United States
Environmental Protection
Agency •
Office of Water
4304
EPA-822-B-98-005
July 1998
AMBIENT WATER QUALITY
CRITERIA DERIVATION
METHODOLOGY HUMAN HEALTH
Technical Support Document
Final Draft
-------
-------
AMBIENT WATER QUALITY CRITERIA DERIVATION METHODOLOGY FOR THE
PROTECTION OF HUMAN HEALTH - TECHNICAL SUPPORT DOCUMENT
1. INTRODUCTION , 1
1.1 Background 1
1.2 Need for Revision of the 1980 AWQC National Guidelines 2
1.2.1 Scientific Advances Since 1980 2
1.2.2 EPA Risk Assessment Guidelines Development Since 1980 3
1.2.3 Differing Risk Assessment and Risk Management Approaches for
AWQC and MCLGs 4
1.2.3.1 Group C Chemicals 4
1.2.3.2 Consideration of Non-Water Sources of Exposure 5
1.2.3.3 Cancer Risk Ranges 6
1.3 Purpose of this Document 6
1.4 Criteria Equations 7
1.5 Glossary/Acronyms 9
List of Acronyms Used 9
2. ELEMENTS OF METHODOLOGY REVISIONS AND ISSUES BY TECHNICAL
AREA 11
2.1 Cancer Effects 11
2.1.1 Background on EPA Cancer Assessment Guidelines 11
2.1.1.1 1980 AWQC National Guidelines 11
2.1.1.2 1986 EPA Guidelines for Carcinogenic Risk Assessment 14
2.1.1.3 Scientific Issues Associated with the Current Cancer Risk
Assessment Methodology for the Development of AWQC 16
2.1.2 Proposed Revisions to EPA's Carcinogen Risk Assessment
Guidelines 17
2.1.3 Revised Carcinogen Risk Assessment Methodology for Deriving
AWQC 19
2.1.3.1 Weight-of-Evidence Narrative 19
2.1.3.2 Dose Estimation (by the Oral Route) 20
2.1.3.3 Dose-Response Analysis 22
2.1.3.4 AWQC Calculation 29
2.1.3.5 Risk Characterization 31
2.1.3.6 Use of Toxicity Equivalence Factors (TEF) and Relative
Potency Estimates 31
2.1.4 Case Study (Compound Y, a Rodent Bladder Carcinogen) 32
2.1.4.1 Background and Evaluation for Compound Y 32
2.1.4.2 Conclusion and Use of the MOE Approach for Compound Y .. 33
2.1.4.3 Use of the Default Linear Approach for Compound Y 36
2.1.4.4 Use of the LMS Approach for Compound Y 38
-------
-2.1.4.5 Comparison of Approaches and Results for Compound Y 38
2.1.5 References 39
2.2 Noncancer Effects 40
2.2.1 Introduction 40
2.2.2 Hazard Identification 41
2.2.3 Dose-Response Assessment 42
2.2.4 Selection of Critical Data 43
2.2.4.1 Critical Study 43
2.2.4.2 Critical Data and Endpoint 44
2.2.5 Deriving RfD Using the NOAEL/LOAEL Approach 44
2.2.5.1 Selection of Uncertainty Factors and Modifying Factors 45
2.2.5.2 Confidence in NOAEL/LOAEL-Based RfD 48
2.2.5.3 Presenting the RfD as a Single Point or as a Range 49
2.2.6 Deriving an RfD Using a Benchmark Dose Approach 52
2.2.6.1 Overview of the Benchmark Dose Approach 53
2.2.6.2 Calculation of the RfD Using the Benchmark Dose Method 54
2.2.6.3 Limitations of the BMD Approach 61
2.2.6.4 Example of the Application of the BMD Approach 61
2.2.7 Categorical Regression 66
2.2.7.1 Summary of the Method 66
2.2.7.2 Steps in Applying Categorical Regression 66
2.2.8 Chronic, Practical Nonthreshold Effects 67
2.2.9 Acute, Short-Term Effects 68
2.2.10 Mixtures 68
2.2.11 References 70
2.3 Exposure Analyses 73
2.3.1 Role of Exposure Data in Setting AWQC 73
2.3.2 Exposure Factors in AWQC Algorithms 74
2.3.2.1 Body Weight 77
Chronic Exposure Scenarios 77
Developmental Effects Exposure Scenarios 78
2.3.2.2 Drinking Water Intake 81
Chronic Exposure Scenarios 82
Developmental Effects Exposure Scenarios 84
Inhalation and Dermal Exposure 85
2.3.2.3 Fish Intake Rates 85
Preference #1: Use of Local Information 87
Preference #2: Use of Surveys from Similar Geographic Areas
and Population Groups 89
Preference #3: Use of Distributional Data from National
Food Consumption Surveys 103
Preference #4: Use of Default Intake Rates from the CSFII... 122
2.3.2.4 Incidental Ingestion 123
-------
2.3.3 Quantification of Exposure 125
2.3.4 Consideration of Non-Water Sources of Exposure When Setting
AWQC 127
2.3.4.1 Exposure Decision Tree Approach 129
Problem Formulation 130
Data Adequacy 132
Regulatory Actions 134
Allocation Decisions 134
2.3.4.2 Notes on Use of the Exposure Decision Tree Approach for
Setting AWQC 136
2.3.4.3 Setting AWQC for Chemical X Using the Decision
Tree Approach 140
Sources and Uses of Chemical X 140
Population of Concern 140
Data Used to Assess Exposure to Chemical X 141
Adequacy of Exposure Data 150
Setting AWQC 153
2.3 A A Unavailability of Substances from Different Routes of
Exposure 157
2.3.5 References 158
2.4. Use of BAFs in the Derivation of AWQC 165
2.4.1 Introduction 165
2.4.1.1 Bioaccumulation and Bioconcentration Concepts 166
2.4.2 Definitions 167
2.4.3 Determining BAFs for Nonpolar Organics 172
2.4.4 Estimating Baseline BAFs 173
2.4.4.1 Field-Measured Baseline BAF 173
2.4.4.2 Baseline BAF Derived from Biota-Sediment
Accumulation Factors (BSAFs) 186
2.4.4.3 Baseline BAF Derived from a Laboratory-Measured BCF and
Food-Chain Multiplier 219
2.4.4.4 Baseline BAF from Predicted BCF and Food-Chain
Multiplier 235
2.4.4.5 Metabolism 237
2.4.4.6 Mixtures 237
2.4.5 BAFs Used in Deriving AWQC 238
2.4.5.1 General Equation for an AWQC BAF 238
2.4.5.2 Baseline BAF 238
2.4.5.3 Lipid Content of Aquatic Species Eaten by Humans 239
2.4.6 Determining BAFs for Inorganic Substances 258
2.4.7 Example Calculations 259
2.4.7.1 Example 1: Field-Measured BAF for Chemical M 259
2.4.7.2 Example 2: Laboratory-Measured BCF for Chemical R 263
in
-------
2.4.8 Trophic Level-Specific Fish Consumption Rates 267
2.4.9 References 270
3. MINIMUM DATA CONSIDERATIONS 276
3.1 Background 276
3.1.1 Threshold Effects Guidelines 276
3.1.2 Non-Threshold Effects 277
3.1.2.1 Animal Studies 277
3.1.3 Exposure Assumptions 278
3.2 Minimum Data Considerations in the Federal Register Notice 278
3.2.1 Noncancer - Data Suggestions 278
3.2.1.1 RfD Development (Minimal Data) 278
3.2.1.2 RfD Development (Ideal Situation) 279
3.2.2 Cancer - Data Suggestions 279
3.2.2.1 Minimum Data 279
3.2.2.2 Ideal Situation 280
3.2.3 Exposure - Data Suggestions 280
3.3 Site-Specific Criterion Calculation 280
3.4 Organoleptic Criteria 281
3.5 Criteria for Chemical Classes 281
3.6 Criteria for Essential Elements 282
Appendix A Average Fish Consumption A-l
Appendix B Evaluation of the Quality of Data Set(s) for Use in Deriving an RfD B-l
Appendix C Derivation of Basic Equations Concerning Bioconcentration and Bioaccumulation
of Organic Chemicals C-1
Appendix D Derivation of the Equation Defining ffd D-l
Appendix E Derivation of the Equation to Predict BAF from the BSAF E-l
Appendix F EPA New Draft Protocol for Determining Octanol-water Partition Coefficients (K^)
for Compounds with Log Kow Values > 5 F-l
Appendix G Amount of Commercial Food Items Consumed and Intake of Chemical X from
Commercial Food Items G-l
Appendix H Summary of Criteria Documents for Acrylonitrile,
1,3-Dichloropropene, and Hexachlorobutadiene H-l
IV
-------
1.
INTRODUCTION
1.1 Background
EPA published the availability of ambient water quality criteria (AWQC) documents for 64
toxic pollutants and pollutant categories identified in Section 307(a) of the Clean Water Act (CWA
or the Act) in the Federal Register on November 28, 1980 (45 FR 79318). The November 1980
Federal Register notice also summarized the criteria documents and discussed in detail the methods
used to derive the AWQC for those pollutants. The AWQC for those 64 pollutants and pollutant
categories were published pursuant to Section 304(a)(l) of the CWA:
"The Administrator, . . . shall develop and publish, . . . , (and from time to time
thereafter revise) criteria for water quality accurately reflecting the latest scientific
knowledge (A) on the kind and extent of all identifiable effects on health and welfare
including, but not limited to, plankton, fish, shellfish, wildlife, plant life, shorelines,
beaches, esthetics, and recreation which may be expected from the presence of
pollutants in any body of water, including ground water; (B) on the concentration and
dispersal of pollutants, or their byproducts, through biological, physical, and
chemical processes; and (C) on the effects of pollutants on the biological community
diversity, productivity, and stability, including information on the factors affecting
rates of eutrophicationand rates of organic and inorganic sedimentation for varying
types of receiving waters."
The AWQC published in November 1980 provided two essential types of information: (1)
discussions of available scientific data on the effects of the pollutants on public health and welfare,
aquatic life, and recreation; and (2) quantitative concentrations or qualitative assessments of the
levels of pollutants in water which, if not exceeded, will generally ensure adequate water quality for
a specified water use. Water quality criteria developed under Section 3 04(a) are based solely on data
and scientific j udgments on the relationship between pollutant concentrations and environmental and
human health effects. The 304(a) criteria do not reflect consideration of economic impacts or the
technological feasibility of meeting the chemical concentrations in ambient water. As discussed
below, 304(a) criteria are used by States and Tribes to establish water quality standards, and
ultimately provide a basis for controlling discharges or releases of pollutants.
The 1980 AWQC were derived using guidelines and methodologies developed by the
Agency for calculating the impact of waterborne pollutants on aquatic organisms and on human
health. Those guidelines and methodologies consisted of systematic procedures for assessing valid
and appropriate data concerning a pollutant's acute and chronic adverse effects on aquatic organisms,
nonhuman mammals, and humans. The guidelines and methodologies were fully described in
Appendix B (for protection of aquatic life and its uses) and Appendix C (for protection of human
health) of the November 1980 Federal Register notice.
-------
The focus of the current Federal Register notice, which this document accompanies, is the
draft revisions to the methodology for the development of A WQC to protect human health; a similar
process to revise the methodology for deriving AWQC for the protection of aquatic life is currently
underway at the Agency. Once the draft revisions are finalized, the Agency will use the revised
AWQC methodology to both develop new AWQC for additional chemicals and to revise existing
AWQC. The notice includes summaries of three criteria developed using the draft revised
methodology which are also included in this document (Appendix H). The full criteria document:?
for these three chemicals are available through the National Technical Information Service (NTIS)
or on EPA's Internet web site. These AWQC were developed to demonstrate the different risk
assessment and exposure approaches presented in the Federal Register notice. In addition, EPA
intends to derive AWQC for the protection of human health for several chemicals of high priority,
including but not limited to, PCBs, lead, mercury, arsenic, and dioxin, within the next several years.
EPA anticipates that the focus of 304(a) criteria development will be criteria for bioaccumulative
chemicals and chemicals considered highest priority by the Agency. EPA's prioritization process
for developing and revising AWQC is discussed in Appendix II of the Federal Register notice. It
is important to emphasize that the Draft AWQC Methodology Revisions presented here are also
intended to provide States and Tribes flexibility in setting water quality standards by providing
scientifically valid options for developing their own water quality criteria that consider local
conditions. States and Tribes are encouraged to use the methodology once it is finalized to derive
their own AWQC. However, the draft methodology in the Federal Register also defines the default
factors EPA will use in evaluating and determining consistency of State water quality standards with
the requirements of the CWA. These default factors will also be used by the Agency to calculate
304(a) criteria values when promulgating water quality standards for a State or Tribe under Section
303(c) of the Act.
1.2 Need for Revision of the 1980 AWQC National Guidelines
1.2.1 Scientific Advances Since 1980
Since 1980, EPA risk assessment practices have evolved significantly, particularly in the
areas of cancer and noncancer risk assessments, exposure assessments, and bioaccumulation.
In cancer risk assessment, there have been advances with respect to the use of mode of action
information to support both the identification of carcinogens and the selection of procedures to
characterize risk at low, environmentally relevant exposure levels. Related to this is the
development of new procedures for quantifying cancer risk at low doses to replace the current
default linear multistage model (LMS).
In noncancer risk assessment, the Agency is moving toward the use of statistical models,
such as the benchmark dose approach and categorical regression, to derive reference doses (RfDs)
in place of the traditional NOAEL- (no observed adverse effect level) based method.
-------
In exposure analysis, several new studies have addressed water consumption and fish-tissue
consumption. These exposure studies provide a more current and comprehensive description of
national, regional, and special-population consumption patterns; these are reflected in the Draft
AWQC Methodology Revisions presented in the Federal Register notice accompanying this
technical support document. In addition, more formalized procedures are now available to account
for human exposure from multiple sources when setting health goals that address only one exposure
source.
With respect to bioaccumulation,the Agency has moved to ward the use of a bioaccumulation
factor (BAF) to reflect the uptake of a contaminant by fish from all sources rather than just from the
water column as reflected by the use of a bioconcentration factor (BCF) in the 1980 methodology.
The Agency has also developed detailed procedures and guidelines for estimating BAF values.
1.2.2 EPA Risk Assessment Guidelines Development Since 1980
When the 1980 AWQC National Guidelines were developed, EPA had not yet developed
formal cancer or noncancer risk assessment guidelines. Since then EPA has published several risk
assessment guidelines documents. In 1996, the Agency published Proposed Guidelines for
Carcinogen Risk Assessment (61 FR17960), which, when finalized, will supersede the carcinogenic
risk assessment guidelines published in 1986 (51 FR 33992). In addition, guidelines for
mutagenicity assessment were also publishedin 1986(51 FR34006). With respect to noncancer risk
assessment, the Agency published guidelines in 1988 for assessing male and female reproductive
risk (53 FR 24834) and in 1991 for assessing developmental toxicity (56 FR 63798). In 1991, the
Agency also developed an external review draft of revised risk assessment guidelines for noncancer
health effects.
In addition to these risk assessment guidelines, EPA also published the Exposure Factors
Handbook in 1990, which presents commonly used Agency exposure assumptions and the surveys
from which they are derived. In 1992 EPA published the Guidelines for Exposure Assessment (57
FR 22888), which describes general concepts of exposure assessment, including definitions and
associated units, and provides guidance on planning and conducting an exposure assessment. Also,
in the 1980'sthe Agency published the Total Exposure Assessment Methodology (TEAM), which
presents a process for conducting comprehensive evaluation of human exposures. Finally, the
Agency has recently developed the Relative Source Contribution Policy, which is currently
undergoing Agency review, for assessing total human exposure to a contaminant and allocating the
RfD among the media of concern.
Additionally, since 1980 work groups have been established at EPA, specifically, CRAVE
and the RfD/RfC Work Group, to support the consistent evaluation of the carcinogenic and non-
carcinogenic effects of chemicals.
-------
1.2.3 Differing Risk Assessment and Risk Management Approaches for AWQC and
MCLGs
There are some differences that have arisen in the risk, assessment and risk management
approaches used by EPA's Office of Water for the derivation of AWQC under the authority of the
Clean Water Act and MCLGs (Maximum Contaminant Level Goals) under the Safe Drinking Water
Act. Two notable differences are with respect to the treatment of chemicals designated as Group C
carcinogens and the consideration of non-water sources of exposure when setting an AWQC or
MCLG for a noncarcinogen.
1.2.3.1 Group C Chemicals
Chemicals are typically classified as Group C—i.e., possible human carcinogens—underthe
existing EPA cancer classification scheme for any of the following reasons:
• Carcinogenicity has been documented in only one test species and/or only one cancer
bioassay, and the results do not meet the requirements of "sufficient evidence."
• Tumor response is of marginal significance due to inadequate design or reporting.
• Benign, but not malignant, tumors occur with an agent showing no response in a
variety of short-term tests for mutagenicity.
• There are responses of marginal statistical significance in a tissue known to have a
high or variable background rate.
The 1986 Guidelines for Carcinogenic Risk Assessment specifically recognized the need for
flexibility with respect to quantifying the risk of Group C carcinogens. The guidelines noted that
agents judged to be in Group C may generally be regarded as suitable for quantitative risk
assessment, but that case-by-case judgments may be made in this regard.
The EPA Office of Water has historically treated Group C chemicals differently under the
CWA and the SD WA. It is important to note that the 1980 AWQC National Guidelines for setting
AWQC under the CWA predated EPA's carcinogen classification system, which was proposed in
1984 (49 FR 46294) and finalized in 1986 (51 FR 33992). The 1980 AWQC National Guidelines
did not explicitly differentiate among carcinogens with respect to the weight-of-evidence for
characterizing them. For all pollutants judged as having adequate data for quantifying carcinogenic
risk— including those now classified as Group C—AWQC were derived based on carcinogenic risk
data. In the November 1980 Federal Register notice, EPA emphasized that the AWQC for
carcinogens should state that the recommended concentration for maximum protection of human
health is zero. At the same time, the criteria published for specific carcinogens presented water
concentrations for these pollutants correspondingto individual lifetime cancer risk levels in the range
oflO-'tolO'5.
-------
In the development of national primary drinking-water regulations under the SDWA, EPA
is required to promulgate a health-based MCLG for each contaminant. The Agency policy has been
to set the MCLG at zero for chemicals with strong evidence of carcinogenicity associated with
exposure from water. For chemicals with limited evidence of carcinogenicity, including many
Group C carcinogens, the MCLG is usually obtained using the RfD for that chemical based on its
noncancer effects with the application of an additional uncertainty factor (UF) of 1 to 10 to account
for its possible carcinogenicity. If valid noncancer data for a Group C carcinogen are not available
to establish an RfD but adequate data are available to quantify the cancer risk, then the MCLG is
based upon a nominal lifetime excess cancer risk calculation in the range of 10~5 to 10"6 (ranging
from one case in a population of 100,000 to one case in a population of one million). Even in those
cases where the RfD approach has been used for the derivation of the MCLG for a Group C
carcinogen, the drinking water concentrations associated with excess cancer risks in the range of 10"5
to 10"6 are also provided for comparison.
It should also be noted that EPA's pesticides program has applied both of the previously
described methods for addressing Group C chemicals in actions taken under the Federal Insecticide,
Fungicide, and Rodenticide Act (FIFRA) and finds both methods applicable on a case-by-case basis.
Unlike the drinking water program, however, the pesticides program does not add an extra UF to
account for potential carcinogenicity when using the RfD approach.
1.2.3.2 Consideration of Non-Water Sources of Exposure
The 1980 AWQC National Guidelines for setting AWQC recommended the use of the
following equation to derive the criterion:
C = [ADI - (DT + IN)] -s- [2 + 0.0065R]
(Equation 1.2.1)
where C is the criterion value; ADI is the acceptable daily intake (mg/kg-day); DT is the non-fish
dietary intake (mg/kg-day); IN is the inhalation intake (mg/kg-day); 2 is the assumed daily water
intake (L/day); 0.0065 is the assumed daily fish consumption (kg); and R is the bioconcentration
factor (L/kg). As implied by this equation, the contributions from non-water sources, namely air and
non-fish dietary intake, were to be subtracted from the ADI, thus reducing the amount of the ADI
"available" for water-related sources of intake. In practice, however, when calculating human health
criteria, these other exposures were generally not considered because reliable data on these exposure
pathways were not available. Consequently, the AWQC were usually derived such that drinking
water and fish ingestion accounted for the entire ADI (now called RfD).
In the drinking water program, a similar "subtraction" method was typically used in the
derivation of MCLGs proposed and promulgated in drinking water regulations through the mid-
1980s. More recently, the drinking water program has consistently used a "percentage" method in
the derivation of MCLGs for noncarcinogens. In this approach, the percentage of total exposure
-------
typically accounted for by drinking water, referred to as the relative source contribution (RSC), is
applied to the RfD to determine the maximum amount of the RfD "allocated" to drinking water. In
using this percentage procedure, the drinking water program also applies a ceiling level of 80 percent
of the RfD and a floor level of 20 percent of the RfD. That is, the MCLG cannot account for more
than 80 percent of the RfD, nor less than 20 percent of the RfD.
The drinking water program usually takes a conservative approach of applying an RSC factor
of 20 percent to the RfD when adequate exposure data do not exist, assuming that the major portion
(80 percent) of the total exposure comes from other sources, such as diet.
1.2.3.3 Cancer Risk Ranges
In addition to the different risk assessment approaches discussed above for deriving AWQC
and MCLGs for Group C carcinogens, different risk management approaches have arisen between
the drinking water and ambient surface water programs for using upper bound lifetime excess risk
values when setting health-based criteria for carcinogens.1 As indicated previously, the surface
water program derives AWQC for carcinogens that generally correspond to lifetime excess cancer
risk levels of 10'7 to 10'5. The drinking water program has set MCLGs for Group C carcinogens
based on a slightly less stringent risk range of 10'6 to 10'5, while MCLGs for chemicals with strong
evidence of carcinogenicity are set at zero.
It is also importantto note that under the drinking water program, for those substances having
an MCLG of zero, enforceable Maximum Contaminant Levels (MCLs) have generally been
promulgated to correspond with cancer risk levels ranging from 10'6 to 10"4. Unlike AWQC and
MCLGs, which are strictly health-based criteria, MCLs are developed with consideration given to
the costs and technological feasibility of reducing contaminant levels in water to meet those
standards.
1.3 Purpose of this Document
This document is meant to add technical detail to the principles and recommendations
presented in the Federal Register notice for the Proposed Revisions to the Methodology for Deriving
Ambient Water Quality Criteria for the Protection of Human Health. This document includes
detailed examples of many of the ideas presented in the Federal Register in an effort to explain the
thought process behind many of the new risk assessment directions being taken by the Agency. For
instance, there is an example of how to apply the new cancer guidelines to a chemical which causes
cancer but may not be genotoxic or mutagenic. In addition, three sample criteria have been derived
applying the new cancer guidelines; these are included in Appendix H and should be read together
with this document and the Federal Register notice. On the noncancer side, an example is included
on how to use the benchmark dose approach. To supplement the discussion in the Federal Register
1 Throughout this document, the term "risk level" regarding a cancer assessment endpoint specifically refers to an upper bound estimate of
excess lifetime cancer risk.
-------
on exposure, many datasets on fish consumption rates (both nationally and regionally) have been
incorporated into this document. In addition, a detailed discussion on deriving relative source
contributions is presented. To support the understanding of bioaccumulation, the data used to
calculate the percent lipid by fish species has been added.
As noted above, three sample criteria (actual 307(a) list toxic chemicals) have been updated
using the revised methodology to (1) illustrate the changes that can be expected (numerically) when
applying the revised methodology; and (2) to demonstrate the logic behind the revised methodology
and the judgments required to fulfill the recommendations of the guidance. As noted on the criteria
documents themselves, the Agency is proposing to develop streamlined criteria with a focus on
critical lexicological and exposure studies only. Due to limited resources and a need to update
criteria as quickly as possible, EPA has decided to develop more abbreviated versions of criteria
documents with an emphasis on existing risk assessments (IRIS or other EPA health assessment
documents) where available and still relevant, focusing to a greater degree on pertinent exposure and
toxicological studies which may influence the development of a criterion. EPA will continue to
conduct comprehensive reviews of the literature for the latest studies but will not provide a summary
or evaluation of those studies which are deemed less significant in the criterion development process.
1.4 Criteria Equations
The following equations for deriving AWQC include toxicological and exposure assessment
parameters which are derived from scientific analysis, science policy, and risk management
decisions. For example, parameters such as a field-measured BAF or a point of departure from an
animal study (in the form of a LOAEL/NOAEL/LED10) are scientific values which are empirically
measured, whereas the decision to use animal effects as a surrogate for human effects involves
judgment on the part of the EPA (and other agencies) as to the best practice to follow when human
data are lacking. Such a decision is, therefore, a matter of science policy. On the other hand, the
choice of default fish consumption rates for protection of a certain percentage (in this case, 90
percent and 95 percent respectively) of the general population, is clearly a risk management decision.
In many cases, the Agency has selected parameters using its best judgment of the overall protection
afforded by the resulting AWQC when all parameters are combined. For a longer discussion of the
differences between science, science policy, and risk management, please refer to Appendix I,
Section E of the Federal Register notice. Section E also provides further details with regard to risk
characterization as related to this methodology, with emphasis placed on explaining the uncertainties
hi the overall risk assessment.
-------
The generalized equations for deriving AWQC based on noncancer and cancer effects are:
Noncancer Effects
AWQC = RfD • RSC •
BW
DI+(FI • BAF),
(Equation 1.4.1)
Nonlinear Cancer Effects
AWQC =
Pdp
RSC
BW
SF
(Equation 1.4.2)
' BAF>,
Linear Cancer Effects
AWQC = RSD •
BW
DI + (FI • BAF)x
(Equation 1.4.3 )
where:
AWQC
RfD
Pdp
SF
RSD
RSC
BW
DI
FI
BAF
Ambient Water Quality Criterion (mg/L)
Reference dose for noncancer effects (mg/kg-day)
Point of departure for nonlinear carcinogens (mg/kg-day), usually a
LOAEL, NOAEL, or LED,0
Safety Factor for nonlinear carcinogens (unitless)
Risk-specific dose for linear carcinogens (mg/kg-day)
(Dose associated with a target risk, such as 10"5)
Relative source contribution factor to account for non-water sources
of exposure. (Not used for linear carcinogens.) May be either a
percentage (multiplied) or amount subtracted, depending on whether
multiple criteria are relevant to the chemical.
Human body weight (proposed default = 70 kg for adults)
Drinking water intake (proposed default = 2 L/day for adults)
Fish intake (proposed defaults = 0.0178 kg/day for general population
and sport anglers, and 0.039 kg/day for subsistence fishers)
Bioaccumulation factor, lipid normalized (L/kg)
-------
1.5 Glossary/Acronyms
List of Acronyms Used
ADI
ASTM
AWQC
BAF
BCF
BMD
BMR
BSAF
BW
Clg
CDC
CR
CSFII
CWA
DI
DNA
DOC
DT
ED10
EMAP
EPA
FCM
FDA
FEL
FI
FIFRA
FR
FSTRAC
GI
GLI
IARC
II
ILSI
IN
IRIS
kg
Acceptable Daily Intake
American Society of Testing and Materials
Ambient Water Quality Criteria
Bioaccumulation Factor
Bioconcentration Factor
Benchmark Dose
Benchmark Response
Biota-Sediment Accumulation Factors
Body Weight
Carbon-18
U.S. Centers for Disease Control and Prevention
Consumption Rate
Continuing Survey of Food Intake by Individuals
Clean Water Act
Drinking Water Intake
Deoxyribonucleic Acid
Dissolved Organic Carbon
Non-Fish Dietary Intake
Dose Associated with a 10 Percent Extra Risk
Environmental Modeling and Assessment Program
Environmental Protection Agency
Food Chain Multiplier
Food and Drug Administration
Frank Effect Level
Fish Intake
Federal Insecticide, Fungicide, and Rodenticide Act
Federal Register
Federal State Toxicology and Risk Analysis Committee
Gastrointestinal
Great Lakes Water Quality Initiative
International Agency for Research on Cancer
Incidental Intake
International Life Sciences Institute
Inhalation Intake
Integration Risk Information System
kilogram
Octanol-Water Partition Coefficient
Liter
-------
LEDIO
LMS
LOAEL
LR
MCL
MCLG
MF
mg
ml
MLE
MoA
MoE
MoS
NCHS
NHANES
NIEHS
NOAEL
NOEL
NPDES
NTIS
NTR
ODES
PAH
PBPK
PCB
PCS
Pdp
POC
q,*
RDA
RfC
RfD
RPF
RSC
RSD
SAR
SAB
SDWA
SF
STORET
TCDD-dioxin
TEAM
The Lower 95 Percent Confidence Limit on a Dose Associated with a 10
Percent Extra Risk
Linear Multistage Model
Lowest Observed Adverse Effect Level
Lifetime Risk
Maximum Contaminant Level
Maximum Contaminant Level Goal
Modifying Factor
Milligrams
Milliliters
Maximum Likelihood Estimate
Mode of Action
Margin of Exposure
Margin of Safety
National Center for Health Statistics
National Health and Nutrition Examination Survey
National Institute of Environmental Health Sciences
No Observed Adverse Effect Level
No Observed Effect Level
National Pollutant Discharge Elimination System
National Technical Information Service
National Toxics Rule
Ocean Data Evaluation System
Polycyclic Aromatic Hydrocarbon
Physiologically Based Pharmacokinetic
Polychlorinated Biphenyl
Permits Compliance System
Point of Departure
Particulate Organic Carbon
Cancer Potency Factors
Recommended Daily Allowance
Reference Concentration
Reference Dose
Relative Potency Factor
Relative Source Contribution
Risk Specific Dose
Structure-Activity Relationship
Science Advisory Board
Safe Drinking Water Act
Safety Factor
STOrage and RETrieval U.S. Waterways Parametric Data Base
Tetrachlorodibenzo-/?-dioxin
Total Exposure Assessment Methodology
10
-------
TEF
TSD
USDA
UF
Toxicity Equivalency Factor
Technical Support Document
United States Department of Agriculture
Uncertainty Factor
2. ELEMENTS OF METHODOLOGY REVISIONS AND ISSUES BY TECHNICAL
AREA
2.1 Cancer Effects
This section provides a discussion of the current status of the cancer risk assessment
methodology employed by EPA and modifications in that methodology, which are based on recent
scientific developments and the Agency's experience in this field.2 A discussion is provided of:
• Background information on the origins of current cancer risk assessment methods
and limitations associated with those methods.
• New approaches recommended in the Proposed Guidelines for Carcinogen Risk
Assessment (61 FR 17960, April 23, 1996), which revises the 1986 Cancer
Guidelines.
• Modifications in the A WQC methodology for carcinogens proposed by EP A's Office
of Water.
• An example showing the application of the new methodology to an organo-
phosphonate pesticide.
2.1.1 Background on EPA Cancer Assessment Guidelines
2.1.1.11980 AWQC National Guidelines
When EPA published the 1980 AWQC National Guidelines,3 formal Agency guidelines for
assessing carcinogenic risk from exposure to chemicals had not yet been adopted. The methodology
for assessing carcinogenic risk used by EPA in the 1980 AWQC National Guidelines is based
primarily on the Interim Procedures and Guidelines for Health Risks and Economic Impact
Assessment of Suspected Carcinogens published by EPA in 1976(41 FR21402). Although the 1980
AWQC National Guidelines recommended the use of both human epidemiological and animal
2See also: Notice of Draft Revisions to the Methodology for Deriving Ambient Water Quality Criteria for the Protection of Human Health,
in the Federal Register. Herein after referred to as EPA, 1998).
'The term "1980 AWQC National Guidelines" refers to material presented in Appendix C of the November 1980 FR notice describing
EPA's method for deriving AWQC for the protection of human health.
11
-------
studies to identify carcinogens, potential human carcinogens were primarily identified as those
substances causing a statistically significant carcinogenic response in animals. It was also assumed,
for risk assessment purposes, that chemical carcinogenesis was a non-threshold phenomenon.
Two types of data are used for quantitative cancer risk estimates:
• Lifetime animal studies.
• Human studies where excess cancer risk is associated with exposure to the agent.
(Human data with sufficient quantification to carry out risk assessment are not
available for MoE agents.)
The scaling of doses from animals to humans uses a conversion factor of body weight raised
to the 2/3 power (BW20). The specific equation for converting an animal dose to a human equivalent
dose using the BW273 scaling factor is:
Human Equivalent Dose (mg/kg-day)
= Animal Dose (mg/kg-day) x
Animal BW
Human BW
2/3
Animal BW
2/3
Human BW
(Equation 2.1.1)
This approach is based on the assumption that doses between species are related to surface
area. Exposure is defined in mg of contaminant/(body weights/day (Mantel and Schneiderman,
1975). This assumption is more appropriate at low concentrations, where sources of non-linearity,
such as saturation or induction of enzyme activity, are less likely to occur.
The estimation of cancer responses typically uses animal bioassay data extrapolated to low
doses approximating human exposure. Extrapolation is usually carried out using the linearized
multi-stage model (LMS). The LMS model is used to fit the tumor data with computer programs
(e.g., GLOBAL 86) that calculate the 95th percentile upper confidence limit on the linear slope in
the low dose range. The slope which is obtained is referred to as the q,*, or cancer potency.
When animal data are used for these calculations, the body weights are scaled using B W2/3,
as discussed above. The q,* values obtained using the LMS model are expressed in the form of the
upper bound estimate of lifetime risk per (mg/kg-day). These values are often used to estimate the
upper bound of the lifetime cancer risk for long-term low level exposure to agents.
The risk assessments carried out with this model are generally considered conservative,
representing the most plausible 95th percentile upper bound for risk. The "true risk" is considered
unlikely to exceed the risk estimate derived by this procedure, and could be as low as zero. The
12
-------
LMS approach was endorsed by four agencies in the Interagency Regulatory-feiaison Group and was
characterized as less likely to under-estimaterisk at the low doses typical of environmental exposure
than other models and approaches that were available.
Because of the uncertainties associated with dose-response evaluations, EPA believed that
it was prudent to use the LMS to estimate cancer risk for the AWQC. These uncertainties include:
• The need for animal-to-human extrapolation;
• The use of average exposure assumptions; and
• The serious public health consequences that could result if risk were underestimated.
In deriving water quality criteria, the slope factors are currently estimated using the LMS
model under most circumstances. When human (epidemiological) data are available, other
approaches have been used.
Basic assumptions which are used to calculate the AWQC include:
• An "average" daily consumption rate of 2 liters of water per person per day (from all
sources).
• An average daily fish consumption rate of 6.5 grams per day.
• An average body weight of 70 kilograms (kg) (154 pounds).
The maximum lifetime cancer risk generated by waterborne exposure to the agent is targeted
in the range of one in one hundred thousand to one in ten million (lO'5 to 10'7). The formula for
deriving the AWQC in milligrams per liter (mg/L) for carcinogens presented in the 1980 AWQC
National Guidelines is:
where:
10
AWQC (mg/L) -
(70)
(qi)(2+0.0065R)
(Equation 2.1.2)
Target cancer risk level; the 1980 AWQC National Guidelines recommended
risk levels in the range of 10"5 to 10'7
13
-------
70
"Assumed body weight of an adult human being (kg)
q,* = Carcinogenic potency factor for humans derived from LMS model (mg/kg-
day)'1
2 = Assumed daily water consumption of an adult human (L/day)
.0065 = Assumed daily consumption of fish (kg)
R = Bioconcentration factor (L/kg) from water to food (e.g., fish, birds)
2.1.1.2 1986 EPA Guidelines for Carcinogenic Risk Assessment
Since 1980, EPA risk assessment practices have evolved significantly. In September 1986,
EPA published its Cancer Risk Assessment Guidelines (referred to subsequently in this document
as the 1986 Cancer Guidelines) in the Federal Register (51 FR 33992, EPA, 1986). The 1986
Cancer Guidelines were based on the publication by the Office of Science and Technology Policy
(OSTP, 1985) that provided a summary of the state of knowledge in the field of carcinogenesis and
a statement of broad scientific principles of carcinogen risk assessment on behalf of the federal
government.
The 1986 Cancer Guidelines established a classification scheme to describe the nature of the
cancer data base and evidence supporting the carcinogenicity of an agent. This classification system
is based on a similar scheme developed by the International Agency for Research on Cancer (I ARC).
This scheme is described briefly below. More detailed information can be obtained from the 1986
Cancer Guidelines (EPA, 1986).
The classificationscheme utilizes several alpha-numerical groups for classifying chemicals
with respect to the evidence available regarding their carcinogenic potential for humans:
Group A: Human carcinogen; sufficient evidence from epidemiological studies.
Group B: Probable human carcinogen; sufficient evidence in animals or limited
evidence in humans.
Group C: Possible human carcinogen; limited evidence of carcinogenicity in animals
in the absence of adequate human data.
Group D: Not classifiable; inadequate data or no data.
Group E: No evidence of carcinogenicity in adequate studies in at least two species or
in both epidemiological and animal studies.
14
-------
Within Group B there are two subgroups: Bl and B2. According to the 1986 Cancer
Guidelines: "Usually Group Bl is reserved for agents for which there is limited evidence of
carcinogenicity from epidemiologic studies. It is reasonable, for practical purposes, to regard an
agent for which there is 'sufficient' evidence of carcinogenicity in animals as if it presented a
carcinogenic risk to humans. Therefore, agents for which there is 'sufficient' evidence from animal
studies and for which there is 'inadequate evidence' or 'no data' from epidemiologic studies would
usually be categorized under Group B2." (USEPA, 1986)
The 1986 Cancer Guidelines also include guidance on the definition of sufficient or limited
evidence. The weight-of-evidence for human studies is evaluated as sufficient when a causal
relationship is indicated by the study. When animal studies are used in the evaluation of
carcinogenicity, sufficient evidence includes agents which have been demonstrated to cause:
• an increased incidence of malignant tumors; or
• an increased incidence of combined malignant and benign tumors:
1) in multiple species or strains; or
2) in multiple experiments (e.g., with different routes of administrationor using
different dose levels); or
3) to an unusual degree in a single experiment with regard to high incidence,
unusual site or type of tumor.
• an early age at onset.
Additional evidence may be provided by data on dose-response, from short-term tests, or on
chemical structure.
Evidence is considered limited when a causal interpretation is credible but alternative
explanations are not sufficiently excluded. Limited evidence indicates that the data base for an agent
can be placed into one of three categories:
• Studies involve a single species, strain, or experiment and do not meet criteria for
sufficient evidence;
• Experiments are restricted by inadequate dosage levels, inadequate duration of
exposure, inadequate period of follow-up, poor survival, too few animals, or
inadequate reporting; or
• An increase in benign but not in malignant tumors.
15
-------
Subsequent refinements of this designation have included agents which do not demonstrate
positive responses in a variety of short-term tests for mutagenicity and those with responses of
marginal statistical significance in a tissue known to have a high or variable background rate.
For cancer risk quantification, the 1986 Cancer Guidelines continued the recommended use
of the linearizedmultistage model (LMS) as the only default approach. The 1986 Cancer Guidelines
also state that low-dose extrapolation models and approaches other than the LMS model might be
considered more appropriate based on biological grounds. However, no guidance was given in
choosing other approaches; thus, departures from the LMS procedure have been rare in practice. The
1986 Guidelines continued to recommend the use of BW273 as a dose scaling factor between species.
2.1.1.3 Scientific Issues Associated with the Current Cancer Risk Assessment
Methodology for the Development of AWQC
In reviewing the current approach for the development of Water Quality Criteria for Human
Health, EPA believes that there is not sufficient flexibility in the 1986 Cancer Guidelines. In
addition, insufficient attention is given to critical information including:
• The mode of action.
• Relevance of animal bioassay data to humans.
• Route of exposure.
• The duration and magnitude of exposure.
• Additional difficulties are associated with the following:
• Many agents fall between groups (e.g., between B2 and C) and may be
difficult to assign to a specific group;
• Effects may be greatly modulated by the conditions of exposure.
• The use of linear extrapolation may not be appropriate for all agents,
including some which appear to induce tumors at high but not low doses and
which do not interact directly with DNA.
All of these issues have been considered in the development of new guidelines.
After the 1992 National Workshop on Revision of the Methods for Deriving National
Ambient Water Quality Criteria for the Protection of Human Health, EPA requested its Scientific
Advisory Board (SAB) to review the Workshop report. The SAB recommended against the interim
adoption of the 1986 Guidelines into the AWQC methodology, indicating that it might create
16
-------
considerable confusion in the future, once new Cancer Guidelines were formally proposed and
implemented. EPA was encouraged by both groups to incorporate new approaches into the AWQC
methodology. As recommended by these two groups, EPA is proposing revisions to the cancer risk
assessment methodology for the development of AWQC by incorporating new approaches discussed
in the EPA Proposed Cancer Risk Assessment Guidelines dated April 23,1996 (61 FR 17960).
2.1.2 Proposed Revisions to EPA's Carcinogen Risk Assessment Guidelines
EPA has recently published Proposed Guidelines for Carcinogen Risk Assessment (EPA,
1996), which contain proposed revisions to the 1986 Cancer Guidelines. These revisions are
designed to ensure that the Agency's cancer risk assessment methods reflect the most current
scientific information.4 Although many fundamental aspects of the 1986 Cancer Guidelines have
been retained, there are a number of key changes proposed, some of which address the specific issues
mentioned in the preceding section. Proposed changes to the Cancer Guidelines are discussed here
because many of the agency-wide principles that are proposed are incorporated into the proposed
revisions to the AWQC methodology.
The key changes in the Proposed Cancer Guidelines include:
a) Hazard assessmentpromotes the analysis of all biological information rather than
just tumor findings.
b) An agent's mode of action in causing tumors is emphasized to reduce the
uncertainty in describing the likelihood of harm and in determining the dose-response
approach(es).
c) Increased emphasis on hazard characterization to integrate the data analysis of
all relevant studies into a weight-of-evidence conclusion of hazard, to develop a
working conclusion regarding the agent's mode of action in leading to tumor
development, and to describe the conditions under which the hazard may be
expressed (e.g., route, pattern, duration and magnitude of exposure).
d) A weight-of-evidencenarrative with accompanying descriptors (listed in Section
2.1.3.2 below) replaces the current alphanumeric classification system. The
narrative is intended for the risk manager and lays out a summary of the key
evidence, describes the agent's mode of action, characterizes the conditions of hazard
expression, and recommends appropriate dose-response approach(es). Significant
strengths, weaknesses, and uncertainties of contributing evidence are highlighted.
The overall conclusion as to the likelihood of human carcinogenicity is given for
each route of exposure.
4They are referred to hereafter as the Proposed Cancer Guidelines.
17
-------
g)
e) Biologically-based extrapolation models are the preferred approach for
quantifying risk. It is anticipated, however, that the necessary data for the parameters
used in such models will not be available for most chemicals. The new guidelines
allow for alternative quantitative methods, including several default approaches.
f) Dose-response assessment is a two-step process. In the first step, response data are
modeled in the range of observation, and in the second step, a determination of the
point of departure or range of extrapolation below the range of observation is made.
