruption of cellular processes
through multiple metabolites and
mechanisms in liver, kidney, fetus
Initial body weight loss, moderate
nephrotoxicity
47% decrease in erythrocyte
superoxide dismutase
concentration (adult females after
10-week exposure)
Primary Effect
LOAEL
(mg/kg-day)
50
1.8-3.2
11-20 ppm
0.028
Not established
2.8
0.91
Oral RfD
(mg/kg-day)
0.02
1.6
0.1
0.00007
Not
established
0.003
0.3
Oral RfD Combined
Uncertainty Factor
300
1
10
100
Not established
1000
3
3 Source: U.S. EPA (2007). Note: users should always check with IRIS for current toxicological qualitative evaluations and reference values
b The benchmark dose is a BMD10 value, i.e., the dose at the 95% confidence limit of the dose-response model corresponding to a 10% increase
in incidence of these effects compared with controls.
Acronyms and abbreviations are defined as follows: LOAEL = lowest-observed-adverse-effect level; mg/kg-day = milligram per kilogram body
weight per day; NOAEL = no-observed-adverse-effect level; RfD = reference dose.
B-32
-------
TABLE B-2
Example Noncancer Data Table: Primary Effects from Inhalation Exposures3
Chemical
Arsenic (inorganic)
Beryllium
Cadmium
Chromium III
(insoluble salts)
Chromium VI
(dissolved aerosols,
chromic acid mists)
Chromium VI
(particulates)
Copper
Mercury
Nickel (soluble salts)
Nitrate
Primary System/
Organ Affected
Not established
Lung
Not established
Not established
Respiratory system
Respiratory system
Not established
Central nervous
system
Not established
Not established
Primary Noncancer Effect
No observed effect
Beryllium sensitization,
progression to chronic
beryllium disease
No observed effect
No observed effect
Atrophy of the nasal
septum
Lactate dehydrogenase in
bronchoalveolar lavage
fluid, indicating
inflammation and injury
No observed effect
Hand tremor, increases in
memory disturbance
No observed effect
No observed effect
LOAEL for Primary Effect
(mg/m3)
Not established
0.0002
Not established
Not established
0.000714
Not established
(benchmark dose is
0.034)b
Not established
0.009
Not established
Not established
Inhalation RfC
(mg/m3)
Not established
0.00002
Not established
Not established
0.000008
0.0001
Not established
0.0003
Not established
Not established
Inhalation RfC
Combined
Uncertainty Factor
Not established
10
Not established
Not established
90
300
Not established
30
Not established
Not established
B-33
-------
TABLE B-2 cont.
Chemical
Nitrite
Trichloroethylene
Uranium (soluble
salts)
Zinc
Primary System/
Organ Affected
Not established
Central nervous
system, liver and
endocrine system
Not established
Not established
Primary Noncancer Effect
No observed effect
Adverse effects on central
nervous system
No observed effect
No observed effect
LOAEL for Primary Effect
(mg/m3)
Not established
Not established
Not established
Not established
Inhalation RfC
(mg/m3)
Not established
Not established
Not established
Not established
Inhalation RfC
Combined
Uncertainty Factor
Not established
Not established
Not established
Not established
a Source: U.S. EPA (2007). Note: users should always check with IRIS for current toxicological qualitative evaluations and reference values
b The benchmark dose is a BMD10 value, i.e., the dose at the 95% confidence limit of the dose-response model corresponding to a 10% increase
in incidence of these effects compared with controls.
Acronyms and abbreviations are defined as follows: LOAEL = lowest-observed-adverse-effect level. In some cases this reflects an adjusted value
(e.g., for beryllium, the study LOAEL was adjusted to account for inhalation rate and days exposed); mg/m3 = milligram per cubic meter (air); RfC =
reference concentration.
B-34
-------
TABLE B-3
Example Noncancer Data Table: Comparison of Selected Secondary Effect Levels to Reference Doses for Oral Exposures3
Chemical
Arsenic
(inorganic)
Type of
Level
RfD
Lowest
human
NOAEL
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
0.0003
0.0004
0.0004
0.0008
0.025
0.8
Ratio
to RfD
1
1.3
1.3
2.7
83
2670
Study Basis
NOAEL of 0.0008 mg/kg-day;
LOAEL of 0.014 mg/kg-day;
human study; UF 3;
(inorganic)
Chronic drinking water study,
continuous exposure
(inorganic)
Chronic drinking water study;
continuous exposure
(pentavalent arsenic)
Chronic drinking water study
(test compound not reported)
Rat gavage study (7 months)
(arsenic solution)
Dog oral study (26 weeks)
(trivalent)
Organ/System Effect
Skin - hyperpigmentation, keratosis;
possible vascular complications
Skin - lesions; abnormal nerve
conduction
Skin - pigmentation changes,
hyperkeratosis; Gl system -
nausea, diarrhea
Skin - hyperpigmentation,
hyperkeratosis
No increased embryonic effects;
infrequent slight expansion of
ventricles of the cerebrum, renal
pelvis, urinary bladder
Liver- mild increase in serum
ALT/AST
Reference
U.S. EPA, 2007
Cebrian etal., 1983
(cited in U.S. EPA,
2007 and ATSDR,
2000a)
Cebrian etal., 1983
(cited in ATSDR,
2000a)
Foy etal., 1992
(cited in ATSDR,
2000a)
Nadeenko et al.,
1978 (cited in U.S.
EPA, 2007 and
ATSDR, 2000a)
Neigerand Osweiler,
1989 (cited in
ATSDR, 2000a)
B-35
-------
TABLE B-3 cont.
Chemical
Beryllium
Type of
Level
RfD
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
0.002
Not
reported
Not
reported
0.7
0.7
12
Ratio
to RfD
1
NA
NA
350
350
6000
Study Basis
BMD10 of 0.46 mg/kg-day;
dog oral study; in food;
UF 300; (sulfate tetrahydrate)
NA
NA
Rat oral study; in water
(3 years) (sulfate)
Rat oral study; drinking water
(91 days) (sulfate)
Dog oral study; in food
(172 weeks) (sulfate)
Organ/System Effect
Multiple target organs; small
intestinal lesions.
NA
NA
Various organ systems (e.g.,
cardiovascular, endocrine, hepatic,
renal, respiratory)
Whole body - no effects
Gl system - ulcerative, inflammatory
lesions; hematopoetic system -
erythroid hypoplasia of bone
marrow; whole body -weight loss,
increased mortality
Reference
U.S. EPA, 2007
NA
NA
Schroederand
Mitchener, 1975
(cited in ATSDR,
2002b)
Freundt and Ibrahim,
1990 (cited in
ATSDR, 2002b)
Morgareidge et al.,
1976 (cited in
ATSDR, 2002b)
B-36
-------
TABLE B-3 cont.
Chemical
Cadmium
Cadmium
Type of
Level
RfD-
water
RfD-
food
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
0.0005
0.001
0.0021
0.0078
0.0081
0.001
Ratio
toRfD
1
1
2.1
7.8
16
2
Study Basis
NOAEL of 0.005 mg/kg-day
(water); human study; UF 10
NOAEL of 0.01 mg/kg-day
(food); human study; UF 10
Chronic lifetime exposure in
food (test compound not
reported)
Chronic oral study (25 years)
(inorganic)
Rat chronic oral study
(5 months);
in water (chloride)
Rat chronic oral study
(18 months);
in water (acetate)
Organ/System Effect
Kidney - proteinuria (note:
supporting data have been derived
from many animal and human
studies, renal effects, proteinuria
and calcium pharmacokinetic
parameters)
Kidney - proteinuria (note:
supporting data have been derived
from many animal and human
studies, renal effects, proteinuria
and calcium pharmacokinetic
parameters)
Kidney - no effects
Kidney - renal tubule interstitial
lesions
Whole body - no effects
Cardiovascular system-
hypertension;
increase in systolic blood pressure
Reference
Data from U.S. EPA,
2005b (effect type
note from RAIS,
1991)
Data from U.S. EPA,
2007 (effect type
note from RAIS,
1991)
Nogawaet al., 1989
(cited in ATSDR,
1999b)
Shiwen etal., 1990
(cited in ATSDR,
1999b)
Perry etal., 1989
(cited in ATSDR,
1 999b)
Koppetal., 1982
(cited in ATSDR,
1 999b)
B-37
-------
TABLE B-3 cont.