In addition to modeling tumor data, the new guidelines call for the use and modeling
of other kinds of responses if they are considered to be more informed measures of
carcinogenic risk.
Three default approaches are provided—linear, nonlinear, or both. Curve
fitting in the observed range would be used to determine a point of departure. A
standard point of departure is proposed as the effective dose corresponding to the
lower 95 percent limit on a dose associated with 10 percent extra risk (LEDIO). The
linear default is a straight line extrapolation from the response at the LEDIO to the
origin (zero dose, zero extra risk). The nonlinear default begins with the identified
point of departure and provides an MoE analysis rather than estimating the
probability of effects at low doses. The MoE analysis is used to compare the point
of departure with the human exposure levels of interest (Pdp/exposure). The key
objective of the MoE analysis is to describe for the risk manager how rapidly
responses may decline with dose. Other factors are also considered in the MoE
analysis (nature of the response, human variation, species differences,
biopersistence).
h) The approach used to calculate oral human equivalent dose when assessments
are based on animal bioassays has been refined to include a change in the default
assumption for interspecies dose scaling (using body weight raised to the 3/4 power).
With recent proposals to emphasize mode of action understanding in risk assessment and
to model response data in the observable range to derive points of departure or HMDs for both
cancer and noncancer endpoints, EPA health risk assessment practices are beginning to come
together. The modeling of observed response data to identify points of departure in a standard way
will help to harmonize cancer and noncancer dose-response approaches and permit comparisons of
cancer and noncancer risk estimates.
It is importantto note that the cancer risk assessmentprocess outlined in the Proposed Cancer
Guidelines is not limited just to the quantitative aspects. Extensive guidance is provided in the
Proposed Cancer Guidelines regarding hazard assessment and risk characterization (EPA, 1996).
The Proposed Cancer Guidelines should be consulted for detailed information regarding the
new methodology and the scientific basis for the proposed changes. All of the above listed changes,
18
-------
as well as other methodological issues discussed in the Proposed Cancer Guidelines, have a direct
bearing on the proposed methods for deriving AWQC discussed in this TSD. Rather than including
a summary in mis document that would provide only limited detail, the reader is urged to review the
guidelines, as provided in the Federal Register notice in their original form (USEPA, 1996).
2.1.3 Revised Carcinogen Risk Assessment Methodology for Deriving AWQC
The revised methodology for deriving numerical AWQC for carcinogens is consistent with
the principles included in the Proposed Cancer Guidelines. This discussion of the Draft AWQC
Methodology Revisions for carcinogens focuses primarily on the quantitative aspects of deriving
numerical AWQC values. However, the Proposed Cancer Guidelines emphasize the importance of
qualitative information as critical to the cancer risk evaluation process. Consequently, the proposed
guidelines also recommend that a numerical AWQC value derived for a carcinogen is to be
accompanied by appropriate hazard assessment and risk characterization information.5
This section contains a discussion of the weight-of-evidencenarrative, describing information
relevant to a cancer risk evaluation. This is followed by a discussion of the quantitative aspects of
deriving numerical AWQC values for carcinogens. It is assumed that data from an appropriately
conducted animal bioassay provide the underlying basis for deriving the AWQC value. The
discussion focuses on the following topics:
• Dose estimation.
• Characterizing dose-response relationships in the range of observation and at low,
environmentally relevant doses.
• Calculating the AWQC value.
• Risk characterization.
• Use of Toxicity Equivalent Factors (TEF) and Relative Potency Estimates.
The first three listed topics encompass the quantitative aspects of deriving AWQC for carcinogens.
2.1.3.1 Weight-of-Evidence Narrative
As stated in the EPA Proposed Cancer Guidelines, the new method for cancer risk assessment
includes a weight-of-evidence narrative which is based on an overall weight-of-evidence of
biological, chemical, and physical considerations. The weight-of-evidence narrative lays out key
evidence and includes a discussion of tumor data, information on the mode of action, and its
implications for human hazard and dose-response evaluation. Emphasis will also be focused on the
5 See also EPA, 1996.
19
-------
route and level of exposure and relevance to humans. In addition, a discussion of the strengths and
weaknesses of the data base is included. The hazard assessment emphasizes analysis of all
biological information rather than just tumor findings.
The weight-of-evidence narrative is written for the risk manager, and thus explains in
nontechnical language the key data and conclusions, as well as the conditions for hazard expression.
Conclusions about potential human carcinogenicity are presented by route of exposure. Contained.
within this narrative are simple likelihood descriptors that essentially distinguish whether there is
enough evidence to make a projection about human hazard (i.e., known human carcinogen, should:
be treated as if known, likely to be a human carcinogen, or not likely to be a human carcinogen) or
whether there is insufficient evidence to make a projection (i.e., the cancer potential cannot be
determined because evidence is lacking, conflicting, inadequate, or because there is some evidence
but it is not sufficientto make a projection to humans). Because one encounters a variety of data sets
on agents, these descriptors are not meant to stand alone; rather, the context of the weight-of-
evidence narrative is intended to provide a transparent explanation of the biological evidence and
how the conclusions were derived. Moreover, these descriptors should not be viewed as
classification categories (like the alphameric system), which often obscure key scientific differences
among chemicals. The new weight-of-evidence narrative also presents conclusions about how the
agent induces tumors and the relevance of the mode of action to humans, and recommends a dose-
response approach based on the mode-of-action understanding.
2.1.3.2 Dose Estimation (by the Oral Route)
Determining the Human Equivalent Dose
An important objective in the dose-response assessment is to use a measure of internal or
delivered dose at the target site when sufficient data are available. This is particularly important in
those cases where the carcinogenic response information is being extrapolated to humans from
animal studies. Generally, the measure of dose provided in the underlying human studies and animal
bioassays is the applied dose, typically given in terms of the unit mass per unit body weight per unit
of time, (e.g., mg/kg-day). When animal bioassay data are used, it is necessary to make adjustments
to the applied oral dose values to account for differences in pharmacokinetics between animals and
humans that affect the relationship between applied dose and delivered dose at the target organ.
In the estimation of a human equivalent dose, the Proposed Cancer Guidelines recommend
that when toxicokinetic data are available, they are used to convert the doses used in animal studies
to equivalent human doses. However, in most cases, there are insufficient data available to compare
dose between species. In these cases, the estimate of a human equivalent dose is based on science
policy default assumptions. In the past, body weight raised to the 2/3 power was used (as discussed
in Section 2.1.1.1). To derive an equivalent human dose from animal data, the new default
procedure is to scale daily applied oral doses experienced over a lifetime in proportion to body
weight raised to the 3/4 power.
20
-------
The 3/4 adjustment factor is used because metabolic rates, as well as most rates of
physiological processes that determine the disposition of a dose, scale this way. Thus, the rationale
for this factor rests on the empirical observation that rates of physiological processes consistently
tend to maintain proportionality with body weight raised to 3/4 power. Based on this assumption,
the "human equivalent" of the applied oral dose in an animal study is obtained from the following
algorithm where the doses are in mg/kg-day:
Human Equivalent Dose =
Animal Dose x
Animal BW
Animal BW 3/4
>
(Equation 2.1.3)
Human BW
3/4
Human BW
1/4
This equation can be simplified to:
Human Equivalent Dose = (Animal Dose)[(Animal BW)/(Human BW)]
(Equation 2.1.4)
This procedure does not calculate the delivered dose, but rather adjusts the applied dose (e.g.,
exposure) to account for interspecies differences in delivered doses.
This change in approach yields an estimate of delivered dose which is larger than that
obtained using body weight raised to the 2/3 power in cases where the animals used in the study have
a lower body weight than humans (e.g. rodents, dogs, rabbits, and most animals used for
toxicological testing). Since a larger dose is estimated using this approach, the cancer potency which
is estimated using the 3/4 scaling approach is slightly lower than the potency which is calculated
using body weight raised to the 2/3 power.
A more extensive discussion of the rationale and data supporting the Agency's change in
scaling factors from 2/3 to 3/4 is in USEPA (1992b) and the Proposed Cancer Guidelines.
Dose Adjustments for Less-than-Lifetime Exposure Periods
In the 1980 AWQC National Guidelines, two other dose-related adjustments were discussed.
The first addressed situations where the experimental dosing period (le) is less than the duration of
the experiment (Le). In these cases, the average daily dose is adjusted downward by multiplying by
the ratio (le/Le) to obtain an equivalent average daily dose for the full experimental period. This
adjustment would also be used in situations where animals are dosed fewer than seven days per
week. If, for example, "daily" dosing is done only five days each week, the lifetime daily dose
would be calculated as 5/7 of the actual dose given on each of the five days.
21
-------
The second dose adjustment addresses situations where the experimental duration (Le) is
substantially less than the natural lifespan (L) of the test animal. For example, for mice and rats the
natural lifespans are defined as 90 weeks and 104 weeks respectively. If the study duration is less
than 78 weeks for mice, or less than 90 weeks for rats, applied doses are adjusted by dividing by a
factor of (L/Le)3. (Alternatively,the cancer potency factor obtained from the study could be adjusted
upward by multiplying by the factor of (L/Le)3.)
This adjustment is considered necessary because a shortened experimental duration does not
permit the full expression of cancer incidence that would be expressed during a lifetime study. In
addition, most carcinogenic responses are manifest in humans and animals at higher rates later in
life. Age-specific rates of cancer increase as a constant function of the background cancer rate
(Anderson, 1983) by the 2nd or higher power of age (Doll, 1971). In the adjustment recommended
here, it is assumed that the cumulative tumor rate will increase by at least the 3rd power of age. It
is important to note that although both dose adjustments discussed in this section were included in
the 1980 AWQC National Guidelines, the second adjustment has not been commonly used in
practice.
2.1.3.3 Dose-Response Analysis
Dose-response analysis addresses the relationship of dose to the degree of response observed
in an animal or human study. Extrapolations are necessary when environmental exposures are
outside of the range of study observations. Past observations of response have focused on the
observation of tumors. The Proposed Cancer Guidelines suggest that responses may include tumors
or other effects related to carcinogenicity. Non-tumor effects may include changes in DNA,
chromosomes, or other key macromolecules; effects on growth signal transduction, induction of
physiological or hormonal changes, effects on cell proliferation, or other effects that play a role in
the carcinogenic process. Non-tumor effects are referred to as "precursor data" in the following
discussion.
Specific guidance regarding the use of animal data, presentation of study results, and
selection of the optimal data for use in a dose-response analysis is discussed in detail in the Proposed
Cancer Guidelines. It includes recommendations that multiple data arrays be presented including:
combined data from different experiments, ranges of results from more than one data set, tumors
generated by different modes of action, and combined tumors at more than one site within a single
experiment.
Characterizing Dose-Response Relationships in the Range of Observation
The first quantitative component in the derivation of AWQC for carcinogens is the dose-
response assessment in the range of observation. Two options are available for the assessment in
the observed range:
• Development of a biologically-based or case-specific model.
22
-------
• Curve-fitting of the tumor or precursor data.
A biologically-basedmodel is one whose parameters are calculated independently of curve-
fitting of tumor data.6 If data on the agent are sufficient to support the parameters of a biologically-
based or case specific model and the purpose of the assessment is to justify investing resources
supporting its use, this type of model is the first choice for both the observed tumor and related
response data and for extrapolation below the range of observed data in either animal or human
studies. Extensive data are required to both build the model and to estimate how well it conforms
with observed tumor development specific to the agent. Case-specific models are based on general
concepts of mode of action and data on the agent. The Proposed Cancer Guidelines contain more
detail on these approaches. There is not sufficient data to utilize these types of models for most
agents.
In the absence of adequate data to generate a biologically-based model or case-specific
model, dose-response relationships in the observed range can be addressed through curve-fitting
procedures for tumor or precursor data. The models should be appropriate to the type of response
data in the observed range.
The Proposed Cancer Guidelines recommend employing the lower 95 percent confidence
limit on a dose associated with an estimated 10 percent extra risk of tumor or relevant nontumor
response (LED,0). The LED10 (the lower 95 percent confidence limit on a dose associated with 10
percent extra risk) is a standard point of departure,7 adopted as a matter of science policy to remain
as consistent and comparable from case to case as possible. It is also a comparison point for
noncancer endpoints.
The rationale supporting its use is that a 10 percent response is at or just below the limit of
sensitivity for discerning a significant difference in most long-term rodent studies. The lower
confidence limit on dose is used to appropriately account for experimental uncertainty (Barnes et al.,
1995); it does not provide information about human variability. Uncertainties include such factors
as number and spacing of doses, sample sizes, the precision and accuracy of dose measurements, the
accuracy of pathological findings, and the selection of low dose extrapolation (discussed below).
For some data sets, a choice of the point of departure other than the LED10 may be
appropriate. The objective is to determine the lowest reliable part of the dose-response curve for the
beginning of the second step of the dose-response assessment—determiningthe extrapolation range.
Therefore, if the observed response is below the LED10, then a lower data point may be a better
6An example of a biologically-based model is applied in the case of diesel exhaust emission (See Chen, CW. and G. Oberdorster. 1996.
Selection of Models for Assessing Dose-Response Relationship for Particle-Induced Lung Cancer. Inhalation Toxicol., 8:259-278).
7 Use of the LEDIO as the point of departure is recommended with this methodology, as it is with the Proposed Cancer Guidelines. Public
comments were requested on the use of the LED,0, EDIO, or other points. EPA is currently evaluating these comments, and any changes in the
Cancer Guidelines will be reflected in the final AWQC methodology.
23
-------
choice. Moreover, some forms of data may not be amenable to curve-fitting estimation, but can be
evaluated using an estimation of a LOAEL or NOAEL, e.g., certain continuous data.
Analysis of human studies in the observed range is designed on a case by case basis
depending on the type of study and how dose and response are measured in the study. In some cases
the analysis may incorporate consideration of an agent's interactive effects with other agents. The
use of population risk rather than individual risk may be appropriate in some cases, depending on
the nature of the data set (e.g., human epidemiological data).
Extrapolation to Low, Environmentally Relevant Doses
In most cases, the derivation of an A WQC will require an evaluation of carcinogenic risk at
environmental exposure levels substantially lower than those used in the underlying bioassay.
Various approaches are used to extrapolate risk outside the range of observed experimental data.
In the Proposed Cancer Guidelines, the choice of extrapolation method is largely dependent on the
mode of action. The Proposed Guidelines also indicate that the choice of extrapolation procedure
follows the conclusions developed in the hazard assessment about the agent's carcinogenic mode
of action, and it is this mode of action understanding that guides the selection of the most appropriate
dose-response extrapolation procedure. It should be noted that the term "mode of action" is
deliberately chosen in the new guidelines in lieu of the term "mechanism" to indicate the use of
knowledge that is sufficient to draw a reasonable working conclusion without having to know the
processes hi detail as the term mechanism might imply. The proposed guidelines preferred the choice
of a biologically-basedmodel, if the parameters of such models can be calculated from data sources
independent of tumor data. It is anticipated that the necessary data for such parameters will not be
available for most chemicals. Thus, the new guidelines allow for several default extrapolation
approaches (low-dose linear, nonlinear, or both).
Biologically-Based Modeling Approaches. If a biologically-basedor case-specificmodel has
been used to characterize the dose-response relationships in the observed range, and the confidence
in the model is high, it may be used to extrapolate the dose-response relationship outside the
observed data range. Although biologically-basedand case-specific approaches are appropriate both
for characterizing observed dose-response relationships and extrapolating to environmentally
relevant doses, it is not expected that adequate data will be available to support such approaches for
most substances. In the absence of such data, the default linear approach, the non-linear (margin of
exposure) approach, or both linear and non-linear approaches are used.
Default Linear Extrapolation Approach. The default linear approach proposed here Is
essentially a replacement of the linearized multistage (LMS) approach that has served as the default
approach for EPA cancer risk assessments. This new approach is used in the derivation of AWQC
for:
• Agents with a mode of action of gene mutation resulting from reactivity with DNA;
24
-------
• Agents, with evidence that supports a mode of action other than DN A reactivity, that
are better supported by the assumption of low dose linearity; and
• Carcinogenic agents lacking information on the mode of action.
As this suggests, the linear default is used for carcinogens which lack information supporting
the use of a non-linear approach. The proposed default linear approach is considered generally
health-conservative. Evidence of effects on cell growth control via direct interaction with DNA
constitutes an expectation of a linear dose-response relationship in the low dose range, unless there
is information to the contrary.
The procedures for implementing the default linear approach begin with the estimation of
a point of departure (LED10). The point of departure value incorporates the interspecies conversion
to the human equivalent dose and the other adjustments for less-than-lifetimeexperimental duration.
In most cases, the extrapolation for estimating response rates at low, environmentally relevant
exposures is accomplished by drawing a straight line between the response at the "point of
departure" (LED10) and the origin (i.e., zero dose, zero response). This is mathematically represented
as:
where:
y
m
x
y = mx
(Equation 2.1.5)
Response or incidence
Slope of the line (cancer potency factor)
Dose
The slope of the line, "m" (i.e., Ay/Ax, the estimated cancer potency factor at low doses), is
computed as:
m =
0.10
(Equation 2.1.6)
When an LED10 isn't used, the standard equation for the slope of a line may be used:
25
-------
m =
x2 -x,
(Equation 2.1.7)
where:
Yi
X2
X!
Response at the point of departure
Response at the origin (zero)
Dose at the point of departure
Dose at the origin (zero)
Due to the use of the origin for y, and x,, the equation simplifies to:
(Equation 2.1.8)
The risk-specific dose (RSD) is then calculated for a specific incremental targeted lifetime
cancer risk (in the range of 10"4 to 10"6) as:
RSD =
Target Incremental Cancer Risk
m
(Equation 2.1.9)
where:
RSD
Target Risk8
m
Risk-specific dose (mg/kg-day)
Value typically hi the range of 10"4 to 10'6
Cancer potency factor (mg/kg-day)"1
* In 1980, the target lifetime cancer risk range was set at 10"' to 10'3. However, both the expert panel for the AWQC workshop (1992) and
SAB recommended that EPA change the risk range to 10* to 10"4, to be consistent with drinking water.
26
-------
The use of the RSD to compute the AWQC is described below in the section titled "AWQC
Calculation."
Default Non-Linear Approach. As discussed hi the Proposed Cancer Guidelines, the use of
a non-linear approach for risk assessment is recommended where there is no evidence for linearity
and there is sufficient evidence to support an assumption of non-linearity. As noted above, this
would NOT be used for agents with:
• A mode of action of gene mutation resulting from reactivity with DNA;
• Evidence that supports another mode of action that is anticipated to be linear; or
• Carcinogenic agents lacking information on the mode of action.
A definitive determination regarding an agent's mutagenicity may not be possible, since
many agents yield mixed results in mutagenicity assays. Mode of action data are used in a case
study provided in Section 2.1.4 of this document. The chemical discussed is not mutagenic but
causes stone formation in male rat bladders, leading to tumor formation at high doses. Stone and
subsequent tumor formation are not expected to occur at doses lower than those that induced the
physiologic change that leads to stones, based on the mode of action data.
The non-linear approach is indicated for agents having a mode of action that may lead to a
dose-response relationship that is non-linear, with response falling much more quickly than linearly
with dose, or those being most influenced by individual differences in sensitivity. Alternatively, the
mode of action may theoretically have a threshold (e.g., the carcinogenic response may be a
secondary effect of toxicity or of an induced physiological change (that is itself a threshold
phenomenon). EPA does not generally try to distinguish between modes of action that might imply
a "true threshold" from others with a non-linear dose-response relationship, because there is usually
not sufficient information to determine empirically.
The Proposed Cancer Guidelines recommend that non-linear probability functions NOT be
fitted to the response data to extrapolate quantitative low-dose risk estimates. Different models can
lead to a very wide range of results. Also, there is currently no basis to choose among the different
models. If there is sufficient information to choose a model, a biologically-based or case-specific
model should be used.
The Proposed Cancer Guidelines recommend use of a margin of exposure (MoE) approach
to evaluate concern for various levels of exposure. This entails the comparison of a minimum effect
dose level such as the LED,0, NOAEL, or LOAEL environmental exposures of interest. In the
context of deriving AWQC, the environmentally relevant exposures are targets rather than actual
exposures. A Safety Factor (SF) is then applied to account for various types of uncertainty. This
approach is similar to the benchmark dose approach described in the noncancer section of this TSD.
27
-------
The MoE approach used here is similar to the analysis carried out by EPA to accompany
estimates of RfD or concentrations for noncancerendpoints. However, a threshold of carcinogenic
response is not necessarily assumed. If the evidence for an agent indicates a threshold, (e.g., when
carcinogenicityis secondary to another toxicity that has a threshold) the MoE analysis is similar to
what has been done for a noncancer endpoint, and an RfD for that toxicity may also be estimated and
considered in the cancer assessment.
To support the use of the MoE approach, information is provided in the risk assessment about
the current understanding of the phenomena that may be occurring as dose (exposure) decreases
substantially below the observed data. This provides information about the risk reduction that is
expected to accompany a lowering of exposure. Information regarding the various factors which
influence the selection of a SF are also included in the discussion below.
There are two main steps in the MoE approach:
• The first step is the selection of a point of departure (Pdp) that is a "minimum effect
dose level." As noted above, the Pdp may be the LED10 for tumor incidence, or in
some cases, it may also be appropriate to use a NOAEL or LOAEL value from a
precursor, such as a response that is a precursor to tumors. When animal data are
used, the Pdp is a human equivalent dose or concentration arrived at by interspecies
dose adjustment (as discussed above) or toxicokinetic analysis.
• The second step in using MoE analysis to establish an AWQC is to conduct an
analysis to derive a SF to apply to the Pdp. (This is supported by analysis in the
MoE discussion provided in the risk assessment). The following issues are to be
considered when establishing the SF for the derivation of AWQC using the MoE
appioach (others may be found appropriate in specific cases):
The slope of the observed dose-response relationship at the point of departure
and its uncertainties and implications for risk reduction associated with
exposure reduction (e.g., a steep slope implies an apparent greater reduction
hi risk as exposure decreases that may support a smaller margin).
Variation in sensitivity to the phenomenon involved, among members of the
human population.
Variation in sensitivity between humans and the animal study population.
The nature of the response used for the dose-response assessment, for
instance, a precursor effect, or tumor response. The latter may support a
greater margin.
28
-------
Persistence of the agent in the body. Greater persistence argues for a greater
MoE. This persistence issue is particularly relevant when precursor data
from less than lifetime studies are the response data being assessed.
As a default assumption for two of the factors listed above, the Proposed Cancer Guidelines
recommend that a factor of no less than 10-fold each be employed to account for human variability
and for interspecies differences in sensitivity when humans may be more sensitive than animals.
When data indicate that humans are less sensitive than animals, a default factor of no smaller than
1/10 may be employed to account for this. If information about human variability or interspecies
differences is available, it is used.
The size of the overall SF is a matter of policy. The rationale for selection of the SF should
be fully explained and related to the toxicity and other data presented in the weight-of-evidence
narrative discussed previously.
The SF is used to modify the Pdp in the final equation. This is shown below in the Section
2.1.3.4 on AWQC calculation.
Both Linear and Non-Linear Approaches. In some cases both linear and non-linear
procedures may be used. When data indicate that there may be more than one operant mode of
action for cancer induction at different tumor sites, an appropriate procedure is used for each site.
The use of both the default linear approach and the non-linear approach may be appropriate to
discuss implications of complex dose-response relationships, and may be decoupling analysis of
regions of the overall dose response that reflect differing modes of action.
2.1.3.4 AWQC Calculation
Linear Approach
The following equation is used for the calculation of the AWQC for carcinogens where a
RSD is obtained from the default linear approach:
BW
AWQC = RSD x
(Equation 2.1.10)
The AWQC calculation shown above is appropriate for water bodies that are used as sources
of drinking water (and for other uses). If the water bodies are not used as drinking water sources the
approach is modified. The drinking water value (DI in the equation shown above) is substituted with
29
-------
an incidental ingestion value (II) of 0.01 L/day. The incidental intake is assumed to occur from
swimming and other activities. The fish intake value is assumed to remain the same.
Non-Linear Approach
In those cases where the non-linear, MoE approach is used, a similar equation is used to calculate
theAWQC:
AWQC =
BW
SF ^ DI+(FI x BAF),
(Equation 2.1.11)
x RSC%
where:
AWQC
RSD
Pdp
SF
BW
DI
FI
BAF
RSC%
Ambient water quality criterion (mg/L)
Risk-specific dose (mg/kg-day)
Point of departure (mg/kg-day)
Safety factor (unitless)
Human body weight (kg)
Drinking water intake (L/day)
Fish intake (kg/day)
Bioaccumulation factor (L/kg)
Relative source contribution (%)
As noted above for the linear approach, the AWQC calculation shown above is appropriate
for water bodies that are used as sources of drinking water (and for other uses). If the water bodies
are not used as drinking water sources DI is substituted with an incidental ingestion value (II) of 0.01
L/day.
A difference between the AWQC values obtained using the linear and non-linear approaches
is that the AWQC value obtained using the default linear approach corresponds to a specific
estimated incremental lifetime cancer risk level in the range of 10"4 to 10'6. In contrast, the AWQC
value obtained using the non-linear approach does not describe or imply a specific cancer risk.
The actual AWQC chosen is based on a review of all relevant information, including cancer,
noncancer, ecological, and other critical data. The AWQC may, or may not, utilize the value
obtained from the cancer analysis, if it is less protective than that derived from the noncancer
endpoint.
30
-------
2.1.3.5 Risk Characterization
Risk characterizationinformationis included with the numerical AWQC value and addresses
the major strengths and weaknesses of the assessment arising from the availability of data and the
current limits of understanding of the process of cancer causation. Key issues relating to the
confidence in the hazard assessment and the dose-response analysis (including the low dose
extrapolation procedure used) are discussed.
Whenever more than one interpretation of the weight-of-evidence for carcinogenicity or the
dose-response characterization can be supported, and when choosing among them is difficult, the
alternative views are provided along with the rationale for the interpretation chosen in the derivation
of the AWQC value. Where possible, quantitative uncertainty analyses of the data are provided; at
a minimum, a qualitative discussion of the important uncertainties is presented.
Important features of the risk characterizationinclude significant scientific issues, significant
science and science policy choices that were made when alternative interpretations of data exist, and
the constraints of the data and the state of knowledge. The assessments of hazard, dose-response,
and exposure are summarized to generate risk estimates for the exposure scenarios of interest.
The Proposed Cancer Guidelines contain more detailed guidance regarding the development
of risk characterization summaries and analyses.
2.1.3.6 Use of Toxicity Equivalence Factors (TEF) and Relative Potency
Estimates
The 1996 Proposed Guidelines for Carcinogen Risk Assessment (USEP A, 1991; 1996) state:
"A Toxicity Equivalence Factor (TEF) procedure is one used to derive quantitative dose-response
estimates for agents that are members of a category or class of agents. TEFs are based on shared
characteristics that can be used to order the class members by carcinogenic potency when cancer
bioassay data are inadequate for this purpose. The ordering is by reference to the characteristics and
potency of a well-studied member or members of the class. Other class members are indexed to the
reference agent(s) by one or more shared characteristics to generate their TEFs." In addition, the
Proposed Cancer Guidelines (USEPA, 1996) state that TEFs are generated and used for the limited
purpose of assessment of agents or mixtures of agents in environmental media when better data are
not available. When better data become available for an agent, its TEF should be replaced or
revised. To date, adequate data to support use of TEFs has been found in only one class of
compounds (dioxins) (USEP A, 1989).
The uncertainties associated with TEFs are explained when this approach is used. This is a
default approach to be used when tumor data are not available for individual components in a
mixture. Relative potency factors (RPFs) can be similarly derived and used for agents with
carcinogenicity or other supporting data. These are conceptually similar to TEFs, but are less firmly
based on science and do not have the same levels of data to support them. TEFs and relative potency
31
-------
factors are used only when there is no better alternative. When they are used, uncertainties
associated with them are discussed. As of today, there are only three classes of compounds for
which relative potency approaches have been examined by EPA: dioxins, polychlorinatedbiphenyls
(PCBs), and polycyclic aromatic hydrocarbons (PAHs).
2.1.4 Case Study (Compound Y, a Rodent Bladder Carcinogen)
This section illustrates an application of the non-linear method (MoE) for a rodent bladder
carcinogen (Compound Y). A brief summary of the data set is provided below with conclusions
regarding the weight-of-evidence. The AWQC obtained using the default linear and LMS
approaches are included for purposes of comparison only and would not be used for agents with the
characteristics described for Compound Y. In addition, considerably more detail would be provided
in a weight-of-evidence narrative.
2.1.4.1 Background and Evaluation for Compound Y
Compound Y is an organophosphonate which has been tested in subchronic, chronic,
reproductive, and carcinogenic assays in multiple species. Tumors were observed only in rat studies.
No human data are available. Based on a review of the toxicity, mechanistic, metabolic, and other
data summarized below for this agent, it was concluded that a non-linear approach is most
appropriate for establishing AWQC based on carcinogenicity.
Lifetime cancer bioassays of Compound Y identified bladder tumors and hyperplasia in male
and female rats at doses of 1500 mg/kg-day and higher in the diet. These effects were not observed
at 100 and 400 mg/kg-day. The rates of bladder cancer observed in females were lower than those
observed in males. In a 90-day study designed to evaluate the mechanisms of tumor induction, the
following sequence was identified as critical to bladder tumor formation in rats:
1) Large doses of Compound Y produce urinary calcium/potassiumimbalance followed
by
2) Diuresis, a sharp drop in urine pH, formation of urinary calculi, and
3) Appearance of transitional cell hyperplasia in the renal pelvis, ureter, and urinary
bladder.
These effects occurred within two weeks of exposure onset, persisted to the end of exposure,
and were reversible upon cessation of the 90-day exposure.
The pathological events caused by Compound Y are believed to result from prolonged
mechanical irritation by bladder calculi that developed in response to the exposure. At high but not
lower subchronic doses in the male rat, Compound Y leads to elevated blood phosphorus levels; the
body responds by releasing excess calcium into the urine. The calcium and phosphorus combine in
32
-------
the urine and precipitate into multiple stones in the bladder. The stones are very irritating to the
bladder; the bladder lining is eroded, and cell proliferation occurs to compensate for the loss of the
lining. This leads to development of hyperplasia, with subsequent tumor formation. A prolonged
increase in the rate of proliferation of cells of the urinary bladder has been proposed to be an
important step in the induction of urinary bladder tumors (Cohen and Ellwein, 1989; 1990). Thus,
the association of cell proliferation, hyperplasia, and subsequent cancer induction as a result of
urinary stone formations due to exposure to Compound Y is proposed as one mode of action which
may justify, after a review of all relevant data, the use of a non-linear approach, such as the MoE
approach.
Studies of the components of this agent yield no evidence of carcinogenicity in the bladder.
In metabolic studies in animals, the metallic component in isolation from the parent molecule was
not absorbed to a significant extent from the gastrointestinal tract.
Compound Y has been assessed via a battery of mutagenicity assays that have yielded
negative results, and a review of the chemical structure does not suggest potential genotoxicity. The
metabolites of Compound Y have also yielded negative results in mutagenicity assays and yielded
no evidence of carcinogenicity. The negative genotoxicity results for Compound Y and structurally
related agents provide further support for the use of a non-linear approach, such as the MoE
approach, to establish AWQC.
2.1.4.2 Conclusion and Use of the MoE Approach for Compound Y
Compound Y, a metal aliphatic phosphonate, is likely to be carcinogenic to humans only
under high-exposure conditions following oral and inhalation exposure that lead to bladder stone
formation, but is not likely to be carcinogenic under low-exposure conditions. It is not likely to be
a human carcinogen via the dermal route, given that the compound is a metal conjugate that is
readily ionized and its dermal absorption is not anticipated. The weight-of-evidence is based on (1)
bladder tumors only in male rats at high exposure; (2) the absence of tumors at any other site in rats
or mice; (3) the formation of calcium-phosphorus-containingbladder stones in male rats at high, but
not low, exposure. The bladder stones erode bladder epithelium and result in profound increases hi
cell proliferation and cancer; and (4) the absence of carcinogenic structural analogues or mutagenic
activity.
There is a strong mode of action basis for the requirements of high doses of Compound Y,
which leads to excess calcium and increased acidity in the urine, resulting in the precipitation of
bladder stones and subsequent increase in cell proliferation and tumor hazard potential. Lower doses
fail to perturb urinary constituents, lead to stones, produce toxicity, or give rise to tumors.
Therefore, dose-response assessment should assume non-linearity.
A major uncertainty is whether the profound effects of Compound Y may be unique to the
rat. Even if Compound Y produced stones in humans, there is only limited evidence that humans
with bladder stones develop cancer.
33
-------
Based on the progression of pathology leading to tumors, in which hyperplasia is an early
critical step, hyperplasia was selected as the sentinel precursor effect which was used as the basis
for the calculation of AWQC using the MoE approach. Hyperplasia incidence data in a lifetime rat
study are available for Compound Y. Tumor data from the same lifetime rat study were used to
calculate AWQC using the default linear and LMS approaches for purposes of comparing methods
and results. The data used for all three approaches are summarized in Table 2.1.1 below.
Table 2.1.1: Study Results from a Lifetime Exposure of Male Rats to Compound Y
Animal Dose in mg/kg-day
(scaled human equivalent doses)
0
400
(BW3/4 = 106.4)a
(BW273 = 68.4)b
1500
(BW3/4 = 398.9)a
(BW2/3 = 256.5)b
Number
in Group
•
73
78
78
Number Responding
tumors (combined
papilloma &
carcinoma)
3
2
21*
hyperplasia
5
5
29*
a. The 3/4 scaling factor is the new proposed method and is used with the new linear
model in this case study for comparison purposes.
b. The 2/3 scaling factor is presently in use and is used with the LMS method later in
this section for comparative purposes.
* There were statistically significant (p<0.05) increases in bom tumor incidence and
hyperplasia in the treated group compared to the control group.
Identification of the Point of Departure (Pdp)for Compound Y
The point of departure (Pdp) chosen for the MoE calculations was 400 mg/kg-day, which is
the maximum animal dose yielding no observable hyperplastic effects (the NOAEL shown in Table
34
-------
2.1.1).9 The study found males to be more sensitive to tumor induction than females and the
hyperplasiaresults in male rats were used for AWQC calculations. The human equivalent dose for
the NOAEL of 106.4 mg/kg-day was calculated using the new scaling factor of body weight raised
to the 3/4 power (as shown in Equation 2.1.3).
Discussion of the Points Affecting Selection of the SFfor Compound Y
Intraspecies Variability. There is variability within the human population in responses to
xenobiotic agents which may result from a variety of factors including health status, diet, age, and
genetic composition. Research on Compound Y did not identify a common health or genetic
condition which would yield a subpopulation who are particularly susceptible to the carcinogenic
effects of Compound Y nor did it indicate an exceptionally high or low level of intraspecies
variability.
Interspecies Variability. Animals and humans may vary widely in their responses to agents
due to their differing physiologies and metabolism. A review of human case studies and
epidemiological studies indicate that humans may be significantly less susceptible to the influence
of bladder irritation, stone formation, and subsequent tumor formation than male rodents. This
would suggest a smaller factor for interspecies variability.
Confidence in the Study (Dose Selection). There is a wide range in dose levels between the
NOAEL and LOAEL in the selected study. The hyperplastic response rate at the LOAEL is 37
percent (i.e., 29/78), which is high for the initial response measurement. Additional data would help
to refine the NOAEL and better describe the dose-response dynamic in the low response range.
Exposure Duration. This exposure scenario is chronic, so there is no need to apply an
additional safety factor.
Persistence. This chemical is not persistent in the body, so there is no need to apply an
additional safety factor.
Shape of the Dose-Response Curve. The data available indicate a steep slope at the point of
departure (at 400 mg/kg-day animal dose). This would suggest a rapid reduction in risk with lower
doses, or a smaller SF.
In summary, an overall SF of 30 is used in the MoE calculation. The selection of the SF is
based on a consideration of all the factors discussed above, such as intraspecies variability (10),
interspecies variability (3 is used here because animal dose has already been adjusted to a human
equivalent dose), and the adequate data base on this chemical. This factor of 30 is sufficient for
human health protection. The risk may decline considerably with doses lower than the point of
departure; the male rat is a very sensitive model (mice do not respond). Physiological phenomenon
This is based on a dietary conversion factor for rats from ppm to mg/kg-day of .05.
35
-------
is likely to fall off sharply with dose as shown by the dose-response curve. Further, bladder stone
and subsequent tumor formation is not a common phenomenon in humans.
AWQC Calculations for Compound Y
Equation 2.1.11 shown in Section 2.1.3 was used to calculate the AWQC for Compound Y:
Pdp
AWQC = —- x
BW
x RSC%
SF ^ DI+(FI x BAF),
(Equation 2.1.11)
The following input parameters were used:
Pdp = Point of departure (106.4 mg/kg-day (NO AEL))
SF = Safely factor of 30
BW = Body weight for adult (70 kg)
DI = Drinking water intake (2 L/day)
FI = Fish intake (0.01780 kg/day)
BAF = Assumed bioaccumulation factor (BAF) (300 L/kg)
RSC = 20% (assumed)
This calculation yields an AWQC of 6.7 mg/L. The body weight, water intake, fish intake.,
and RSC% values used in the above calculation are the currently proposed default values for adults
(see the exposure section of this document). The BAF, which accounts for the accumulation of
Compound Y from water through the food chain and into fish tissue, has been arbitrarily chosen for
purposes of this case study.
The AWQC calculations shown above is appropriate for water bodies that are used as sources
of drinking water (and for other uses). See Section 2.1.3.4 for additional information on
modifications for non-drinking water sources.
2.1.4.3 Use of the Default Linear Approach for Compound Y
This section is provided for purposes of illustrating the use of the default linear approach for
deriving AWQC based on carcinogenicity and to compare the resulting AWQC to that obtained
above using the MoE approach. As discussed in Section 2.1.4.1 above, it is important to note that
the default linear method would most likely not, in practice, be recommended as an approach for
quantifying the risk and deriving the AWQC for Compound Y given the hazard characteristics
described for this substance.
36
-------
Computing the Human Equivalent Dose for Compound Y
The doses used in the study were adjusted to obtain a human equivalent dose, as shown in
Table 2.1.1. In the absence of pharmacokinetic data, this was done using a scaling factor of BW3/4,
with a male rat weight of 0.35 kg and a human weight of 70 kg (as shown in Equation 2.1.3).
Calculation of A WQCfor Compound Y
To describe the dose-response of tumor incidence data in the observed range, a curve-fitting
model such as the multistage or other approach appropriate for the data can be used. In the case of
Compound Y, three data points (at doses of 0,400, and 1500 mg/kg-day) were used in the multistage
model (GLOBAL 86) to calculate the LED10 (the 95 percent lower confidence limit on a dose
associated with a 10 percent increase in response). The value obtained for the LED10 is 204 mg/kg-
day.
The cancer slope factor (m) is calculated by dividing 0.1 by the LED10 using Equation 2.1.6:
m=-
0.10
LED
10
(Equation 2.1.6)
This yields an estimated cancer slope factor of 4.9 x 10"4 per mg/kg-day. The cancer slope factor is
then used in Equation 2.1.9 with a specified risk level (in this case 10'6) to calculate a RSD:
_ Target Incremental Cancer Risk
—
m
(Equation 2.1.9)
This yields an RSD of 2.0 xlO'3 mg/kg-day.
The RSD is used in Equation 2.1.10 with the same input parameters (body weight, drinking
water intake, fish intake, and BAF) as those used for the MoE approach:
AWQC = RSD x
BW
x BAF)x
(Equation 2. 1.10)
37
-------
This yields an AWQC of 0.019 mg/L (rounded from 0.0189 mg/L) for a target risk of 10'6 (As noted
above, this approach is appropriate for water bodies used as drinking water sources. See Section
2.1.3.4 for non-drinking water sources).
2.1.4 A Use of the LMS Approach for Compound Y
This section is provided strictly for purposes of comparing the use of the MoE approach with
the traditional linearized multistage (LMS) method for deriving AWQC for carcinogens. As
discussed above, the LMS approach would not be used in practice to quantify risk and derive the
AWQC for Compound Y given the hazard characteristics described for this substance.
First, the LMS approach was used to fit the male rat tumor data shown in Table 2.1.1 with
the computer program GLOB AL86. This program calculates the 95th percentile upper confidence
limit on the linear slope (i.e., the q,*) in the low dose range. A human equivalent dose was
calculated using the BW2'3 interspecies dose scaling factor for purposes of illustrating the results
obtained applying the 1980 AWQC derivation methodology. The human equivalent doses obtained
using this scaling factor are shown in Table 2.1.1 above. (The same data set, using differently scaled
doses, was employed for both the new linear and LMS approaches.) The q,* value obtained using
the LMS approach is 6 x 10"4 (mg/kg-day)"1.
Equation 2.1.9 was used with a reference incremental cancer risk of 10"6 to calculate an RSD
of 1.7 x 1O'3. Equation 2.1.10 was then used to calculate the AWQC with the same input parameters
(body weight, drinking water intake, fish intake, and B AF) as those used for the MoE approach. (As
noted above, this approach is appropriate for water bodies used as drinking water sources. See
Section 2.1.3.4 for non-drinking water sources.) The AWQC was calculated to be 0.016 mg/L and
was rounded from 0.0157 mg/L.
2.1.4.5 Comparison of Approaches and Results for Compound Y
The results of the three approaches used for Compound Y are summarized in Table 2.1.2.
The AWQC calculated using the MoE approach is substantially higher than that obtained using the
default linear and LMS approaches. If larger or smaller SFs were used in the MoE calculations, the
AWQC obtained using the MoE approach would decrease or increase accordingly. The quantitative
relationship between AWQC derived using different methods will vary depending on the nature of
the data set and the SFs and Pdp selected for use in the MoE approach.
38
-------
Table 2.1.2: Comparison of AWQC Obtained for Compound Y
Using the MoE, Default Linear, and LMS Approaches
Method
AWQC (mg/L)
MoE
Using hyperplasia as a precursor for determining the Point of
Departure (Pdp) and a SF of 30.
6.7
Default Linear
Using linear extrapolation from the LED10 with a 10"6 risk level and
an interspecies scaling factor based on BW3/4.
0.019
LMS
Using the linearized multistage approach with a 10'6 risk level and
an interspecies scaling factor based on BW273.
0.016
2.1.5 References
Anderson, E.L. 1983. Quantitative Approaches in Use to Assess Cancer Risk. Risk Analysis. 3(4):
227-295.
Barnes, D.G., G.P Daston, J.S. Evans, A.M. Jarabek, R.J. Kavlock, C.A. Kimmel, C. Park, and H.L.
Spitzer. 1995. Benchmark Dose Workshop: Criteria for Use of a Benchmark Dose to
Estimate a Reference Dose. Regul. Toxicol. Pharmacol. 21: 296-306.
Doll, R. 1971. Weibull Distribution of Cancer: Implications for Models of Carcinogenesis. J. Roy.