Chemical
Carbon
tetrachloride
Type of
Level
RfD
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
0.0007
Not
reported
Not
reported
1.0
10
Ratio
to RfD
1
NA
NA
1430
14,300
Study Basis
NOAEL of 0.71 mg/kg-day;
LOAEL of 7.1 mg/kg-day; rat
gavage study (12 weeks);
UF1000
NA
NA
Rat gavage study (12 weeks)
Rat gavage study (12 weeks)
Organ/System Effect
Liver- lesions (mild centrilobular
vacuolization and increases in
serum sorbitol
dehydrogenase activity)
NA
NA
Liver- substantially elevated
sorbitol dehydrogenase; mild
centrilobular vacuolization
Liver- substantially elevated
sorbitol dehydrogenase; mild
centrilobular vacuolization
Reference
U.S. EPA, 2007
NA
NA
Bruckner et al., 1986
(cited in ATSDR,
2003a)
Bruckner etal., 1986
(cited in ATSDR,
2003a)
B-38
-------
TABLE B-3 cont.
Chemical
Chromium III
(insoluble
salts)
Type of
Level
RfD
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
1.5
Not
reported
Not
reported
0.46
5.0
Ratio
to RfD
1
NA
NA
0.31
3.3
Study Basis
NOAEL of 1468 mg/kg-day;
rat chronic oral study;
UF 1000 (chronic oxide)
NA
NA
Rat chronic drinking water
study
(2-3 years) (trivalent)
Mouse drinking water study
(12 weeks) (trivalent)
Organ/System Effect
Liver and spleen - decreased organ
weights
NA
NA
Cardiovascular system, liver,
kidney, whole body - no effects
Reproductive system - increased
testes, decreased preputial gland
weights; decreased number of
implantations and viable fetuses;
increased ovarian, decreased
uterine weights; whole body -
decrease in body weight gain
Reference
U.S. EPA, 2007
NA
NA
Schroederet al.,
1965 (cited in
ATSDR, 2000b)
Elbetieha and
AI-Hamood, 1997
(cited in ATSDR,
2000b)
B-39
-------
TABLE B-3 cont.
Chemical
Chromium VI
Chromium VI
Type of
Level
RfD
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
0.003
Not
reported
0.57
1.1
3.5
Ratio
to RfD
1
NA
190
367
1170
Study Basis
NOAEL of 2.5 mg/kg-day; rat
chronic drinking water study
(1 year); UF 1000 (potassium
chromate)
NA
Unspecified environmental
exposure (hexavalent)
Mouse oral study, in food
(9 weeks) (hexavalent)
Mouse oral study, in food
(9 weeks) (hexavalent)
Organ/System Effect
No effects
NA
Gl system - oral ulcers, diarrhea,
vomiting abdominal pain;
hematopoetic system -
leukocytosis, immature neutrophils
Liver- cytoplasmic vacuolization of
hepatocytes
Liver- cytoplasmic vacuolization of
hepatocytes
Reference
U.S. EPA, 2007
NA
Zhang and Li, 1987
(cited in ATSDR,
2000b)
NTP, 1996 (cited in
ATSDR, 2000b)
NTP, 1996 (cited in
ATSDR, 2000b)
B-40
-------
TABLE B-3 cont.
Chemical
Mercury
Type of
Level
RfD
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
0.0003
0.0005
0.0012
0.05
0.05
Ratio
to RfD
1
1.67
4
167
167
Study Basis
LOAEL of 0.317 mg/kg-day;
rat study; UF1000;
(mercuric chloride)
Oral study (methyl mercury)
Oral study, food
(methylmercuric chloride)
Rat oral study, food (52 days)
(methylmercuric chloride)
Monkey oral study, water
(328-907 days)
(methylmercury hydroxide)
Organ/System Effect
Kidney - autoimmune
glomerulonephritis; assumes the
oral absorption of divalent mercury
is 7% and absorption from
subcutaneous exposure is 100%
Developmental - no effects
Developmental -
delayed walking, abnormal motor
scores
Developmental - increased
incidence of eye defects in fetuses
Developmental - impaired visual
recognition memory in offspring
Reference
U.S. EPA, 2007
Myers etal., 1997
(cited in ATSDR,
1999c)
Cox etal., 1989
(cited in ATSDR,
1999c)
Khera and
Tabacova, 1973
(cited in ATSDR,
1999c)
Gunderson et al.,
1 988 (cited in
ATSDR, 1999c)
B-41
-------
TABLE B-3 cont.
Chemical
Nickel
(soluble
salts)
Type of
Level
RfD
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
0.02
0.02
Not
reported
0.97
0.23
Ratio
to RfD
1
1
NA
48
12
Study Basis
NOAEL of 5 mg/kg-day;
LOAEL of 50 mg/kg-day; rat
study; in food; UF 300
Oral study; water (178 days)
(sulfate)
NA
Rat oral study; water (28
days) (chloride)
Rat oral study; water (28
days) (chloride)
Organ/System Effect
Multiple target organs;
changes in body and organ weights
Dermal - no effects
NA
Hematopoetic system - no effects;
liver- no effects
Whole body - decreased body
weight gain;
metabolic system effects
Reference
U.S. EPA, 2007
Santucci etal., 1994
(cited in ATSDR,
2003b)
NA
Weischeretal., 1980
(cited in ATSDR,
2003b)
Weischeretal., 1980
(cited in ATSDR,
2003b)
B-42
-------
TABLE B-3 cont.
Chemical
Nitrate
Type of
Level
RfD
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
1.6
3.7
3.2
20
60
Ratio
to RfD
1
2.3
2
12
38
Study Basis
NOAEL of 1 .6 mg/kg-day;
human study; LOAEL of 1 .8-
3.2 mg/kg-day; (infants,
drinking water in formula);
UF1
Oral study, 1- to 6-month-old
infants; nitrate in formula
Oral study, 8-day to 5-month-
old infants; nitrate in formula
Oral rat drinking water study
(2 years) (sodium nitrite)
Rat oral study, drinking water
(2 years) (sodium nitrite)
Organ/System Effect
Hematopoetic system -
methemoglobinemia
Hematopoetic system -
no methemoglobinemia clinical
signs
Hematopoetic system -
cyanosis, methemoglobinemia
Respiratory system -
dilated bronchi, fibrosis,
emphysema
Lung - dilated bronchi, fibrosis and
emphysema,
Circulatory/cardiovascular system -
fibrosis, degenerative foci
Reference
U.S. EPA, 2007
Simon etal., 1964
(cited in U.S. EPA,
2007)
Bosch etal., 1950
(cited in U.S. EPA,
2007)
Shuval and Gruener,
1972 (cited in
U.S. EPA, 2007)
Shuval and Gruener,
1972 (cited in RAIS,
1995)
B-43
-------
TABLE B-3 cont.
Chemical
Nitrite
Trichloro-
ethylene
Type of
Level
RfD
Lowest
human
NOAEL
Lowest
human
LOAEL
RfD
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
0.1
1.0
Not
reported
NA
Not
reported
Not
reported
18
0.18
Ratio
to RfD
1
10
NA
NA
NA
NA
60,000
600
Study Basis
NOEL of 1 .0 mg/kg-day;
LOAEL of 1.1-2.0 mg/kg-day;
human study;
UF 10 (from nitrate)
Oral study, infants, nitrate in
formula
NA
NA
NA
NA
Mouse drinking water study
(6 months)
Rat drinking water study,
gestational (3 months)
Organ/System Effect
Hematopoetic system -
methemoglobinemia
Hematopoetic system -
methemoglobinemia above 10%
NA
Various effects - liver; kidney;
developing fetus
NA
NA
Gl - gas pockets in the intestinal
coating;
blood in the intestines
Developmental -
increased fetal heart abnormalities
Reference
U.S. EPA, 2007
Walton, 1951 (cited
in U.S. EPA, 2007)
NA
ATSDR, 1997c
U.S. EPA, 2000I
NA
NA
Tucker etal., 1982
(cited in ATSDR,
1 997c)
Dawson etal., 1993
(cited in ATSDR,
1997c)
B-44
-------
TABLE B-3 cont.