Stat. Soc.A. 13: 133-166.
Mantel, N. and M. A. Schneiderman. 1975. Estimating "Safe Levels," a Hazardous Undertaking.
Cancer Res. 35: 1379.
Office of Science and Technology Policy (OSTP). 1985. Chemical Carcinogens: Review of the
Science and its Associated Principles. Federal Register 50: 10372-10442.
USEPA. 1976. Interim Procedures and Guidelines for Health Risks and Economic Impact
Assessment of Suspected Carcinogens. Federal Register 41:21402-21405.
39
-------
USEPA. 1986. Guidelines for Carcinogen Risk Assessment. Federal Register 51:33992-34003.
USEPA. 1989. Interim Procedures for Estimating Risks Associated with Exposures to Mixtures of
Chlorinated Dibenzo-p-dioxins and Dibenzofurans (CDDs and CDFs) and 1989 Update.
Washington, DC: Risk Assessment Forum. EPA/625/3-89/016.
USEPA. 1991. Workshop Report on Toxicity Equivalency Factors for Polychlorinated Biphenyl
Congeners. Washington, DC: Risk Assessment Forum. EPA/625/3-91/020.
USEPA. 1992a. Report of the National Workshop on Revision of the Methods for Deriving
National Ambient Water Quality Criteria for the Protection of Human Health. Washington,
DC.
USEPA. 1992b. Draft Report: A Cross-Species Scaling Factor for Carcinogen Risk Assessment
Based on Equivalence of Mg/Kg3/4/Day. Federal Register 57: 24152-24173.
USEPA. 1996. Proposed Guidelines for Carcinogen Risk Assessment. Federal Register 61: 17960.
April 23.
USEPA. 1998. Notice of Draft Revisions to the Methodology for Deriving Ambient Water Quality
Criteria for the Protection of Human Health. Federal Register Notice.
2.2 Noncancer Effects
2.2.1 Introduction
The evaluation of risks from noncarcinogenic chemicals traditionally has been based on the
assumption that noncarcinogens have a dose or level below which no adverse effects are expected
to occur. The risk parameter developed by EPA for noncarcinogens is called the Reference Dose
(RfD). The Integrated Risk Information System (IRIS) Background Document entitled Reference
Dose (RfD): Description and Use in Health Risk Assessments (USEPA, 1988) defines an RfD as "an
estimate (with uncertainty spanning approximately an order of magnitude) of a daily exposure to the
human population (including sensitive subgroups) that is likely to be without appreciable risk of
deleterious effects over a lifetime." The RfD is acknowledgedly an estimate and, thus, may not be
completely protective of every individual within a highly variable population, conversely, neither
are exposures above the RfD necessarily unsafe. Some "individuals may have better adaptive or
protective capacities than others and responses may vary with age and state of health; thus,
individuals respond differently to toxicant exposure (Barnes and Dourson, 1988).
The key step in deriving water quality criteria for the protection of human health from
noncancer effects is the determination of the RfD. As described in Section 1.4, the RfD is used in
concert with additional information regarding exposure and the bioaccumulation potential of the
40
-------
substance to derive an AWQC for noncancer effects. The procedures presented in USEPA (1988)
for deriving the RfD using an experimentally derived No Observed Adverse Effect Level
(NOAEL)/Lowest Observed Adverse Effect Level (LOAEL) approach are incorporated into this
chapter. The Agency is also investigating alternative methods for estimating the RfD; thus, this
guidance document contains information on two alternative methods, the Benchmark Dose (BMD)
and Categorical Regression approaches. The Agency continues to conduct research on the utility of
both of these methods in the noncancer risk assessment process and recommends their application
in circumstances where the data are sufficient. The Agency used the BMD approach to derive a
RfD for methyl mercury (USEPA, 1994a).
This section begins with a discussion of hazard identification and dose-response
characterization. This is followed by a description of factors to be considered in the selection of
critical data sets for use in the risk assessment evaluation. The procedures for deriving an RfD for
a substance using the traditional NOAEL/LOAEL approach are presented as the accepted current
risk assessment practice used by USEPA. Next, the BMD method for deriving an RfD is discussed
and an example of its application is provided for illustrative purposes. A brief discussion of
Categorical Regression is also included, with references to the relevant literature. The chapter
concludes with specific sections on several issues relevant to noncancer risk assessment, including
practical nonthreshold effects and risks from short-term exposures and mixtures.
While the intent of this guidance is to provide sufficient information to apply methods for
deriving RfDs, this document does not detail all relevant issues and underlying theory associated
with these methods. For further information, the reader is referred to the sources cited in the
reference list (in particular, USEPA, 1988; Crump et al., 1995; and Hertzberg and Miller, 1985).
2.2.2 Hazard Identification
The first step in the risk assessment involves preparing a hazard identification, based on a
review of data available to characterize the health effects associated with chemical exposure. The
RfD Background Document (USEPA, 1988) outlines considerations for choosing data upon which
to base a hazard identification for noncancer health effects.10 Assessors should prepare a hazard
identification document that describes the nature of exposure, the type and severity of effects
observed, and the quality and relevance of data to humans. Well-conducted human studies are
considered the best for establishing a link between exposure to an agent and manifestation of an
adverse effect. In the absence of adequate human data, the Agency relies primarily on animal
studies. In such cases, the principle studies are drawn from experiments conducted on laboratory
mammals, most often rat, mouse, rabbit, guinea pig, dog, monkey, or hamster. Well-designed
animal studies offer the benefit of controlled chemical exposures and definitive toxicological
analysis. Supporting evidence provides additional information for dose-response assessment and
may come from a wide variety of sources, such as metabolic and pharmacokinetic studies. In vitro
"The Agency has also developed guidelines that explain the process of hazard identification for developmental (USEPA, 1991a)
and reproductive (USEPA, 1994b) effects. Please refer to these EPA documents for guidance in these areas.
41
-------
studies seldom provide definitive hazard identification data, but they can often provide insight into
the compound's potential for human toxicity.
Important to the hazard identification is consideration of the biological and statistical
significance of observed effects. The determination of whether an effect is adverse requires
professional judgment. For guidance, adverse health effects are those deleterious effects which are
or may become debilitating, harmful, or toxic to the normal functions of an organism, including
reproductive and developmental effects. Adverse effects do not include such effects as tissue
discoloration without histological or biochemical effects, or the induction of the enzymes involved
in the metabolism of the substance. Guidelines for defining the severity of adverse effects have been
suggested by Hartung and Durkin (1986). EPA has also developed guidelines for the ranking of
observed effects (USEPA, 1995) and a ranking scheme for slight to severe effects. Distinguishing
slight effects such as reversible enzyme induction and reversible subcellular change from more
serious effects is critical in distinguishing between a NOAEL and LOAEL.
It is also important to evaluate the reversibility of an effect. Reversibility refers to whether
or not a change will return to normal or within normal limits either during the course of or following
exposure. However, even a reversible effect may be adverse to an organism. In performing a hazard
identification, irreversible effects should be distinguished from less serious, but still adverse,
reversible changes.
The exposure conditions for toxicity tests, including the route (e.g., inhaled versus
ingested), source (e.g., water versus food), and duration, should be discussed in the hazard
identification. The hazard identification should also include an evaluation of the quality of studies.
Elements that affect the quality of studies include the soundness of the study protocol, the adequacy
of data analysis, the characterization of the study compound, the types of species used, the number
of individuals per study group, the number of study groups, dose spacing, the types of observations
recorded, sex and age of animals, and the route and duration of exposure (USEPA, 1988).
The hazard identification should conclude with a weight-of-evidencediscussion. In general,
the discussion should review the results of different studies and develop an overall picture of the
chemical's toxicity. Evidence for possible toxicity in humans is supported by similar results across
species and across investigators. A plausible mechanism of action for the effect, as well as similar
toxic activity in chemicals of similar structure, also add to the weight-of-evidence.
2.2.3 Dose-Response Assessment
The dose-response assessment involves the evaluation of toxicity data to identify doses at
which statistically and/or biologically significant effects occur and identify NOAEL and/or LOAEL
values. The effects data are also evaluated to see if there is a quantitative relationship between dose
and the magnitude of the effect. Dose-response relationships can be linear, curvilinear or U-shaped.
The RfD is traditionally estimated by identifying the most appropriate NOAEL for the critical
42
-------
effect. The LOAEL may be used to estimate the RfD if no appropriate NOAELs have been
identified.
2.2.4 Selection of Critical Data
2.2.4.1 Critical Study
Ideally, the scientific data for noncancer effects should include sufficient information to
characterize quantitatively the incidence and severity of response as dose increases. However,
complete data are frequently lacking. Instead, the Agency bases the derivation of the RfD on the
NOAEL or LOAEL from a critical study or collection of critical studies. The choice of the critical
study or studies to use in the derivation of the chronic RfD requires professional judgment
concerning the quality of the studies, the definition of adverse effects and their level of occurrence.
As part of the hazard identification, all relevant toxicity data on a chemical should be evaluated to
support the establishment of the RfD. Those studies representing the best quality and most
appropriate data should be considered for defining adverse effects and their level of occurrence.
In choosing a study on which to base the RfD, the Agency recommends a hierarchy of
acceptable data. Most preferable is a well-conducted epidemiologic study that demonstrates a
positive association between a quantifiable exposure to a chemical and human disease. Use of
acceptable human studies avoids the problems of interspecies extrapolation, and thus, confidence
in the estimate is often greater. At present, however, human data adequate to serve as a basis for
quantitative risk assessment are available for only a few chemicals. Most often, inference of adverse
health effects for humans must be drawn from toxicity information gained through animal
experiments with human data serving qualitatively as supporting evidence. Under this condition,
health effects data must be available from well-conducted animal studies and relevant to humans
based on a defensible biological rationale, e.g., similar metabolic pathways. In the absence of data
from a more "relevant" species, data from the most sensitive animal species tested, i.e., the species
demonstrating an adverse health effect at the lowest administered dose via a relevant route of
exposure, shall generally be used as the critical study.
The route of administrationmust be considered when choosing the critical study from among
quality toxicity tests. The vehicle in which the chemical is administered is also relevant. For
example, within the oral route of exposure, the bioavailability of a chemical ingested from one
source (e.g., food) may differ from when it is ingested from another source (e.g., water). Usually,
the toxicity data base does not provide data on all possible routes, sources, and/or durations of
administration. In general, the preferred exposure route is that which is considered most relevant
to environmental exposure. For example, when developing drinking water standards, the Agency
has placed greater weight on oral studies in experimental animals, especially those studies in which
the contaminant is administered via water. However, in the absence of data on the exposure route
and/or source of concern, it is the Agency's view that the potential for the toxicity manifested by one
route and/or source of exposure may be relevant to other exposure routes and/or sources. EPA
guidelines for the development of interim inhalation reference concentrations (USEPA, 1989)
43
-------
discuss specific issues relevant to route-to-route extrapolation. These include issues of portal-of-
entry effects, available pharmacokineticdata for the routes of interest, measurements of absorption
efficiency by each route of interest, comparative excretion data when the associated metabolic
pathways are equivalent by each route of interest, and comparative systemic toxicity data when such
data indicate equivalent effects by each route of interest.
Preference should be given to studies involving exposure over a significant portion of the
animal's lifespan since this is anticipated to reflect the most relevant environmental exposure.
Studies with shorter time frames can rniss important effects. In selected cases, studies of less than
90 days can be used for quantification but the study must be of exceptionally high quality. In
general short-term tests should not be used for anything other than interim RfDs or for
developmental RfDs. However, developmental effects can sometimes be the critical effect and serve
as the basis of an RfD. The duration of a developmental study is generally less than 15 days.
2.2.4.2 Critical Data and Endpoint
The experimental exposure level representing the highest dosage level tested at which no
adverse effects were demonstrated in any of the species evaluated should be used for criteria
development. By basing criteria on the critical toxic effect, it is assumed that all toxic effects are
prevented (USEPA, 1988). In the absence of such data, the lowest LOAEL dosage may be used for
criteria development and an additional uncertainty factor for LOAEL to NOAEL extrapolation is
applied. When two or more studies of equal quality and relevance exist, the geometric means of the
NOAELs or LOAELs may be used.
Often a chemical may elicit multiple effects, each with a different NOAEL and LOAEL.
From among these effects, the Agency selects a critical endpoint. The critical endpoint is the effect
that exhibits the lowest LOAEL (USEPA, 1988).
2.2.5 Deriving RfD Using the NOAEL/LOAEL Approach
The IRIS background document (USEPA, 1988) describes methods used to derive an RfD
for a given chemical and criteria for selection of the critical NOAEL or LOAEL. Appropriate
uncertainty factors (UF) and modifying factors (MF) are then applied to the selected endpoint to
derive the RfD.
The general equation for deriving the RfD is (USEPA, 1988):
^ „ J NOAEL LOAEL
RfD (mg/kg-day) = or
UF*MF
(Equation 2.2.1)
UF*MF
44
-------
where:
NOAEL = An exposure level at which there are no statistically or biologically
significant increases in the frequency or severity of observed adverse
effects between the exposed population and its appropriate control;
some effects may be produced at this level, but they are not
considered as adverse, nor precursors to specific adverse effects.
LOAEL = The lowest experimental exposure level at which there are
statistically or biologically significant increases in frequency or
severity of observed adverse effects between the exposed population
and its appropriate control group. The LOAEL may be used if the
NOAEL cannot be determined.
UF = An uncertainty factor which reduces the dose to account for several
areas of scientific uncertainty inherent in most toxicity data bases.
Standard UFs are used to account for variation in sensitivity among
humans, extrapolation from animal studies to humans, and
extrapolation from less than chronic NOAELs to chronic NOAELs.
An additional UF may be employed if a LOAEL is used to define the
RfD.
MF = A modifying factor, to be determined using professional judgment.
The MF provides for additional uncertainty not explicitly included in
UF, such as completeness of the overall data base and the number of
species tested. (The value for MF must be greater than zero and less
than or equal to 10; the default value for the MF is 1).
The RfD is generally expressed in units of milligrams per kilogram of body weight per day
(mg/kg-day).
2.2.5.1 Selection of Uncertainty Factors and Modifying Factors
The choice of appropriate UFs and MFs must be a case-by-case judgment by experts and
should account for each of the applicable areas for uncertainty and nuances in the available data that
impact uncertainty. Several reports describe the underlying basis of UFs (Zielhuis and van der
Kreek, 1979; Dourson and Stara, 1983) and research into this area (Calabrese, 1985; Hattis et al.,
1987; Hartley and Ohanian, 1988; Lewis et al., 1990; Dourson et al., 1992).
The uncertainty factors (UFs) summarized in Table 2.2.1 account for five areas of scientific
uncertainty inherent in most toxicity databases: inter-human variability (H) (to account for variation
in sensitivity among the members of the human population); experimental animal-to-human
extrapolation (A); subchronicto chronic extrapolation(S) (to account for uncertainty in extrapolating
45
-------
from less-than-chronic NOAELs (or LOAELs) to chronic NOAELs); LOAEL to NOAEL
extrapolation (L); and data base completeness (D) (to account for the inability of any single study
to adequately address all possible adverse outcomes). Each of these five areas is generally addressed
by the Agency with a factor of 1,3, or 10. The default value is 10.
In addition, a modifying factor (MF) may be used to account for areas of uncertainty that are
not explicitly considered using the standard UF. This value of the MF is greater than zero and less
than or equal to 10, but it should generally be used on a log 10 basis (i.e., 0.3, 1, 3, 10) as are the
standard UFs. The default value for this factor is 1.
Uncertainty Factor
UFH
Table 2.2.1: Uncertainty Factors and the Modifying Factor
Definition
Use a 1-, 3-, or 10-fold factor when extrapolating from valid data in studies using long-term
exposure to average healthy humans. This factor is intended to account for the variation in
sensitivity (intraspecies variation) among the members of the human population.
UFA Use an additional 1-, 3-, or 10-fold factor when extrapolating from valid results of long-term
studies on experimental animals when results of studies of human exposure are not available or
are inadequate. This factor is intended to account for the uncertainty involved in extrapolating
from animal data to humans (interspecies variation).
UFS Use an additional 1-, 3-, or 10-fold factor when extrapolating from less-than-chronic results on
experimental animals when there are no useful long-term human data. This factor is intended
to account for the uncertainty involved in extrapolating from less-than-chronic NOAELs to
chronic NOAELs.
UFL Use an additional 1-, 3-, or 10-fold factor when deriving an RfD from a LOAEL, instead of a
NOAEL. This factor is intended to account for the uncertainty involved in extrapolating from
LOAELs to NOAELs.
UFD Use an additional 1-, 3-, or 10-fold factor when deriving an RfD from an "incomplete" data
base. Missing studies, e.g., reproductive, are often encountered with chemicals. This factor is
meant to account for the inability of any study to consider all toxic endpoints. The
intermediate factor of 3 QA log unit) is often used when there is a single data gap exclusive of
chronic data. It is often designated as UFD.
Modifying Factor
Use professional judgment to determine the MF, which is an additional uncertainty factor that is greater than
zero and less than or equal to 10. The magnitude of the MF depends upon the professional assessment of
scientific uncertainties of the study and data base not explicitly treated above (e.g., the number of species tested).
The default value for the MF is 1.
Note: With each UF or MF assignment, it is recognized that professional scientific judgment must be used.
46
-------
The Agency's reasoning in its use of the MF is that the areas of scientific uncertainty labeled
H, A, S, L, or D do not represent all of the uncertainties in the estimation of an RfD. For example,
the fewer the number of animals used in a dosing group, the more likely it is that no adverse effect
will be observed at a dose point which may have had an effect in a larger population. Such a case
might-argue for modifying the usual 10-fold factors—a 100-fold UF might be raised to 250 if too
few animals were used in a chronic study. While this increase is scientifically reasonable, it
introduces two difficulties: the adjustments applied could differ between risk assessors, and the
applied precision of the result might not be justified by the data. For example, a UF of 250 has an
implied precision of 2 digits and is not appropriate in relation to the variability of the biological
response. The Agency intends to avoid these difficulties through limiting the options for the
modifying factor (1, 3, 10).
In practice, the magnitude of the overall UF is dependent on professional judgment as to the
total uncertainty in all areas. When uncertainties exist in one, two or three areas, the Agency
generally uses 10-, 100-, and 1,000-fold UF respectively. When uncertainties exist in four areas, the
Agency generally uses an UF no greater than 3,000. It is the Agency's opinion that toxicity data
bases that are weaker and would result in UFs in excess of 3,000 are too uncertain as a basis for
quantification. In such cases, the Agency does not estimate an RfD, and additional toxicity data are
sought or awaited. For a few chemicals, an UF of 10,000 was applied. However, in such cases, the
risk assessment was completed before current policies for the maximum UF were in place.
The Agency occasionally uses a factor of less than 10 or even a factor of 1, if the existing
data reduce or obviate the need to account for a particular area of uncertainty. For example, the use
of a 1-year rat study as the basis of an RfD may suggest the use of a 3-fold, rather than 10-fold,
factor to account for subchronic to chronic extrapolation, since it can be empirically demonstrated
that 1-year rat NOAELs are generally closer in magnitude to chronic values than are 3-month
NOAELs (Swartout, 1990). Lewis et al. (1990) more fully investigate this concept of variable
uncertainty factors through an analysis of expected values.
The modification of UFs from their standard values should follow the general guidelines for
composite UFs and the overall precision of one digit for UFs. The composite uncertainty factor to
use with a given data base is again strictly a case-by-case judgment by experts. It should be flexible
enough to account for each of the applicable five areas of uncertainty and any nuances in the
available data that might change the magnitude of any factor. The Agency describes its choice for
the composite UF and sub-components for individual RfDs on its Integrated Risk Information
System (IRIS). Because of the high degree of judgment involved in the selection of uncertainty and
MFs, the risk assessment justification should include a detailed discussion of the selection of
uncertainty factors, along with the data to which they are applied.
47
-------
2.2.5.2 Confidence in NOAEL/LOAEL-Based RfD
As stated previously, when available, adequate data from acceptable human studies should
be used as the basis for the RfD. Use of good epidemiology studies generally give the highest
confidence in RfDs. In the absence of such data, RfDs are estimated from studies in experimental
animals.
The Agency generally considers a "complete" data base for calculating a chronic RfD for
noncancer health effects to include the following:
• Two adequate mammalian chronic toxicity studies, by the appropriate route in
different species, one of which must be a rodent.
• One adequate mammalian multi-generation reproductive toxicity study by an
appropriate route.
• Two adequate mammalian developmental toxicity studies by an appropriate route in
different species.
For a "complete" database, the likelihood that additional toxicity data may change the RfD
is low. Thus, the Agency usually has confidence in such an RfD because additional toxicity data are
not likely to change the value.
The Agency considers aNOAEL from a well-conducted, mammalian subchronic (90-day)
study by the appropriate route as a minimum data base for estimating an RfD. However, for such
a data base, additional toxicity data may change the RfD. Thus the Agency generally has less
confidence in such an RfD.
For some chemicals, an acute health hazard is the critical effect of concern. These could
include neurotoxic or immunotoxic effects of acute exposures at environmental levels of
contaminant. In such cases, longer term studies (subchronic or chronic) that would typically be
included in a review of the toxicity literature may not capture the critical endpoint. Under such
circumstances, greater emphasis should be placed on characterizing the acute threshold as opposed
to the potential chronic effects.
Developmental toxicity data, if they constitute the sole source of information, are not
considered an adequate basis for chronic RfD estimation. This is because such data are often
generated from short-term chemical exposures, and, thus, are of limited relevance in predicting
possible adverse effects from chronic exposures. However, if a developmental toxicity endpoint is
the critical effect established from a "complete" data base, a chronic RfD can be derived from such
data, applying the uncertainty and MFs normally required. Developmental data are the basis for
48
-------
developmental reference doses (RfDDT).'' The term RfDuj is used to distinguish the developmental
value from the chronic RfD which refers to chronic exposure situations. Uncertainty factors for
developmental toxicity include a 10-fold factor for interspecies variation and a 10-fold factor for
intraspecies variation; in general, an uncertainty factor is not applied to account for duration of
exposure. In some cases, additional factors may be applied due to a variety of uncertaintiesthat exist
in the data base. For example, the standard study design for developmental toxicity study calls for
a low dose that demonstrates a NOAEL, but there may be circumstances where a risk assessment
must be based on the results of a study in which a NOAEL for developmental toxicity was not
identified. For details regarding risk assessment for developmental toxicants, refer to EPA risk
assessment guidelines (USEPA, 1991 b).
2.2.5.3 Presenting the RfD as a Single Point or as a Range
Although the RfD has traditionally been presented and used as a single point estimate, its
definition contains the phrase "... an estimate (with uncertainty spanning perhaps an order of
magnitude)..." (USEPA, 1988). Underlying this concept is the reasoning that during the derivation
of the RfD, the selection of the critical effect and~of the total uncertainty factor is based on the "best"
scientific judgment of the Agency Work Group and that other groups of competent scientists
examining the same database would reach a similar conclusion, within an order of magnitude. For
example, although EPA recently verified a single number as the RfD for arsenic (0.3 ug/kg-day),
there was not a clear consensus on the oral RfD. Applying the Agency's RfD methodology, "strong
scientific arguments can be made for various values within a factor of 2 or 3 of the currently
recommended RfD value, i.e., 0.1 to 0.8 ug/kg/day" (USEPA, 1993).
Presenting the RfD as a range may be more appropriate than expressing it as a point estimate
because rarely are sufficient data available to precisely determine a lifetime threshold for a human.
Even when there are good, reliable data, the variability of response in the human population argues
for expressing the RfD as a range. However, although EPA supports the use of a range that spans
one order of magnitude for most RfDs, there are a number of potential interpretations of the term
"order of magnitude" as described below:
• Range = x to lOx. (where point estimate of RfD=x). This view is supported by those
who believe that the risk assessment process is so inherently conservative that the
RfD should be considered to be the lowest estimate, with the range of imprecision
all resting above this point estimate.
• Range = 0.3x to 3x. This view is held by many EPA scientists who have developed
RfDs. The RfD point estimate, x, is the midpoint of a range that spans an order of
magnitude.
"A RfD for developmental toxicity (RfDOT) is discussed in USEPA (1991a).
49
-------
• Range = O.lx to x. This is the view held by many risk managers. Regulatory
decisions (e.g., setting of standards or cleanup levels) are made based on the
assumption that standards or cleanup levels are protective as long as they do not
exceed the RfD.
• Range = O.lx to lOx. This range represents the assumption that the order of
magnitude range could be on either side of the point estimate x.
The Agency is considering a risk management approach where the upper and lower bounds
of the range are correlated to the uncertainty. Because the uncertainty around the dose response
relationship increases as extrapolation below the observed data increases, the use of a range for the
RfD may be more appropriate in characterizing risk than the use of a point estimate. Therefore, as
a matter of risk management policy, it is proposed that if the product of the UFs and MF used to
derive the RfD is 100 or less, there would be no consideration of a range. When greater than 100
but less than 1,000, the maximum range that could be considered would be one half of a log,0 (3-
fold) or a number ranging from the point estimate divided by 1.5 to the point estimate multiplied by
1.5. At 1,000 and above the maximum range would be a log,0 (10-fold) or a number ranging from
the point estimate divided by 3 to the point estimate multiplied by 3.
EPA advocates the use of the point estimate of the RfD as the default to derive the AWQC.
The use of another number within the range defined by the uncertainty would then have to be
justified. As used in this document, justificationmeans that there are scientific data which indicate
that some value in the range other than the point estimate may be more appropriate than the point
estimate, based on human health or environmental fate considerations. Table 2.2.2 gives examples
of some factors to consider when determining whether to use the point estimate of the RfD or values
higher or lower than the point estimate. The factors presented in Table 2.2.2 should be considered
in making the decision as to whether or not to use a value other than the point estimate within a
range; the uncertainty will influence the magnitude of the range.
50
-------
Table 2.2.2: Some Scientific Factors to Consider
When Using the RfD Range
Use point estimate RfD
Use lower range of RfD
Use upper range of
RfD
- Default position
- Total UF/MF product is 100 or less
- Essential nutrient
- Increased bioavailability from medium
- The seriousness of the effect and whether or not it is reversible
- A shallow dose-response curve in the range of observation
- Exposed group contains a sensitive population (e.g., children or
fetuses)
- Decreased bioavailability with humans
- RfD based on minimal LOAEL and a UF/MF of 1,000 or greater
- A steep dose-response curve in the range of observation
- No sensitive populations identified
The use of an order of magnitude may not be appropriate for all chemicals. There are many
factors that can affect the degree of "precision" of the RfD, and thereby affect the magnitude of the
RfD range. The completeness of the data base plays a major role. Observing the same effects in
several animal species, including humans, can increase confidence in the RfD point estimate and
thereby narrow the range of uncertainty. Other factors that can affect the precision are the slope of
the dose-response curve, seriousness of the observed effect, spacing of doses, and the route of
exposure. For example, a steep dose-response curve indicates that relatively large differences in
effect occur with a given change in dose; thus, there will be a greater chance that the data will allow
scientists to distinguish clearly (i.e., statistically) between doses that produce an effect and those that
do not. For a situation where the RfD is derived from a LOAEL for a serious effect, an additional
uncertainty factor is often used in the RfD derivation to protect against less serious effects that could
have occurred at lower doses had lower doses been evaluated. Dose spacing and the size of the study
groups used in the experiment can also affect the confidence in the RfD. The "true" NOAEL can
fall anywhere between the experimentally determined NOAEL (the highest dose administered
without an adverse effect) and the LOAEL (the lowest dose administered causing an observable
adverse effect). The wider the dose spacing, the greater the margin of uncertainty about where the
"true" NOAEL may fall. Finally, for some RfDs, the route of exposure in the experiment may not
match the route of exposure in the environment, and interroute extrapolation may be considered
using assumptions about differences in absorption rates between routes.
There are cases when a range should not be used. For example, the RfD for zinc (USEPA,
1992) is based on consideration of nutritional data, a minimal LOAEL, and a UF of 3. If the factor
51
-------
of 3 were used to bound the RfD for zinc, then the upper-bound level would approach the minimal
LOAEL. This situation must be avoided, since it is unacceptable to set a standard at levels that may
cause an adverse effect. The risk manager must be informed of those specific cases when it is not
scientifically correct to use the RfD range. Table 2.2.2 provides managers with guidelines on the
scientific basis for using the range.
2.2.6 Deriving an RfD Using a Benchmark Dose Approach
A number of issues have been raised regarding the development of the RfD based on the
traditional NOAEL/LOAEL approach. These concerns include the following:
• The traditional approach does not incorporate information on the shape of the dose-
response curve, but focuses only on a single point (the NOAEL or LOAEL).
• The value of the NOAEL depends on the number and spacing of the doses in the
experiment. The possible NOAEL values are limited to the discrete values of the
experimental doses. Theoretically, the experimental no adverse effect level could be
any value between the experimental NOAEL and the LOAEL, and typically the true
NOAEL is below the observed NOAEL.
• Data variability is not directly taken into account. For example, studies based on a
larger number of animals may detect effects at lower doses than studies with fewer
animals; as a result, the NOAEL from a small study may be higher than the NOAEL
from a similar but larger study in the same species. The traditional approach does not
have a mechanism to account for such data variability.
• The determination of the NOAEL is dependent on the background incidence of the
effect in control animals; therefore, statistically significant differences between the
dose groups and the control group are more difficult to detect if background
incidence is relatively high, even if biologically significant effects occur.
• In conjunction with exposure data, the NOAEL-based RfD can be used to estimate
the size of the population at risk, but not the magnitude of the risk.
In response to these concerns, alternative approaches have been developed that attempt to
address some of these shortcomings. One such alternative, the BMD approach, has been the subject
of extensive research over the past decade (Crump 1984,1995; Gaylor, 1983, 1989; Dourson et al.,
1985; Brown and Erdreich, 1989; Kimmel, 1990; Faustman et al., 1994; Allen et al., 1994a, 1994b).
The following discussion presents the general methods for calculation of a RfD using the BMD
approach; for more extensive discussion, the reader is referred to Crump et al. (1995). To date, the
Agency has used the BMD approach for deriving the RfD for methyl mercury (USEPA, 1994a) and
the RfC for several compounds.
52
-------
2.2.6.1 Overview of the Benchmark Dose Approach
A benchmark dose (BMD) or concentration (BMC) is defined as a statistical lower
confidence limit on the dose producing a predetermined level of change in response (the benchmark
response-BMR) relative to controls The BMD/BMC is intended to be used as an alternative to the
NOAEL in deriving a point of departure for low dose extrapolations. The BMD/BMC is a dose
corresponding to some change in the level of response relative to background and is not dependent
on the doses used in the study. The BMR is based on a biologically significant level of response or
on the response level at the lower detection limit of the observable dose range for a particular
endpoint in a standard study design. The BMD/BMC approach does not reduce uncertainty inherent
in extrapolating from animal data to humans (except for that in the LOAEL to NOAEL
extrapolation), and does not require that a study identify a NOAEL, only that at least one dose be
near the range of the response level for the BMD/BMC.
The central step in deriving a BMD is to calculate lower bounds directly on the dose
estimate. This modeling process is limited to the experimental range and no attempt is made to
extrapolate to doses far below the experimental range. Generally, the models used in the BMD
approach are statistical rather than biologically-based models; thus, they cannot be reliably used to
extrapolate to low doses without incorporating detailed information on the mechanisms through
which the toxic agent causes the particular effect being modeled.
Once a mathematical dose-response curve and its corresponding curve of confidence limits
are established, the assessor selects a point on the lower confidence dose curve corresponding to the
chosen BMR (e.g., 1 percent, 5 percent, or 10 percent increase in the incidence of an effect). This
point on the lower confidence curve is the lower confidence bound of the effective dose for that
BMR (denoted as the BMD) (see Exhibit 2.2.1). A BMD may be calculated for each agent for which
there is an adequate data base.
The BMD approach offers a number of advantages over the traditional approach for the
derivation of the RfD from the NOAEL/LOAEL divided by uncertainty factors. The advantages of
the BMD approach are that it considers the dose-response curve, including its shape; better accounts
for statistical variability in the data; is not overly sensitive to dose spacing and, thus, is not limited
to experimental doses for determining the effect level. In addition, studies with small group sizes
and evaluation of a limited number of endpoints, which may identify artificially high NOAELs, will
tend to yield lower BMD values because the confidence bands will be wider. The BMD analyses
for developmental effects shows that the NOAELs from studies are actually at about a 5 percent
response level (Faustman et al., 1994). Therefore, the BMD approach provides an incentive to
conduct more robust studies, since better studies give narrower confidence bands.
53
-------
Exhibit 2.2.1
Derivation of RfD Using BMD Approach
G>
V)
Q
a.
W
100%-
90% 1
80% i
70% -
60% _
50% _
40%-
30% -
20% -
10% _
0% _.
0
Dose Response
Modeling uses animal
or human data
50
100
150
Dose
200
250
300
2.2.6.2 Calculation of the RfD Using the Benchmark Dose Method
The determination of an RfD using the BMD approach involves four basic steps. The first
step involves the selection of the experiments and responses that will be used for modeling the
BMD. Second, BMDs are calculated for the selected responses; BMD values should be calculated
for all endpoints that have the potential for yielding the critical BMD. Third, a single BMD is
selected from among those calculated. Finally, the RfD is calculated by dividing the chosen BMD
by appropriate uncertainty factors. The decision points associated with these steps are outlined in
Table 2.2.3. The discussion below summarizes the critical issues unique to the BMD approach. The
following discussion of the issues largely incorporates the information from Crump et al. (1995).
54
-------
Table 2.2.3: Steps and Decisions Required in the BMD Approach
Step
1. Selection Study/Response
2. Model dose-response
3. Select BMD(s)
4. Calculate RfD
Decisions
1. Experiments to include
2. Responses to model
1. Format of data
2. Mathematical model(s)
3. Handling model fit
4. Measure of altered response
1. Critical BMR
2. Confidence limit calculation
1. Uncertainty factors
Source: Crump et al., 1995
Selection of Response Data to Model
The selection of experiments and responses suitable for BMD modeling involves
considerations similar to those for identifying the appropriate studies upon which to base a NOAEL.
There may be several appropriate studies and relevant health effects that could be modeled for a
chemical. Ideally, BMD calculations would be performed for the complete set of relevant effects.
However, utilizing all relevant responses for the calculation of BMDs may be resource-intensive.
Further, it is difficult to interpret results from a large number of dose-response analyses. When
selecting the data to model it is often considered appropriate to limit attention to those responses for
which there is evidence of a dose-response relationship. Statistically, such a relationship may be
indicated by significant trends (either increasing or decreasing) in the response as dose level
increases. Considerations of biological significance may also be warranted. Another alternative is
to focus efforts on modeling the most critical effects as seen at the LOAEL. However, limiting the
number of responses modeled may potentially misrepresent the minimum BMD.12
l2This is due to the fact that an effect seen only at doses above the LOAEL but having a shallow dose-response could produce a lower BMD
than an effect seen at the LOAEL, which has a steeper dose-response.
55
-------
Use of Categorical Versus Continuous Data
A central issue in the selection of data to model concerns the form of the data used.
Categorical data, particularly quantal data, are relatively straightforward to use in the BMD
approach, since the data are expressed as the number (or percent) of subjects exhibiting a defined
response at a given dose. Data may also be of the continuous form, where results are expressed as
the measure of a continuous biological endpoint, such as a change in organ weight or serum enzyme
level. With continuous data the results are generally presented in terms of means and standard
deviation for dose groups but are most valuable when data for individual animals are available. To
perform dose-response modeling of such data, the way the data are expressed must be decided.
Continuous data can be modeled by looking at the mean response for each dose group as a fraction
of the mean response of the control group or as the percentage of animals showing an adverse
response at each dose level. (Gaylor and Slikker, 1990; Crump, 1995). Such approaches take
advantage of the continuous nature of the response data, but express the results in terms that are
directly comparable to those derived from analysis of categorical data, i.e., in terms of additional or
extra risk, rather than hi terms of changes in mean response. In particular, Crump (1995) has
extended those considerations so that the model used for analysis of a continuous endpoint yields
the same model type as used for analysis of any quantal endpoints. Such developments have
enhanced the consistency of results across different endpoints for any particular chemical. In any
case, application of the BMD approach to continuous data requires professional judgment in order
to determine what level or category of response constitutes an abnormal (adverse) effect. The BMD
approach is not recommended for routine use but may be used when data are available and justify
the extensive analyses required.
Choice of Mathematical Model
Various mathematicalapproacheshave been proposed for determiningthe BMD. Table 2.2.4
shows a number of dose-response models that may be used for estimating the BMD with quantal or
continuous data.
Information generally required for application of dose-response models for categorical
(including quantal) data includes the experimental doses, the total number of animals in each
dose group, and the number of these whose responses are in each of the categories of response. For
continuous data, the experimental doses, number of animals in each dose group, mean response in
each group, and sample variance of response in each dose group are needed.
The BMD approach should not be applied to data sets with only two experimental groups (a
control and one positive dose). In such cases, much of the advantage of the BMD approach with
respect to consideration of the dose-response shape will be lost; such data supply little information
about the shape of the dose-response curve. The more doses available, especially at lower doses, the
greater the expected benefit of the BMD approach as compared to the NOAEL-based approach.
56
-------
Table 2.2.4: Dose-Response Models Proposed for Estimating BMDs
Model
Formula
Quantal Data
Quantal linear regression (QLR)
Quantal quadratic regression (QQR)
Quantal polynomial regression (QPR)
Quantal Weibull (QW)
Log-normal (LN)
= c + (l-c){l-exp[-q1(d-d0)]}
P(d) = c + (l-c){.l-exp[-qi(d-d0)2]}
P(d) = c + (l-c){l-exp[-qidr...qkdk]}
P(d) = c + (l-c){l-exP[-qidk]}
P(d) = c + (l-c)N(a+b log d)
Continuous data
Continuous linear regression (CLR)
Continuous quadratic regression (CQR)
Continuous linear-quadratic regression
(CLQR)
Continuous polynomial regression
(CPR)
Continuous power (CP)
m(d) = c + q,(d-do)
m(d) = c + q,d+q2d2
m(d) = c + q,d+...+qkdk
m(d) = c + q,(d)k
Notes: P(d) is the probability of a response at the dose, d; m(d) is the mean response at the dose, d.
In all models, c, qj,...qk, and d0 are parameters estimated from the data. For the quantal models,
0<;c<; 1 and q^O. For the CPR model proposed by Crump (1984), all the q{ have the same sign. In
the CLQR model discussed by Gaylor and Slikker (1990), q, and q2 were not constrained to have
the same sign. For all models, d02:0, ks> 1. N(x) denotes the normal cumulative distribution
function.
SOURCE: Crump et al., 1995.
57
-------
Handling Model Fit
Fitting the models to experimental data gives estimates of the parameters that describe each
model. This fitting, usually accomplished through maximum likelihood methods, estimates the
probability of response (for quanta! data) or the mean response (for continuous data) for each dose
level. Goodness-of-fit tests can be used to determine if a model adequately describes the dose-
response data.
In many cases, several models may fit the data well. In these cases, other considerations can
be used to select an appropriate model. For example, the statistical assumptions underlying the
model should be reasonable for the given data. Quantal results, for example, Eire assumed to follow
a binomial distribution around a dose-dependent expected value. This assumption requires that each
subject responds independently and all have an equal probability of responding. Continuous
responses for each dose level are assumed to follow a normal distribution and are also assumed to
be independent. When biological factors may be important (e.g., intralitter correlation for
developmental toxicity data) they may also be used to select appropriate models. Another biological
considerationmay be whether or not a threshold is assumed to exist. If a threshold is expected for
the given effect, then a model that allows for a threshold dose may be chosen for modeling. The
biological plausibility of the dose-response curve shape should also be a consideration in model
selection.
Even with these considerations, several different models may often adequately describe the
data. In these cases, the choice of the model may not be critical, especially since the estimation will
be confined to the observed dose range. Thus, any model that suitably fits the empirical data is
likely to provide a reasonable estimate of a BMD.
In certain data sets, none of the standard models may provide a reasonable fit to the data. Fit
is assessed statistically by comparing the model predictions to the observations. Goodness-of-fit
statistics formalize that comparison and pro vide p-values, ranging between 0 and 1, as a measure of
fit. When using a x2 statistic, larger p-values are indicative of good fit; smaller p-values of poorer
fit. Sufficiently small p-values (e.g., less than 0.01 or 0.05) are typically viewed as an indication
that the model was not adequate for describing the observed dose-response pattern.
Poor fit is often due to reduced responses at higher doses that are inconsistent with the dose-
response trend for lower doses, perhaps due to competing toxic processes or saturation of metabolic
systems related to the toxic response of interest. Several procedures can be used to adjust the
modeling process in these circumstances. For example, responses at the highest doses could be
eliminated, since those doses are usually least informative of responses in the lower dose region of
interest. In the case of saturated metabolic pathways, pharmacokinetic data can be used to estimate
delivered dose to the organ of interest. The BMD modeling can then be conducted on the delivered
dose. (Andersen et al., 1987,1993; Gehring et al., 1978).
58
-------
A particularly valuable exercise with respect to all of these fit issues is a visual (graphical)
examination of the model predictions in relation to the observations. This supplements the formal
statistical assessment of fit and may, in fact, be equally or more informative. The statistical test
assesses overall fit. For the purposes of BMD estimation, fit is most crucial in and around the
response level used to define the BMD (i.e., the BMR). Thus, for example, models that have similar
fits to the entire data set may differ with respect to their predictions near the BMR, and it may be
possible to select one over another on the basis of that more local behavior.
Measure of Altered Response
Crump (1984) proposed two measures of increased response for quantal data. These are
additional risk and extra risk. Additional risk is simply the probability of response at dose d, P(d),
minus the probability of response at zero dose (control response), P(0). It describes the additional
proportion of animals that respond in the presence of a dose. Extra risk is additional risk divided by
[l-P(O)]. It describes the additional proportion of animals that respond in the presence of a dose,
divided by the proportion of animals that would not respond under control conditions. These
measures are distinguished in the way they account for control responses. For example, if a dose
increases a response from 0 to 1 percent, both the additional risk and the extra risk is 1 percent.
However, if a dose increases risk from 90 to 91 percent, the additional risk is still 1 percent, but the
extra risk is 10 percent. The choice of extra risk versus additional risk is based to some extent on
assumptions about whether an agent is adding to the background risk. Extra risk is viewed as the
default because it is more conservative.
Analogous measures of risk have been proposed for continuous data (Crump, 1984). First,
altered response can be expressed as the difference between the mean response to dose d minus the
mean control response. The second measure is simply the difference between dose and control
means divided by (i.e., normalized by) the control mean response. The second measure expresses
change as a fraction of the control response rather than as an absolute change.
More recent consideration of BMDs for continuous endpoints have suggested other
alternatives. Allen et al. (1994a, 1994b) and Kavlock et al. (1995) determined that normalizing
changes in mean responses by a multiple of the background standard deviation produced BMDs that
were comparable, on average to NOAELs. For the developmental endpoints that those investigators
studied, the preferred multiple for the standard deviation was 0.5.