Chemical
Uranium
(soluble
salts)
Type of
Level
RfD
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
0.003
Not
reported
Not
reported
0.06
0.05
Ratio
to RfD
1
NA
NA
20
17
Study Basis
LOAEL of 2.8 mg/kg-day,
rabbit dietary study; UF 1000;
(30 days) (uranyl nitrate
hexahydrate; soluble salt)
NA
NA
Rat drinking water study
(91 days) (uranyl nitrate
hexahydrate)
Rabbit drinking water study
(91 days) (uranyl nitrate
hexahydrate)
Organ/System Effect
Kidney - moderate nephrotoxicity;
whole body - initial body weight loss
NA
NA
Endocrine system - multi-focal
reduction of follicularsize; increased
epithelial height in thyroid;
decreased amount and density of
colloid in males only
Kidney - anisokaryosis, nuclear
vesiculation
Reference
U.S. EPA, 2007
NA
NA
Oilman etal., 1998a
(cited in ATSDR,
1999d)
Oilman etal., 1998b
(cited in ATSDR,
1999d)
B-45
-------
TABLE B-3 cont.
Chemical
Zinc
Type of
Level
RfD
Lowest
human
NOAEL
Lowest
human
LOAEL
Lowest
human
LOAEL
Lowest
animal
NOAEL
Lowest
animal
LOAEL
Value
(mg/kg-
day)
0.3
0.06
0.71
0.71
3.5
0.5
Ratio
to RfD
1
0.2
2.4
2.4
12
1.7
Study Basis
LOAEL of 1 .0 mg/kg-day;
human dietary supplement
study; UF 3
Dietary supplement study
(1 1 weeks) (aspartate)
Dietary supplement study
(12 weeks) (gluconate)
Dietary supplement study
(6 weeks) (gluconate)
Rat gavage study; in water
(20 months) (chloride)
Mouse oral study, in water
(60 days) (acetate)
Organ/System Effect
Hematopoetic system - 47%
decreased ESOD cone, (in adult
females after 10-week exposure)
Developmental - no effects
Liver - decreased serum
HDL-cholesterolb
Hematopoetic system -
decreased ESOD activity
Reproductive effects -
decreased live pups per litter
Nervous system -
increase in latency in inhibitory
avoidance test
Reference
Yadricketal., 1989
(cited in U.S. EPA,
2007)
Kynast and Saling,
1986 (cited in
ATSDR, 2003c)
Black etal., 1988
(cited in ATSDR,
2003c)
Fischer etal., 1984
(cited in ATSDR,
2003c)
Khan etal. ,2001
(cited in ATSDR,
2003c)
De Oliveira et al.,
2001 (cited in
ATSDR, 2003c)
Note: users should always check with IRIS for current tox qualitative evaluations and reference values. This table presents information for
15 chemicals selected for study at a contaminated site. The form of the chemical or compound used in the toxicity study that served as the
basis for the indicated level is given in parentheses; where not listed here, the chemical itself was identified as the test chemical. Selected
acronyms are defined as follows; others (e.g., EPA acronyms) are included in the notation at the front of this report. ALT/AST = alanine
aminotransferase/aspartate aminotransferase; BMD10 = benchmark dose, at the 95% confidence limit of the dose-response model
corresponding to a 10% increase in incidence of the effect compared with the control; ESOD = erythrocyte superoxide dismutase; Gl =
gastrointestinal system; HDL = high-density lipid; LOAEL = lowest-observed-adverse-effect level; mg/kg-day = milligram per kilogram per day;
NA = not available/not applicable; NOAEL = no observed adverse effect level; RfD = reference dose; UF = uncertainty factor.
Low levels of low-density lipoprotein (LDL) cholesterol put a person at a high risk of heart disease. Taken from The American Heart Association
"What are Healthy Levels of Cholesterol?" See http://www.americanheart.org/presenter.ihtml?identifier=183.
B-46
-------
APPENDIX C
SEVERITY OF TOXIC EFFECT
C.1. OVERVIEW
In 1980, EPA incorporated the judgment of severity of toxic effect into the
Ambient Water Quality Criteria (AWQC) methods (U.S. EPA, 1980). Four categories of
severity were proposed by EPA as distinguished by the terms:
No-Observed-Effect Level (NOEL)
No-Observed-Adverse-Effect Level (NOAEL)
Lowest-Observed-Adverse-Effect Level (LOAEL)
Frank-Effect Level (FEL).
EPA (1980) also described the choice of uncertainty factor when a LOAEL was used as
the basis of the Acceptable Daily Intake (ADI) (now referred to as the Reference Dose
(RfD), where more severe LOAELs yielded larger safety factors (now called uncertainty
factors). This judgment also determined whether a LOAEL was really an FEL. If the
LOAEL was judged to be an FEL, then an RfD was not estimated because the data
base was then judged to not have fully explored the threshold region of toxicity, that is
the region where the severity of toxic effect was minimal.
As it was originally envisioned the choice of uncertainty factor (UF) with the use
of a LOAEL was to be greater than 1 and up to and including 10, where minimal effects
at the LOAEL would be associated with generally a UF of >1 to 5 and more severe or
extensive tissue damage at the LOAEL would generally warrant a >5 to 10-fold UF. In
practice, however, EPA scientists generally restricted their choices to either 3 or 10 (see
U.S. EPA, 2007 for numerous examples). This is because of the difficultly in
distinguishing the relative nature of severity within a given organ and among organs,
and the corresponding difficulty in being precise with choices of UF.1
This early EPA (1980) distinction is evident today, as more fully discussed in
Haber et al. (2001) where dose-response processes for noncancer toxicity depend in
part on professional judgment as to whether an effect or collection of effects observed
1 Current discussions of severity often broaden the concept beyond that related to the scientific judgment
in the types of endpoints at the LOAEL. This broader concept of severity includes relative judgments of
severity among target organs (an exceedingly difficult task), and societal judgments about what is or is
not a severe effect and whether or not extra uncertainty factors should be used for certain endpoints.
While this broader interpretation is important, the development of noncancer health risk assessment
values, such as an RfD, are based on the more limited interpretation and use of the severity conceptas
in the 1980 guidelines. While the severity among organ systems is an area of present study and scientific
debate, the societal interpretation of severity is best left to the Risk Characterization step of risk
assessment/risk management.
C-1
-------
at any given dose of a chemical constitutes an adverse response. As before, such
judgment may not be easily rendered, and requires experts trained in the area. For
example, Figure C-1 taken from Haber et al. (2001) shows individual disability as a
function of organ system impairment, and the overlapping areas of adverse and
nonadverse effects. Table C-1, also taken from Haber et al. (2001), more clearly
describes some of the terms shown in Figure C-1, as well as some other key terms for
hazard identification.
While this figure and table are useful tools to show the broad concept of
adversity, the analysis of adversity for a given chemical or situation is strictly a case-by-
case analysis by experts. For example, a chemical often elicits more than one toxic
effect, even in one species, or in tests of the same or different duration. After assessing
the quality of each study, identifying the biological and statistical significance of
observed effects (discussed by Haber et al., 2001 and U.S. EPA, 2002e), and
distinguishing between reversible and irreversible endpoints (also discussed in Haber et
al., 2001), risk assessment scientists often identify the critical effect(s).
The critical effect(s) is the first adverse effect(s) or its known and immediate
precursor that occur as dose rate increases in a study. When several studies are
compared, the critical effect is generally the lowest one that occurs collectively. Current
dose-response methods described in this text and elsewhere use the critical effects as a
basis for the dose-response assessment. The critical effects may change among
toxicity studies of different durations, may be influenced by toxicity in other organs, and
may differ depending on the availability of data on the shape of the dose-response
curve.
Where specific guidance on hazard identification is not available, some general
considerations regarding the types of toxicity evidence and adversity of effect are
needed. Towards this end, risk assessment scientists look at the available data in
several different ways, as outlined below. The following considerations illustrate some
broad concepts of hazard identification applicable for all organ systems.
C.2. RANKING TOXIC EFFECT
Several schemes are available for ranking the severity of toxic effects. One of
the first schemes for noncancer toxicity was developed by EPA scientists in the
evaluation of reportable Quantities under the Superfund legislation (DeRosa et al.,
1985). These scientists ranked the increasing severity of noncancer health effects as
shown in Table C-2.