It is not clear when measures of risk expressed relative to the background (e.g., extra risk)
are preferable to measures expressed as absolute changes. Additional research is required to provide
guidance regarding the measure of altered response that is most appropriate in particular
circumstances.
59
-------
Selection of the BMR
A critical decision for deriving the BMD is the selection of the Benchmark level of risk
(BMR). Since the BMD is used like a NOAEL in the derivation of the RfD, the BMR should be
selected near the low end of the range of increased risks that can be detected in a bioassay of typical
size. The ED10 is frequently chosen as the BMR. The ED10 is the dose predicted to cause a 10
percent increase in the incidence of the effect in the test population. For some data, it may be
possible to adequately estimate the ED05 or ED01, which are closer to a true no-effect dose. Levels
between the ED01 and the ED 10 are usually the lowest levels of risk that can be estimated adequately
for binomial endpoints from standard toxicity studies (Crump, 1984). Another consideration is the
goal of model independence. Different dose-response models may fit the data equally well yet give
very different estimates of risk far below the observable range (Crump, 1984). This argues for use
of a BMR close to the range of responses that can be reliably measured in typical studies.
During a BMD Workshop, sponsored by EPA, participants generally agreed that the
appropriate BMR should either be 5 percent or 10 percent, but acknowledged that future research
might demonstrate the advisability of selecting one value over another (ILSI, 1993). Research by
Allen et al. (1994a, 1994b) and Faustman et al. (1994) indicates that BMDs defined in terms of 10
percent increases in probability of response tend to be, on average, similar to corresponding
NOAELs for quantal developmental toxicity studies. For the purposes of water quality criteria
derivation, EPA recommends the use of the ED05 or ED10 when deriving a BMD.
Calculating the Confidence Interval
The BMD is defined to be the lower confidence bound on the dose corresponding to the
selected BMR. A statistical lower confidence limit is used rather than a maximum likelihood
estimate (MLE) for several reasons. The use of confidence limits accounts for the sample size of
the experiment; the fact that NOAELs do not account for sample size is a major criticism of
NOAEL-based derivation of the RfD. Furthermore, a lower confidence limit is more stable to minor
changes in data and, rarely, may be estimable where a MLE is not.
To calculate the upper confidence bound on response, and subsequently, the lower bound on
effective dose, decisions must be made regarding the selection of the procedure for calculating
confidence limits and the size of the confidence limits.
The recommended method used to calculate the confidence bounds on the curve relies on
maximum likelihood theory. This approach is the same one used by EPA in the computer program
for cancer dose-response modeling. The approach can be applied to BMD modeling using
commercially available software. A detailed explanation of theory supporting this approach is found
in Crump (1984).
By convention, the size of the statistical confidence limits can range from 90 to 99 percent.
The methods of confidence limit calculation and choice of confidence limits are critical. The
60
-------
Agency recommends the use of one-sided 95th percentile confidence limits for BMD modeling.
This is consistent with the size of the confidence limits used in cancer dose-response modeling.
Selection of the BMD As the Basis for the RfD
An important decision is the choice of the appropriate BMD to use in the RfD calculation
when multiple BMDs are calculated. Multiple BMDs can be calculated when different models fit
the response data for a single study, when more than one response is modeled in a single study, and
when there are different BMDs from different studies. When multiple BMDs exist because several
models fit a single data set, the analyst may select the smallest BMD or combine BMDs by using
the geometric average. When multiple BMDs are calculated due to different responses or different
studies that examine the same endpoint, the choice among BMDs may also involve the selection of
the "critical effect" and the most appropriate species, sex, or other relevant feature of experimental
design.
Use of Uncertainty Factors with BMD Approach
Once a single or averaged BMD is selected, the RfD can be calculated by dividing the BMD
by one or more uncertainty factors. It is still necessary to use uncertainty factors with a BMD,
because the BMD can miss sensitive subpopulations and is still subject to interspecies extrapolation
uncertainties. As a default, all applicable uncertainty factors used in the traditional NOAEL-based
RfD approach, except for the LOAEL-NOAEL extrapolation factor, should be retained. Other
factors, such as the size of the BMR and confidence bounds, biological considerations (such as the
possibility of a threshold), severity of the modeled effect, and the slope of the dose-response curve,
may affect the choice and magnitude of uncertainty factors (see Crump et al., 1995, for more
detailed discussion).
2.2.6.3 Limitations of the BMD Approach
The BMD approach has been proposed as an alternative procedure that can be used until
biologically motivated approaches are available for some or all effects. It provides specific
improvements over NOAEL-based approaches, but by no means does it resolve all issues or
difficulties associated with noncancer risk assessment. The BMD approach allows for objective
extrapolation of animal response data to human exposures across the different study designs
encountered in noncancer risk assessment.
2.2.6.4 Example of the Application of the BMD Approach
The followingprovides a simple exampleof the application of the BMD approach to quantal
toxicity data. The example given is taken from Crump et al. (1995) for acrylamide. The purpose
of presenting this example is to illustrate the method only; no actual risk value nor AWQC for
acrylamide is derived.
61
-------
Selection of Data to Model
This example takes the approach of identifying a critical study rather than modeling all
endpoints seen in valid studies. For this example, a 6-month dietary study of neurological effects
in rats is used as the critical study for acrylamide (Johnson et al., 1986, as cited in Crump et al.,
1995). The endpoint examined in this study was tibial nerve degeneration. The researchers recorded
the occurrence of nerve degenerationin two categories: none or mild, and moderate or severe. Since
mild nerve degeneration occurs spontaneously in older rats, and because mild degeneration showed
no dose-responserelationship,only moderate and severe degeneration were recorded as responses.
The data are presented in quantal form, with no or mild degeneration considered "no response," and
moderate to severe degeneration recorded as a response. The dose levels and number of animals
responding in each dose group are shown in Table 2.2.5.
Table 2.2.5: Rats Experiencing Moderate or Severe Nerve Degeneration in Response to
Acrylamide Dose
Dose (mg/kg-day)
0
0.01
0.1
0.5
2.0
Number affected
9
6
12
13
16
Number tested
60
60
60
60
60
Choice of Mathematical Model
From Table 2.2.4, we can select from among the various models available for quantal data.
Fitting is accomplished through the use of maximum likelihood estimation to estimate the
probability of a response at each dose level. The actual fitting exercise is done through the use of
computer software.
Results of Information Above
All of the models can be tried to see which achieves the best fit. The following Exhibits
illustrate the best-fit modeling of the study data for the Weibull model (Exhibit 2.2.2) and the
quadratic model (Exhibit 2.2.3). Table 2.2.6 provides the best-fit model parameters for the two
equations.
62
-------
Exhibit 2.2.2 Quantal Weibull Regression - Extra Risk
0.50
0.40 -:
P(d) Modeled
P(d) Observed
P(d)99th
P(d)95th
P(d) 90th
-0.10
Dose (M9/kg-d)
0.50
Exhibit 2.2.3 Quantal Quadratic Regression - Extra Risk
0.40 ..
0.30 . „
R(d) Modeled
+ P(d) Observed
P(d)99th
.\ ,' x«- V «
? Z<* *^>v^V
<•* Xi- •**> '-
« ^ 'v-*^.*
0.10
0.00
-0.10
• *'""*• - ~ \t^**» "'*•'* * ^"-^X, -- I Ji^ "^ "**• "^*^ "t''**!-*'*^ " *%" »v« i^ ''" ^-^***!-** • >• .> % ^---<^^>
v^••> ".- Ai\::;...: V^^ /A^i^^ii'-;
r^t*-'3f%,._^ ; ^;- •^,•^.'.^'1 -./^-f-^^^;'-, -" ^*¥^<^
Dose (MQ/kg-d)
63
-------
Table 2.2.6: Best-Fit Model Parameters from Modeling of the Acrylamide Data
Model
Quantal Weibull
Quantal quadratic
Background rate
0.15 .
0.16
ql
0.08
0.034
k
1
—
chi-square p value
0.48
0.34
Note that in example given here, the measure of altered response is extra risk, which is
defined as:
ER(d) = [P(d)-(P(0)]/[l-P(0>]
(Equation 2.2.2)
Extra risk is the fraction of animals that respond when exposed to a dose, d, among animals
who otherwise would not respond.
Both models fit the data adequately as shown in Table 2.2.6. In both cases the chi-squared
goodness of fit yields P-values greater than 0.05. Therefore, either model can be used for derivation
of BMD. Neither model, as fitted to this data set, suggests a threshold for this response. However,
both models do indicate a background rate in the absence of exposure to acrylamide.
Selection of the BMR
For the data set discussed above, the BMDs were calculated using the quantal Weibull and
the quantal quadratic models for 1, 5, and 10 percent extra risk (Table 2.2.7 estimates are in units
ofmg/kg-day):
64
-------
Table 2.2.7: BMD Values Calculated Using Quantal Weibull and Quadratic Models
Model
Weibull
Quadratic
BMR
10
5
1
10
5
1
BMD (mg/kg-day) for Confidence Limit:
90th percentile
0.73
0.35
0.07
1.28
0.89
0.39
95th percentile
0.64
0.3 li
0.06 !
1.19;
0.83;
0.37;
99th percentile
0.52
0.25
0.05
1.06
0.74
0.33
The calculated HMDs are about a factor of two apart for the BMDIO values, but are about a factor
of six apart for the BMDj. This demonstrates the model dependence of the BMD values when low
BMR levels are selected. '
Calculating the Confidence Interval
As shown in Table 2.2.7, the BMDs were calculated for 90th, 95th, and 99th percentile
confidence limits. The effect of the confidence limit on the estimated BMD was slightly less for the
quantal quadratic than for the quantal Weibull. Model results were most comparable for the 90th
and 95th percentile confidence limits and least comparable for the 99th percentile confidence limits.
These results demonstrate that the BMD tends to be more model-dependent for wider (higher
percentile) confidence intervals. For the remainder of the .example, the 95th percentile confidence
limit estimate is used.
Selection of the BMD as the Basis for the RfD
The example above yields different 95th percentile BMD,0 values based on the two models.
Since there is no basis upon which to eliminate one of th;e BMDs (i.e., goodness of fit, statistical
assumptions and biological considerations), both must be; considered. Either the smaller estimate
may be used, or a geometric average may be used. In this case, the selection of which BMD to use
is a risk management decision. In the example, the lower of the two BMDs (0.64) was chosen for
the RfD calculation.
65
-------
Use of Uncertainty Factors with BMD Approach
Once the BMD is chosen, the RfD is Derived by dividing the BMD by uncertainty factors.
The same uncertainty factors applied to a NOAEL are used. In this case a factor of 10 was selected
for interspecies extrapolation and a factor of 10 for human interspecies variability. Using an total
UF of 100 and applying it to the 95th percentile confidence limit BMD for 10 percent response
derived with the quanta! Weibull model yields an RfD of 0.006 mg/kg-day.
2.2.7 Categorical Regression
2.2.7.1 Summary of the Method
Categorical regression is another method under investigationto estimate risks associated with
systemic toxicity (Dourson et al., 1997; Guth et al., 1997). In this approach, health effects are
grouped into ordered severity categories (ranging from no effect to severe effect). This
simplificationallows for the incorporation into the analysis of both quantal and continuous data, as
well as data that are reported qualitatively rather than quantitatively. Furthermore, information on
many health effects can be considered together. Logistic regression analysis techniques are then
applied to the data: the cumulative odds of falling into severity categories is the dependent variable,
and exposure concentration, exposure duration, and other parameters are the independent variables.
Using the regression results, the RfD is then specified as the dose at which the probability of adverse
effects is sufficiently small at some level of confidence, modified, as in the NOAEL and BMD
approaches, by appropriate uncertainty factors. For example, the dose of interest, D, might be
defined as that dose for which one could conclude with 95 percent certainty that the probability of
an adverse effect was less than 0.01. The value D would then be adjusted by uncertainty factors to
derive the RfD.13
2.2.7.2 Steps in Applying Categorical Regression
The categorical regression approach begins with a review of the toxicological data base for
the chemical. For each valid study, the toxic responses observed are assigned to one of several
ordered severity categories, based on biological and statistical considerations. For example,
responses may be grouped into four categories: (1) no effect; (2) no adverse effect; (3) mild-to-
moderate adverse effect; and (4) severe or lethal effect. These correspond to the dose categories used
in setting the RfD, namely the No Observed Effect Level (NOEL), NOAEL, LOAEL, and Frank
Effect Level (FEL), respectively.
Since all response data are used in categorical regression analysis, there is no need to specify
the lowest dose showing "mild-to-moderate" adverse effects. Accordingly, a more general term,
"Note that the logistic regression could be used to estimate the response to exposures greater than the RfD. BMD models could be used
similarly, but caution is warranted when doing so in either case.
66
-------
adverse-effect level (AEL), is generally used in categorical regression in place of the term LOAEL
to describe mild-to-moderate effects.
The probability of observing a response in a category at a given dose level is estimated by
dividing the number of responses observed in that severity category divided by the total number of
observations recorded for that dose level. Sufficient numbers of dose groups in each of several
categories are required for the categorical regression. Judgment is required to define the types of
effects that correspond to the severity categories.
The log odds for each dose and severity level is calculated, and then regressed against dose.
The resulting regression equation can be used to calculate the probability of an effect of given
severity for any dose.
Several model structures (logistic, Weibull, or others) may be used to perform the categorical
regression. Logistic regression on the ordered categories (Harrell, 1986; Hertzberg, 1989) allows
the dependent variable (e.g., severity parameter) to be categorical and the independent variables to
be either categorical or continuous.
The goodness of the fit of the model to the data can be judged using several statistical
measures, including the overall %2, model parameter standard errors and their %2 significance levels,
concordance statistics and correlation coefficients for the overall model, and the model covariates
(Hertzberg and Wymer, 1991). A variety of criteria are currently being investigated.
Some advantages of using the categorical regression to derive the RfD include the following:
data concerning different health effects can be incorporated; it allows for refinement through
improved data and statistical methodology; and several indicators of uncertainty in the estimates are
provided. In addition, the categorical regression approach can be used to evaluate likely responses
to exposures above the RfD.
2.2.8 Chronic, Practical Nonthreshold Effects
Noncarcinogens are generally assumed to exhibit a threshold below which adverse effects
are unlikely to occur. There are, however, exceptions to this general rule. Of particular concern are
teratogenic and reproductive toxicants that may act through a genetic mechanism. EPA has
recognized the potential for genotoxic teratogens and germline mutagens and discussed this issue
in the 1991 Amendments to Agency Guidelines for Health Assessments of Suspect Developmental
Toxicants (USEPA, 199 la) and in the 1986 Guidelines for Mutagenicity Risk Assessment (USEPA,
1986a). Various statements within these guidelines raise concern for the potential for future
generations inheriting chemically induced germline mutations or suffering from mutational events
occurring in utero. An awareness of the potential for such teratogenic/mutagenic effects should be
established in order to deal with such data. At this time, genotoxic teratogens and germline
mutagens should be considered an exception and the traditional uncertainty factor approach the rule
for calculating criteria or values for chemicals demonstrating developmental/reproductive effects.
67
-------
In the absence of adequate data to support a genetic or mutational basis for developmental or
reproductive effects, the default becomes an uncertainty factor approach. For such chemicals, this
guidance recommends the procedures described above for noncarcinogens assumed to have a
threshold.
Where evidence for a genetic or mutational basis does exist, a nonthreshold mechanism shall
be assumed for genotoxic teratogens and germline mutagens. Since there is no well established
mechanism for calculating criteria protective of human health from the effects of these agents,
criteria will be established on a case-by-case basis.
2.2.9 Acute, Short-Term Effects
States may choose to derive criteria that correspond to acute or short-term exposures. These
criteria should correspond to a level of exposure that is "without appreciable risk of deleterious
effects during some relatively short period of time" (USEPA, 1991 c). The derivation of such values
follows the same general approaches described above for criteria based on chronic effects. The
primary difference lies in the type of toxicity data used as the basis for the evaluation. Generally,
studies that mimic the exposure pattern and duration of interest will be considered more relevant to
the development of acute or short-term criteria. This is especially important where acute or short-
term effects are of a substantially different nature than low-level chronic effects. Where toxicity data.
do not match the exposure of interest, professional judgment is required to evaluate the relevance
of the available data. Factors such as the pharmacokinetics,potential recovery periods, and potential
for bioaccumulation should be considered in judging the relevance of the data.
The Office of Water has establishedprocedures for deriving Health Advisories (HAs) for one
day, ten days, and longer-term. In general HAs are developed by using NOAELs or LOAELs from
studies with similar duration to the exposure period of concern, though there is some flexibility in
this regard. Studies used for HAs should provide information on the critical endpoint. Studies that
identify only frank toxic responses should not be used since these levels are far above the protective
level targeted by HAs. More information on the derivation of HAs is given in Ware (1988).
2.2.10 Mixtures
Exposures to multiple noncarcinogens may occur simultaneously. Possible interactions
among chemicals hi a mixture are usually placed in one of three categories:
• Antagonistic, where the chemical mixture exhibits less toxicity than is suggested by
the sum of the toxic effects of the components.
• Synergistic, where the chemical mixture exhibits greater toxicity than is suggested
by the sum of the toxic effects of the components.
68
-------
• Additive, where the toxicity of the chemical mixture is equal to the sum of the
toxicities of the components.
In only a few instances have the interactive effects of chemical mixtures been specifically
studied. Where data on the effects of chemical mixtures exist, they should be used to characterize
risk. Using the available data is especially important in cases where the resulting toxic effect from
the mixture has been demonstrated to be greater than the sum of the individual effects. Where
specific data are not available on the interactive effects of particular chemical mixtures, the methods
described below can be used by states to characterize risks from chemical mixtures. When risks from
multiple chemicals are added, the quality of experimental evidence that supports the assumption of
dose addition should be stated clearly (USEPA, 1986b).
In cases where the chemicals in the mixture induce the same effect by similar modes of
action, contaminants may be assumed to contribute additively to risk (USEPA, 1986b), unless
specific data indicate otherwise. To characterize risks from multiple chemical exposure to
noncarcinogens, the dose for each chemical with a similar effect first is expressed as a fraction of
its RfD. These ratios are added for all chemicals to obtain the chemical mixture hazard index:
HI . =
mix
m=l
RfD
(Equation 2.2.3)
where HImix is the hazard index of the mixture (unitless), Em is the exposure to chemical m, RfDm is
the reference dose for chemical m, and n is the number of chemicals in the mixture. A hazard index
greater than one implies that the individual is at some risk of the non-carcinogenic effect, and the
concern is the same as if exposure from an individual chemical exceeded the acceptable level by the
same amount (USEPA, 1986b). However, the numerical value of the hazard index does not indicate
the magnitude and severity of the risk.
Some chemical mixtures may contain chemicals that cause dissimilar health effects.
Methods currently do not exist for combining dissimilar health effects to characterize overall health
concerns from chemical mixtures. Instead, States should characterize and present the risks from
these contaminants separately.
69
-------
2.2.11 References
Allen, B.C., R.J. Kavlock, C.A. Kimmel and E.M. Faustman. 1994a. Dose Response Assessments
for Developmental Toxicity: II. Comparison of Generic Benchmark Dose Estimates with
NOAELs. Fundam. Appl. Toxicol. 23:487-495.
Allen, B.C., R. J. Kavlock, C. A. Kimmel and E.M. Faustman. 1994b. Dose Response Assessments
for Developmental Toxicity: III. Statistical Models. Fundam. Appl. Toxicol. 23:496-509.
Andersen M., H. Clewell, M. Gargas, F.A. Smith, R.H. Ritz. 1987. Physiologically Based
Pharmacokinetics and Risk Assessment Process for Methylene Chloride. Toxicol. Appl.
Pharmacol. 87:185-205.
Andersen, M., J. Mills, M. Gargas, L. Kedderis, L. Birnbaum, D. Neubert, and W. Greenlee. 1993.
Modeling Receptor-mediatedProcess with Dioxin: Implications for Pharmacokinetics and
Risk Assessment. Risk Analysis 13:25-26.
Brown, K.G. and L.S. Erdreich. 1989. Statistical Uncertainty in the No-Observed-Adverse-Effect-
Level. Fund. Appl. Toxicol. 13(2): 235-244.
Calabrese,E. 1985. Uncertainty Factors and Interindividual Variation. Regulatory Toxicology and
Pharmacology. 5:190-196.
Crump, K.S. 1984. A New Method for Determining Allowable Daily Intakes. Fund. Appl. Toxicol.
4:854-871.
Crump, K. 1995. Calculation of Benchmark Doses from Continuous Data. Risk Analysis. 15:79-
89.
Crump, K.S., Allen, B., and E. Faustman. 1995. The Use of the Benchmark Dose Approach in
Health Risk Assessment. Prepared for USEPA Risk Assessment Forum. EPA/630/R-94-
007.
Dourson, M.L. and J. Stara. 1983. Regulatory History and Experimental Support of Uncertainty
(Safety) Factors. Regulatory Toxicology and Pharmacology 3: 224-239.
Dourson, M.L., R.C. Hertzberg, R. Hartung and K. Blackburn. 1985. Novel Approaches for the
Estimation of Acceptable Daily Intake. Toxicol. Ind. Health 1(4): 23-41.
Dourson, M.L., Knauf, L.A. and J.C. Swartout. 1992. On Reference Dose (RfD) and its Underlying
Toxicity Data Base. Toxicol. Ind. Health 8(3): 171-189.
70
-------
Dourson, M.L., L.K. Teuschler, P.R. Durkin, and W.M. Stiteler. 1997. Categorical Regression of
Toxicity Data, A Case Study Using Aldicarb. Regulatory Toxicity and Pharmacology 25:
121-129.
Faustman, E.M., B.C. Allen, R.J. Kavlock and C.A. Kimmel. 1994. Dose Response Assessment
for Developmental Toxicity: I. Characterization of Data Base and Determination of
NOAELs. Fundam. Appl. Toxicol. 23:478-486.
Gaylor, D.W. 1983. The Use of Safety Factors for Controlling Risk. J. Toxicol. Environ. Health
11:329-336.
Gaylor, D.W. 1989. Quantitative Risk Analysis for Quantal Reproductive and Developmental
Effects. Environ. Health Perspect. 79:243-246.
Gaylor, D.W. and W. Slikker, Jr. 1990. Risk Assessment for Neurotoxic Effects. Neurotoxicol.
11:211-218.
Gehring, P J., P. J. Watanabe, and C.N. Park. 1978. Resolution of Dose-Response Toxicity Data for
Chemicals Requiring Metabolic Activation: Example — Vinyl Chloride. Toxicol. Appl.
Pharmacol. 44:581-591.
Guth, D.J., R.J. Carroll, D.G. Simpson, and H. Zhou. 1997. Categorical Regression Analysis of
Acute Exposure to Tetrachloroethylene. Risk Analysis 17(3): 321-332.
Harrell, F. 1986. The Legist Procedure. SUGI Supplemental Library Users Guide, Ver. 5th. Ed.
Gary, NC: SAS Institute.
f
Hartley, W.R. and E.V. Ohanian. 1988. The Use of Short-term Toxicity Data for Prediction of
Long-term Health Effects. In: Trace Substances in Environmental Health - XXII. D.D.
Hemphil, (ed). University of Missouri, May 23-26, pp. 3-12.
Hartung, R. and P.R. Durkin. 1986. Ranking the Severity of Toxic Effects: Potential Applications
to Risk Assessment. Comments on Toxicology. 1(1): 49-63.
Hattis, D., L. Erdreich and M. Ballew. 1987. Human Variability in Susceptibility to Toxic
Chemicals — A Preliminary Analysis of Pharmacokinetics Data from Normal Volunteers.
Risk Analysis. 7(4): 415-426.
Hertzberg. 1989. Fitting a Model to Categorical Response Data with Applications to Species
Extrapolation of Toxicity. Health Physics 57:405-409.
Hertzberg, R.C. and M.E. Miller. 1985. A Statistical Model for Species Extrapolation using
Categorical Response Data. Toxicol. Ind. Health l(4):43-63.
71
-------
Hertzberg, R.C. and L. Wymer. 1991. Modeling the Severity of Toxic Effects. Presentation at the
84th Annual Meeting of the Air and Waste Management Association. June 16-21,1991.
International Life Sciences Institute (ILSI). 1993. Report on the Benchmark Dose Workshop.
Washington, DC: International Life Sciences Institute, Risk Science Institute.
Kavlock, R.J., B.C. Allen, E.M. Faustman and C.A. Kimmel. 1995. Dose Response Assessment
for Developmental Toxicity: IV. Benchmark Doses for Fetal Weight Changes. Fundam.
Appl. Toxicol. 26:211-222.
Kimmel, C.A. 1990. Quantitative Approaches to Human Risk Assessment for Noncancer Health
Effects. Neurotoxicology 11:189-198.
Lewis, S.C., J.R. Lynch and A.I. Nikiforov. 1990. A New Approach for Deriving Community
Exposure Guidelines from No-Observed-Adverse-Effect-Levels. Reg. Toxicol. Pharmacol.
11:314-330.
Swartout. 1990. Personal Communication to M.L. Dourson of the Office of Technology Transfer
and Regulatory Support on January 12 . Washington, DC.
USEPA. 1986b. Guidelines for the Health Risk Assessment of Chemical Mixtures. Federal
Register 51: 34014-34025. September 24.
USEPA. 1988. Reference Dose (RfD): Description and Use in Health Risk Assessments.
Integrated Risk Information System (IRIS). Online. Intra-Agency Reference Dose (RfD)
Work Group. Cincinnati, OH: Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office. February.
USEPA. 1989. Interim Methods for Development of Inhalation Reference Doses. Office of Health
and Environmental Assessment. EPA/600/8-88-066F.
USEPA. 199la. Amendments to Agency Guidelines for Health Assessments of Suspect
Developmental Toxicants. Federal Register 56: 63798-63826. December 5.
USEPA. 1991b. Final Guidelines for Developmental Toxicity Risk Assessment. Federal Register
56: 63798-63826. December 5.
USEPA. 1991c. General Quantitative Risk Assessment Guidelines for Noncancer Health Effects.
Technical Panel for Development of Risk Assessment Guidelines for Noncancer Health
Effects. Second External Review Draft. ECAO CIN-538.
72
-------
USEPA. 1992. Reference Dose (RfD) for Oral Exposure for Inorganic Zinc. Integrated Risk
Information System (IRIS). Online. (Verificationdate 10/1/92). Cincinnati, OH: Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office.
USEPA. 1993. Reference Dose (RfD) for Oral Exposure for Inorganic Arsenic. Integrated Risk
Information System (IRIS). Online. (Verificationdate 2/01/93). Cincinnati, OH: Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office.
USEPA. 1994a. Reference Dose (RfD) for Oral Exposure for Methylmercury. Integrated Risk
Information System (IRIS). Online. (Verificationdate 11/23/94). Cincinnati, OH: Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office.
USEPA. 1994b. Guidelines for Reproductive Toxicity Risk Assessment. External Review Draft.
EPA/600/AP-94/001. February.
USEPA. 1995. RQ Document for Solid Waste. Report on the Benchmark Dose Peer Consultation
Workshop: Risk Assessment Forum. Office of Research and Development.
EPA/630/R96/011. November.
USEPA. 1998. Notice of Draft Revisions to the Methodology for Deriving Ambient Water Quality
Criteria for the Protection of Human Health. Federal Register Notice.
Ware, G.W. (ed). 1988. Reviews of Environmental Contamination and Toxicology: U;S.
Environmental Protection Agency Office of Drinking Water Health Advisories. Vol. 104.
New York: Springer-Verlag, Inc.
Zielhuis, R.L. and F.W. van der Kreek. 1979. The Use of a Safety Factor in Setting Health Based
Permissible Levels for Occupational Exposure. Int. Arch. Occup. Environ. Health. 42:191 -
201.
2.3 Exposure Analyses
2.3.1 Role of Exposure Data in Setting AWQC
The AWQC are primarily established to protect individuals from adverse health effects
caused by pollutants in United States' inland and estuarine waters. To achieve this goal, exposure
factors representative of the population to be protected should be used in the equations to derive the
criteria (see Section 2.3.4.2 for the equations to derive criteria). In addition, exposures from other
non-water sources such as air and food should be taken into account so that the criteria are protective
of individuals who may be exposed to a particular pollutant from multiple exposure routes.
The following sections describe data available to determine exposure factors and discuss
methods for incorporating non-water sources of exposure, including EPA's recommended default
73
-------
values and methods. In addition, Table 2.3.1 presents sources of exposure-related information.
While not intended to be a comprehensive list, Table 2.3.1 does indicate some of the readily
available sources for contaminant concentration data and exposure intake parameters.
2.3.2 Exposure Factors in AWQC Algorithms
Several exposure factors are included in the equations to derive AWQC. These factors
include (1) body weight of the individuals exposed; (2) drinking water ingestion rates; (3) fish
consumption rates; (4) incidental ingestion of water; and (5) the relative source contribution factor
to account for other exposures. Body weights and fish intake assumptions are used in each criterion.
The uses of and values for these factors differ based on several considerations. One
consideration in the choice of values for a specific exposure parameter depends on whether the water
body has been designated as a drinking water supply source or as a recreational, non-potable source.
The drinking water ingestion rate is recommended for use for those waters designated as public
drinking water supply sources. This rate represents the amount of water an individual drinks per day
(see Section 2.3.2.2 for further discussion of the general policy to include a drinking water ingestion
rate when setting AWQC). For waters that are used for recreational purposes, an individual may
incidentally swallow water when swimming or waterskiing. Thus, an incidental ingestion rate would
be applied in these circumstances. Although it is possible that individuals would incidentally ingest
water from drinking water sources, incidental ingestion is not included for these sources of intake
because the incidental ingestion rate is negligible when the assumed daily drinking water ingestion
rate is utilized.
Another consideration in determining the exposure factors used is whether health effects
result from chronic exposure or whether developmental health effects are being evaluated. For
example, if a chemical causes both developmental and chronic health effects, a State or Tribe may
wish to evaluate the chemical using relevant chronic or developmental exposure factors associated
with those health effects, respectively, to determine whether to set criteria based on chronic or
developmental effects. For chronic health effects, intake rates and body weights of adults or rates
relevant to lifetime exposures are the most applicable because the health effects are associated with
a long period of exposure. However, for pollutants that may cause health effects after shorter-term
exposure to a chemical, exposure factors for children may be most useful when setting criteria for
RfJDs based on health effects in children, because children often have a higher intake per body
weight than adults. In addition, children may be more susceptible to certain contaminants than
adults, and may have less capability to detoxify contaminants (USEPA, 1994a). Thus, for such
potential situations, EPA default values include intake rates for children.
74
-------
Table 2.3.1: Sources of Contaminant Concentration and Exposure Intake Information
Name of Source
Aerometric Information
Retrieval System (AIRS)
Continuing Survey of Food
Intake by Individuals (CSFII)
Exposure Factors Handbook
Inventory of Exposure-Related
Data Systems Sponsored by
Federal Agencies
National Food Consumption
Survey (NFCS)
National Health and Nutrition
Examination Survey
(NHANES)
National Inorganics and
Radionuclides Survey
(NIRS)/National Pesticides
Survey (NFS)
National Sediment Inventory
(NSI)
Safe Drinking Water
Information System (SDWIS)
Total Diet Study (TDS) - also
known as Market Basket Survey
Type of Information
An updated air data base of many different sites
(from rural to urban/industrial) that includes
Federally required information, as well as data
submitted voluntarily by States.
A national food consumption survey conducted
approximately annually.
Summarization of studies and data bases to provide
statistical data on factors used in assessing
exposure.
Compilation of information on Federally managed
data systems that contain exposure information.
A national food consumption survey conducted
each decade by USDA. The last survey was
conducted in 1987-88.
A national health and nutrition survey conducted
each decade. Based on a probability sample of
noninstitutionalized people residing in the U.S.
Most recent Federal surveys (mid to late 1980s)
conducted to characterize occurrence of a series of
inorganic and radionuclide chemicals (NIRS) and
pesticides (NFS) in public drinking water supplies
from ground water sources (and rural domestic
wells with the NFS).
Compilation of available data bases on sediment
contamination/sediment chemistry data. These
include data on fish tissue residues of chemical
contaminants.
Compiled data that includes monitoring required
and provided under the program for unregulated
contaminants (Section 1445 of the Safe Drinking
Water Act).
Contaminant concentrations in foods purchased
from supermarkets or grocery stores throughout the
U.S. four to five times a year. Food items in the
TDS are of similar type included in the NFCS and
the second NHANES (both described above).
Agency/Author
Office of Air Quality
Planning and
Standards, EPA
U.S. Department of
Agriculture (USDA)
National Center for
Environmental
Assessment, EPA
Agency for Toxic
Substances and
Disease Registry;
Centers for Disease
Control; EPA
USDA
National Center for
Health Statistics
EPA
EPA
EPA
Food and Drug
Administration
75
-------
Table 2.3.1: Sources of Contaminant Concentration and Exposure Intake Information
Name of Source
Total Water and Tap Water
Intake in the United States:
Population-Based Estimates of
Quantities and Sources
Type of Information
Presents estimates of total water (includes water
intrinsic to foods) and tap water intake in the
population of the continental U.S. Data used are
based on the NFCS (described above).
Agency/Author
Ershow and Cantor
(NCI/NIH)
Shorter-term exposures may also pose risks to other people with special susceptibilities due
to illness (e.g., persons with kidney, liver, or other diseases may be especially vulnerable to toxins
which attack those systems). When States and Tribes assess intake from pollutants that cause
toxicity resulting from such exposures, they may wish to investigate intakes for these other
population groups that may also have a high intake per body weight, and/or may be highly subject
to adverse effects from these toxicants. It may be appropriate to calculate criteria using
developmental and chronic toxicity and exposure assumptions to see which criterion is more
stringent.
Developmental effects resulting from prenatal exposure to contaminants have become an area
of significant concern (USEPA, 1994a). Thus, in addition to considering use of exposure factors
specific to adults or children, States and Tribes may wish to use exposure factors specific to women
of childbearingage in cases where developmental health effects may be of concern. These exposure
factors are described below.
Fish consumption intake rates may also differ based on the target population to be protected.
Some states may have a large population of recreational fishers who may fish a few times a year or
during a fishing vacation. Other States may have populations that subsist on fish for a large portion
of a year. Thus, the fish intake exposure factor may differ depending on the population that is to be
protected. Different types of fishers are discussed in greater detail, below, in the section that
describes fish intake rates.
When setting AWQC, it is preferable to use exposure information reflective of individuals
who actually use the water body for which AWQC are to be determined. When dealing with such
diverse populations as those throughout the United States, extreme ranges of behaviors and activities
are likely. Therefore, EPA explicitly recommends that, for certain exposure factors that may be
highly variable (e.g., fish intake rates), States use available local data. These data should be used
especially in cases that result in AWQC that are more stringent than criteria derived using default
exposure assumptions suggested by EPA. In many situations, local exposure data may not be
available. Therefore, EPA also recommends default values for each of the exposure values discussed
below.
The following sections discuss available data and describe some of the above issues in
greater detail. In addition, the sections discuss EPA's recommendations for use of the exposure
76
-------
factors and present EPA's suggested default values. The incorporation of these exposure factors into
equations to derive criteria are described in detail in Section 2.3.4.2.
2.3.2.1 Body Weight
The 1980 AWQC National Guidelines used a body weight of 70 kg in the derivation of
AWQC, which represents EPA's Agency-wide adult body weight assumption used in its risk
assessments and approximates the average adult body weight of 71.8 kg from an analysis of the
National Health and Nutrition Examination Survey (NHANES II), as reported in the Exposure
Factors Handbook (USEPA, 1997a). In the current, updated guidance, EPA recommends several
default body weight values, depending on whether chronic effects or acute effects are being
evaluated. The use of these data in equations to derive AWQC are described in Section 2.3.4.2.
Chronic Exposure Scenarios
For chemicals that cause chronic effects, EPA recommends using a default body weight of
70 kilograms. This value approximates the mean for adults from two sets of data. The first set of
data comes from NHANES II, which was conducted from 1976 through 1980 and for which
information on a variety of health and nutritional characteristics of individuals were collected
(adapted from NCHS, 1987). The National Center for Health Statistics compiled body weight data
from NHANES II for over 20,000 individuals aged 6 months to 74 years. Weighted mean body
weights were determined from this data. The mean body weight value for men and women ages 18
to 74 years old, using data from NHANES II, is 71.8 kg. The median body weights for men and
women from this study are 76.9 and 62.4 kilograms, respectively. Table 2.3.2 includes a distribution
of mean and median adult body weights, by age group, based on NHANES II data. Body weights
are presented for men, women, and both sexes combined.
The second set of body weight data come from Ershow and Cantor (1989). These authors
used data collected during the 1977-1978 Nationwide Food Consumption Survey (NFCS), which
surveyed 30,770 individuals who constituted a stratified random sample designed to represent the
noninstitutionalizedU.S. population living in households (USDA, 1988). Body weights were self-
reported by participants. The mean value for body weight for adults ages 20 - 64 years old is 70.5
kg. Means and percentile values of body weight from Ershow and Cantor (1989) listed by sex and
age are presented in Table 2.3.3. The revised EPA Exposure Factors Handbook (USEPA,'l997a)
recommends a value of 71.8 kg for adults, based on the NHANES II data. However, the Handbook
also acknowledges the 70 kg value commonly used in EPA risk assessments and cautions assessors
on the use of values other than 70 kg. Specifically, the point is made that the 70 kg value is used in
the derivation of cancer slope factors and unit risks that appear on IRIS. Consistency is advocated
between the dose-response relationship and exposure factors assumed.
77
-------
Table 2.3.2: Body Weight (in kilograms) of Adults from NHANES II
Age
18 -24 years
25 - 34 years
35 - 44 years
45 - 54 years
55 - 64 years
65 74 years
Overall:
18 -74 years
Men
Mean
73.7
78.7
80.9
80.9
78.8
74.8
78.1
Median
72.0
77.5
79.9
79.0
77.7
74.2
76.9
Women
Mean
60.6
64.2
67.1
68.0
67.9
66.6
65.4
Median
58.0
60.9
63.4
65.5
65.2
64.8
62.4
Men and
Women
Mean
67.2
71.5
74.0
74.5
73.4
70.7
71.8
Source: Adapted from NCHS( 1987)
Developmental Effects Exposure Scenarios
In certain cases, pregnant women may represent a more appropriate target population for
consideration when setting water quality criteria than all adults in cases where developmental effects
may be of concern (USEPA, 1994b). In these types of cases, body weights representative of women
of childbearing age may be appropriate to adequately protect offspring from such health effects. For
example, in the Great Lakes Water Quality Initiative, EPA chose women of childbearing age as the
target population for development of the mercury criterion (USEPA, 1995). To determine a mean
body weight value appropriate to this population, separate body weight values for women in
individual age groups within the range of 15 to 44 years old, taken from NHANES II (adapted from
NCHS, 1987), were combined and weighted by current population percentages (U.S. Bureau of the
Census, 1996) to obtain a value applicable to the current population. The resulting mean body
weight value for this age group is 63.8 kg.
Ershow and Cantor (1989) also present data on mean and median body weights for pregnant
women, of 65.8 and 64.4 kilograms, respectively. Based on these data, States may wish to use a
value of 65 kilograms in combination with relevant developmental toxicity data when assessing risks
for pregnant women and for setting AWQC.
Likewise, for some contaminants, RfDs based on health effects in children may be of primary
concern. As stated in the Federal Register notice, because children generally eat more fish and drink
more water per body weight than adults, higher intake rates per body weight may be more
appropriate in the derivation of AWQC to provide adequate protection for these individuals. In
78
-------
addition, because children may be more susceptible to the effects of some pollutants than adults
(USEPA, 1994b), they should be especially considered when assessing adverse effects that occur
following such exposures. Information on children's body weights (from NHANESII) are included
in Tables 2.3.3 and 2.3.4.
To protect children against health effects from water and fish intake when RfDs are based
on health effects in children, EPA recommends a default body weight of 28 kilograms, which
represents a mean body weight for children 0 to 14 years old. This is only recommended for
chemicals for which adverse effects for children are the most critical endpoint in the chemical's
lexicological profile. This body weight can be used with fish intake rates for children in the same
age group when deriving criteria for protection against such health effects from eating fish. The
default recommendationis made, in part, due to the limitations of the default fish consumption data.
Specifically, the limited sampling base prohibits the use of finer age group divisions due to
unacceptable confidence intervals with such finer fish intake divisions. However, finer age divisions
are provided in Tables 2.3.3 and 2.3.4 for States and Tribes to consider using along with more robust
fish consumption data. As with other recommended body weight values, the default estimate is
based on information from analyses of NHANES II data (adapted from NCHS, 1987). Current
population estimates (U.S. Bureau of the Census, 1996) were used to weight information on body
weights for individuals in several age groups up to age 14 years, using body weight information from
NCHS (1987) to represent values applicable to the current population. This calculation resulted in
weighted mean body weight values of 28 kg for this age group. A similar analysis using body
weights for separate age groups within the 0-14 year range from Ershow and Cantor (1989), and
weighting by current population estimates also resulted in a mean body weight of 28 kilograms.
If States wish to specifically evaluate infants and toddlers, EPA recommends a lower default
body weight of 10 kilograms, as has been used in previous water program guidance and regulations.
The body weight is representative of children up to three years old. EPA recommends using data
from this particular age group because these children may be particularly susceptible to acute effects
from water-based formula intake (e.g., nitrate). Data used to determine this body weight value come
from Ershow and Cantor (1989) and the analysis of NHANES II data (adapted from NCHS, 1987).
The analysis of NHANES II data indicate 10th, 25th, and 50th percentile values for children less
than three years old as 8.5, 9.6, and 11.3 kilograms for females, and 9.1, 10.3, and 11.8 kilograms
for males, respectively. Mean body weights from NHANES II are 9.1 for children ages 6-11
months, 11.3 for 1-year-olds, and 13.3 for 2-year-olds (adapted from NCHS, 1987). From the
Ershow and Cantor study, the 10th, 25th, and 50th percentile values for children 1-3 years old are
10.4, 11.8, and 13.6 kg, respectively, with a mean value of 14.1 kg (Ershow and Cantor 1989).
States and Tribes may instead wish to consider certain general developmental ages (e.g., pre-
school, pre-adolescent, adolescent, etc.) or certain specific developmental landmarks (e.g.,
neurological development in the first four years, etc.) depending on the chemical of concern. EPA
encourages States and Tribes to use Tables 2.3.3 and 2.3.4 to choose a body weight intake, if they
believe a particular age subgroup is more appropriate due to these developmental ages or landmarks.