C-2
-------
Death
3 Disease
T3
"ro
3
5
"> Disturbed
T3 Function
Health
Adverse effect
Increasing severity
Non-adverse effect
Homeostasis Compensation Breakdown Failure
and adaptation
Organ System Impairment
Adapted from Patty's Toxicology, 2001, Chapter 5
FIGURE C-1
Individual Disability as a Function of Organ System Impairment
C-3
-------
TABLE C-1
Some Key Definitions for Hazard Identification
'ADAPTIVE EFFECT enhances an organism's performance as a whole and/or its
ability to withstand a challenge. An increase in liver weight due to an increase in
hepatic smooth endoplasmic reticulum is an example of an adaptive effect, if hepatic
metabolism reduces the chemical's toxicity.
* COMPENSATORY EFFECT maintains overall function without enhancement or
significant cost. Increased respiration due to metabolic acidosis is an example of a
compensatory effect.
'CRITICAL EFFECT is the first adverse effect, or its known precursor, that occurs as
dose rate or exposure level increases. One or more effects may be critical.
*ADVERSE EFFECT is a biochemical change, functional impairment, or pathological
lesion that impairs performance and reduces the ability of an organism to respond to
additional challenge. The determination of such effects may require special tests or
observation, such as preparation of slides for histological analysis.
*FRANK EFFECT is an unmistakable adverse effect, such as convulsions or mortality.
The determination of frank effects can be done by clinical observation and normally
does not require special tests.
'SEVERITY connotes the toxicological significance attached to the continuum of
effects, including adaptive, compensatory, critical, adverse, and frank effects, potentially
associated with exposure of xenobiotics.
Source: Haberetal. (2001).
C-4
-------
TABLE C-2
Rating Values for NOAELs, LOAELs and FELs Used to Rank Chronic Toxicity
Rating Effects
1 Enzyme induction or other biochemical change with no pathologic change and no
change in organ weights.
2 Enzyme induction and subcellular proliferation or other changes in organelles but no
other apparent effects.
3 Hyperplasia, hypertrophy, or atrophy, but no change in organ weights.
4 Hyperplasia, hypertrophy, or atrophy with changes in organ weights.
5 Reversible cellular changes: cloudy swelling, hydropic change, or fatty changes.
6 Necrosis, or metaplasia with no apparent decrement in organ function. Any
neuropathy without apparent behavioral, sensory, or physiologic changes.
7 Necrosis, atrophy, hypertrophy, or metaplasia with a detectable decrement in organ
functions. Any neuropathy with a measurable change in behavioral, sensory, or
physiologic activity.
8 Necrosis, atrophy, or metaplasia with definitive organ dysfunction. Any neuropathy
with gross changes in behavior, sensory, or motor performance. Any decrease in
reproductive capacity. Any evidence of fetotoxicity.
9 Pronounced pathologic changes with severe organ dysfunction. Any neuropathy
with loss of behavioral or motor control or loss of sensory ability. Reproductive
dysfunction. Any teratogenic effect with maternal toxicity.
10 Death or pronounced life-shortening. Any teratogenic effect without signs of
maternal toxicity.
Source: DeRosa et al. (1985).
C-5
-------
Scientists with the ATSDR use a scheme for ranking of severity of toxic effect
that is intermediate between the EPA (1980) version and the one used for Reportable
Quantities (DeRosa et al., 1985). This scheme by ATSDR (Pohl and Abadin, 1995) has
the following 5 severity rankings:
No-Observed-Effect Level (NOEL)
No-Observed-Adverse-Effect Level (NOAEL)
Minimal-Lowest-Observed-Adverse-Effect Level (LOAEL!)
Moderate-Lowest-Observed-Adverse-Effect Level (LOAEL2)
Frank-Effect Level (PEL).
In addition, it is still the current practice to use a varying uncertainty factor with a
LOAEL used to estimate an RfD/RfC. The choice of uncertainty factor to be
extrapolated to the NOAEL generally depends on the severity of the effect at the
LOAEL. Sometimes this uncertainty factor is used with the choice of a benchmark dose
(BMD) of moderate severity, although this latter use is not uniform nor universally
accepted. In either case, more severe effects should be judged to need a larger
uncertainty factor because the expected NOAEL is further away from the LOAEL or
BMD. Less severe effects would not require a large factor, because, presumably, the
LOAEL or BMD is closer to the unknown NOAEL (Dourson et al., 1996).
For cancer toxicity, it is recognized that tumors vary in severity, with the most
common distinction made between malignant and benign categories. However, these
distinctions in severity do not generally affect the quantitative dose response
assessment, other than to categorize the evidence that a chemical causes the cancer
endpoint or not. Older categories of cancer evidence from EPA (1986b) are:
Group A: Known human carcinogen. Substances for which "sufficient"
evidence from human epidemiologic studies supports a causal connection
between exposure to the substance and cancer.
Group B1: Probable human carcinogen (limited human evidence). Weight of
evidence of human carcinogenicity based on epidemiologic studies is
"limited."
Group B2: Probable human carcinogen (no human evidence). Substances
for which there is "no data," or "no evidence" from human epidemiologic
studies, but for which the weight of evidence of carcinogenicity based on
animal studies is "sufficient."
Group C: Possible human carcinogen. Substances with "limited" evidence of
carcinogenicity in animals, and "inadequate evidence," "no data" or "no
evidence" from human epidemiologic studies.
Group D: Not evaluated. Not classifiable for human carcinogenicity
(insufficient data).
Group E: Noncarcinogenic. Evidence of noncarcinogenicity in humans (U.S.
EPA, 1986b).
C-6
-------
EPA's newer classification (U.S. EPA, 2005d) also emphasizes categories of
cancer evidence, rather than severity of endpoint. These categories are:
Carcinogenic to humans
Likely to be carcinogenic to humans
Suggestive evidence of carcinogenic potential
Inadequate information to assess carcinogenic potential
Not likely to be carcinogenic to humans.
As with the early guidelines of 1986 these categorizations do not make
distinctions of severity, and thus severity does not generally affect the quantitative dose
response assessment.
C.3. EXAMPLES OF SEVERITY USE FOR QUANTITATIVE DOSE RESPONSE
ASSESSMENT
Work by EPA over a number of years indicates that the regression of toxicity data
viewed as categories of pathological staging is useful for exploring the likely health risk
at doses above the RfD/RfC. This categorical regression has both theoretical support
by Hertzberg and colleagues and practical application for individual chemicals (e.g.,
Hertzberg and Miller, 1985; Hertzberg and Wymer, 1991). For example, Dourson et al.
(1997) used categorical regression for developing risks above the RfD for aldicarb. In
this paper the severity of cholinergic effects was described using the EPA (1980)
breakdown of severity:
Frank effects (FEL): Abdominal pain, nausea and/or vomiting, diarrhea,
seizures, disorientation or confusion, excitation, or mortality
Adverse effects (LOAEL): Brain, whole blood or RBC cholinesterase inhibition
of more than 20%, muscular weakness or twitching, blurred vision and/or
watery eyes, pinpoint pupils, excess salivation, sweating or clamminess,
hyperactivity or altered patterns of locomotion
Non-adverse effects (NOAEL): Whole blood, RBC or plasma cholinesterase
inhibition of less than 20%
No effects (NOEL).
Teuschler et al. (1999) also expanded the use of categorical regressions to
compare the likely risks at exposures above the RfDs for multiple chemicals with
different critical effects. Existing health risk data for diazinon, disulfoton, EPTC,
fenamiphos and lindane were analyzed. As expected, the estimated risks of adverse
effects above the RfD varied among the chemicals. For example, at 10-fold above the
RfD these risks were modeled to be 0.002, 0.0001, 0.0007, 0.002, 0.02, respectively.
The results and impacts of this analysis further indicate that categorical regression is a
useful screening tool to analyze risks above the RfD for specific chemicals, and suggest
C-7
-------
its application in evaluating comparative risks where multiple chemical exposures exist.
Not surprisingly, categorical regression has been cited in a number of methods texts
since these publications and is actively used in a number of situations, including air
toxics evaluations (e.g., Elder etal., 2002; Guthetal., 1997).
C-8
------- |