79
-------
Table 2.3.3: Self-Reported Body Weight (kilograms) for Both Sexes from Ershow and Cantor (1989)"'
Sex
Both
Males
Females
Age (yr)
<0.5
0.5-0.9
1-3
4-6
7-10
11-19
20-64
65+
All
<0.5
0.5-0.9
1-3
4-6
7-10
11-19
20-64
65+
All
<0.5
0.5-0.9
1-3
4-6
7-10
11-19
20-64
65+
All
Mean
5.8
9.2
14.1
20.3
30.6
55.2
70.5
68.6
59.3
6.2
9.6
14.4
20.5
31.0
58.3
79.4
74.4
63.8
5.5
8.8
13.7
20.0
30.2
52.1
64.1
64.5
55.7
Standard
Deviation
1.8
2.0
3.2
4.6
7.8
13.4
15.2
13.1
22.6
1.8
2.1
3.3
4.5
7.9
14.9
13.0
11.5
25.3
1.8
1.7
3.0
4.6
7.6
11.0
13.4
12.6
19.5
Percentile Distribution
1
c
c
8.6
12.7
18.1
28.6
44.5
40.8
9.1
c
c
9.1
13.6
18.1
29.0
54.4
49.9
9.1
c
c
8.6
12.7
18.1
28.1
43.1
39.9
9.1
5
3.2
6.8
10.0
13.6
20.4
34.0
49.9
48.5
15.9
3.6
6.8
10.0
14.5
20.9
34.0
61.2
56.7
15.4
2.7
6.8
9.5
13.6
20.4
34.0
47.6
45.4
15.9
10
3.6
7.3
10.4
15.4
22.7
38.6
52.2
52.2
22.7
3.6
7.3
10.9
15.9
22.1
38.6
64.0
61.2
20.9
3.6
6.8
10.0
15.0
21.8
38.6
49.9
49.9
24.9
25
4.5
8.2
11.8
17.2
24.9
45.4
59.0
59.0
48.5
5.0
8.2
12.2
18.1
24.9
47.2
70.3
67.1
49.9
4.1
7.7
11.3
17.2
24.9
45.4
54.4
55.8
48.1
50
5.4
9.1
13.6
20.0
29.5
54.4
68.0
68.0
61.2
6.4
9-1.
13.6
20.4
29.5
59.0
78.0
73.9
70.3
5.4
8.6
13.6
19.1
29.5
52.2
61.2
63.5.
56.7
1 Does not include pregnant women, lactating women, or breast-fed children.
b Individual values may not add to totals due to rounding.
c Value not reported due to insufficient number of observations.
80
-------
Table 2.3.4: Mean Body Weights (kilograms) of Children from NHANES II
Age
6-11 months
1 year
2 years
3 years
4 years
5 years
6 years
7 years
8 years
9 years
10 years
11 years
12 years
13 years
14 years
Boys
9.4
11.8
13.6
15.7
17.8
19.8
23.0
25.1
28.2
31.1
36.4
40.3
44.2
49.9
57.1
Girls
8.8
10.8
13.0
14.9
17.0
19.6
22.1
24.7
27:9
31.9
36.1
41.8
46.4
50.9
54.8
Boys and Girls
9.1
11.3
1.3.3
15.3
17.4
19.7
22.6
24.9
28.1
31.5
36.3
41.1
45.3
50.4
56.0
Source: Adapted from NCHS (1987)
2.3.2.2 Drinking Water Intake
The 1980 AWQC National Guidelines used a value of 2 liters/day to represent the drinking
water intake of an individual. In these updated guidelines, EPA recommends the same value when
setting chronic criteria. In addition, EPA recommends a drinking water intake specific to children
for protecting against certain health effects (with those chemicals for which the critical effect of
concern is based on children) because young children may intake a large amount of water per body
weight.
81
-------
Chronic Exposure Scenarios
To protect against health effects due to chronic exposure, EPA recommends an adult-specific
drinking water intake of 2 liters per day. This value has been used as a nationwide estimate of adult
daily water consumption in the drinking water program for setting Maximum Contaminant Level
Goals (MCLGs) and Maximum ContaminantLevels, to be protective of a majority of the population
over the course of a lifetime. The value is also suggested by the Exposure Factors Handbook to be
supported by studies analyzed hi the Handbook for use as an upper-percentile intake rate (USEP A,
1997a). In addition, the value was recommended in the Technical Support Document for setting
water quality criteria for human health in the Great Lakes Region (USEPA, 1995). Based on the
study data from Ershowand Cantor (1989, summarized below), EPA also recommends a 2 liters per
day intake for women of childbearing age.
The value of 2 liters/day has been estimated as being somewhere between the 75th and the
100th percentiles, as reported by different studies of drinking water intake. Thus, using this higher
than average value in combination with recommended default body weights and fish intake rates
would protect most individuals in the population. However, certain individuals who work or
exercise in hot climates may consume water at rates significantly higher than 2 liters/day. Some of
the most highly exposed individuals, such as migrant workers, may not be captured in national
surveys of drinking water intake. Several studies that have estimated drinking water intake are
described below.
One study by the National Cancer Institute (NCI) estimated intake from tap water (which
includes water directly from the tap and tap water added to foods and beverages during preparation)
using data from the NFCS (Ershow and Canter, 1989). For 11,700 adults ages 20 - 64 years old, this
study reports 50th, 75th, and 90th percentile tap water intakes of 1.3, 1.7, and 2.3 liters/day.,
respectively. Table 2.3.5 includes the distribution of intake values by age from this study.
NCI determined drinking water intake values from a study in which 9,000 individuals were
questioned in a population-based, case-control study investigating a possible relationship between
bladder cancer and drinking water (Cantor et al., 1987). This study estimated an overall average tap
water consumption rate of 1.39 liters of water per day. The 100th percentile consumption rate was
estimated to be about 1.96 liters per day as shown in Table 2.3.6.
A survey of drinking water literature by the National Academy of Sciences (NAS) has
calculated the average per capita water consumption to be 1.64 liters per day. NAS estimates that
daily water consumption may vary with physical exercise and fluctuations in temperature and
humidity. It is reasonable to assume those living in arid, hot climates will consume higher levels
of water. However,NAS adopted the 2 liters/day volume to represent the intake of the majority of
water consumers (NAS, 1977). In another survey, the Food and Drug Administration's (FDA) Total
Diet Study estimated rates for water and water used to make drinks and soups for two groups of
adults to be 1.04 and 1.26 liters per day with an average of 1.15 liters per day. Finally, EPA
estimates based on the U.S. Department of Agriculture's (USDA) 1977-78 Nationwide Food
82
-------
Table 2.3.5: Tap Water Intake (g/day) for Both Sexes from Ershow and Cantor (1989)
ab
Sex
Both
Sexes
Males
Females
Age (yr)
<0.5
0.5-0.9
1-3
4-6
7-10
11-19
20-64
65+
All
0.5
0.5-0.9
1-3
4-6
7-10
11-19
20-64
65+
All
<0.5
0.5-0.9
1-3
4-6
7-10
11-19
20-64
65+
All
Mean
272
328
646
742
787
965
1366
1459
1193
250
322
683
773
802
1050
1460
1570
1250
293
333
606
709
772
882
1297
1382
1147
Standard
Deviation
247
265
390
406
417
562
728
643
702
232
249
406
414
437
605
798
704
759
259
281
368
395
395
503
664
584
648
Percentile Distribution
50
240
268
567
660
731
867
1252
1367
1081
240
264
606
693
738
942
1339
1448
1123
240
278
532
622
726
799
1207
1309
1049
75
332
480
820
972
1016
1246
1737
1806
1561
320
408
867
1033
1046
1364
1841
1952
1634
358
500
783
930
992
1147
1655
1687
1505
90
640
688
1162
1302
1338
1701
2268
2287
2092
569
634
1228
1336
1391
1856
2485
2460
2205
672
712
1114
1231
1299
1540
2147
2167
1988
95
800
764
1419
1520
1556
2026
2707
2636
2477
757
871
1464
1530
1609
2179
2949
2790
2673
800
759
1339
1491
1475
1825
2491
2472
2316
99
c
c
1899
1932
1998
2748
3780
3338
3415
c
c
2061
1900
2055
2967
4083
3712
3760
c
c
1806
1932
1888
2424
3359
3071
3097
a Does not include pregnant women, lactating women, or breast-fed children.
b Individual values may not add to totals due to rounding.
c Value not reported due to insufficient number of observations.
83
-------
Consumption Survey identified daily beverage intakes ranging from 1.48 to 1.73 litersperday. Both
the FDA and USDA studies were cited in USEPA (1997a). Based on these studies, EPA estimated
an average adult drinking water consumption rate to be 1.41 liters per day and the 90th percentile
value to be 2.35 liters per day (USEPA, 1997a).
Table 2.3.6: Frequency Distribution of Tap Water Consumption Rates*
Consumption Rate (L/day)
0.80
0.81-1.12
1.13-1.33
1.45 - 1.95
1.96
Cumulative Frequency (%)
19.2
39.6
59.7
79.9
100.0
*Represents consumption in a "typical" week.
Source: Cantor et al. (1987)
Developmental Effects Exposure Scenarios
As noted above, for some contaminants, RfDs based on health effects in children are of
primary concern. Because infants and small children have a higher water consumption per body
weight compared to adults, a higher water consumption rate per body weight may be needed for
comparison with doses from relevant toxicity studies. Use of these higher water consumption rates
when setting criteria based on health effects associated with children should result in adequate
protection for infants and children. Estimating a mean drinking water intake for children ages 0-14
years old, which combines drinking water intake for five age groups within the larger age group of
0-14 years from Ershow and Cantor (1989) and weighting by current population estimates (from
U.S. Bureau of the Census, 1996) results in a drinking water intake of approximately 750 ml. As
a slightly more protective measure than using 750 ml, EPA proposes a drinking water intake of 1
liter. This value is equivalent to about the 75th percentile value, of 960 ml for children ages 1-10
years old (Ershow and Cantor, 1989). The distribution of drinking water intakes for age groups
within the 0-14 year-old group from Ershow and Cantor (1989) is included in Table 2.3.5. This
value is also appropriate to use for evaluating smaller children ages 1-3 years old and has been used
as a default by EPA's water program office for small children in past regulatory efforts.
84
-------
Inhalation and Dermal Exposure
A number of water contaminants are volatile and thus diffuse from water into the air where
they may be inhaled. In addition, drinking water is used for bathing and ambient waters for
swimming and thus, there is at least the possibility that some contaminants in water may be dermally
absorbed.
Dermal absorption and the inhalation of volatilized drinking water contaminants may be
responsible for significant increases in exposure over and above that due to ingestion. However, this
issue is quite complicated. A significant fraction of the water that is ingested is either boiled or
allowed to stand prior to ingestion. In both cases, it is reasonable to assume that volatilization will
decrease the concentration of volatile contaminants in the water that is actually ingested. In addition,
because volatilization can decrease the concentration of volatile contaminants in the water that
comes in contact with the skin, it follows that volatilization can decrease the extent of dermal
absorption.
Thus, volatilization may increase exposure via inhalation and decrease exposure via
ingestion. The net effect of volatilization and dermal absorption upon total exposure to volatile
contaminants in water is unclear. Although several approaches can be found in the literature,
including various models that have been used by EPA, the Agency currently does not have a
proposed methodology for explicitly incorporating inhalation (i.e., from volatilization) and dermal
absorption exposures from household water uses in the derivation of health-based criteria (i.e.,
MCLGs or AWQC). The Agency is currently exploring the effect of volatilization and dermal
absorption upon exposure to drinking water contaminants. For example, the Agency has a joint
agreement with the International Life Sciences Institute (ILSI) to develop guidance on estimating
exposures of inhalation and dermal absorption from contaminants in water. It is anticipated that this
guidance would be incorporated into this methodology when it is available.
2.3.2.3 Fish Intake Rates
Fish intake rates (expressed in grams/day) are used in the equations to derive AWQC.
Throughout this section, the terms "fish intake" or "fish consumption" are used. They generally
refer to the consumption of finfish and shellfish, and the national survey described in this section
(the CSFII) includes both. States and Tribes should ensure that when selecting local or regionally-
specific studies, both types are included when the population exposed are consumers of both types.
If the population of concern are also believed to consume aquatic plants from the water body, this
source should be accounted for with the estimate of other exposures (i.e., the relative source
contribution analysis). Ideally, fish intake rates should be representative of individuals who eat fish
from a given water body for which AWQC are to be set. In addition, priority should be given to
identifying and protecting the most highly exposed fish eaters in the area. Although highly exposed
populations cannot be precisely defined and may differ depending on the water body to be protected,
such fish eaters may generally be separated into two groups: (1) sportfishers, defined generally as
the group of anglers who eat the fish they catch recreationally; and (2) subsistence fishers,
85
-------
individuals who rely on fish for a large part of their protein intake. A more detailed description, as
well as examples, of these highly exposed groups follows.
Sportfishers may vary widely in their catch and consumption rates. Some may eat fish for
short periods throughoutthe year or during certain fishing seasons. Others may fish for much longer
periods during a year. Although sportfishers may primarily fish recreationally and only supplement
their regular diets with the fish they catch, some sportfishers may eat large amounts of fish
throughout the year.
Populations which have been identified as eating a larger portion of sport-caught fish than
the general population (e.g., Native Americans) yet are not recreational fishers are distinguished
from the above group of sportfishers. Such fishers (called subsistence fishers here) may rely on
catching and eating fish to meet nutritional needs or because of cultural traditions. Subsistence
fishers may catch fish year round (CRITFC, 1994) or preserve fish to eat throughout the year. Some
of these fishers, such as Asian-Americans, often consume portions of the fish that recreational fishers
do not often consume (including liver, kidneys, brains, and eyes). In addition, fish may be prepared
whole, providing greater exposure to contaminants (e.g., organs and remains are often used as soup
stock) (Pestana, 1994; Shubat, 1994; Allbright, 1994; Cung, 1994; Nehls-Lowe, 1994; University
of Wisconsin SeaGrant, 1994; Den, 1994; Young, 1994; Lorenzana, 1994). Subsistence fishers are
often (although not always) low income individuals, and may reside in either urban or rural areas.
Several ethnic groups have been identified as having members who subsist on fish. Several
specific groups of Native American fishers have been identified in the Northwest and the Great
Lakes Region (Kmiecik, 1994; CRITFC, 1994; Den, 1994; Young, 1994; Eng, 1994). Asian-
American fishers are a group that includes numerous populations such as Laotian, Hmong,
Cambodian, and Vietnamese, each with differing consumption patterns and cultural traditions.
Asian-American fishers hi particular may eat a larger portion of the fish than generally
recommended, including consumption of additional organs or the whole fish (Pestana, 1994; Shubat,
1994; Allbright, 1994; Cung, 1994; Nehls-Lowe, 1994; University of Wisconsin SeaGrant, 1994;
Den, 1994; Young, 1994; Lorenzana, 1994).
When estimating fish intake for the population of concern, EPA recommends that central
tendency values (i.e., median or mean values) or higher percentile values from studies of fish
consumption relevant to the identified group be used in the derivation of criteria. Values lower than
the median or mean should not be used because the identified populations would not be adequately
protected. Furthermore, when considering median values from fish consumption studies, States need
to ensure that the distribution is based on survey respondents who reported consuming fish because
surveys based on both consumers and non-consumers typically result in median values of zero.
Because fish consumption habits may vary among different types of populations and among
States, EPA prefers that States use information on fish consumption rates directly relevant to the
population being addressed. However, such information may not always be available. Thus, EPA
86
-------
proposes the hierarchy of preferences for consideration offish consumption data described below.
Although Preferences #1 and #2 are likely to result in higher intake rates than the default
recommendations of Preference #4, if the converse is true (i.e., if site-specific or similar
geographic/population studies indicate lower intake rates than the recommended defaults), States
may choose the lower intake rates determined from the first two preferences. However, if a State
chooses values (whether central tendency or high end) that particularly target highly exposed
consumers, they should be compared to high-end fish intake rates for the general population to make
sure that the highly exposed consumers within the general population would also be protected by the
chosen intake rates. As discussed in the Federal Register presentation of the Methodology, it is
recommended that cooked weight values of intakes be used.
Preference #1: Use of Local Information
Once a State has identified the particular population, which, if a protected subgroup, will also
afford acceptable protection to the entire population, EPA recommends that States use results from
fish intake surveys conducted in the geographic area where the State is located to estimate fish intake
rates (measured in grams/day) that are likely to most closely represent the defined populations being
addressed. Generally, the more specific the data are to the individuals who use the water body of
interest, the better the data are considered to be for estimating accurate fish intake rates.
Information on local fish consumption habits may not be already available to States. Thus,
if time and money permit, States are encouraged to conduct their own surveys in order to obtain
estimates offish consumption (in grams/day)and to characterize fisher populations within the State,
and specifically, the locality of interest. The EPA guidance manual entitled Guidance for
Conducting Fish and Wildlife Consumption Surveys (USEPA, 1997b) may be useful in planning and
conducting surveys. This guidance document reviews five methods of obtaining fish consumption
data:
• Recalled information collected by telephone.
• Recalled information collected by in-person interviews.
Recalled information from self-administered mailed questionnaires.
• Diaries maintained by anglers.
• On-site creel censuses (obtaining harvest data collected on-site from single
anglers).
87
-------
The advantages and disadvantages of each method are addressed, and suggestions about
procedures to solve problems associated with each survey method are given.
In addition, Consumption Surveys for Fish and Shellfish lists several suggestions regarding
the type of information to collect when conducting these surveys. Examples of this information
include: (1) sociodemographic characteristics such as age of the angler, number of household
members, and pregnancy or lactation status of women in the household; (2) fishing activities
including seasonal and temporal distribution of fishing activities, whether the angler fishes for sport
or consumption, and the type of fish captured (whether bottom feeders or pelagic); and (3)
preparation and consumption patterns including portions of the fish consumed, methods of
preparation prior to cooking, and procedures for cooking. Such information can be used to
investigate patterns of consumption among high-risk groups, and can aid in most accurately
characterizing consumption rate information using as few assumptions as possible.
The document presents a variety of guidelines for conducting surveys, and is intended to
provide methods for efficiently and cost-effectively collecting information necessary for valid
statistical analyses of risks to subsistence and recreational anglers. States should refer to USEPA
(1997b) for more detailed information on methods of conducting fish consumption surveys.
A couple of issues not addressed in detail in USEPA (1997b) should be emphasized when
planning a fish consumption survey. The first issue involves identification of subsistence fisher
populations. Because it may be difficult to identify subsistence fisher populations solely through
traditional approaches such as mail or phone surveys, it may be necessary for surveyors to use other
methods to target these populations. A couple of methods may be of use. One method involves
contact with community organizations that represent these populations (e.g., Indian tribal
organizations) that have already established a relationship with community members. In addition,
creel clerks (those who interview fishers at specific fishing locations) may be good sources of
information on fisher demographics because they have direct contact with individuals at fishing sites
(Shubat, 1993). It is importantto anticipate cultural and language requirements of each ethnic group
and to try and followthe community-based approach indicated above. Asians and Pacific Islanders
are currently the fastest growing minority population in the U.S. For many first and second
generation immigrants and refugees, survey methods which utilize creel, mail -in, telephone or door-
to-door techniques are ineffective in obtaining reliable data characterizing fish and seafood
consumption patterns (Nakano, 1996; USEPA, 1996). Informal studies indicate a preference for
bottom dwelling fish; therefore, Asian and Pacific Islander surveys should include an appropriate
species list (Soukhaphonh et al., 1996).
A second issue important to emphasize is that, if States intend to consider health effects
resulting from acute exposures when setting AWQC, surveyors may wish to obtain information
regarding maximum amounts of fish that may be eaten at a meal. Because many surveys are
designed to obtain information simply on the number of fish meals eaten by an individual over a
specific time period, rather than the size of the fish meals, maximum meal sizes may not generally
be obtained by a fish consumption survey. As noted above, such large acute exposures may be
88
-------
especially problematic to children, people with special susceptibilities, and pregnant women. In
addition, such high doses may be more likely to occur at specific times during the year, such as
periods when certain types of fish are available or during specific events (e.g., summer vacation,
Native American religious festivals, or fishing tournaments). Thus, surveyors may wish to consider
obtaining information on such maximum intake rates and determine whether these rates are likely
to occur during specific times during the year.
Preference #2: Use of Surveys from Similar Geographic Areas and Population
Groups
For those States and Tribes that do not have resources available to conduct a survey of
consumption rates of local populations and when such information is otherwise not available, EPA's
second preference in determining fish intake rates is for States and Tribes to use results from existing
fish intake surveys that reflect similar geography and/or population groups. For instance, States or
Tribes with subsistence fisher populations may wish to use consumption rates from studies that focus
specifically on these groups, or, at minimum, use rates that represent high-end values from studies
that measured consumption rates for a range of types of fishers (e.g., recreational/sport fishers,
subsistence, minority populations). A State or Tribe in a particular region of the country may
consider using rates from studies that surveyed the same region; for example, a State or Tribe that
has a climate that allows year-round fishing may underestimate consumption if rates are used from
studies taken in regions where individuals fish for only one or two seasons per year. A State or
Tribe that has a high percentage of a particular age group (such as elderly individuals, who have been
shown to have higher rates in certain surveys) may wish to use age-specific consumption rates,
which are available from some surveys.
Fish intake rates estimated from available surveys that have investigated the fish
consumption habits of individuals are described below and presented in Tables 2.3.7 and 2.3.9.
These surveys are divided into the two previously identified groups of highly exposed fisher
populations (sportfishersand subsistence fishers) described above. Although the surveys are divided
into these two categories for ease of presentation, it should be noted that these two categories cannot
always be strictly delineated. In particular, there may be individuals included in the sportfisher
surveys that exhibit habits indicative of subsistence fishers (i.e., eating fish as a large part of their
diet). Also, some members of identified subsistence populations may not subsist on fish as a major
portion of their diets.
These surveys use a variety of methods. The methods used in the surveys to estimate the fish
intake rates are included in Tables 2.3.8 and 2.3.10. A few points should be made about the methods
used in these studies to estimate the consumptionrates. One major issue regarding these rates is that
although they are presented as grams per day in Tables 2.3.7 and 2.3.9, they should be considered
to be approximations of actual gram/day amounts only. For example, the estimates are generally
obtained by memory recall, not strict daily log-keeping of grams eaten per day. In addition, surveys
generally ask respondents to estimate the number of meals they have eaten over a given period of
89
-------
time. Although some surveys include questions about approximate size of the meals, others do not
ask any questions about the actual size of the meals eaten during that time and, instead, assume all
meals are a given size (most often 227 grams, or a half pound).
A second major issue to be addressed is that the estimates offish intake may vary across
surveys for reasons which depend on the type of fish included in the survey. For instance, surveys
may report consumption of only certain types of fish. Some surveys have focused primarily on
either freshwater or saltwater fish, whereas others have collected information on both types. In
addition, some surveys have queried individuals about whether they have eaten recreational fish
only, whereas others have questioned respondents about intake of commercial fish, or both.
Methods of averagingfish consumptioninformationalso differ among studies. Some studies
average the consumption rates over all individuals, regardless of whether they ate fish, while other
surveys average the information only for those individuals who reported eating fish. For example,
Cox, Vaillancourt, and Hayton (1993) report consumption rates averaged for the fish-eating
population, whereas the Alabama Department of Environmental Management (1993) report a rate
averaged for both individuals who eat fish and those who do not eat fish.
As discussed in the Federal Register notice, fish consumption surveys also vary in terms of
whether reported rate values are for cooked fish, uncooked fish, or whether the study is unclear as
to which is reported. States and Tribes should check to" see if the survey study clearly identifies
whether weights represent cooked or uncooked fish.
Many of the differences in survey methods are highlighted in the text and accompanying
Tables 2.3.8 and 2.3.10. However, States should consult the individual surveys to obtain the most
complete descriptions of the study and resulting consumption rates. USEPA's Guidance for
Conducting Fish and Wildlife Consumption Surveys (USEP A, 1997b) includes detailed descriptions
of the various methods for conducting surveys, including their strengths and limitations.
Sportfishers
As noted above, sportfishers differ with respect to their catch and consumption habits.
Surveys of the general sportfishing population may include those who primarily fish for recreational
purposes or eat fish for a small portion of the year but may also include some individuals who eat
fish as a main staple in their diets. Results of sportfisher surveys are described in the following
paragraphs, and are included in Tables 2.3.7 and 2.3.8.
Alabama Fishers. The Alabama Department of Environmental Management (1993)
conducted a survey from August 1992 to July 1993 on-site at various fishing locations. In this
survey, 1,586 individuals were interviewed and asked about the number offish that were caught and
kept for consumption. Demographic information on age, gender, income, and region was also
collected. Two survey methods, which differed in determining meal size of the fish catch that was
to be eaten, were used to estimate consumption rates. Mean and 95th percentile consumption rates
90
-------
for the harvest method were 45.8 and 50.7 grams/day, respectively, and the rates for the serving
method were 43.1 and 50.9 grams/day, respectively. Both were averaged over a year. Although
consumption rates were not found to vary across major ethnic groups, some specific subpopulations
had higher than mean consumption as a function of age and income. Black anglers with incomes
less than $15,000 ate a mean of 63 grams/day, and anglers over 50 years old consumed a mean of
76 grams/day of sport-caught fish (Alabama DEM, 1993).
California Fishers. The Santa Monica Bay Restoration Project contracted with the Southern
California Coastal Water Research Project and MBC Applied Environmental Sciences to conduct
a seafood consumption study from September 1991 to August 1992 (SCCWRP and MBC, 1994).
The purpose of the study was to characterize recreational anglers fishing in the Santa Monica Bay,
including identifying ethnic subgroups of the population with the highest consumption. Information
on household income was also evaluated. The survey form included a census and a questionnaire.
Twenty-nine sites were surveyed on 99 days of sampling, with 2,376 anglers included in the census
and over 1,200 interviews (71%). Of these, 555 anglers (45%) provided enough information to be
used to derive consumption rates. The overall median and mean consumption rates were 21.4 and
49.6 grams/day, respectively, with the highest median and mean consumption rates (85.7 and 137.3
grams/day) in the Other category (i.e., primarily Pacific Island origin). Among ethnic groups, the
90th percentile rates ranged from 64.3 to 173.6 grams/day, with the Hispanic category having the
lowest and the Other category having the highest rates. With respect to income, the study showed
that the lowest income group (<$5,000/year) had the highest median consumption rate (32.1
grams/day) but the highest income group (>$50,000/year) had the highest mean consumption rate
(58.9 grams/day) and the highest 90th percentile (128.6 grams/day). However, it should be noted
that two-thirds of the survey population was comprised of higher income anglers.
Louisiana Fishers. The Louisiana Department of Environmental Quality conducted a
seafood consumption survey in 1993 (Dellenbarger,etal., 1993). A telephone survey was conducted
of 1,100 households in Houma, LA, a coastal community. Households sampled were stratified by
ethnic characteristics;however,the households were otherwise randomly selected. Other high-end
consumers were individuals over 50 years old, who consumed a mean value of 40 grams/day. Rates
for all types offish consumed were 65 grams/day, comprised of 17 grams/day of fresh water fish,
15 grams/day of saltwater fish, and 33 grams/day of shellfish. These rates include both sport-caught
and commercial fish and are averaged for only those people who ate fish and seafood.
New York Fishers. Based on a survey of 4,530 anglers, the New York State Department of
Environmental Conservation (Connelly, et al., 1990) estimated that consumption offish (all types)
by New York State anglers averaged about 45.2 meals per year, or 28.1 grams per day (assuming
227 grams per meal). Averages are also listed by age of the angler, income group, and the area
within New York State where the angler lives. The highest average was recorded for the two Long
Island counties of Nassau and Suffolk, whose populations consumed a combined mean value of 37
grams/day offish. Other high-end consumers were individuals over 50 years old, who consumed
a mean value of 40 grams/day.
91
-------
Table 2.3.7: Sportfishers" Fish Intake Data
Fisher Groupb
Alabama fishers'
Louisiana (coastal) fishers2
New York fishers'
New York (Hudson River) fishers4
Michigan fishers5
Michigan fishers6
Michigan fishers7
Wisconsin fishers (10 counties)8
Wisconsin fishers (10 counties)8
Ontario fishers9
Ontario fishers10
Los Angeles Harbor fishers"
Washington State
(Commencement Bay) fishers12
Washington State (Columbia
River) fishers"
Maine fishers (inland waters)14
Washington State (Columbia
River) fishers15
Fish Intake Rates (g/day)
mean
45.8
28.1
23 (typical)
14.5
18.3
44.7
12.3
26.1
22.5
31 (average)
7.7
6.4
1.8
median
65
37
23
2.0
80%ile
30
120.8
90%ile
62
50 (approx.)
225
54
13
95%ile
50.7
80
70
(approx.)
37.3
63.4
338.8
26
Fish Type'
F+S
F+S
F
F+S
F+S
F
F
F
F
S
F+S
R+C
R+C
R
R
R+C
R
R
R+C
R
R
R+C
NOTES:
'Sportfishers may include individuals who eat fish as a large portion of their diets
bFisher groups refer to the same headings as those that appear in the text .
Tish Type: F= Freshwater, S = Saltwater (may indicate either estuarine or marine waters), R = Recreationally caught
SOURCES: 8c. t . ,OQQ
.., _ _ ,, ,„, 5Fiore, et al., 1989
r^,, f • ?*?',£? 9C°*. Vaillancourt, and Hayton, 1993
Dellenbarger et a 1993 .oSonstegard, 1985
Connelley et al., 1990 8
Barclay, 1993 12
W«t,ctd., 993 ;
we£N f -,o« '4Ebertetal.,1993
'Humphrey, 1976
92
-------
Table 2.3.8: Sportfishers" Survey Methods
Fisher Groupb
Alabama fishers
Louisiana
(coastal) fishers2
New York
fishers3
New York
(Hudson River)
fishers4
Michigan
fishers5
Michigan
fishers6
Michigan
Ishers7
Wisconsin
Ishers (10
counties)8
Wisconsin
fishers (10
counties)8
Ontario fishers9
Ontario fishers10
,os Angeles
larbor fishers"
Washington
State
Commencement
Bay) fishers12
Number
Surveyec
1,586
1,100
4,530
336
2,684
1,104
182
801
801
494
1,059
508
Methods (See Key)c
Contac
Method
on-site
randomd
license
on-site
icense
icense
icense
icense
icense
icense
n-site
cense
Instrumen
int
tele
mail/follow
up by tele
int
mail
mail
mail
mail
mail
nt
nt/follow
p by tele
Reporting
Method
log
recall
recall
recall
recall
recall
log
ecall
ecall
ecall
ecall
ecall
Catch vs
Consump
tion
catch
consump-
tion
catch
consump-
tion
consump-
tion
consump-
tion
catch
onsump-
ion
onsump-
lon
onsump-
on
atch
atch
Individua
vs.
Househol
d
individual
household
individual
lousehold
lousehold
individual
ndividual
ndividual
ndividual
ndividual
ndividual
Data
Available
age, eth,
inc, reg, sex
age, edu,
eth, inc, oth
age, inc, reg
age, edu,
eth, inc, reg
sex
age, edu,
eth, inc, reg,
sex
age, edu,
eth, reg, sex
age, edu,
eth, reg, sex
ge, reg,
ex
ge, eth
Duration
12mos
1 mos
12 mos
12 mos
6 mos
24 mos
summer,
fall
12 mos
summer,
fall
93
-------
Fisher Groupb
Table 2.3.8: Sportfishers" Survey Methods
Methods (See Key)'
Number
Surveyed
Contact
Method
Instrument
Reporting
Method
Catch vs.
Consump-
tion
Individual
vs.
Househol
d
Data
Available
Duration
Washington
State (Columbia
River) fishers"
10,900
license
int
recall
consump-
tion
household
12mos
Maine fishers
(inland
waters)14-*
Washington
State (Columbia
River) fishers15-'
KEY:
Contact Method:
Instrument:
Log/Recall:
Catch/Consumption:
Individual/Household:
Data Available:
Census/Random/Fish Licenses/On-Site/Tribal Members
Personal Interview/Mail Survey/ Telephone Survey
Respondents recorded consumption information in a log or recalled consumption
information during interview
Catch: Original data from catch rates extrapolated to consumption rates
Consumption: Data obtained on consumption patterns
Consumption information obtained either for individuals or for households
Study may have data on: Age/Education/Ethnicity/Income/Region/Sex/Other
NOTES:
'Sportfishers may include some individuals who eat fish as a large portion of their diets.
bFisher groups refer to the same headings as those that appear in the text.
c Blank cells indicate information is not available.
dA "stratified random" approach was used to obtain information with adequate representation of the population of
interest.
•Data available only from draft documents. Consequently, detailed information was not available at the time of
publications. Additional information will be provided in future revisions of this document.
SOURCES:
lALDeptEnvMgt, 1993
2Dellenbarger, et al., 1993
'Connelley, et al., 1990
4Barclay, 1993
'West, etal., 1993
'West, etal., 1989
'Humphrey, 1976
8Fiore, et al., 1989
9Cox, Vaillancourt, and Hayton, 1993
10Sonstegard, 1985
"Puffer, etal., 1982
12Pierce, et al., 1981
"Honstead, et al, 1971
HEbert etal., 1993
15Soldat, 1970
94
-------
Barclay (1993) conducted direct interviews with 336 shore-based anglers on the Hudson
River at sites including the upper Hudson, mid-Hudson, and lower Hudson sites, at both urban and
rural sites. These surveys were conducted between June and November of 1991 and April and July
of 1992. Because the survey did not reach anglers in boats or all river areas, the authors of the
survey note that the results cannot be directly extrapolated to the entire population of Hudson River
anglers. Over 58 percent of the individuals eat their catch. The survey reports that the average
frequency of fish consumption reported was 3 meals over the previous month, but did not ask
respondents about the size of their fish meals. Assuming 227 grams (8 ounces) offish would be
eaten per meal and assuming 4.3 weeks per month, the results translate to an average fish
consumption rate of 23 grams/day.
Michigan Fishers. West et al. (1993) completed a survey of Michigan fishers over a one year
period. For this survey, 2,684 individuals who purchased fishing licenses responded to mailed
surveys. Consumption of commercial and sport-caught fish was estimated through a 7-day recall,
and data were separated demographically by age, education, ethnicity, income, region and gender.
Mean consumption was estimated to be 14.5 grams/day. The 80th percentile was 30 grams/day, 90th
percentile consumption rate was 62 grams/day and the 95th percentile rate was 80 grams/day.
Several specific subpopulations surveyed in this study had higher than average consumption rates.
Minority fishers (primarily black and non-reservation Native Americans) with annual incomes less
than $25,000 averaged the highest consumption rate of all Michigan angler groups surveyed,
consuming a mean of 43.1 grams/day of sport-caught fish.
An older survey by West et al. (1989) evaluated Michigan fishers as part of a revision of
exposure pathways for the Michigan Toxic Substance Control Commission. This earlier study was
only conducted over a six month period, but its results were corroborated by the more recent survey
data. The population studied was sport anglers, and consumption of both self-caught and
commercial fish was considered. The survey relied on seven-day recall in order to estimate mean
fish consumption; the percentage of respondents consuming no fish was high (56.6 percent). The
study concluded that mean fish consumption for Michigan sport anglers and their families is 16.1
grams/day, after adjustment for non-response bias. The 90th percentile consumption is
approximately 50 grams/day, the 95th percentile is about 70 grams/day, and the maximum reported
fish consumption is over 200 grams/day.
The Michigan Department of Natural Resources conducted a survey of 3 81,000 sport-fishers
in 1974 (Humphrey, 1976, as cited in Rupp et al., 1979). This survey obtained mean catch of 36 Ibs.
offish per year (44.7 grams per day) for consumption.
Wisconsin Fishers. In a survey of anglers in Wisconsin, the annual mean number of sport-
caught meals was 18. Using the assumed fish meal size of 8 ounces (227 grams) from this survey,
the estimated mean daily consumption of sport-caught fish in Wisconsin is about 11 grams/day!
When respondents who consumed no sport-caught fish were excluded, the mean daily sport-caught
fish intake was 12 grams/day (Fiore et al., 1989). The 95th percentile value, counting only those
individuals who consumed any sport-caught fish, was determined to be 37 grams/day.
95
-------
Ontario Fishers. Another study was completed by Ontario sports fishers in 1992 (Cox,
Vaillancourt,andHayton, 1993). Questionnaires were inserted randomly into 10,000 copies of 1992
Guide to Eating Ontario Sports Fish, and 494 replies were received. Questions regarding fish
preferences and catch rate, consumption rate and portion sizes, and use of consumption advisories
were asked. A mean daily consumption of 22.5 grams/day was calculated based on estimated
average meal size and frequency of eating sport-caught sportfish. Anecdotal evidence provided by
one researcher studying the Ontario sportfishers during an earlier survey from 1985 (Sonstegard,
1985 as cited in Kleiman, 1985) found that an average sportfisher consumed a mean of 31
grams/day, while a high-end consumer ate 62 grams/day (the percentile value was not specified).
The maximum amount consumed was over 310 grams/day.
Idaho Fishers. One study was conducted in the Lake Coeur d'Alene region in Idaho (West,
1993; RichterandRondinelli, 1989). 933 individuals were surveyed, including Native Americans
living both on and off reservations, individuals selected randomly from individuals with fishing
licenses in Idaho, and volunteers who were recruited for the study. Tribal members were surveyed
in person, while others were surveyed primarily by telephone. All respondents were asked to recall
fish consumption patterns. This study was conducted over a period of three months, so data must
be extrapolatedto the rest of the year. Consumption rates for the licensed fisher population ranged
from 16 to 27 grams/day.
Los Angeles Harbor Fishers. From January to December of 1980, 1059 interviews with
sportfishers were conducted in several fishing areas of the Los Angeles Harbor area (Puffer et al.,
1982). No fisher was sampled more than once. Data was collected on the following: amount of fish
caught on the day of the interview, the primary use of the fish (whether eaten by the fisher's family,
given away, thrown back, etc.), frequency of fishing, and other variables. Based on this data and
assuming that only an edible portion (1/4 to 1/2) of the caught fish would be eaten, median and 90th
percentile consumption rates of 37 grams per day and 225 grams per day were determined. The 95th
percentile was 338.8 grams/day. Consumption rates were also estimated by age, race, and species
caught This study indicatesthat median consumption rates for Orientals/Samoansare 71 grams/day
and 113 grams/day for individuals over 65 years old.
Washington State Fishers. Interviews were conducted with fishers in Commencement Bay,
Washington from July to late November; 304 interviews were conducted in summer and 204 were
conducted in the fall (Pierce, 1981). Data were collected on size and amount of specific species
caught, size of the fishers' families, frequency of fishing, and planned use of the fish. The fishers
were later called about whether the fish had been eaten. USEP A (1989a) used these data to estimate
a median consumptionrate value of 23 grams per day and a 90th percentile of 54 grams per day (also
reported in USEP A, 1997a). The authors note that although a survey of night/dawn fishing was
conducted only once, fish caught at this time could represent a significant part of the total fish caught
from the bay. Therefore, these values may underestimate fish consumption.
Honstead et al. (1971, as cited in Rupp et al., 1979) conducted a study of the consumption
patterns of sportfishers on the Columbia River in the Tri-City area of Hanford, Washington. This
96
-------
survey monitored 10,900 persons, each of which were members of households where a Columbia
River angler resides. The surveyors required respondents to recall the number of fish meals
consumed over a 12-month period and, using an estimate of 200 grams per meal, calculated the mean
annual consumption to be 2.8 kg per year (7.7 grams/day).
Lake Ontario Fishers. Connelly et al. (1996) surveyed 1,202 Lake Ontario anglers through
mail questionnaires, diaries, and telephone interviews. The mail questionnaires were based on a 12
month recall of 1991 fishing trips; the diaries involved self-recording of 1992 fishing trips. Of the
1,202 participants, 853 returned a diary or provided diary information by telephone. Participants
were instructed to record in the diary the species offish eaten, meal size, method by which fish was
acquired (sport-caught or other), fish preparation and cooking techniques used, and the number of
household members eating the meal. Due to changes in health advisories for Lake Ontario which
resulted in less Lake Ontario fishing, only 43 percent, or 366 persons indicated that they fished Lake
Ontario in 1992. The mean fish intake from all sources was 17.9 grams/day and from sport-caught
sources was 4.9 grams/day. The median rates were 14.1 grams/day for all sources and 2.2 grams/day
for sport-caught; the 95th percentiles were 42.3 grams/day and 17.9 grams/day for all sources and
sport-caught, respectively. Residents of large cities and younger people had lower intake rates on
average. The authors note that although diaries tend to provide more accurate information than
studies based on 12 month recall, a considerable portion of the respondents participated in the study
for only a portion of the year and some errors may have been generated in extrapolating the results
to an entire year.
Alaska Communities. Wolfe and Walker (1987) analyzed a data set from 98 Alaska
communities (four large urban population centers and 94 small communities) for harvests offish,
land mammals, marine mammals, and other wild resources. The data set was developed by various
researchers in the Alaska Department of Fish and Game, Division of Subsistence, between 1980 and
1985. Respondents were asked to estimate the quantities of particular species that were harvested
and used by members of their households during the previous 12 month period. Urban sport fish
harvests were derived from a survey that was mailed to a randomly selected statewide sample of
anglers. For the four urban centers, fish harvests ranged from 6.2 grams/day to 26.2 grams/day. The
range for the 94 small communities was 31 grams/day to 1,541 grams/day. For the 94 communities,
the median per capita fish harvest was 162 grams/day. Dressed weight, the portion brought into the
kitchen for use, varied by species and community, but in general was 70 to 75 percent of total fish
weight. The authors used a factor of .5 to convert harvest to intake rates, yielding a median per
capita consumption rate in the 94 small communities of 81 grams/day, and a range of 15.5 to 770
grams/day.
Savannah River Fishers. Turcotte (1983) estimated fish consumption from the Savannah
River in Georgia based on total harvest, population studies, and a Georgia fishery survey. The angler
survey data, which included the number of fishing trips per year as well as the number and weights
of fish harvested per trip, were used to estimate the average consumption rate in the angler
population. The study found an average consumption rate of 31 grams/day and a maximum rate of
58 grams/day.
97
-------
Alabama Fishers. Meredith and Malestuto (1996) studied anglers in 29 locations in Alabama
to estimate freshwater fish consumption. The purpose of their study was to compare two methods
of estimating fish consumption: the harvest or krill survey, and the serving-size method. The two
techniques yielded comparable estimates of mean fish intake: 43 and 46 grams/day, respectively.
Florida Fishers Who Receive Food Stamps. As part of a larger effort, the Florida Department
of Environmental Regulation attempted to identify fish consumption rates of anglers who were
thought to consume higher rates offish. Interviews with twenty-five households' primary seafood
preparers were conducted at each of five food stamp centers per quarter for an entire year. The
respondents were asked to recall fish consumption at home within the previous seven days. Sekerke
et al. (1994) found that adult males in the study consumed 60 grams/day of finfish and 50 grams/day
of shellfish; adult females consumed 40 grams/day and 30 grams/day, respectively, of finfish and
shellfish.
Subsistence Fishers
Subsistence fishers consume fish as a major staple of their diet. As noted above, subsistence
fishers often have higher consumption rates than other fisher groups; however, consumption rates
vary considerably among subsistence fishers. Consequently, generalizations should not be made
about this fisher group. If studies contained in this section are used to estimate exposure patterns
for a subsistence population of concern, care should be taken to match the dietary and population
characteristics of the two populations as closely as possible. Several surveys evaluating the
consumption patterns of subsistence fishers have been initiated in the last several years. Some of
these have been completed and many more are currently being carried out, with results expected in
the near future. Althoughmany of these surveys provide only a range of consumption rates, a great
deal of qualitative information has been gained through these surveys, both about the individual
populations that were studied and about effective survey methods for different groups of subsistence
fishers. The consumption rates reported by these surveys are presented below. Results of these
surveys and the methods used to collect the data are summarized in Tables 2.3.9 and 2.3.10.
Great Lakes Tribes. The Great Lakes Indian Fish and Wildlife Commission conducted a
survey of spear fishing for walleye among Native Americans living on reservations in the Great
Lakes Region (Kmiecik, 1994). This study was designed to evaluate the concern about mercury
among spear fishers hi the tribes of the Great Lakes Region. The results of this study showed that
people were modifying their behavior about where to fish and types and sizes offish to keep based
on concerns about mercury. Although consumption rates had no baseline for comparison prior to
mercury concerns, many respondents indicated that they modified their consumption offish due to
concerns about mercury contamination. Despite these possible decreases in fish consumption, the
rates of consumption of walleye were still extremely high; the mean value was 351 grams/day, while
the maximum amount consumed was 1,426 grams/day. These daily consumption rates were
calculated by multiplying the average portion size, as reported by the respondents, by the
respondents' average consumption of 2.75 meals per week (that is the average of each season's
98
-------
meals/wk), and then dividing by 7 days/wk. Assuming individuals may have been eating other fish
in addition to walleye, the rates may be higher than these values.
Idaho Fishers. As described above, a study conducted in the Lake Coeur d'Alene region in
Idaho surveyed Native Americans, individuals with fishing licenses, and volunteers (West, 1993;
Richter and Rondinelli, 1989). This study was conducted over a period of three months, so data
must be extrapolated to the rest of the year. Consumption rates of tribal members ranged from 28
to 49 grams/day.
Columbia River Tribes. One of the most comprehensive surveys of fishing patterns among
Native Americans has been conducted by the Columbia River Inter-Tribal Fisheries Commission.
The study surveyed four of the tribes living in the Columbia River Basin (CRITFC, 1994). From
four tribes both on and off the reservation, 717 individuals were surveyed in person regarding their
consumption patterns of self-caught fish, wild game, and wild rice. The responses were based on
memory recall, and the survey was conducted over a full year. Mean consumption from this study
was calculated as 58.7 grams/day and the 95th percentile is 170 grams/day.
American Samoan Fishers. A number of surveys have been conducted in American Samoa
that have attempted to assess the potential risk of industrial development in the major harbor on the
main island of American Samoa where most of the population also lives and fishes (Den, 1994;
Young, 1994; Eng, 1994). The local EPA conducted a pilot scoping survey to assess the extent of
the contamination in the fisheries resources and heavy metal poisoning in blood and urine of sample
populations; a brief survey of consumption rates was included in this survey. The results of this
study (the toxicity of the harbor and the fisheries resources) encouraged the local EPA to apply for
a study to be conducted by the Centers for Disease Control (CDC). Results indicated fish
consumption at a rate of approximately 12 grams/day (Ponwith, 1991; ATSDR, 1995).
Wisconsin Chippewa Indians. Peterson et al. (1994) investigated the extent of exposure of
Chippewa Indians who consume fish caught in northen Wisconsin lakes to methylmercury. The
study, conducted in May 1990, included 175 randomly selected and 152 nonrandomly selected
participants. The authors reported that both groups had similar fish consumption rates. Participants
were asked to complete a questionnaire describing their routine fish consumption and, more
extensively, their fish consumption during the previous two months. Results from the survey
showed a mean fish consumption of 1.2 meals per week. This includes fish from all sources. The
consumption figure translates to a fish intake of 20 grams/day, using 117 grams/meal as the average
weight offish consumed per fish meal in the general population. Consumption varied seasonally,
with the highest consumption during April and May, the spearfishing season for walleye. During
peak consumption months, males and respondents under 35 consumed more fish than females and
respondents 35 and over.
Miccousukee Indian Tribes. The Centers for Disease Control (1993) administered dietary
questionnaires to 2 children and 183 adults from the Miccousukee Indian Tribes of South Florida.
The survey found that 31 percent ate fish from the Everglades during the previous six months; 57
99
-------
percent consumed marine fish during the previous six months. The median consumption of local
fish was 3.5 grams/day; the maximum consumption was 168 grams/day. Blue gill was the most
common species of local fish consumed; largemouth bass were consumed in greatest quantity.
Wisconsin Tribes. A 1992 EPA report entitled Tribes at Risk (The Wisconsin Tribes
Comparative Risk Project) reported an average total fish intake for Native Americans living in
Wisconsin of 35 grams/day. The average daily intake of locally harvested fish was 31.5 grams.
Tribes ofPuget Sound. In November 1994 Toy et al. (1995) completed a study of fish
consumption among 190 adult members of the Tulalip and Squaxin Island Tribes of Puget Sound.
The study was conducted between February and May 1994. Fish consumption practices were
assessed using dietary recall methods, food models, and a food frequency questionnaire. Fish
consumed were categorized into anadromous fish (e.g., king salmon and sockeye salmon), pelagic
fish (e.g., cod and pollock), bottom fish (e.g., halibut and sole), and shell fish (e.g., manila clams,
scallops, and mussels). Anadromous fish and shell fish were consumed in greatest quantities.
The 50th percentile consumption rate for all fish combined for the Tulalip Tribe was 0.55
grams/kg body weight/day and 0.52 grams/kg body weight/day for the Squaxin Tribe. If an average
body weight is assumed to be 70 kg, the daily fish consumption rate for adults in the Tulalip Tribe
was 38.5 grams/day and 36.4 grams/day for the Squaxin Tribe. The weighted combined median
daily fish consumption for both tribes was 37.1 grams.
Native Americans near Clear Lake, California. Harnly et al. (1997) found that Native
Americans living near Clear Lake, California consumed an average of 84 grams/day of fish (60
grams/day of sport fish plus 24 grams/day of commercial fish). The most popular species of
sportfish were: catfish, perch, hitch, bass, and carp. Commercial species most commonly eaten
were: snapper, tuna, salmon, crab, and shrimp.
Hawaiian Islands. The Mercury Study Report to Congress (1997) cites a number of studies
on the commercial utilization of seafood [i.e., Higuchi and Pooley (1985) and Hudgins (1980)] and
analyses of epidemiology [i.e., Wilkens and Hankin (1996)] which provide a basis to describe
general patterns of fish consumption among Hawaiians. These studies indicate that, on average,
Hawaiians consumed 30.5 grams/day of fish in 1972 and 24.0 grams/day in 1974. A 1987 State of
Hawaii study of 400 residents cited by the authors found that shrimp and mahimahi were the most
popular seafoods.
Alaska Natives. Nobmann et al. (1992) performed a nutrient analysis of the food consumed
in eleven communities that represented different ethnic and socioeconomic regions of Alaska. The
survey sample included 351 adults aged 21-60 years. Information was obtained using 24 hour
dietary recalls during five seasons over an 18-month period. The mean daily intake of fish and
shellfish of Alaska Natives was 109 grams/day.
100
-------
Table 2.3.9: Subsistence Fishers8 Consumption Data
Fisher Group6
Great Lakes tribes1
Columbia River tribes2
Florida residents receiving food stamps3
Florida Asian residents3
High-end Caucasian consumers on Lake
Michigan4
Wisconsin tribes5
Chippewa tribes in Wisconsin6
Native Alaskan Adults7-0
Fish Intake Rates (g/day)
mean
351
58.7
23
59
54
31.5
55
109
95%ile
170
max
1426
132
Fish Type0
F
F
F+S
F+S
F+S
R
R
R+C
NOTES:
"Subsistence fishers include groups (such as the Florida residents receiving food stamps) that may eat sport-caught fish at high
rates but do not subsist on the fish as a large part of their diet
bFisher groups refer to the same headings as those that appear in the text
°Fish Type: F= Freshwater, S = Saltwater (may indicate either estuarine or marine waters). R = Recreationally caught
SOURCES:
'Kmiecik, 1994
2CRITFC, 1994
'Degner et al., 1994
"Hovinga, 1992; 1993
SUSEPA, 1992
"Peterson et al., 1995
7Nobmann et al., 1992
101
-------
Table 23.10: Subsistence Fishers' Survey Methods
Fisher Type*
Great Lakes tribes'
Columbia River
tribes2
Florida residents
receiving food
stamps'
Florida Asian
residents'
High-end Caucasian
consumers on Lake
Michigan4-11
Wisconsin tribes"
Chippewa tribes in
Wisconsin*"1
Native Alaskan
Adults"
Methods (See Key)'
Number
Surveye
d
69
717
500
120
115
323
351
Contact
Method
tribe
tribe/
random
random'
tribe/
random
Instrument
mail
interview
telephone
interview
Reporting
Method
recall
recall
recall
recall
recall
Catch vs.
Consumption
consumption
consumption
consumption
consumption
consumption
Individual vs.
Household
individual
individual
individual
individual
Data
Available
NA
age. eth. reg.
sex
age, eth, reg,
sex, income
sex,
employmt.,
age, education
Duration
2 mos
12mos
12 mos (of
total study,
not recall)
1 mo
18 mos
KEY:
Contact Method: Census/Random/Fish Licenses/On-Site/Tribal Members
Instrument: Personal Interview/Mail Survey/Telephone Survey
Log/Recall: Respondents recorded consumption information in a log or recalled consumption information during interview
Catch/Consumption: Catch: Original data from catch rates extrapolated to consumption rates. Consumption: Data obtained on
consumption patterns
Individual/Household: Consumption information obtained either for individuals or for households
Data Available: Study may have data on: Age/Education/Ethnicity/Income/Region/Sex/Other
NOTES:
'Subsistence fishers include groups (such as the Florida residents) that may eat sport-caught fish at high rates but not subsist on fish as a large
part of their diets.
bFishcr groups refer to the same headings as those that appear in the text.
'Blank cells indicate information is not available.
'Data available only from draft documents.
•Number sampled per county was proportionate to population in county compared to the entire State.
SOURCES:
'Kmiecik, 1994
JCR1TFC, 1994
'Degneretal., 1994
'Hovinga, 1992; 1993
'USEPA, 1992
'Peterson etal., 1995
'Nobmann et al., 1992
102
-------
Surveys in Progress
The Wisconsin Department of Health is currently conducting a study of the Hmong
populations of Sheboygan and Manitowac (Nehls-Lowe, 1994; University of Wisconsin Sea Grant,
1994). These surveys have resulted from a larger project that was designed to inform the Asian-
American populations of the potential dangers of eating too much contaminated fish.
The EAGLE project (EAGLE, 1991; Cole, 1994) is a Canadian collaboration of the
Assembly of First Nations and Health and Welfare Canada. This project is a several year study of
the health of First Nation communities throughout the northern Great Lakes region. Consumption
patterns by these communities of local food sources have been obtained, and preliminary results of
this project have been compiled. The results will not be ready, however, until sometime in 1997
(Wheatley, 1996).
Another study is underway among the Ojibway peoples (Chippewa) of the upper Great Lakes
region (Bellinger, 1993 and 1996) and is currently being finalized. This study is designed primarily
to study the correlation between fish consumptionhabits, body burdens, and neurobehavioral effects.
EPA Regions 9 and 10 have begun studies of the Asian-American/Pacific Islander
communities in Washingtonand California (Den, 1994; Young, 1994; Eng, 1994; Lorenzana, 1994).
In order to most effectively reach the communities that they wish to survey, Region 10 awarded the
project to a local Asian-American community. Specifically called the Asian Pacific American
Seafood Consumption Study and conducted via the Refugee Federation Service Center and the
University of Washington, the community group designed the study with input from technical
advisors (statisticians, toxicologists, epidemiologists), agency representatives and various
community groups (USEPA, 1996). Personal interview surveys conducted in the King County area
of Seattle, Washington were completed during the of summer 1997, and a report is expected by
September 1998 (Lorenzana, 1998). A study of the Laotian community in the San Francisco Bay
Area (specifically, west Contra Costa County) was funded by EPA Region 9 and conducted via the
Asian Pacific Environmental Network, a non-profit organization that coordinates environmental
health projects. This represents a community-based survey where the survey questions were
designed with input from the Pacific Asian community, interested agencies and academia, and where
the actual survey was conducted by an elder and junior person from each of the various ethnic groups
composing the Laotian community. Data collection was completed during the Summer of 1997, and
a draft report developed. This underwent subsequent peer review and the report was finalized in
March 1998 (Den, 1998).
Preference #3: Use of Distributional Data from National Food Consumption
Surveys
If information from existing regional studies is not relevant to a given State, EPA's third
preference is that States use distributional information for intake of fresh/estuarine species for
different population groups from national food consumption surveys. EPA has analyzed one such
103
-------
national survey, the combined 1989, 1990, and 1991 Continuing Survey of Food Intake by
Individuals (CSFH). EPA recommends use of the CSFII (see Tables 2.3.11 through 2.3.22), but
believes similar nationally-based surveys are appropriate for consideration (see Table 2.3.1 on
sources of exposure information). The 1989 through 1991 CSFII data are the most recent analyzed
by EPA for developing estimates. As more current data become available, these estimates may be
revised by EPA.
In addition to providing nationally-based infornlation, which offers a greater quantity of data
points, the CSFII information is also presented here for regional break-outs from the same data set.
States may wish to consider these regional values if they have at least some information as indicated
with Preferences #1 and #2, and if they believe that the consumption rates of the particular
population of concern differ from the national rates. However, if a State has not identified a separate
well-defined population of highly-exposed consumers and believes that the national data from the
CSFII are representative, EPA recommends these national data. Although the regional break-outs
are provided for the States to consider, EPA believes that there is less confidence in the regional
estimates due, in part, to the relatively small sample size that results when these break-outs are made
(see Tables 2.3.12 through 2.3.20). For example, the Mountain Region is indicative of a very small
sample size with few respondents who reported consumption during the study period and is skewed
by a few high consumers (e.g., a mean of 3.23 grams/day and a 90th percentile of 0.48 grams/day).
It is, therefore, not recommended that these breakouts be used by themselves to represent
regional/State intakes. In addition, the geographic divisions, as created, may not accurately reflect
the consumption patterns of each State within a given division. For example, whereas the Pacific
geographic division (i.e., California, Oregon, Washington) may be reasonable to use for West Coast
populations, it is doubtful that the values for the West South Central division apply equally to each
State (Table 2.3.18). That is, fish consumption in Louisiana may be vastly different than fish
consumption in Oklahoma.
A detailed set offish consumption tables from the CSFII is presented in Appendix A of this
document. The tables indicate consumption rates for adults, children under 14, women of child-
bearing age (considered to be ages 15-44), as well as per capita values. Both average and acute
values are presented for the adults and per capita groups, whereas only acute values are given for
children under 14 and women of childbearing age. The procedures for determining average and
acute consumption are described on pages 107 and 117, respectively. Appendix A includes the
regional breakouts that are also listed in Tables 2.3.12 to 2.3.20. All of the aforementioned tables
are presented in both grams/day and in mg/kg/day. Finally, the Appendix includes species breakouts
by mean consumption for each of the four major groups in grams/day.
The U.S. Department of Agriculture conducts the CSFHs, through which dietary intake data
is collected for selected years from April of one year to March of the next (USEPA, 1998). About
25 percent of the interviews are conducted in a calendar quarter. These data are collected from the
48 conterminous States over 3 consecutive days. On the first day of the survey, participants give
information to an in-home interviewer, and on the second and third days, data are taken from self-
administered dietary records. Meals consumed both at home and away from home are recorded.
104
-------
However, it was not possible to distinguish between the intake offish locally caught and that which
was not. Although the assumption that all freshwater/estuarine fish consumed comes from a
particular water body is somewhat conservative, EPA believes that this is a reasonable assumption
to ensure adequate protection from such fish subject to contamination.
The CSFII 1989-1991 did not draw samples from Alaska or Hawaii. As these two States
could potentially contain a larger percentage of subsistence fishers than the population from the 48
conterminous States, the absence of data from these two States could result in a slight underestimate
of per capita fish consumption for the entire population. This underestimate is probably insignificant
given that the populations of Alaska and Hawaii are quite small compared with that of the total
conterminous States (USEPA, 1998). However, as indicated above, Alaska and Hawaii are
encouraged to make decisions on fish intake via Preferences #1 and #2, if possible, to ensure the
most accurate estimates.
The CSFII survey is a national multi-stage, stratified-cluster area probability sample. The
48 conterminous States were divided into 60 strata. Within these strata, counties, cities, and areas
within cities were grouped into relatively homogeneous units called primary sampling units (PSUs).
Two of these units were sampled, with replacement, from each of the strata. Each PSU was sampled
with probability proportional to its 1985 projected population. These units were further divided into
area segments, from which predetermined numbers of households were selected for participation.
Each household within an area segment had equal probability of selection (USEPA, 1998). To allow
the data for the three survey years to be combined, the area segments for each of the years were
drawn from the same PSUs (USEPA, 1998).
Each of the surveys consists of "basic" and "low-income" samples. Individuals in all
households, regardless of income, were eligible for inclusion in the basic sample. In the low-income
sample, only households with gross income at or below 130 percent of the Federal poverty threshold
were eligible for inclusion. Both samples are included in the distributional estimates using data from
the three survey years.
Response rates for the three survey years and for the low-income and basic surveys varied
from 40 to 53 percent. USDA corrected the survey weights for non-response.
Of the 6,000 food categorizationsin the CSFII surveys, 465 relate to fish. Survey respondents
with 3 days of dietary intake data reported consumption across 284 of these fish-related food codes.
The amount of a fish-related food code reported was adjusted according to information in the USDA
recipe file to reflect the proportion offish in the recipe. For example, if fish was 80 percent of the
recipe, then consumption of 100 grams of the food code was adjusted to 80 grams to represent the
amount offish consumed.
Food codes were assigned to either a freshwater/estuarine or marine habitat. Food codes
containing flatfish (i.e., flounder, smelt, halibut, plaice, and sole), clams, scallops, crabs (with the
exception of king crab) and salmon were apportioned between the freshwater/estuarine and marine
105
-------
habitats based on the proportions of freshwater/estuarine and marine species landed during 1989,
1990, and 1991 reported by the National Marine Fisheries Service (NMFS, 1995-96).
In some cases habitat assignments are based on NMFS data and life-cycle considerations.
If a particular species is listed by NMFS as commercially harvested in marine waters, but is known
to spend at least part of it's life-cycle in estuarine or freshwater habitats, further evaluation was
undertaken to determine the significance of a species' life-cycle with respect to exposure to
chemical contaminants in freshwater and estuarine waters. Species with life-cycles utilizing both
freshwater/estuarine and marine habitats identified as contributing significantly to the CSFII fish
consumption rate determination include shrimp and salmon.
Shrimp, which are harvested in both marine waters and freshwater/estuarine waters, spend
their juvenile years up to sexual maturity in freshwater/estuarine habitats. At sexual maturity,
shrimp are nearly adult size and generally begin migration to marine waters where they spend the
remainder of their adult life. Once shrimp reach sexual maturity, they are nearly full grown and
available for commercial harvesting from both estuarine and marine waters. Though shrimp are
harvested from both estuarine and marine waters, they have been assigned to the freshwater/estuarine
habitat designation. This is because shrimp are harvestable from estuarine waters or immediately
after migrating to marine waters (Zein-Eldin and Renaud, 1986; Kutkuhn, n.d.).
The six species of anadromous salmon found in North American waters spend their first 3
months to 3 years in freshwater/estuarine habitats before migrating to marine waters. At the time
of out migration to marine waters, juveniles measure up to 5 inches in length. All six species will
then spend 1-5 years maturing as adults in open sea before migrating back to freshwater lakes, rivers
and streams for spawning. Depending on the species, spawning may occur from one to eight weeks
after entering freshwater habitats (some unique populations of sockeye salmon may spend up to 6
months in lakes prior to spawning). Additionally (with the exception of these sockeye populations)
most salmon fast, thus spending their energy making the trip to their spawning destination. Because
these six species of salmon spend essentially their entire adult life in open seas prior to commercial
harvesting from marine waters, all salmon have been designated marine habitat with the exception
of 1 percent of the total U.S. harvest which accounts for salmon which are farmed raised or harvested
from landlocked populations (Groot and Margolis, 1991).
Consumption for the given USDA food code with an unknown habitat designation was
allocated across the freshwater/estuarine and marine habitat types in the same percentages as those
observed across food codes from known habitat types.
Average daily individual consumptions for a given fish-by-habitat category were calculated
by summing the amount offish eaten by the individual across three reporting days for all fish-related
food codes in a given fish-by-habitat category. The total individual consumption was then divided
by three to obtain an average daily consumption. The three-day individual food consumption data
collection period is one during which a majority of sampled individuals did not consume any finfish
or shellfish. The non-consumptionof finfish or shellfish by a majority of individuals, combined with
106
-------
consumption data from high-end consumers, resulted in a wide range of observed fish consumption.
This range offish consumption data would tend to produce distributions offish consumption with
larger variances than would be associated with a longer survey period, such as 30 days. The larger
variances would reflect greater dispersion, which results in larger upper-percentile estimates, as well
as wider confidence intervals associated with parameter estimates. It follows that the estimates of
the upper percentiles of per capita fish consumption based on three days of data will be conservative
with regards to risk (USEPA, 1998).
For each type of criteria (chronic, acute, or consideration of developmental effects),
percentile values from distributional data on intakes are presented below for consumption of
freshwater and estuarine fish, as well as for consumption of all fish (including marine species).
Chronic Criteria
Table 2.3.11 and Exhibit 2.3.1 include distributional data on intake rates of fresh/estuarine
fish for adults 18 years and older. These intake values represent "as consumed" weights; that is, they
are primarily cooked weight intakes but also include any raw fish consumption (e.g., raw shellfish)
reported. These estimates were determined by averaging information from both consumers and non-
consumers offish over the three days of the survey. This survey did not specifically ask questions
on whether a respondent eats fish or how often and, therefore, it is not possible to identify consumers
from non-consumers. Since the CSFII reporting period is only three days, long-term consumption
distributions cannot be well characterized using the CSFII data. EPA is recommending adult intakes
(i.e., specifically based on individuals age 18 years and over) for the general, sport fisher, and
subsistence fisher populations to be consistent with the fact that the assumptions used for drinking
water intake and body weight are also based on adults. These values represent reasonable intake
rates for long-term exposure that result in chronic effects. As shown in Table 2.3.11, the arithmetic
mean for adults is 5.59 g/day; the median is 0 g/day. The 90th percentile value is 17.80 g/day, the
95th percentile is 39.04 g/day, and the 99th percentile is 86.30 g/day. Ninety percent confidence
intervals for the mean and 90 percent bootstrap intervals for the median and percentile values are
also recorded in Table 2.3.11. EPA determined confidence interval estimates for the percentile
estimates by using Efron's percentile bootstrap technique (USEPA, 1998). Exhibit 2.3.1 shows the
cumulative distribution (via histogram), which States may wish to use to estimate intake rates at
different percentile values from those values presented in Table 2.3.11.
107
-------
Table 2.3.11: Daily Estimates of Fish Consumption (Finfish and Shellfish): Individuals of
Age 18 and Over in the U.S. Population (g/day)
Statistic
Estimate
90% Interval*
Lower Bound
Upper Bound
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
5.59
0
17.80
39.04
86.30
4.91
0.00
14.89
36.13
81.99
6.28
0.00
20.63
42.16
96.67
All Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
18.01
0.00
60.64
86.25
142.96
16.85
0.00
57.06
80.29
134.23
19.17
0.00
64.63
91.00
154.15
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications
Source: CSFII (1989-1991)
Geographic data are included for areas of the United States, as determined by the U.S.
Department of Commerce for the 1980 Census of Population. Because of small sample size, these
data are provided for all individuals rather than for those only > 18 years old. The regions are
broken out as follows:
New England:
Middle Atlantic:
South Atlantic:
Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island,
Vermont
New Jersey, New York, Pennsylvania
Delaware, District of Columbia, Florida, Georgia, Maryland, North
Carolina, South Carolina, Virginia, West Virginia
East North Central: Illinois, Indiana, Michigan, Ohio, Wisconsin
108
-------
East South Central: Alabama, Kentucky, Mississippi, Tennessee
West North Central: Iowa, Kansas, Minnesota, Missouri, Nebraska, North Dakota, South
Dakota
West South Central: Arkansas, Louisiana, Oklahoma, Texas
Mountain:
Pacific:
t Arizona, Colorado, Idaho, Montana, Nevada, New Mexico, Utah,
Wyoming
California, Oregon, Washington
As stated on page 106, States and Tribes should consider these data in combination with
other regional studies and not by themselves because of the lack of confidence due to the small
sample size.
Tables 2.3.12 through 2.3.20 include these breakouts for finfish and shellfish, again
representing as consumed intakes:
Table 2.3.12: Distribution of Finfish and Shellfish Consumption: New England
Statistic
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
Estimate (g/day)
4.94
0.00
17.58
30.11
72.04
All Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
21.90
0.00
73.56
91.33
145.86
109
-------
Table 2.3.13: Distribution of Finfish and Shellfish Consumption: Middle Atlantic
Statistic
Estimate (g/day)
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
3.79
0.00
12.13
25.29
61.72
AH Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
18.68
0.00
61.26
80.54
153.23
Source: CSFII (1989-1991)
110
-------
Table 2.3.14: Distribution of Finfish and Shellfish Consumption: South Atlantic
Statistic
Estimate (g/day)
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
4.92
0.00
16.72
30.45
77.54
AH Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
16.33
0.00
57.62
83.39
130.78
111
-------
Table 2.3.15: Distribution of Finfish and Shellfish Consumption: East North Central
Statistic
Estimate (g/day)
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
2.88
0.00
5.10
18.24
58.24
All Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
13.21
0.00
47.50
72.05
114.31
Source: CSFII (1989-1991)
112
-------
Table 2.3.16: Distribution of Finfish and Shellfish Consumption: East South Central
Statistic
Estimate (g/day)
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
10.66
0.00
37.64
58.41
165.12
All Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
16.63
0.00
52.26
69.94
165.12
Source: CSFII (1989-1991)
113
-------
Table 2.3.17: Distribution of Finfish and Shellfish Consumption: West North Central
Statistic
Estimate (g/day)
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
4.48
0.00
4.41
25.84
104.32
All Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
12.85
0.00
42.99
63.05
141.07
Source: CSFH (1989-1991)
114
-------
Table 2.3.18: Distribution of Finfish and Shellfish Consumption: West South Central
Statistic
Estimate (g/day)
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
7.04
0.00
23.85
55.46
112.68
All Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
13.06
0.00
49.69
74.77
114.22
Source: CSFII (1989-1991)
115
-------
Table 2.3.19: Distribution of Finfish and Shellfish Consumption: Mountain
Statistic
Estimate (g/day)
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
3.23
0.00
0.48
20.90
78.60
AH Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
11.20
0.00
• 39.32
58.55
95.84
Source: CSFII (1989-1991)
116
-------
Table 2.3.20: Distribution of Finfish and Shellfish Consumption: Pacific
Statistic
Estimate (g/day)
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
3.93
0.00
10.16
26.46
68.74
All Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
16.81
0.00
55.87
83.44
122.64
Source: CSFII (1989-1991)
Developmental Criteria
Table 2.3.21 presents as consumed weight distributional data for children ages 0 to 14 who
are "acute" consumers offish. Exhibit 2.3.2 graphically shows a more complete distribution of
values for these consumers. The term "acute consumer" does not refer to adverse health effects or
toxicity studies. It refers to the subset of survey responses where fish was actually consumed. That
is, the distributional data from the CSFII for these "acute consumers" was determined by using only
data from the individuals who ate fish during the survey period. "Acute consumers" were defined
as individuals who reported consumption of a fish-related food code at least once in the three-day
reporting period. In addition, if an individual consumed fish for two of the three days, the average
daily consumption for that individual was calculated by summing the two daily consumption values
and dividing by two. As noted above, these data may be most appropriate to use for evaluating
exposures for children because they generally have higher intake rates per body weight than adults.
EPA did not generate intervals around the estimates because variance estimation algorithms require
data from at least two primary sampling units (PSUs) per stratum, and this criterion was not always
met for these acute consumers (USEPA, 1998).
117
-------
Table 2.3.21: Daily Estimates of Fish Consumption: Finfish and Shellfish - Acute
Consumers,14 Children 0 to 14 Years Old in the U.S. Population
Statistic
Estimate (g/day)
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
45.73
28.35
108.36
136.24
214.62
All Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Peicentile
74.80
56.49
153.70
178.08
337.46
Source: CSFII (1989-1991)
14 Acute consumer refers to respondents who reported consuming fish during the 3-day survey period.
118
-------
e
VI
O
p
O
( 90l x). sjenpiAipuijo jgqumM
119
-------
The CSFII data indicate that for median, mean, and upper percentile intakes, rates for
children are higher per body weight than rates for adults, with differences of up to 8.6 g/kg-day at
the 99th percentile (See Appendix A).
For in-utero developmental effects, intake rates for women of childbearing age may be most
appropriate. Thus, Table 2.3.22 presents the distribution of as consumed fish intake values for
women ages 15-44 years old who are acute consumers, as described above. Exhibit2.3.3 graphically
shows a more complete distribution of values for these acute consumers.
Table 2.3.22: Daily Estimates of Fish Consumption: Acute Consumers,15 Women 15 to 44
Years Old in the U.S. Population (g/day)
Statistic
Estimate
Fresh/Estuarine
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
61.40
35.22
148.83
185.44
363.56
AH Fish (including marine)
Mean
50th Percentile
90th Percentile
95th Percentile
99th Percentile
88.80
69.95
170.01
212.56
361.04
Source: CSFII (1989-1991)
" Acute consumer refers to respondents who reported consuming fish during the 3-day survey period.
120
-------
w
I
tw
t>
"^
K— t
3
HISTOGRAM OF D,
1
C^*!
CO
cs
•O
1
w
.fc
0
E
5
CO
§
U
ffi
CO
E
o
0~t
w
P-,
e
*p^
rt
&.
O
PL,
CO
J3 H
_w
E ^
•S e
s •-
g-5
§ 2
0 «n
<* tu
^ bfi
O
&o
13
S
(U
fc
•S
S
"o
CO
§
1
rt
rs - Fresh/Estuarine Fii
CD
3
en
O
U
B
3
9OI x)
121
-------
Preference #4: Use of Default Intake Rates from the CSFII
The 1980 AWQC National Guidelines recommended a fish intake rate of 6.5 grams/day,
based on the mean consumption rate of freshwater and estuarine finfish and shellfish from 30-day
diary results reported in the 1973-74 National Purchase Diary Survey. These updated guidelines
recommend several default intake rates depending on the population and type of effect (chronic or
developmental) that is being considered.
Default Intake Rate for Chronic Effects
Although EPA prefers that States use one of the above methods to determine fish intake, this
section presents two default intake rates that EPA believes represent appropriate fish intake values
for different population groups, and are appropriate for determining intake related to contaminants
that may cause chronic effects. EPA recommends using the following intake rates (of freshwater
and estuarine finfish and shellfish) based on information for all adults from the CSFII: 17.80 g/day
for the general adult population and sport fishers, and 86.30 g/day for subsistence fishers. These
values represent the intake of freshwater/estuarine finfish and shellfish as consumed. By applying
17.80g/day as a default for the general adult population, EPA intends to select an intake rate that is
protective of a majority of the population. EPA further considers that, although these rates are
reflective of high-end consumers in the general population and do not directly reflect intakes specific
to sportfishers and subsistence fishers, they are indicative of the average consumption among sport
fishers and subsistence fishers, respectively. Specifically, comparison of the CSFII intake rates with
results from state and regional surveys indicate that these rates may be appropriate for the defined
sportfisher and subsistence fisher populations (refer to study summaries for sportfishers and
subsistence fishers under Preference #2). As noted above, however, sportfisher and subsistence
fisher populations are generalized terms and each group may encompass a variety of types of
individuals. Thus, States should try to use intake rates more specific to the population addressed
before considering the default intake rates suggested here (ensuring that the rates chosen meet the
minimum discussed on page 90).
Default Intake Rate for Developmental Effects
For a few fish contaminants, health effects in children are of primary concern (e.g.,
cholinesterase inhibitors). Because children have a higher fish consumption rate per body weight
compared to adults, using a higher fish consumption rate per body weight may be necessary for
setting AWQC to assure adequate protection for children from toxicants that cause such effects.
EPA advises that in absence of local data or other data approximating local information on values
appropriate for children's intake, States use a value of 108.36 g/day. This value represents the 90th
percentile from the combined 1989-1991 CSFII surveys for acute consumption (defined above under
Preference #3, hi the Developmental Criteria subsection) of freshwater/estuarinefinfish and shellfish
for children ages 0 to 14. The value represents only those children from the CSFII survey who ate
fish during the 3-day survey period, and the intake was averaged over the number of days during
which fish was actually consumed. It is recommended that this value be used with a body weight
122
-------
of 28 kilograms (discussed above) to protect individuals from adverse effects of eating fish when
RfDs are based on health effects in children.
Developmental effects may be of concern for children or women of childbearing age. To
provide additional protection from adverse effects when pregnant women are of particular concern,
a default intake rate of 148.83 g/day, specific to women of childbearing age (15-44 years old), is
suggested for setting AWQC to protect against such developmental effects. This value represents
approximately the 90th percentile of acute consumption of freshwater/estuarinefinfish and shellfish
for women in this age group from the CSFII survey. As with the rate for children, this value
represents only those women who consumed fish during the 3-day survey period.
2.3.2.4 Incidental Ingestion
The drinking water ingestion rate of 2 liters/day is used only for setting AWQC for those
water bodies designated as public water supply sources. Individuals exposed to water from water
bodies that are not listed as public water supply sources would not be likely to ingest 2 liters/day
from these waters. However, even if a water body is not used for public drinking water supplies, it
is possible that an individual may incidentally ingest some amount of water if he or she swims,
fishes or boats in the water body. Literature on recreational exposure combined with assumptions
about the average mouthful of water ingested for every hour of total body contact can be used to
determine an incidental ingestion rate. EPA recommends an incidental ingestion rate of 10 ml/day
based on data from studies below when developing chronic criteria. The criteria that would be
calculated using incidental ingestion would include water bodies that are designated to be used for
recreational purposes only.
Incidental ingestion can be determined by estimating the number of hours that an individual
may be in contact with water during recreation and multiplying this value by an average mouthful
of water assumed to be swallowed during each hour of recreational exposure which results in total
body contact with water. The estimate of 10 ml/day is based on an assumption that an individual
may be hi total contact with water for 123 hours a year (which represents an hour of exposure per
day throughout four summer months) and may ingest 30 ml of water per hour of total contact (State
of Michigan, 1985).16 The value is calculated by multiplying 30 ml/hour by 123 hours a year and
dividing by 365. This value has been proposed for use in the proposed Water Quality Guidance for
the Great Lakes (58 FR 20869).
Other studies of recreational exposure suggest much variation in this rate, with some similar
estimates of exposure as a result of water skiing, swimming, boating, and fishing activities. These
studies are discussed below.
16 Superfund guidance suggests that an average mouthful of water may be 50 ml [SUPERFUND Risk Assessment Guidelines (USEPA,
1989b)J.
123
-------
EPA has reported exposure durations for swimming, water skiing, boating, and fishing. EPA
recently estimated a national average frequency of swimming of 7 days/year with a 2.6 hour duration
(USEPA, 1989b). This value may be compared with an earlier EPA publication (USEPA, 1979),
which estimated an average annual frequency of 9 days/year with a 2 hour duration of exposure per
day. EPA estimated that approximately 20 million individuals participated in water skiing (which,
like swimming, involves total body contact with water) for a total of 260 million hours per year.
This averages to 14 hours of exposure per participant. USEPA (1979) also listed individuals that
participated in other water activities. Sixty-eight million people were involved nationally in boating
with an average duration of 24 hours per participant per year and 54 million people fished with 122
hours per participant per year. Total body contact with water during boating and fishing was
identified as 40 percent and 20 percent, respectively (USEPA, 1979). These values were used to
adjust the exposure duration for participants in these activities to yield exposures on a total contact
basis. The resulting total contact exposures per year for swimming, water skiing, boating and fishing
were calculated as 18,14,10, and 24 hours, respectively. Adding all exposures yields 66 hours of
total body contact exposure from recreational activities.
Several recreational surveys have been conducted in Michigan which indicate up to 105
hours of total water exposure. Estimating total exposure suggests total hours of exposure per year
that are higher than the national average. The calculation of these exposures involves assuming an
individual participates in all activities for the number of days listed in the 1981 Michigan Travel and
Recreation Survey and for the duration of hours per participation as identified in the 1976 Recreation
survey. In addition to these assumptions, exposure was adjusted by the percentage of total body
contact exposure involved in the activity. This adjustment was made assuming the same percentages
for total body contact used in USEPA (1979). The calculation of total body exposure resulting from
these activities is indicated in Table 2.3.23.
Although the default value of 10 ml/day for chronic ingestion is appropriate for situations
in which exposure occurs daily for about four months, States and Tribes in warmer climates may
wish to use higher incidental ingestion rates for chronic criteria to protect individuals who may swim
in lakes or rivers for a greater portion of the year. For example, Louisiana uses 89 ml/day to account
for exposure due to incidental ingestion when developing criteria for non-drinking water supplies.
The assumptions used by the Louisiana Department of Environmental Quality in determining the
89 ml/day are described in Louisiana DEQ (1989,1994).
In addition to chronic values for incidental ingestion, States and Tribes may wish to use an
incidental ingestion rate for evaluating contaminants that cause adverse health effects from shorter-
term exposures based on the amount of water that may be ingested in a given hour of recreational
activity.
124
-------
Table 2.3.23: Yearly Total Hours of Total Body Contact as Determined by Michigan
Recreational Surveys
Swimming
Fishing
Power Boating
Water Skiing
Sailing
Canoeing
Activity Days
per Participant
13.3
14.3
24.5 (total)
9.6
10.4 (total)
4.8
Hours per
Participation
2.1 (ave.)
3.7 (ave.)
3.2
1.5
3.2
3.9
Body Contact
Adjustment
1.0
0.2
0.4
1.0
0.4
0.4
TOTAL
Hours of
Exposure
27.9
10.6
31.4
14.4
13.3
7.5
105.1
2.3.3 Quantification of Exposure
In typical exposure assessments, the magnitude, frequency, and duration of exposure is
quantified for a given population and specifically selected exposure pathways. After selecting
exposure concentration values in each environmental medium to be addressed (e.g., water, food),
pathway-specific intake rates are subsequently selected. Given an assumption that exposure occurs
over a period of time, dividing the exposure assumption by the period of time will give an average
exposure rate per unit time. Alternatively, exposure can be estimated by normalizing both the time
and body weight factors, expressed in units of mg chemical/kg body weight-day. (This is discussed
in the Federal Register notice and comment is requested on this alternative.)
The term "intake" used with this methodology describes the daily exposure estimate
(normalized for a lifetime) and is expressed in units of mg chemical/liter. Specifically, for purposes
of establishing AWQC which are, by and large, based on chronic health effects data and are intended
to be protective of the general population over a lifetime of exposure, the criteria calculations are
made in terms of a person's daily exposure. That is, the AWQC represent an acceptable daily
exposure over a lifetime for which no adverse health effects associated with that chemical are
expected to occur. Hence, the expression of the AWQC is in mg chemical/day. The AWQC
calculation includes an assumption of body weight.
When selecting contaminant concentration values in environmental media and exposure
intake values for the Relative Source Contribution (RSC) analysis, it is important to realize that each
value selected (includingthose intakes recommended as default assumptions in the AWQC equation)
is associated with a distribution of values for that parameter. Determining how various subgroups
125
-------
fall within the distributions of overall exposure and how the combination of exposure variables
defines what population is being protected is a complicated and, perhaps, unmanageable task,
depending on the amount of information available on each exposure factor included. Many times,
the default assumptions used in EPA risk assessments are derived from the evaluation of numerous
studies and are generally considered to represent a particular population group or some national
average. Therefore, describing with certainty the exact percentile of a particular population that is
protected with a resulting criteria is often not possible.
General recommendations for selecting values to be used in exposure assessments for both
individual and population exposures are discussed in EPA's Guidelines for Exposure Assessment
(USEPA, 1992). The ultimate choice of the contaminant concentration values used in the RSC
estimate and the exposure intake rates requires the use of professional judgment. In particular, when
combining variable values for the AWQC estimate, the basis of the health effect (e.g., chronic) and
the population (e.g., general population) must be kept in mind; for example, combining a 90th
percentile intake with a 5th percentile body weight is not appropriate because it is not likely that the
smallest person would have the highest intake and it would not be appropriate with such a chronic
effect, general population scenario. Similar judgments must be made for less-than-lifetime health
effects and different target population groups. The following are general recommendations. States
and Tribes have the flexibility to consider other parameters based on site-specific information or
other risk management considerations.
Contaminant concentration. The concentration values for all media used in the RSC analysis
are arithmetic means when calculations are made for the general population. These are used to
represent reasonable central tendency estimates for a typically exposed person. Higher concentration
values may be considered when making evaluations for more highly exposed population groups
(e.g., subsistence fishers) whose patterns of exposure with fish consumption are not the same as the
general population. However, higher contaminant concentration values should not be used for all
media (e.g., other dietary or air intake assumptions) unless it is clear that the specific population
group is likely to experience higher concentrations from other media as well. For example, in the
RSC analysis, choosing a higher concentration value for estimating fish exposure with a subsistence
fisher, should not mean automatically using high concentration values for other foods (such as
vegetables, fruit, etc.) or for air exposures.
Body -weight. By and the large, the AWQC will be based on the arithmetic mean of the adult
body weight for the general population. If the health effect of concern is one that specifically occurs
in children, the arithmetic average child body weight is recommended. The same recommendation
of an arithmetic mean is made for women of childbearing age.
Dietary intake (non-fresh/estuarine fish intake) and inhalation. Values recommended for
these assumptions, which are a part of the RSC analysis, are based on the arithmetic means from the
information sources utilized. Specifically, the dietary intake assumptions are taken from the Food
and Drug Administration's Total Diet Study program (Pennington, 1983) and the inhalation rate is
based on a study conducted by the International Commission on Radiological Protection (ICRP,
126
-------
1981), which has been historically used in EPA risk assessments. These studies are representative
of the overall U.S. population.
Fresh/estuarinefish intake and drinking water. The intake rates recommended for these two
parameters are higher than the arithmetic means for the U.S. population. The choice of default
intake assumptions for these parameters represent a risk management decision under the goals of the
Clean Water Act to establish AWQC that are protective of a majority of the population through the
exposure routes of water and fish consumption. The default drinking water intake rate represents
the 84th percentile value from the study on which it is based. The default intake rate for fish
consumption of the general population represents the 90th percentile value from the study on which
it is based. However, it should be kept in mind that the study does not enable accounting for fish
consumers only and, therefore, the intake assumption likely represents less than the 90th percentile
of the population potentially at risk from this exposure route.
EPA considers the national AWQC recommendations to be protective of a majority of the
general population and believes that it has used appropriate professional judgment in recommending
these criteria. EPA encourages States and Tribes to use local or more site-specific exposure intake
and concentration assumptions that they believe would appropriately protect the overall population,
including highly exposed subgroups. The exposure assessmentprocedures used in this methodology,
which includes the RSC Exposure Decision Tree recommendation, do not prohibit the use of Monte
Carlo analysis. States and Tribes may consider using such probabilistic techniques when they have
access to data that are adequate enough to provide meaningful results from such analyses. Again,
the selection of a point off the overall distribution of exposures (which represents a combination of
other distributions) is a decision that involves professional judgment.
2.3.4 Consideration of Non-Water Sources of Exposure When Setting AWQC
In the 1980 AWQC National Guidelines, different approaches for addressing non-water
exposure pathways were used in setting AWQC for the protection of human health depending upon
the toxicological endpoint of concern. For those substances for which the appropriate toxic endpoint
was linear carcinogenicity, only the two water sources (i.e., drinking water consumption and
freshwater/estuarine fish ingestion) were considered in the derivation of the AWQC. Non-water
sources and marine fish ingestion were not considered explicitly. The rationale for this approach is
that in the case of linear carcinogens the AWQC is being determined with respect to the incremental
lifetime risk posed by a substance's presence hi water, and is not being set with regard to an
individual's total risk from all sources of exposure.
In the case of substances for which the AWQC is set on the basis of a nonlinear carcinogen
or a noncancer endpoint where a threshold is assumed to exist, non-water exposures were considered
when deriving the AWQC under the 1980 AWQC National Guidelines. In effect, the 1980 AWQC
National Guidelines specified that the AWQC be calculated based on no more than that portion of
the ADI that remains after contributions from other expected sources of exposure have been
accounted for. The ADI is equivalent to the RfD, which is discussed in Section 2.2. The rationale
127
-------
for this approach has been that for pollutants exhibiting threshold effects, the objective of the AWQC
is to ensure that an individual's total exposure does not exceed that threshold level. It is useful to
note that while the 1980 Guidelines recommended taking inhalation and nonfish dietary sources into
account in setting the AWQC for threshold contaminants, in practice the data on these other sources
were not available. Therefore, the AWQC usually were derived such that they accounted for all of
the ADI (RfD).
EPA is proposing that only a portion of the RfD or Pdp/SF be used in setting AWQC in order
to account for other sources of exposure for threshold toxicants, including both noncarcinogens and
nonlinear carcinogens, lexicological issues related to noncarcinogens and nonlinear carcinogens
are discussed in detail in Sections 2.2 and 2.1, respectively. For carcinogens that act in a linear
fashion, non-water sources would not be taken into account when setting AWQC. The rationale is
the same as that given in the 1980 Guidelines, namely, that the AWQC is being determined for the
incremental lifetime risk posed by a substance's presence in water and not for an individual's total
risk from all exposure sources.
For noncarcinogens for which non-water exposures were considered, the 1980 methodology
included the following general formula for setting the criterion:
AWQC = [70] [ADI-(DT + IN)] - [2 + 0.0065R]
(Equation 2.3.1)
where AWQC is the criterion in units of mg/L; ADI is the Acceptable Daily Intake (now Reference
Dose, RfD) in units of mg/kg-day; DT is non-freshwater and -estuarine fish dietary intake in mg/kg-
day; IN is inhalation intake in mg/kg-day; 70 is human body weight in kg; 2 is the drinking water
consumption in L/day; 0.0065 is fish ingestion in kg/day; and R is the bioconcentration factor in
L/kg. As indicated by the above equation, the 1980 AWQC National Guidelines used a "subtraction"
approach to account for non-water exposure sources when calculating AWQC for noncarcinogenic,
threshold pollutants. That is, the amount of the ADI (RfD) "available" for water sources was
determined by first subtracting out contributions from non-water sources. A similar subtraction
approach was used, albeit inconsistently, in the derivation of drinking water MCLG values in the
early and mid-1980s; more recently, however, the derivation of MCLGs has incorporated what has
been termed the "percentage" approach.
EPA has considered several alternative approaches to account for non-water sources and to
resolve past inconsistencies in its method. All approaches are discussed in detail in a separate
document available in the public docket for this proposal (Borum, unpublished). The result of
discussions on these approaches was a consensus by the Relative Source Contribution Policy
Workgroup to recommend the Decision Tree Approach for internal Agency review. This was
considered the best option of the alternatives presented. To account for exposures from other media
when setting an AWQC, the exposure decision tree for determining proposed RfD or Pdp/SF
allocations represents a method of comprehensively assessing a chemical for regulatory
128
-------
development. This method considers the adequacy of available exposure data, levels of exposure,
relevant sources/media of exposure, and regulatory agendas (i.e., multiple regulatory actions for the
same chemical). The decision tree addresses most of the disadvantages associated with the exclusive
use of either the percentage or subtraction approaches, because they are not arbitrarily chosen prior
to determining the following: specific population(s) of concern, whether these populations are
relevant to multiple-source exposures for the chemical in question (i.e., whether the population is
actually or potentially experiencing exposure from multiple sources), and whether levels of
exposure, regulatory agendas or other circumstancesmake allocation of the RfD or Pdp/SF desirable.
Both subtraction and percentage methods are potentially utilized under different circumstances with
the Decision Tree Approach, and the decision tree is recommended with the idea that there is enough
flexibility to use other procedures if information on the contaminant in question suggests it is not
appropriate to follow the decision tree (e.g., if multiple sources of exposure do not exist for the
population of concern). EPA recognizes that there may be other valid approaches in addition to the
exposure decision tree and the others identified in the Federal Register (FR) notice. EPA is
specifically recommending the Exposure Decision Tree for use with this methodology.
As stated in the FR, current internal policy discussions include the application of this
approach to all program offices to the extent practicable when conducting exposure assessments.
As such, the broader goals are to ensure more comprehensive evaluations of exposure Agency-wide
and consistent allocations of the RfD or Pdp/SF for criteria-setting purposes when appropriate.
2.3.4.1 Exposure Decision Tree Approach
Although the following discussion of the Exposure Decision Tree Approach is included in
the Federal Register notice, it is repeated here for the benefit of the reader, and for use in evaluating
the example below.
When data regarding exposure sources of a given chemical are adequate, the decision tree
is designed to allow for accurate predictions of exposure for the population(s) of concern. When
there are less data, there is an even greater need to make sure that public health protection is
achieved. A series of qualitative alternatives is proposed. Specifically, the decision tree makes use
of chemical information when actual monitoring data are inadequate. It considers information on
the chemical/physicalproperties, uses of the chemical, and environmental fate and transformation,
as well as the likelihood of occurrence in various media. Review of such information, when
available, and concurrence on a reasonable exposure characterization for the chemical would result
in a health-based criterion that is more accurate in predicting exposures than a default of 20 percent.
Although the 20 percent default is still proposed when information is not adequate, the need for
using a default is greatly reduced.
As stated above, the recommendation is made with the understanding that there may be
situations where the decision tree procedure is not practicable or may be simply irrelevant after
considering the properties, uses and sources of the chemical in question. It is important to have the
flexibility to choose other procedures that are more appropriate for setting health-based criteria or,
129
-------
perhaps, allocating the RfD or Pdp/SF, as long as reasons why the regulatory action should follow
a different course are clearly presented. Often, however, the multiple source nature of chemicals is
likely to merit a decision tree evaluation for the purpose of setting human health criteria or standards
for a given chemical. The decision to perform, or not to perform, an allocation could actually be
made at several points during the decision tree process. Working through the whole process may
be most helpful for determining why another approach should be used. While combined exposures
above the RfD (Pdp/SF) may or may not be an actual health risk, a combination of health standards
exceeding the RfD (Pdp/SF) may not be sufficiently protective. Maintaining total exposure below
the RfD (Pdp/SF) is a reasonable health goal and there are circumstances where health-based criteria
for a chemical should not exceed the RfD (Pdp/SF), either alone (if only one criterion is relevant)
or in combination. "Relevancy" here means determining whether more than one criterion, standard,
or other guidance is being planned, performed or is in existence for the chemical in question.
It is clear that this will be an interactive process; input by exposure assessors will be provided
to, and received from, risk managers throughout the process, given that there may be significant
implications regarding control issues (i.e., cost/feasibility), environmental justice issues, etc. In
cases where the decision tree is not chosen, communication and concurrence about the decision
rationale and the alternatively proposed criteria are of great importance.
Exhibit 2.3.4 presents the Exposure Decision Tree. Descriptions of the boxes within the
decision tree are separated by the following process headings to facilitate an understanding of the
major considerations involved.
Problem Formulation
Initial decision tree discussion centers around the first two boxes: identification of
populations) of concern (Box 1) and identification of relevant exposure sources and pathways (Box
2). The term "problem formulation" refers to evaluating the population(s) and sources of exposure
in the manner described above (i.e., the potential for the population of concern to experience
exposures from multiple sources for the chemical in question), such that the data for the chemical
in question consider each source/medium of exposure and its relevancy to the identified
population^). Evaluation includes determining whether the levels, multiple regulatory actions, or
other circumstances make allocation of the RfD or Pdp/SF reasonable. The initial discussion has
also included agreement on the exposure parameters chosen, intakes chosen for each route and any
environmental justice or other social issues that aid in determining the population of concern. The
term "data," as used here and discussed throughout this section, refers to ambient sampling data
(whether from Federal, regional, State or area-specific studies) and not internal human exposure
measurements.
130
-------
Exhibit 2.3.4
Alternate Exposure Decision Tree for Defining Proposed RfD (Pdp/SF) Allocation
2.
3.
4.
5.
Identify population(s) of
concern.
Identify relevant exposure
sources/pathways.
Problem
Formulation
ll.
Are adequate data available
to describe central
tendencies and high-ends
for relevant exposure
sources/pathways?
Yes
I
No
Are exposures from
multiple sources (due to a
sum of sources or an
individual source)
potentially at levels near
(i.e., over 80%), at, or in
excess of the RfD
(Pdp/SF)?
12.
Yes
Describe exposures,
uncertainties, toxicity-
related information,
control issues, and other
information for
management decision.
Are there sufficient data, physical/chemical
property information, fate and transport
information, and/or generalized information
available to characterize the likelihood of
exposure to relevant sources?
No
13.
Is there more than one regulatory action
(i.e., criteria, standard, guidance) relevant
for the chemical in question?
I
No
Can regulatory action be
postponed until better
information is developed?
No
Yes
14.
Use subtraction of appropriate
intake levels from sources other
than source of concern, including
80% ceiling/20% floor.
Are there significant known or
potential uses/sources other
than the source of concern?
Yes
15.
Yes
Is there some information
available on each source
to make a characteri-
zation of exposure?
I
No
Use 20% of the
RfD (Pdp/SF).
IOC.
Yes
Use allocation of the
RfD (Pdp/SF), including
80% ceiling/20% floor.
Option 1: Use
percentage approach
(with ceiling and floor).
Option 2: Subtract
exposure levels from all
sources from the RfD
(Pdp/SF) and apportion
the free space.
Perform allocation as described in Box
14 or 15, with 50% ceiling/20% floor.
131
-------
Data Adequacy
In Box 3, it is necessary that adequate data exist for the relevant sources/pathways of
exposure if one is to avoid using default procedures. In fact, distributional data may exist for some
or most of the sources of exposure. At a minimum, the central tendency and high-end values are
considered necessary to determine an appropriate estimate of exposure when using actual data. It
is critically important to describe and provide guidance for the data adequacy issue, or the approach
could be considered arbitrary.
There are numerous factors to consider in order to determine whether a dataset is adequate.
These include: (1) sample size (i.e., the number of data points); (2) whether the dataset is a random
sample representative of the target population (if not, estimates drawn from it may be biased no
matter how large the sample); (3) the magnitude of the error that can be tolerated in the estimate
(estimator precision); (4) the sample size needed to achieve a given precision for a given parameter
(e.g., a larger sample is needed to precisely estimate an upper percentile than a mean or median); (5)
an acceptable analytical method detection limit; and (6) the functional form and variability of the
underlying distribution, which determines the estimator precision (e.g., whether the distribution is
normal or lognormal and whether the standard deviation is 1 or 10). Lack of information may
prevent assessment of each of these factors; monitoring study reports often fail to include
background information or enough summary statistics (and rarely the raw data) to completely
characterize data adequacy. Thus, a case-by-case determination of data adequacy is likely.
That being stated, there are some criteria, as proposed below, that lead to a rough rule-of-
thumb on what constitutes an "adequate" sample size for exposure assessment. The primary
objective is to estimate an upper percentile point (e.g., say the 90th) and a central tendency value of
some exposure distribution based on a random sample from the distribution. Assuming that the
distribution of exposure is unknown, a nonparametric estimate of the 90th percentile is required. The
required estimate, based on a random sample of n observations from a target population, is obtained
by ranking the data from smallest to largest and selecting the observation whose rank is 1 greater
than the largest integer in the product of .9 times n. For example, in a data set of 25 points, the
nonparametric estimate of the 90th percentile is the 23rd largest observation.
In addition to this point estimate, it is useful to have an upper confidence bound on the 90th
percentile. To find the rank of the order statistic that gives an upper 95 percent confidence limit on
the 90th percentile, the smallest value of r that satisfies the following formula is determined:
r-l
0.95 =
i=o
(Equation 2.3.2)
132
-------
For relatively small data sets, the above formula will lead to selecting the largest observation
as the upper confidence limit on the 90th percentile. However, the problem with using the maximum
is that, in many environmental datasets, the largest observation is an outlier and would provide an
unrealistic upper bound on the 90th percentile. It would, therefore, be preferable if the sample size
n were large enough so that the formula yielded the second largest observation as the confidence
limit.
This motivates establishing the following criterion for setting an "adequate" sample size:
pick the smallest n such that the nonparametric upper 95 percent confidence limit on the 90th
percentile is the second largest value. Application of the above formula with r set to n-l yields n =
45 for this minimum sample size.
For the upper 95 percent confidence limit to be a useful indicator of a maximum exposure
it must not be overly conservative (too large relative to the 90th percentile). It is, therefore, of
interest to estimate the expected magnitude of the ratio of the upper 95 percent confidence limit to
the 90th percentile. This quantity generally cannot be computed, since it is a function of the
unknown distribution. However, to get a rough idea of its value, consider the particular case of a
normal distribution. If the coefficient of variation is between 0.5 and 2.0 (i.e., the standard deviation
divided by the mean) the expected value of the ratio in samples of 45 will be approximately 1.17 to
1.31; i.e., the upper 95 percent confidence limit will be only about 17 to 31 percent greater than the
90th percentile on the average.
It should be noted that the nonparametric estimate of the 95 percent upper confidence limit
based on the second largest value can be obtained even if the data set has only two detects (it is
assumed that the two detects are greater than the detection limit associated with all non-detects).
This is an argument for using nonparametric rather than parametric estimation, since use of
parametric methods would require more detected values. On the other hand, if non-detects were not
a problem and the underlying distribution were known, a parametric estimate of the 90th percentile
would generally be more precise.
As stated above, adequacy is also determined by determining whether the samples are
relevant to and representative of the population at risk. Data may, therefore, be adequate for some
decisions and inadequate for others; this determination requires some professional judgment.
If the answer to Box 3 is no, then the decision tree falls into Box 4. As suggested by the
separate boxes, the available data that will be reviewed as part of Box 4 do not meet the requirements
necessary for Box 3. In Box 4, any data that are available (information about the chemical/physical
properties, uses, and environmental fate and transformation, as well as any other information that
would characterize the likelihood of exposure from various media for the chemical) are evaluated
to make a qualitative determination of the relation of one exposure source to another. Although this
information will always be presented at the outset, it is proposed that this information also be used
to estimate the health-based criteria. The estimate should be rather conservative, given that it is not
based on actual monitoring data (or data that has been considered to be inadequate for a more
133
-------
accurate quantitative estimate). Therefore, there are greater uncertainties, and accounting for
variability is not really possible. With such information, a determination can be made as to whether
there are significant known or potential uses/sources other than the source of concern (Box 8). If
there are not, then it is recommended that 50 percent of the RfD or Pdp/SF can be safely allocated
to the source of concern (Box 9). While this leaves half of the RfD or Pdp/SF unallocated, it is
recommended as the maximum allocation due to the lack of data needed to more accurately quantify
actual or potential exposures. If the answer to the question in Box 8 is yes, and some information
is available on each source of exposure (Box 10 A), apply the procedure in either Box 14 or Box 15
(depending on whether one or more criterion is relevant to the chemical), using a 50 percent ceiling
(Box IOC), again due to the lack of adequate data. If the answer to the question in Box 10A is no,
then use 20 percent of the RfD or Pdp/SF (Box 10B).
If the answer to the question in Box 4 is no; that is, there are not sufficient data/information
to characterize exposure, it may be best to defer action on the chemical until better information
becomes available (Boxes 5 & 6). If this is not possible, then the "default" assumption of 20 percent
of the RfD or Pdp/SF (Box 7) should be used. Box 7 is not likely to be used very much, given that
the information described in Box 4 should be available in most cases. However, EPA intends to use
it as the default value that has also been used in past water program regulations.
Regulatory Actions
If there are adequate data available to describe the central tendencies and high ends from each
exposure source/pathway, then the levels of exposure relative to the RfD or Pdp/SF are compared
(Box 11). If the levels of exposure for the chemical in question are not near (currently defined as
greater than 80 percent), at, or in excess of the RfD or Pdp/SF, then a subsequent determination is
made (Box 13) as to whether there is more than one regulatory action relevant for the given chemical
(i.e., more than one criteria, standard or other guidance being planned, performed or in existence for
the chemical). The subtraction method is considered acceptable when only one criterion is relevant
for a particular chemical. In these cases, other sources of exposure can be considered "background"
and can be subtracted from the RfD (Pdp/SF). When more than one criterion is relevant to a
particular chemical, apportioning the RfD (Pdp/SF) via the percentage method is considered
appropriate to ensure that the combination of criteria, and thus the potential for resulting exposures,
do not exceed the RfD (Pdp/SF).
Allocation Decisions
If the answer to this question (Box 13) is no, then the recommended method for setting a
health-based criterion is to utilize a subtraction calculation (Box 14). Specifically, subtract out
appropriate intake values for each exposure source other than the source of concern, based on the
variability in occurrence levels for that source. This aspect implies that a case-by-case determination
of the variability and the resulting intake chosen will be made, as each chemical evaluated can be
expected to have different variabilities associated with each source of intake. As a default, high-end
intakes (approximating the 90th to 98th percentiles of exposure) could be subtracted out. However,
134
-------
there is concern that an estimate adding 98th percentile values for all sources could be above any
actually exposed population or individual. Therefore, scientific judgment is needed in selecting
intake values, including the appropriateness for the population of concern. The subtraction method
would also include an 80 percent ceiling and a 20 percent floor.
If the answer to the question in Box 13 is yes, then the recommended method for setting
health-based criteria is to allocate the RfD or Pdp/SF among those sources for which health-based
criteria are being set (Box 15). Two main options for allocating the RfD or Pdp/SF are presented
in this box. Option 1 is the percentage approach (with a ceiling and floor). This option simply refers
to the percentage of overall exposure contributed by an individual exposure source. For example,
if for a particular chemical, drinking water were to represent half of total exposure and diet were to
represent the other half, then the drinking water contribution (known as the "relative source
contribution" or RSC) would be 50 percent. The health-based criteria would, in turn, be set at 50
percent of the RfD or Pdp/SF. This option also utilizes an appropriate combination of intake values
for each exposure source based on the variability in occurrence levels of each source. This will also
be determined on a case-by-case basis. Option 2 would involve the subtraction of exposure levels
from all sources of exposure from the RfD or Pdp/SF and apportioning the free space among those
sources for which health-based criteria are being set. There are several ways to do this: 1) divide
the free space among the sources with preference given to the source likely to need the most increase
(e.g., because of intentional uses or because of physical/chemicalproperties like solubility in water);
2) divide the free space in proportion to the "base" amount used (e.g., the source accounting for 60
percent of exposure gets 60 percent of the free space - this is identical to the percentage method; the
outcome is the same); and 3) divide the free space based on current variability of exposure from each
source (i.e., such that more free space is allocated to the source that varies the most). The resulting
criterion would then be equal to the amount of free space allocated plus the amount subtracted for
that source. Note: The allocation options continue to be discussed within EPA as part of an Agency-
wide Pilot Study group. Although some preferences have been discussed, along with strengths and
shortcomings of each option, it is still being deliberated. The Agency -welcomes comments on these
options.
Finally, if the answer in Box 11 is yes, that is, if the levels of exposure for the chemical in
question are near (currently defined as greater than 80 percent), at, or in excess of the RfD or
Pdp/SF, then the estimates of exposures and related uncertainties, potential allocations, toxicity-
related information, control issues, and other information are to be presented to managers for a
decision (Box 12). The high levels referred to in Box 11 may be due to one source contributing that
high level (while other sources contribute relatively little) or due to more than one source
contributing levels that, in combination, approach or exceed the RfD or Pdp/SF. This presentation
may inevitably be necessary due to the control issues (i.e., cost and feasibility concerns) that may
be involved, especially when multiple criteria are at issue. In practice, risk managers are routinely
a part of any decisions regarding regulatory actions and will be involved with any recommended
outcome of the exposure decision tree or, for that matter, any alternative to the exposure decision
tree. However, because exposures that approach or exceed the RfD or Pdp/SF and the feasibility of
135
-------
controlling different sources of exposure are complicated issues, risk managers will need to be
directly involved in formulating any allocation decisions.
Just as with the other outcomes in the exposure decision tree, a recommendation for setting
a health-based value (or values, depending on the number of relevant sources) for chemicals that
apply to Box 12 is also appropriate. It is likely that risk managers will want some input from the
exposure assessors even if exposures are above the RfD or Pdp/SF and control issues apply.
Therefore, in these cases, recommendations can still be offered and should be performed as with
Boxes 13,14, and 15. The recommendation should be made based on health-based considerations
only, just as when the chemical in question was not a Box 12 situation. If the chemical is relevant
to one regulatory action only, the other sources of exposure could be subtracted from the RfD or
Pdp/SF to determine if there is any leftover amount for setting a criterion. If the chemical is a
multiple criteria issue, then a recommended allocation could be made, even though it is possible that
all sources would need to be reduced. Regardless of the outcome of Box 11, all allocations made
(via the methods of Boxes 14 or 15) should include a presentation of the uncertainty in the estimate
and in the RfD or Pdp/SF for a more complete characterization.
The process for a Box 12 situation, versus a situation that is not, differs in that the
presentations for Boxes 14 and 15 are based on a concurrence of allocations (following the review
of available information and a determination of appropriate exposure parameters) in the absence of
control issues that would result in more selective reductions. With Box 12, one or several criteria
possibilities ("scenarios") may be presented for comparison along with implications of the effects
of various control options. It would be most appropriate to present the information in this manner
to risk managers given the complexity of these additional issues, rather than the more definitive
proposals that are not associated with Box 12 situations.
Results of both Boxes 14 and 15 rely on the 80 percent ceiling and 20 percent floor. The 80
percent ceiling was implemented to ensure that the health-based goal will be low enough to provide
adequate protection for individuals whose total exposure to a contaminant is, due to any of the
exposure sources, higher than currently indicated by the available data. This also increases the
margin of-safety to account for possible unknown sources of exposure. The 20 percent floor has
been traditionally rationalized to prevent a situation where small fractional exposures are being
controlled. That is, below a point it is more appropriate to reduce other sources of exposure, rather
than promulgating standards for de minimus reductions in overall exposure. The idea of adding
flexibility with the floor to go lower (perhaps to zero) if necessary in cases where total exposure
exceeds the RfD or Pdp/SF and additional reductions are warranted has also been discussed. The
Agency welcomes comments on this issue.
2.3.4.2 Notes on Use of the Exposure Decision Tree Approach for Setting
AWQC
Because two different types of AWQC are proposed (based on either (1) fish ingestion only
or (2) both fish and water ingestion), special circumstances arise under the decision tree approach
136
-------
when accounting for the drinking water portion of exposure. These circumstances relate to whether
it is a type (1) or (2), and whether one or more health-based criterion is being considered. These four
instances are described below.
When a criterion is being set based on fish ingestion only, and when only one health-based
criterion (i.e., AWQC) is relevant for the chemical, ingestion from drinking water would be
considered a non-ambient water source and would be subtracted from the RfD (or subtracted out
with nonlinear carcinogens; i.e., the Pdp/SF) in the numerator of the equation to determine the
AWQC, as follows:
AWOC-
tRfD ~
BW
FI • BAF
(Equation 2.3.3)
where:
RfD
DW
IN
DT
BW
FI
BAF
Reference dose (mg/kg-day)
Contaminant intake from drinking water (mg/kg-day)
Contaminant intake from air (mg/kg-day)
Contaminant intake from non-fresh/estuarine fish and other dietary intake
(i.e., all other dietary sources) (mg/kg-day)
Body weight (kg)
Fish consumption rate (kg)
Bioaccumulation factor (L/kg)
The terms DW, DT, and IN represent the relative source contribution (RSC) and are indicated here
as separate parameters to facilitate understanding of other common sources that could be subtracted
out (when only one health-based criterion is relevant). In this case, the occurrence of the
contaminant in treated drinking water would be the most relevant concentration data for determining
intake from drinking water, because it is assumed that individuals get their drinking water from the
tap.
When a criterion is being set based on fish ingestion only, and more than one health-based
criterion is being set, then an appropriate RSC allocation procedure, using either Option 1 or Option
2 in Box 15 (Exhibit 2.3.4) would be performed. This calculation is expressed by the following
equation:
AWQC =
RfD • RSC: • BW
fish
FI • BAF
(Equation 2.3.4)
137
-------
where:
RSC
fish
= Relative source contribution for fish as determined by Option 1 or
Option 2 in Box 15 (of Exhibit 2.3.4) and including only the portion
of the intake ascribed to contaminated fish intake
All other parameters are the same as above.
As noted in the definition of RSCfish, only the amount of contaminant intake from eating
contaminated freshwater and estuarine finfish and shellfish would be included in the RSC allocation.
Marine fish intake would normally be accounted for as part of the dietary intake component of the
RSC calculation. Again, intake from treated drinking water would be considered separately as a
non-ambient water source.
If a criterion is being set based on fish and water ingestion, and only one health standard is
being set, then the following equation applies:
AWQC =
[RfD - (DT+IN)] • BW
DI + (FI • BAF)
(Equation 2.3.5)
where:
DI = Drinking water consumption rate (in L/day)
All other parameters are the same as those in Equation 2.3.3.
In this case, drinking water consumptionis not considered in the non-water sources of intake because
the criterion is being set for both fish and water ingestion. Thus, only air and dietary intake are
subtracted from the RfD or Pdp/SF (here, the parameters DT and IN represent the RSC to be
subtracted out).
Finally, in a situation where a criterion is being set for both fish and water ingestion, and
more than one health-based criterion is to be set, then the following equation is applicable:
AWQC =
RfD • RSC • BW
fish + water
DI + (FI • BAF)
(Equation 2.3.6)
138
-------
where:
RSC
fish+water
The relative source contribution for fish and water as determined by
Option 1 or Option 2 in Box 15 (of Exhibit 2.3.4) and including the
portion of the intake for contaminated fish and water intake
DI = Drinking water consumption rate (in L/day)
All other parameters are the same as Equation 2.3.3.
In this case, the concentration of a chemical in ambient water is the relevant exposure source to
include in the RSCfish+water, because use of this criterion assumes that an individual may ingest such
concentrations of water daily.
Guidance has been provided on the type of studies that should be considered for estimating
fish consumption (Preferences #1 through #4) and numerous studies have been summarized.
Recommended values have also been presented for drinking water intakes and body weights.
However, these are just some of the parameters that will be needed in order to perform estimates of
overall exposure to a chemical. While it is not the intention of this document to provide an
exhaustive list of sources of information, Table 2.3.1 does provide suggestions for sources of
information on exposure intake parameters and contaminant data.
Although the consumption of marine species offish is not a direct component of an ambient
water quality criterion, there is certainly a reason to account for ingestion exposures to marine
species, as they may significantly contribute to total human exposure. That is, although the AWQC
derivation may be set for both fish and water ingestion, it is set to protect humans from exposure to
the contaminant in fresh and estuarine species only. Therefore, to protect humans who additionally
consume marine species offish, the marine portion should be considered as part of the "other sources
of exposure" when calculating an RSC value. Specifically, the DT parameter should account for all
non-fresh/estuarine fish dietary intake, thus allowing the common consumption of marine species
to be accounted for as well as all other ingested foods. Regarding the dietary information available
from the Food and Drug Administration's (FDA) Total Diet Study Program (as cited in Table 2.3.1),
EPA believes that the FDA estimates are acceptable to account for exposure to the major marine fish
species in the.typical U.S. diet (e.g., tuna, cod, haddock). However, States may utilize more
comprehensive marine species estimates (e.g., using marine species intake estimates from the CSFII
survey and marine fish contaminant concentration data) provided they ensure that marine fish intake
is not double-counted with the other dietary intake estimate used (e.g., the FDA program).
In all four of the equations above (2.3.3 through 2.3.6), the proposed 80 percent ceiling and
20 percent floor apply. However, if exposures approach or exceed the RfD or Pdp/SF, then
additional risk management decisions will be necessary regarding which exposure sources may most
practically be further reduced (given control and feasibility limitations) beyond the decision tree
approach to RSC.
139
-------
2.3.4.3 Setting AWQC for Chemical X Using the Decision Tree Approach
This example describes the application of the Exposure Decision Tree Approach (described
above and outlined in Exhibit 2.3.4) to account for sources of exposure to a generic Chemical X
when setting AWQC. Two different criteria are evaluated: criteria that include fish intake only (and
are applicable to recreational waters) and criteria that include both fish and drinking water intake
(and are applicable to waters designated as public water supplies). As noted above, different
exposure sources are used, depending on whether criteria are based on assumptions about
consumption of both fish and water or offish only. In the case of estimating a fish-only criterion,
an incidental ingestion rate of water of 0.01 L/day from recreational activities is assumed and
effectively replaces the DI intake assumption of 2 L/day.
The following sections describe the processes of accounting for sources of exposure and the
data needed to apply these processes to the Exposure Decision Tree. Specifically, the sections
describe the sources and uses of the chemical, the population of concern, the data available on
contamination in exposure media and uptake from that media, the adequacy of exposure information,
and the derivation of the AWQC using the decision tree approach. As stated previously, the
underlying objective is to maintain total exposure below the RfD (Pdp/SF) by accounting for other
sources of exposure and, therefore, using only a portion of either the RfD or Pdp/SF in setting
AWQC.
Sources and Uses of Chemical X
There are no known natural sources of Chemical X in the environment. This chemical has
been extensively used as a solvent in many industrial processes and has some uses as a pesticide.
Current releases of Chemical X to the environment may occur from these numerous processes and
from its pesticide use, and possibly from poorly maintained hazardous waste sites, illegal dumping,
and disposal of Chemical X in municipal landfills rather than hazardous waste landfills. In addition,
Chemical X may remain in the environment from past releases. Small amounts may be found in
outdoor and indoor air, soil surfaces, and surface water. Chemical X in surface waters and sediments
bioaccumulates in fish (the determined bioaccumulation factor for fish is 120,000).
Population of Concern
The first step in determining how to set AWQC for Chemical X when considering exposure
contributions from all environmental media is to define the population of concern for the chemical
(see Box 1 in Exhibit 2.3.4). The population of concern may be a group that is either more
toxicologically sensitive or more highly exposed compared with the general population. For
Chemical X, a particular population of concern is subsistence fishers who eat large quantities of self-
caught fish. These fishers may eat fish for a large portion of the year, and may include such groups
as Native Americans, immigrants who rely on fishing (particularly Asian-Americans), and poor
populations (USEPA, 1994b). These individuals, who are highly exposed to self-caught fish, may
have exposures that are much higher than exposures to the general population. In addition to
140
-------
subsistence fishers, other individuals with higher than average exposures are those who engage in
recreational fishing (i.e., sport-caught fish) and eat their catch. For this example, exposures are
evaluated for more highly-exposed fish consumers within the population who may represent
subsistence fishers. Subsistence fishers and sport fishers are compared with exposures of persons
from the general U.S. population. However, use of the default assumptions discussed in Section
2.3.2.3 result in the same estimate for sport fishers and the general population.
Data Used to Assess Exposure to Chemical X
This section discusses data available for the relevant exposure sources and pathways for
Chemical X (Box 2 of Exhibit 2.3.4). Exposure may occur from several environmental media,
including ambient surface water, drinking water, commercial food products, and air. Human
exposures are estimated by combining information on concentrations of Chemical X in
environmental media with intake rates of these environmental media. The largest exposure for
subsistence fishers, sport-fishers, and for the general population appears to be from ingestion of fish.
Exposures from Raw Surface Water
When setting AWQC for protection against intake of pollutants from both fish and water,
ambient concentrations in surface waters (that have not been treated for drinking water) are used to
assess exposure resulting from drinking the water directly from these sources. In addition to the
need for assessing exposures from drinking water from surface water sources, available
concentrations in fish may be used to assess intake from eating contaminated fish.
Exposure Resulting from Drinking Water Directly from Water Bodies. Information on
concentrations in waters as well as intake rates of water are needed to assess exposure from surface
waters. Several surveys have measured concentrations of Chemical X in ambient surface waters.
A majority of these surveys have been conducted in U.S. lakes. In a study of chemical
concentrations in surface water in one lake, average chemical concentrations in 1988, 1990, and
1992 were 0.33,0.32, and 0.18 ng/L. These concentrations are based on both dissolved phase and
paniculate phase concentrations. These authors also show that, from 1980 to 1992, the total
concentrations in the water column decreased with a first order rate constant of 0.20/yr. These
authors note that the lake is relatively unimpacted by point sources of Chemical X and receives most
of its loadings from the atmosphere.
Water samples were collected from another lake from June to October, 1989. In three
periods throughout this time, samples were collected at four or five sites. Taking the arithmetic
mean of these dissolved phase concentrations results in a value of 2.8 ng/L, and a 95th percentile
value of ± 7.2 ng/L. Although the dissolved concentrations were reported at each site, the authors
also give some composite information on Chemical X concentrations in the water. The average of
total Chemical X for sites 18 and 21 was 1.7 ±0.6 ng/L, the average for site 14 was 5.5 ±2.4 ng/L,
and the average for sites 4 and 10 was 15.6 ±11.2 ng/L. Two sites were close to a heavily
industrialized river which is an important source of Chemical X to the lake.
141
-------
Other studies have been conducted in earlier years in several proximally located lakes. One
collected samples in 1980 and reported the average concentration in the first lake to be 1.8 ng/L, with
concentrations of 3.2 ng/L in near shore samples and 1.2 ng/L in open lake samples. Mean
concentrationsranging from 0.63 to 3.3 ng/L were detected in another study of the second lake for
the years 1978 to 1983. Another study reported a mean level of 0.49 ng/L in water columns of a
third lake in 1981. From 1977 to 1981, 373 river samples from 9 locations near one of the lakes
were collected. The overall mean concentration was 300 ng/L, with a detection limit of 50 ng/L.
The authors did not specifically state the number of positives.
Surveys in other areas of the U.S. have also been conducted. Both surface water and
subsurface water drainage were investigated in one area of California during 1977. No samples
contained detectable levels of the chemical, and the detection limit was not reported. Chemical X
was collected from a bay in Texas in an area of suspected contamination. Concentrations ranged
from < 0.01 to 70 ng/L. The authors report an average of 3.1 ng/L but dp not describe whether or
how the average accounted for the non-detected values.
Because the first summarized U.S. lake study indicated a decrease in Chemical X
concentrationsthroughoutthe period from 1980 to 1992, the most recent studies are the most useful
to this analysis, especially where older studies were conducted in areas of suspected contamination.
The information from the studies that measured concentrations in the late 1980s and early 1990s
(from the first two summarized studies) were used to estimate average and high-end values for
Chemical X. The first study reports a value of 0.18 ng/L for 1992, based on the concentration of the
total chemical in the water column; this value is used as an average value for this analysis. As
indicated above, the data came from an uncontaminated lake. We assume that these values are
appropriate, assuming that most water bodies in the United States have not been impacted by point
sources of contamination. However, using this mean value may underestimate the true mean of
concentrations in ambient waters throughoutthe United States. For the high-end estimate, the value
of 15.6 ng/L from the second study is used. This value was not the highest value seen from this
study, because it is an average using data from two sites, but it was taken from the most
contaminated area within the lake. Although the data are reported for both dissolved and total
chemical fractions, data on the total chemical is used to match the data from the first study, which
is based on the total chemical. These estimates are only crude estimates of central and high-end
concentrations in water, because they are based only on information from these lakes.
Combining the above values with the drinking water intake of 2 liters/day yields a central
tendency intake rate of 5.1 x 10'9 mg/kg-day and a high-end value of 4.4 x 10'7 mg/kg-day.
Exposure from Eating Contaminated Fish: Concentrations in Fish. Measurements of
chemical concentration in fish in U.S. waters are available from several surveys. One national study,
begun in 1986, measured Chemical X in fish at nearly 400 sites. A majority of these locations (314)
were affected by a variety of point and nonpoint sources of pollution, 39 locations were from the
United States Geological Survey sites, and 35 areas represented background contaminant levels.
Game fish for human consumption were analyzed as fillets, and bottom feeders were analyzed as
142
-------
whole-body samples. The mean concentration of Chemical X found in samples of this study was
1.89 ug/g (wet weight). The median value determined was 0.21 ug/g and the maximum value found
was 124.0 |ig/g.
National distributions of Chemical X in fish are also available from an ongoing study. The
purpose of this study is to determine differences in concentrations of organochlorines at different
geographic locations, and to estimate changes in these concentrations over time. The most recent
information available is for 1984, in which 112 sites were sampled. These sites were selected to
represent all major river basins in the United States. Eleven sites were common to both national
studies. Composite samples from the ongoing study consisted of five fish and were collected at each
site for two bottom feeder species and one predator species; the whole bodies of these fish were
collected for analysis. The geometric mean value determined from the 1984 survey is 0.39 ng/g, and
the maximum value is 6.7 [ig/g. Earlier data from this survey, collected between 1980 and 1981,
shows a geometric mean chemical residue of 0.53 jig/g, and a maximum value of 11.3 ug/g.
Concentrations in marine fish have been measured in a nationwide shellfish study...Total
chemical concentrations in whole tissue of bivalves (mussels and oysters) collected during 1986
ranged from 0.009 to 6.8 ^g/g (dry weight).
Regional studies of chemical contamination in fish have also been conducted. In New York,
chemical concentrations in standard fillets of striped bass were measured, and were shown to decline
between 1984 and 1990. In 1990, the arithmetic mean was 1.3 fig/g measured on a wet weight basis.
In 1983, levels of 0.6 to 72 ug/g, measured on a lipid basis, were found in fish from major tributaries
and embayments of the Great Lakes. In cooked fish from one, median concentrations ranged from
0.17 to 3.0 ug/g.
Exposure from Eating Contaminated Fish: Consumption Rates of Fish. Consumption rates
of sport-caught fish vary, depending on whether the rate is determined for the general population or
for individuals who receive a large portion of their dietary intake from sport-caught fish. For the
general U.S. population, the proposed default national non-marine fish consumption rate for adults
of 17.80 grams/day has been estimated from information using three years of data from the'
nationally-based Continuing Survey of Food Intake for Individuals (CSFII) conducted by USDA.
The CSFII is conducted annually, and dietary intake data from the 48 conterminous states are
collected over 3-day survey periods (USEPA, 1998). The estimates based on CSFII used
information for both adult consumers and non-consumers of fish, and represent intake of all fish
whether store-bought or sport-caught. This survey is described in more detail under Preference #3
(page 105).
Data on national distributions of fish intake by sportfishers are not available. Although
several surveys have measured consumption of fish by sportfishers in particular areas, these studies
are limited to particular geographic regions and do not approximate a national distribution. Because
of the lack of information specific to national estimates for sportfishers, 17.80 grams/day, which
approximates the 90th percentile from the CSFII, is used here to represent the average consumption
143
-------
rate of the sportfisher population. This value is used to estimate intake for derivation of national
criteria.
Data on national distributions of intake by subsistence fishers are also not available. Some
studies that have specifically targeted subsistence fishers have been conducted in certain geographic
areas. In addition, sportfisher surveys have included information on specific subpopulations who
have high consumption rates and may subsist on fish for a large part of the year. Because of the lack
of national distributions, this example, which is conducted to represent an average intake estimate
of subsistence fishers, uses the default value of 86.30 grams/day, based on the 95th percentile from
the CSFII.17
Combining data on consumption rates with concentration data from the ongoing national
study to estimate exposure from Chemical X yields central tendency and high-end intake estimates
from consumption offish for an average individual from the general, sportfisher, and subsistence
fisher populations, as indicated in Table 2.3.24. Because the high-end values in each case use the
maximum contaminant value from the national study, this high-end intake of Chemical X represents
a value higher than the 90th percentile intake from the chemical in contaminated fish.
Consumption of sport-caught fish may replace consumption of other commercial meats and
fish. Thus, Chemical X intake resulting from consumption of commercial foods was adjusted to
account for this replacement. This adjustment is described below, under the section describing
dietary intake of commercial foods.
Exposure from Treated Drinking Water
In cases where AWQC are set based on fish intake only, drinking water intake is accounted
for as a separate exposure. In these instances, information on treated drinking water, if available,
is the relevant information to use when accounting for other sources of exposure. National and
regional studies have measured Chemical X contamination in both ground-water and surface water
sources of drinking water. Information from these studies can be combined with information on
intake rates of water to estimate total intake of Chemical X from this source.
In a regional study of contamination of drinking water, Chemical X was measured from the
mid-1970s to early 1985 in tap water, raw water, and finished water. Chemical X concentrations
were either not detected, or they were found at levels close to the limit of detection.
17 For simplicity, the example uses the 86.30 g/day subsistence fisher assumption only. The alternative default subsistence fisher intake
value of 39.04 g/day would result in lower estimates of Chemical X intake from fresh/estuarine fish and lower (less stringent) AWQC. These
differences are footnoted in Tables 2.3.24 and 2.3.28.
144
-------
Table 2.3.24: Chemical X Intakes from Eating Fresh/Estuarine Fish for Three Types of
Individuals
General Population
General Sportfishers
Subsistence Fishers18
Central Tendency Estimate
(mg/kg-day)
9.9 xlO'5
9.9 xlO'5
4.8 x 10-4
High-End Estimate
(mg/kg-day)
1.7 xlO'3
1.7 xlO'3
8.3 x lO'3
A discussion of the adequacy of these data for determining exposure estimates follows this
section. Although data on surface water are limited, the detection limit from the first subset of the
first national study summarized is used as a crude estimate of average human exposure. A high-end
estimate may be about 1.4 ug/L, the highest value seen in this study. Concentrations may have
decreased since these data were collected. Because of this possible decrease and because 1.4 ug/L
was the highest value seen, the value may represent a value higher than the 98th percentile, if such
concentrations are still seen.
Ground Water. National studies of Chemical X in ground water showed either no detectable
levels of the chemical, or very few positive values. In the same three-sampling national study, 18
ground-water supplies sampled in the first subset found no positive Chemical X concentrations.
Only one finished ground-water sample out of 18 in each of the other samplings contained detectable
chemical levels, at 0.1 ug/L. The second national study found no detectable chemical levels in
ground water.
Two state studies found detectable chemical concentrations in ground water. One study
measured Chemical X concentrations in 163 wells across the state, including public and private
drinking water wells. Many of the wells sampled were from highly-populated, industrialized areas.
Chemical X concentrations were found in 32 wells, and ranged from 0.06 to 1.27 ug/L. The other
study, a pesticide hazard assessment survey, was conducted from 1983 to 1984. They found
Chemical X in 2 out of 143 samples from 10 counties. The two detectable concentrations were 0.269
and 2.3 |ig/L, and the detection limit was 0.25 ug/L.
Other regional ground-water studies either found no detectable levels of Chemical X, or did
not report the Chemical X concentrations in the detected samples. In one, drinking water wells in
12 towns were sampled for Chemical X in 1984 and 1985. With a detection limit of 3.3 ug/L, no
concentrations of the chemical were found in 42 well locations. In another, a survey of ground-water
supplies found positive chemical concentrations at less than 8 percent of the 96 locations sampled.
18 If the alternate default intake assumption of 39.04 g/day was used instead, the subsistence fisher central tendency and high-end estimates
would be approximately 2.2 x 10"4 mg/kg-day and 3.7 x 10'3 mg/kg-day, respectively.
145
-------
Data on positive samples from the three-sampling national study are very limited.
Information from the first state ground-water study was used to estimate central tendency and high-
end values for Chemical X in ground water. By assuming that the detected concentration values are
equally distributed between 0.06 to 1.27 ug/L and that the non-detected samples had concentrations
of 0.03 ug/L, an average of 0.16 ug/L was determined from this study. To provide a crude estimate
of high exposure, the high value of 1.27 ug/L from the same study may be used. Because these data
come from only one state, the values are not representative of national distributions of Chemical X
in ground-water sources. Thus, these data do not adequately represent a national estimate of risk.
Estimating Exposure from Surface Water and Ground-Water Concentrations. To estimate
exposures from drinking water sources, these chemical concentrations in surface water and ground
water sources are averaged to determine estimates of exposure to Chemical X in drinking water.
To do this, we used the percent of the U.S. population served by systems using surface water and
ground water and determined a weighted average concentration in drinking water by using the
following equation:
DWConc = (0.67) * (SWConc) + (0.33) * (GWConc)
(Equation 2.3.27)
where:
DWConc
0.67
SWConc
0.33
GWConc
Average or high-end drinking water concentration from both surface
water and ground water
The fraction of the U.S. population served by public surface water
supplies
Average or high-end Chemical X concentration in surface waters
The fraction of the U.S. population served by public ground water
supplies
Average or high-end Chemical X concentration in ground waters
The weighted average value determined above was multiplied by an estimate of daily drinking water
intake of 2 liters/day and divided by adult body weight of 70 kilograms to estimate exposure in units
of mg/kg-day. The resulting estimates of intake from drinking water are 3.4 x 10"6 mg/kg-day as a
central tendency value, and 3.9 x 10~5 mg/kg-day as a high-end estimate.
It should be noted that these estimates are larger than the estimates from ambient surface
water sources (by about three orders of magnitude for the central tendency estimate and two orders
146
-------
of magnitude for high-end estimates). This may be because the drinking water samples were
collected a significant number of years prior to the collection of the raw surface water samples. In
addition, the lack of data for both raw surface water and for treated drinking water may also be a
factor in the differences between the intake estimates.
Dietary Intake from Commercial Foods
Estimates of dietary intake of Chemical X from commercial foods combine measurements
of Chemical X concentrations in store-bought foods and daily intake rates of various food items.
Data available on exposure estimates and concentrations in foods are described below.
Two sources of information on dietary intakes of Chemical X from commercial foods and
concentrations of Chemical X in commercial food are available. The estimates of intake used in this
analysis and presented in Table 2.3.25 use the second source of information described here. The first
source of information is an estimate by FDA of the adult dietary intakes, which was determined by
combining Chemical X concentrations detected in 12 Total Diet Studies (TDS) conducted over the
time period between April 1982 and April 1985 with intake rates of different food items. The TDS
measure concentrations of various contaminants in 234 foods purchased from supermarkets or
grocery stores throughout the United States and are collected four or five times a year. Using these
12 TDS samplings, the FDA determined mean daily intakes ranging from 0.038 to 0.054 ng/day for
males and 0.026 to 0.040 ug/day for females. In addition to the FDA estimates of exposure using
information from the 12 TDS samplings, the second source of information includes food
concentrations of Chemical X available from 44 TDS samplings conducted from April, 1982 to
November 1993, including the 12 described above. From the sampling conducted over this full
range of years (1982 to 1993), Chemical X concentrations have been found in 30 of the 234 food
items sampled. For each of the 30 food items, Chemical X was detected in one to three of the 44
TDS sampling collections. However, some of these reported values are trace amounts of Chemical
X, which represent the best estimates of those who analyzed the data, but are below quantifiable
limits. Thus, only eleven samples are above quantifiable limits. These concentrations can be
combined with information on age- and sex-specific intake rates of different food items found in
Pennington (1983) to determine overall exposure.
For this analysis, as noted above, concentration data from the full range of years (1982 to
1993) was used to estimate exposures. For each food item, the detected Chemical X levels were
averaged with the samples in which Chemical X was not detected to estimate an average Chemical
X concentration for a given food item. The nondetected levels were given a value of zero. To
estimate a high-end Chemical X concentration, the highest values seen from each food item were
used.
These concentration data were then combined with information on age- and sex-specific
dietary intakes of different food items to estimate total adult intake of Chemical X from commercial
foods. The daily intake of individual food items was taken from Pennington (1983), which reports
daily intake rates of individual food items for four population groups (males aged 25-30 and 60-65
147
-------
years old and females aged 25-30 and 60-65 years old). To represent the full range of adult ages,
we assumed that the dietary intake of 25-30 year-olds represents dietary habits of individuals aged
18 to 54 years old, and that the consumption rate of 60-65 year-olds represents the consumption rate
for adults aged 55 years and older. The percent of these wider age ranges (18 to 54 years; 55 years
and older) in the United States population and information that half of each age group consists of
males and half consists of females (US DOC, 1992) were then used to determine an age- and sex-
weighted overall average consumption rate for each food item.
Table 2.3.25: Intake of Chemical X from Commercial Food Items by Three Types of
Individuals
General Population
General Sportflshers
Subsistence Fishers
Central Tendency
Estimate (mg/kg-day)
1.13x10*
1.13xlO-6
1.06 xlO"6
High-End Estimate
(mg/kg-day)
4.10 xlO'5
4.1 Ox lO'5
3.86 xlO'5
Source: Based on FDA data.
To estimate total exposure from food, the Chemical X concentrations in the food items
(mg/g) were then multiplied by the consumption rate of each food item (g/day) and then divided by
70 kilograms to determine intake of Chemical X expressed in mg/kg-day. These values for each
food item were then summed, resulting in total mean and high-end intake of Chemical X from the
diet. These total intakes were then further adjusted for average individuals from the general
population, sportfishers, and subsistence fishers to exclude the amount of a chemical that is ingested
through freshwater/estuarinefish intake. For the general population, it was assumed that the amount
of freshwater/estuarine fish intake (17.8 g) that was introduced earlier may replace 17.8 grams of
commercial meat consumed in the diet. Thus, the Chemical X intake from commercial meat was
adjusted by the ratio of freshwater/estuarine fish consumption to total consumption of commercial
meat (17.8g/205g= 0.087). In other words, the Chemical X contribution from commercial meat was
decreased by 8.7 percent for the average individual. Similar adjustments were made for sportfishers
and for subsistence fishers. For sportfishers, the amount of Chemical X intake from commercial
meat was also decreased by 8.7 percent, and for subsistence fishers, the amount was decreased by
19 percent. These assumptions were made with the idea that the "typical" dietary composition based
on the FDA analysis should be adjusted to reflect a fish consumer's diet. It is included in this
example as a reasonable adjustment for exposure assessors to consider. EPA acknowledges that,
with some fish consumer groups, a much more significant adjustment may be more appropriate and
States and Tribes are encouraged to consider the dietary choices of their target population, if
information is available. The resulting intake rates of Chemical X from commercial foods are
included in Table 2.3.25. These intake estimates assume concentrations of Chemical X only in the
food items described above, which equate to approximately six percent of the diet (i.e., assumed
contamination of 161.50 grams of food). These assumptions are presented in Appendix G, which
148
-------
lists the individual food items and Chemical X concentrations. The assumed total daily intake of
foods discussed above, and based on Pennington (1983), is 2,582 g/day.
Intake from Air
Outdoor Air. National data on Chemical X concentration distributions in air are not
available. Monitoring data from EPA, however, are available for several states. From six sites in
one state, the average value is determined to be 2.14 ppb by volume, with a maximum value of 3.9
ppb. Converting from values in ppb to ug/m3 results in an average value of 0.028 ug/m3, and a
maximum of 0.05 ug/m3. Additional studies have reported ambient concentrations in several regions
of the United States. One report summarized ambient air data collected from several studies
conducted in various regions of the United States. These studies, all published in the 1970s, show
a range of concentrations from a low 2.1-9.4 |ng/m3 in one state to a high value of 100 ug/m3, which
is an average value using data from three other states. In a separate report, concentrations of 0.007
ug/m3 in one state and 0.004 ug/m3 in another state were seen during the summer of 1978. Another
state study showed that during the summer of 1985, the ambient concentration was 0.002 ug/m3.
One other state study conducted from 1979 to 1980, showed average atmospheric concentration of
Chemical X as 0.0003 ug/m3. Data from an EPA urban air data base shows a mean value from about
0.002 to 0.007 |ag/m3, with a minimum value of 0.0005 ug/m3 and a maximum value of greater than
0.03 ug/m3.
Data from the more recent studies cited above are combined to estimate exposure. Averaging
the average values from recent studies yields a central tendency air concentration value of 0.008
Hg/m3. The high-end value may be more than 0.03 ug/m3 and may be 0.05 ug/m3. These data are
from limited geographic regions and are not indicative of areas in which Chemical X concentrations
have not been detected. Therefore, the average values are likely to be overestimates of the actual
national average estimates. However, because data are not readily available on the number of areas
where air concentrations have been measured but are below detection, these values are used as crude
estimates of central and high-end values of intake of 2.3 x 10'6 mg/kg-day and 1.4 x 10'5 mg/kg-day,
respectively.19
Indoor Air. Researchalso suggests that indoor air concentrationsmay be significantlyhigher
than outdoor air. One study measured Chemical X levels in seven buildings. All types of buildings
examined had concentrations that were significantlyhigher than outdoor concentrations. A second
study measured the magnitude of the difference between indoor and outdoor air. This study found
that normal indoor air concentrations of Chemical X were at least one order of magnitude higher
than outdoor concentrations. Although these data suggest that indoor air may have higher
concentrations than outdoor air, the study which measured the difference between indoor and
outdoor levels was done in only three buildings. Indoor levels are likely to have decreased via a
"These values were determined by multiplying the air concentrations by a daily air intake of 20 m3 and dividing by the average adult body
weight of 70 kg.
149
-------
reduction in indoor uses of Chemical X. In addition, because this study measured the differences
for only three buildings, the estimate is fairly uncertain. Thus, we have not included separate indoor
air exposures in the current analysis. Instead, the analysis models only outdoor ambient air
concentrations. A more recent study of one building, however, does give an indication of levels of
indoor air. This study indicated levels of 0.018 and 0.017 ng/m3 at two locations were found during
1989-1991.
Adequacy of Exposure Data
The exposure data must be evaluated as to whether they are adequate to estimate central
tendencies and high-end values for each particular exposure medium (see Box 3 of Exhibit 2.3.4).
Although crude exposure estimates have been presented in the previous section, the use of these data
for estimating reliable central and high-end values of exposure are limited. This section outlines the
problems with these data, and indicates why these data are considered inadequate in terms of Box
3 for estimating the total dose from Chemical X in the population of concern to compare with the
RfD for Chemical X. Because data are determined to be inadequate to describe central and high-end
values well for the relevant exposure sources, one of two processes are used to set standards in the
environmental media of concern. Depending on the process, either default values are used as the
allowable dose from a given exposure medium or the available data are used to determine media-
specific allowable doses via a more conservative allocation (starting with Box 4 of Exhibit 2.3.4).
As noted earlier under the description of the Exposure Decision Tree, several factors must
be considered when evaluating data adequacy for allocating the RfD among media. One of the main
factors to consider is the number of samples in the data set being used to describe a particular
exposure medium. Although there are no universal rules about adequate sample sizes, the proposed
rules of thumb discussed earlier on page 146 are used here. For estimating a 90th percentile value
using a non-parametric method, 45 samples are needed, of which at least five must be above
detection limits to determine the 90th percentile value. Fewer samples are usually adequate for
estimating mean and median values. In addition to evaluation of sample size, the other
aforementioned factors should be assessed for a full evaluation of data adequacy [i.e.,
representativenessof the sample, the accuracy in the analytical procedures, and the sensitivity of the
measurement relative to the environmental levels of concern (i.e., whether detection limits are low
enough such that concentrations can be detected in most samples within a data set)].
Intake of Drinking Water from Raw Surface Water Sources
For this analysis, the two most recent studies were used to represent central and high-end
estimates of concentrations. The central tendency estimate was determined using the most recent
data previously indicated. The number of data points that made up this value was five. Because the
high-end estimate of 15.6 ng/L was not determined as a 90th percentile value, the sample size used
to determine the value was not evaluated in the context of the number needed to determine a 90th
percentile.
150
-------
It is important to consider several factors in determining whether the data are adequate. First,
both samples are current and thus more representative of Chemical X concentrationsthan older data.
However, the data are from two lakes only, and thus, do not represent a national distribution of data.
In addition, the data used for the average and the high-end values are taken from two surveys, and
thus, differences exist between these studies such as the number of chemical analogues generally
detected. Neither study reports the detection limits, or whether any of the values were below
detection.
Based on the lack of information regarding detection limits, limited geographic
representation of the data, and low sample sizes, it was determined that data are inadequate to obtain
central tendency and high-end estimates of exposure. Although such estimates are presented, they
represent only crude numbers.
Freshwater/Estuarine Fish Intake
The most recent data appear adequate to use in estimating typical and high-end exposures
from contaminants in fish. The purpose of the national study used was to determine the geographical
extent of chemical contamination. Thus, data were collected from all watersheds in the United States
and were collected in 1984. The sample size of this study seems large enough to adequately
represent the geometric mean value. As noted above, the minimum number of samples needed to
adequately represent the 90th percentile of a given exposure (using non-parametric methods to
estimate acceptable sample size) was determined to be 45. The sample size needed to adequately
determine the median value would be smaller than the size needed to determine the 90th percentile.
Because the geometric mean may be assumed to be equivalent to the median, it is assumed that the
number of samples used in the study is adequate to estimate the geometric mean. In addition to a
minimum sample size needed to estimate the geometric mean, a minimum number of positive values
is also needed to determine the median (or geometric mean). Although it is not known how many
samples are above detection, 91 percent of the stations sampled had Chemical X concentrations
above detection. Thus, for purposes of this example, it was assumed that a majority of samples had
Chemical X concentrations above detection.
For the estimates offish consumption rates, the large sample size and national representation
of the CSFII survey make it a useful survey for measurement of fish consumption, if the assumption
is made that the consumption rates from the CSFII study (which measured consumption of both
sport-caught and commercial fish) apply to consumption of freshwater/estuarine fish.
Intake of Treated Drinking Water from Surface Water Sources
None of the studies of Chemical X concentrations in drinking water from surface water
sources is ideal for estimating exposure through drinking water from surface water sources. The first
national study reviewed may offer the best information to use in estimating exposure because
Chemical X concentrations from surface water sources were taken from many cities across the
country and because the survey reported detection limits. However, because of the large number of
151
-------
nondetected samples, central exposure values cannot be determined without making assumptions
about the concentrations in the undetected values. In addition, the data are older and, therefore, may
not represent current concentrations. The more recent national study did not detect any Chemical
X concentrations and did not report the detection limits. Without knowing the detection limit, it is
impossible to make any assumption about the concentrations in the undetected samples, unless it is
assumed that the concentrations are zero. Because of these problems, these data are considered
inadequate for estimating exposure from surface water supplies of drinking water.
Intake of Treated Drinking Water from Ground Water Sources
Data on exposures from ground water sources of drinking water are difficult to use because
few detected samples have been found. As with the surface water sources, it is impossible to make
assumptions about undetected samples measured in the study chosen. For data from the first national
study, the number of undetected samples and the time period during which the study was conducted
make it difficult to estimate central estimates of exposure from this study with any level of
confidence. The state study chosen collected about 160 samples, of which 30 had concentrations
above the detection limit. However, this study was done at the time that uses of Chemical X stopped
and the concentrations were measured in industrialized settings within this one state. Because some
studies reported so many nondetected values, and others used only regional data, these data are
considered inadequate for estimating reliable central estimates of exposure to Chemical X in ground
water sources of drinking water.
Sources of Food Intake
The limited number of positive samples found in the 44 diet collections make estimating
exposure using these data difficult. Without knowing exact detection limits, it is difficult to make
assumptions about concentrations for the undetected samples. In addition, for any given food item,
generally only one value was above the quantifiable limit for Chemical X. (Two positive samples
were found for one food item.) Thus, neither adequate central or high-end concentrations could be
estimated. Because of these problems, it was determined that the data are not adequate for
estimating national distributions of Chemical X concentrations in food.
Sources of Air Intake
These data are too limited for adequately estimating national exposure to Chemical X from
air. Several of the studies were performed in the 1970s when companies still manufactured
Chemical X. In addition, more recent nationally representative exposure estimates are not available.
Finally, sample size was not available for many of these studies. Thus, it was determined that these
data are inadequate to estimate exposure from air sources of Chemical X.
152
-------
Setting AWQC
Under the Exposure Decision Tree Approach, either the available exposure data or default
values are evaluated against the toxicological dose that should result in no adverse health effects
from exposure to Chemical X. The toxicological dose used to evaluate the exposure and the method
of allocation are described below.
The toxicological dose is determined first, as it is the parameter to which the other factors
are applied. Chemical X has been shown to cause more than one type of toxicity. Two chronic RfDs
have been established for Chemical X, based on the oral route. The lowest value is 2.0 x 10"5 mg/kg-
day and is based on clinical and immunological studies performed on monkeys. The adverse health
effects found in this study include decreased antibody response to injected sheep red blood cells by
three principal cells of the immune system, exudate from the eye, inflammation of eyelid glands, and
changes in finger and toe nails. Because non-water exposures are considered for cases in which
pollutants cause threshold effects, this example uses the chronic RfD value of 2.0 x 10'5 mg/kg-day
to evaluate chronic toxicity effects.
In addition to these chronic effects, Chemical X has been shown to result in adverse health
effects based on results from short duration studies. EPA recommends that where such effects have
been identified, these should be considered and compared with exposure estimates which use intake
rates that may plausibly be ingested within a short time. A literature review shows an acute study
with a LOAEL of 244 mg/kg-day for developmental (fetal) effects. Assuming uncertainty factors
of 10 for animal-to-human,intrahuman, and LOAEL to NOAELs results in a value of 0.244 mg/kg-
day. To determine whether to evaluate the chronic or developmental effects, the differences between
short-term exposures and the "RfDDT" were compared with differences between chronic exposure
and the chronic RfD.
Specifically, health effects and relevant intake assumptions for a target population of
pregnant women were considered for this example due to the fetal developmental effects indicated
above. It was assumed that a pregnant woman might ingest a one-time high dose or multiple high
doses of fish within a short time period. Data show that such doses may be much higher than
average fish ingestion. However, it is unlikely that high-end intakes from other media would occur
simultaneously. Thus, the comparison used high intake assumptions for fish intake only. The
information on "acute" fish intake rates (as defined in Section 2.3.2.3, Preference #3) available from
the CSFII includes assumptions for women of childbearing age (ages 15-44 years old). The 90th
percentile value of "acute" intake (see Section 2.3.2.3 for a discussion of these intake values
determined in the CSFII) is 148.8 g/day, which was used in the comparison. The body weight value
of 65 kilograms from Ershow and Cantor (1989) was used and is applicable to pregnant women.
Comparing,differences between this shorter-term exposure and the developmental effect RfD
"RfDDT" with the differences between chronic exposure and the chronic RfD indicated that shorter-
term exposure compared with the developmental RfD is lower than the chronic exposure versus the
chronic RfD. Chronic exposures using intake from public drinking waters rather than ambient water
153
-------
(1.1 x lO"4 mg/kg-day) are about five and a half times higher than the chronic RfD (2 x 10'5 mg/kg-
day), whereas the shorter-term exposures (1.6 x 10'3 mg/kg-day) are much lower than the
developmental "RfDDT" (0.244 mg/kg-day). Thus, only chronic exposures are considered in this
example.
Using the process to set AWQC outlined in Exhibit 2.3.4, the available exposure data
(although limited) are used to determine allowable doses for each medium. This allocation is
performed because there is more than one source/use of the chemical (Box 8) and because there is
some informationavailableto characterize all sources of exposure (Box 1OA). Because the exposure
estimates are determined not to be adequate enough to represent central and high-end estimates, the
allowable dose for any one source can only be as high as 50 percent of the total allowable dose. The
allowable dose is also limited by the floor 20 percent of the allowable dose (Box IOC). Total
exposures compared with the RfD are included in Table 2.3.26. Each exposure source as a percent
of total exposure, as well as the default values (also as percentages) of these exposures used in
criteria allocation, are included in Table 2.3.27. These allowable doses from each medium are
determined for three types of individuals discussed earlier (average individuals from the general
population, sportfishers, and subsistence fishers). Crude contaminant concentration values are
available for high exposure estimates and for central tendency estimates. Thus, a decision must be
made about whether to use the high-end or central value when setting standards for a given
environmental medium. Guidance is currently being developed to address the use of central
tendency versus high-end values. For this example, central estimates of contaminant concentrations
are used for the general population, and high-end concentration estimates offish are used for sport
and subsistence fishers. [Because the variability in the exposure estimates is based predominantly
on the variability of concentrations in the exposure media, the high-end values do not reflect use of
high-end versus central tendency consumption rates. Rather, differences in consumption rates
(which apply only to fish consumption) are reflective of the defaults used for the different
populations of fishers.] States may wish to use different percentile values for the general population
and other fishers based on concentrations of contaminants in their area.
For the three groups offish consumers being evaluated (subsistence fishers, sportfishers, and
individuals from the general population), two criteria are relevant: the AWQC and health tolerances
set for pesticide use. Thus, an allocation based on the percentage approach (and allocating the free
space) is done. Because total exposure is greater than the RfD, there is no free space to be allocated.
Thus, the percentage approach was used without allocating free space.
154
-------
Table 2.3.26: Total Exposure Compared with the RfD
General population
Sportfisher20
Subsistence fisher20
RfD
Total Exposures with
Ambient water
(mg/kg-day)
l.OxlO'4
1.7 xlO'3
8.3 x lO'3
2.0 x lO'5
Total Exposures with
Drinking water
(mg/kg-day)
1.1 x 10-4
1.7 xlO'3
8.3 x lO'3
2.0 xlO'5
Table 2.3.27: Exposure Information — Percent of Total Exposures and Default Exposure
Percentages for Three Types of Individuals
Fish
Water
Diet
Air
Fish and Water Criterion
Exposure as a Percent of
Total Exposure
Subsist.
99.9
0.01
0.03
Sport
99.8
0.07
0.1
Gen.
96.6
1.1
2.2
Default
Value
50%
20%
20%
Fish-Only Criterion
Exposure as a Percent of
Total Exposure
Subsist.
99.9
0.04
0.01
0.03
Sport
99.6
0.2
0.07
0.1
Gen.
93.5
3.2
1.1
2.2
Default
Value
50%
20%
20%
20%
Total exposure compared with the RfD is included in Table 2.3.26. All exposures are greater
than the RfD. For both criteria and all three types of individuals, the percent of exposure from eating
either (1) fish and water or (2) fish-only is very high (>90 percent of the fisher's total exposure). The
ceiling of 50 percent is used for the RSC allocation for the ambient water quality criterion.
The value of 50 percent is then multiplied by the total allowable dose (2.0 x 10'5 mg/kg-day)
to determine allowable doses for all three types of individuals, as noted in Table 2.3.28. For each
of the criteria, the allowable dose of 1 x 10'5 mg/kg-day ^is used to determine the AWQC. Criteria
20 These estimates were made using high-end values offish exposure (i.e., the reported high-end contaminant concentration, along with the
default consumption rates) for these populations. High-end values were also assumed for the estimates made with ambient water data (i.e.,
high-end contaminant concentrations). If central tendency contaminant concentrations had been used, the subsistence fisher percent of total
exposure attributable to fish and water, and fish-only would be 99.3% and 98.6%, respectively. If central tendency contaminant concentrations
had been used for the sportfisher, the estimates would equal those for the general population.
155
-------
based on (1) fish and water intake, and (2) fish intake only are calculated and presented in this table.
For each type of fisher, the fish and water criterion values do not differ from the fish only criterion
values. Other exposure factors from the equation used in the calculation are: body weight=70 kg;
drinking water intake=2 L/day (or incidental ingestion of 0.01 L/day); fish intake rates of 0.01780
kg/day for the general and sportfisher populations and 0.08630 kg/day for subsistence fishers; and
a bioaccumulation factor of 120,000.
Table 2.3.28: AWQC for Three Types of Individuals
Fish and Water Criterion
Fish Only Criterion
Default Allowable Dose
(50%ofRfD)
1.0xlO-5mg/kg-day
l.Ox 10'5 mg/kg-day
AWQC:
Subsistence Fisher21
Sportfisher
General Population
6.8xlO-8mg/L
3.3xlO-7mg/L
3.3xlO-7mg/L
6.8xlO-8mg/L
3.3xlO-7mg/L
3.3 x lO'7 mg/L
Presenting Information to Risk Managers
Although the above example utilizes the 20 percent floor for the other sources of exposure,
it is clear that a combination of allocations of the RfD, if used, would exceed the 80 percent ceiling
and in the case of the fish-only criterion, would exceed 100 percent of the RfD. This example also
illustrates the potential need for flexibility to lower the floor (perhaps to zero?) due to exposures in
exceedance of the RfD.
Because total exposures from all environmental media are greater than the dose without an
appreciable risk of toxicological effects, several pieces of information can be presented to risk
managers for their review in deciding how to apportion the dose of 2.0 x 10'5 mg/kg-day among
exposure sources. These data include information about the toxicity of the chemical (including
uncertainty in the estimate), information about exposures, and issues involving control of Chemical
X. Because exposures and uncertainties in these estimates were discussed in previous sections, they
will not be discussed again here. Additionally, control technology issues will not be discussed here,
as they are not directly related to estimating exposure via the Exposure Decision Tree Approach.
However, additional toxicity information is presented below.
21 If the alternate default intake assumption of 86.3 g/day was used instead, the subsistence fisher AWQC estimates for fish and water, and
fish only, would be 1.5 x 10'7 mg/L
156
-------
The toxicity data supporting the RfD of 2.0 x 10"5 mg/kg-day should be described in order
to give risk managers an idea about the confidence in the value. Some evaluative information is
available from the Integrated Risk Information System. As noted above, the critical endpoints upon
which the RfD is based include decreased antibody response to injected sheep red blood cells,
exudate from the eye, inflammation of eyelid glands, and changes in finger and toe nails. A total
uncertainty factor of 300 is applied to the Lowest Observed Adverse Effect Level (LOAEL) from
the critical study. The total number of uncertainty factors account for: sensitive individuals within
the population (a 10-fold factor); extrapolation from monkeys to humans (a 3-fold factor); use of a
LOAEL instead of a NOAEL (a partial factor); and use of a subchronic rather than a chronic study
(a 3-fold factor). The overall confidence in the RfD is medium because the confidence in the
principal study based, in turn, on the confidence in the data base are also considered medium.
In addition to toxicity data, information on ways to control exposure in different
environmental media can be presented to risk managers to aid them in determining the relative ease
of controlling exposure from different environmental media. Ideally, this information would include
expected incremental costs of treatment needed per unit decrease of Chemical X and other feasibility
issues associated with controlling Chemical X in different environmental media.
Exposure from consuming freshwater/estuarine fish represents the single largest exposure
to Chemical X. Other sources of exposure represent smaller percentages of total exposure, but all
high-end exposures except air exposures also exceed the RfD individually.
2.3.4.4 Unavailability of Substances from Different Routes of Exposure
For many chemicals, the rate of absorption can differ substantially from ingestion compared
to inhalation. There is also available information for some chemicals which demonstrates
appreciable differences in gastrointestinal absorption depending on whether the chemical is ingested
from water, soil, or food. For some contaminants, plant and animal food products may also have
appreciably different absorption rates. Regardless of the allocation approach used, EPA proposes
using existing data on differences in bioavailability between water, air, soils, and different foods for
estimating total exposure and in allocating the RfD. The Agency has developed such exposure
estimates for cadmium (USEPA, 1994b). In the absence of data, EPA will assume equal rates of
absorption from different routes and sources of exposure.
Information on absorption rates for Chemical X is not available. Available studies only
generally describe varying fractions of the chemical in different media due to varying rates of
volatilization, solubility, and adsorption. Discussions about chemical transformations in the
environment and by human metabolism are similarly vague. In the absence of such data, it is
assumed for this example that Chemical X is fully absorbed and that the rates of absorption from
different routes and sources of exposure are equal.
157
-------
2.3.5 References
Alabama Department of Environmental Management. 1993. Estimation of Daily Per Capita
Freshwater Fish Consumption of Alabama Anglers. Prepared by Fishery Information
Management Systems, Inc., and the Department of Fisheries and Allied Aquacultures,
Auburn University.
Allbright,K. 1994. Minnesota Department of Health, Division of Environmental Health. Personal
communication with Abt Associates. May 27; May 31; July 28.
Agency for Toxic Substances and Disease Registry (ATSDR). 1995. Final Report: Technical
Assistance to the American Samoa Government Department of Health Services, Pago Pago,
American Samoa.
Barclay, B. 1993. Hudson River Angler Survey. Poughkeepsie, NY: Hudson River Sloop
Clearwater, Inc. March.
Borum, D.R. Approaches to Allocating the RfD for Setting Health-Based Criteria. Available in the
Public Docket for The Proposed Revisions to the Ambient Water Quality Criteria Human
Health Methodology (unpublished).
California Environmental Protection Agency. 1997. Chemicals hi Fish: Report No. 1.
Consumption of Fish and Shellfish in California and the United States. Final Draft Report.
Pesticide and Environmental Toxicology Section, Office of Environmental Health Hazard
Assessment. July.
Cantor, K.P., R. Hoover, P. Hartge, et al. 1987. Bladder Cancer Drinking Water Source, and Tap
Water Consumption: A Case-Control Study. J. National Cancer Inst. 79(6): 1269-1279.
Centers for Disease Control. 1993. Mercury Exposure Among the Miccousukee Indian Tribe:
South Florida. Unpublished report. Atlanta, GA.
Chiang, A. 1994. Asian Pacific Environmental Network, Oakland, CA. Personal communication
with Abt Associates. April 15.
Cole, M. 1994. EAGLE Project. Personal Communication with Abt Associates. September 23.
Connelly, N.A., T.L. Brown and B. A. Knuth. 1990. New York Statewide Angler Survey, 1988.
Albany, NY: New York State Department of Environmental Conservation.
Connelly, N.A., B.A. Knuth, and T.L. Brown. 1996. Sportfish Consumption Patterns of Lake
Ontario Anglers. N. Am. J. Fisheries Management (In press at time of citation).
158
-------
Continuing Survey of Food Intakes by Individuals (CSFII). 1989-1991. U.S. Department of
Agriculture. Data and Documentation. Hyattsville, MD: Human Nutrition Information
Service (Cited in Jacobs, 1996).
Cox C., A. Vaillancourt,and A. Hayton. 1993. The Results of the 1992 "Guide to Eating Ontario
Sport Fish" Questionnaire. The Ontario Ministry of Environment and Energy. PIBS 2593E.
Columbia River Inter-Tribal Fish Commission (CRITFC). 1994. A Fish Consumption Survey of
the Umatilla, Nez Perce, Yakama and Warm Springs Tribes of the Columbia River Basin.
Portland, OR: CRITFC. Technical Report 94-3.
Cung, J. 1994. Minnesota Department of Natural Resources, South East Asian Outreach Project.
Personal communication with Abt Associates. July 28.
Degner, R.L., C.M. Adams, S.D. Moss, and S.K. Mack. 1994. Per Capita Fish and Shellfish
Consumption in Florida. Submitted to the Florida Department of Environmental Protection
by the Florida Agricultural Market Research Center. Industry Report 94-2. August.
Dellenbarger, L., A. Schupp, and B. Kanjilal. 1993. Seafood Consumption in Coastal Louisiana.
Louisiana Department of Environmental Quality.
Dellinger, J.A. 1993. Assessment of a Human Population at Risk: The Impact of Consuming
Contaminated Great Lakes Fish on Native American Communities. The Ojibway Health
Study. 1993 Progress and Summary Report (unpublished).
Dellinger, J.A. 1996. Department of Preventative Medicine, Medical College of Wisconsin.
Personal communication with Abt Associates. March 26.
Den, A. 1994. Senior Science Advisor, USEPA Region 9. Personal communication with Abt
Associates. July 21; July 28.
Den, A. 1998. Senior Science Advisor, USEPA Region 9. Personal communication with Denis R.
Borum, Office of Water, USEPA Headquarters, 3/16/98.
EAGLE Project (Effects on Aboriginals from the Great Lakes Environment). 1991. 1991 Report,
Factsheets, and Principles Document. Ottawa, Ontario, Canada.
Ebert, E., N. Harrington, K. Boyle, J. Knight and R. Keenan. 1993. Estimating Consumption of
Freshwater Fish Among Maine Anglers. North American Journal of Fisheries Management
13:737-745.
Eng, G. 1994. Regional Representative, Office of Regional Operations. Agency for Toxic
Substances and Disease Registry. Personal communication with Abt Associates. August 1.
159
-------
Ershow, A.G. and K.P. Cantor. 1989. Total Water and Tap Water Intake in the United States:
Population-based Estimates of Quantities and Sources. Bethesda, MD: National Cancer
Institute. Order #263-MD-810264.
Fiore, B.J., et al. 1989. Sport Fish Consumption and Body Burden Levels of Chlorinated
Hydrocarbons: A Study of Wisconsin Anglers, Archives of Environmental Health. 44: 82-
88.
Groot, C. and L. Margolis. 1991. Pacific Salmon Life Histories. UBC Press. 564pp.
Hamly, M. et al. 1997. Biological Monitoring for Mercury within a Community with Soil and Fish
Contamination. Environ. Health Perspect. 105(4):424-429.
Honstead, J.F., T.M. Beetle and J.K. Soldat. 1971. A Statistical Study of the Habits of Local
Fishermen and Its Application to Evaluation of EnvironmentalDose. Richland, WA: Battelle
Pacific Northwest Laboratories (Cited in Rupp, E. et al.,1980).
Hovinga, M.E., M. Sowers, and H.B. Humphrey. 1992. Historical Changes in Serum PCB and
DDT Levels in an Environmentally Exposed Cohort. Arch. Environ. Contam. Toxicol.
22:362-366.
Hovinga, M.E., M. Sowers, and H.B. Humphrey. 1993. Environmental Exposure and Lifestyle
Predictors of Lead, Cadmium, PCB, and DDT Levels in Great Lakes Fish Eaters. Archives
of Environmental Health 48(2):98-104.
Humphrey, H.E.B. 1976. Evaluation of Changes in the Level of Polychlorinated Biphenyls (PCB)
in Human Tissues. Michigan Department of Public Health. Washington, DC: U.S. FDA,
U.S. Department of Health, Education, and Welfare. FDA 223-73-2209. June.
ICRP. 1981. International Commission on Radiological Protection. Report of the Task Group on
Reference Man. New York: Pergammon Press.
Keill, L. 1996. Washington Department of Ecology. Personal communication with Abt Associates.
March 28.
Kleiman, C.F. 1985. Fish Consumption by Recreational Fishermen: An Example of Lake
Ontario/Niagara River Region. Prepared for USEP A Office of Enforcement and Compliance
Monitoring by Environ Corporation. May 20.
Kmiecik, N. 1994. Great Lakes Indian Fish and Wildlife Commission. Personal communication
with Abt Associates. July 28; August 12.
160
-------
Kutkuhn, J. H. No Date. The Role of Estuaries in the Development of Perpetuation of Commercial
Shrimp Resources. In: Symposium on Estuarine Fishes. American Fisheries Society.
Special Publication 3. 1966. pp. 16-36.
Lorenzana, R. 1994. USEPA Region 10. Personal communication with Abt Associates. July 28.
Lorenzana, R. 1998. Toxicologist, USEPA Region 10. Personal communication with Denis R.
Borum, Office of Water, USEPA Headquarters. March 17.
Louisiana Department of Environmental Quality. 1989. Documentation of Numerical Criteria for
Human Health Protection in the 1989 Water Quality Standards Revision. Office of Water
Resources, Baton Rouge. June.
Louisiana Department of Environmental Quality. 1994. Procedures for Human Health Criteria
Calculation in Louisiana. Office of Water Resources, Baton Rouge. May 11.
Meredith E.K., and S.P. Malvestuto. 1996. Evaluation of Two On-Site Survey Methods for
Determining Daily per Capita Freshwater Fish Consumption by Anglers. American
Fisheries Society Symposium 16:271-276.
Mittlestadt, J. 1994. Tulalip Tribal Department of Natural Resources. Personal communication
with Abt Associates. July 28.
Nakano, C. 1996. Personal communication between Roseanne Lorenzana, USEPA Region 10 and
Connie Nakano, Asian and Pacific Islander Seafood Consumption Study Coordinator.
Seattle, WA: Refugee Federation Service Center.
National Academy of Sciences (NAS). 1977. Drinking Water and Health. Volume I.
Drinking Water Committee, National Research Council. Washington, DC.
Safe
National Center for Health Statistics (NCHS). 1987. Anthropometric Reference Data and
Prevalence of Overweight, United States, 1976-1980. Data from the National Health and
Nutrition Examination Survey, Series 11, No. 238. Hyattsville, MD: U.S. Department of
Health and Human Services, Public Health Service, National Center for Health Statistics.
DHHS Publication No. (PHS) 87-1688.
National Marine Fisheries Service (NMFS). 1995-1996. Fisheries Statistics Division. Personal
communication. (Cited in Jacobs, 1996.)
Nehls-Lowe, H. 1994. Wisconsin Department of Natural Resources. Personal communication with
Abt Associates. July 29.
161
-------
Nobmann, E.D., T. Byers, A.P. Lanier, J.H. Hankin, and M.Y. Jackson. 1992. The Diet of Alaska
Native Adults: 1987-1988. American Journal of Clinical Nutrition 55: 1024-1032.
Pennington, J. A. T. 1983. Revision of the Total Diet Study Food List and Diets. Journal of the
American Dietetic Association, 82: 166-173.
Pestana, E. 1994. Connecticut Commissioner's Office of the Department of Environmental
Protection, Section of Environmental Justice. Personal communication with Abt Associates.
May 18.
Peterson, D. E., M.S. Kanarek, M.A. Keykendall, J.M. Diedrich, H.A. Anderson, P.L. Remington,
and T.B. Sheffy. 1995. Fish Consumption Patterns and Blood Mercury Levels in Wisconsin
Chippewa Indians. Archives of Environmental Health 49(11): 53-58.
Pierce, D., D.T. Novielle and S.H. Rogers. 1981. Commencement Bay Seafood Consumption
Study: Preliminary Report. December.
Ponwith, BJ. 1991. The Shoreline Fishery of American Samoa: A 12 Year Comparison. Pago
Pago, American Samoa: Department of Marine and Wildlife Resources. DMWR Biological
Report Series, No. 2.
Puffer, H. 1981. Consumption Rates of Potentially Hazardous Marine Fish Caught in the
Metropolitan Los Angeles Area. EPA Grant #R807 120010.
Richter, B.S. and R. Rondinelli. 1989. The Relationship of Human Levels of Lead and Cadmium
to the Consumption of Fish Caught In and Around Lake Coeur d'Alene, Idaho. Final Report,
Technical Assistance to the Idaho State Health Department and the Indian Health Service.
Rupp, E., F.L. Miller and C.F. Baes III. 1979. Some Results of Recent Surveys of Fish and
Shellfish Consumption by Age and Region of U.S. Residents. Health Phys. Vol. 39.
SCCWRP and MBC. 1994. Santa Monica Bay Seafood Consumption Study: Final Report.
Southern California Coastal Water Research Project and MBC Applied Environmental
Sciences. Westminster and Costa Mesa, CA. June. (Cited in California, 1997.)
Sekerke, H.L., R.L. Denger, C.M. Adams, and S.D. Moss. 1994. Seafood Consumption Survey in
Florida. Abstract presented at the Society of Toxicology 33rd Annual Meeting (Vol. 14,No.
1). Abstract no. 563. March.
Shubat,P. 1993. Minnesota Department of Health. Personal communication with Abt Associates.
August 25.
162
-------
Shubat, P. 1994. Minnesota Department of Health. Personal communication with Abt Associates.
July 29.
Soldat. 1970. A Statistical Study of the Habits of Fishermen Utilizing the Columbia River below
Hanford. In: Environmental Surveillance in the Vicinity of Nuclear Facilities, W.C. Rienig,
ed. C.C. Thomas, Springfield, IL. pp. 302-308.
Sonstegard, R. 1985. Personal communication as noted in Kleiman (1985), page 9.
Soukhaphonh, S., L. Keill, D. LaFlamme, et al. 1996. Draft Report: A Fish Consumption Survey
of Laotians in Washington State. Washington Department of Ecology.
State of Michigan. 1985. Guidelines for Rule 57(2) 323.1057 of Part 4: Water Quality Standards.
Filed with Michigan Secretary of State. January 2.
Toy, K.A., N.L. Polissar, S. Liao, and G.D. Mittelstaedt. 1996. A Fish Consumption Survey of the
Tulalip and Squaxin Island Tribes of the Puget Sound Region. Marysville, WA: Tulalip
Tribes, Department of the Environment.
Turcotte,S. 1983. Georgia Fishery Study: Implications for Dose Calculations. Technical Division
Savannah River Laboratory. Memorandum to H.P. Olson. DPST-83-319-Rev. 1.
University of Wisconsin, SeaGrant. 1994. Personal communication with Abt Associates. May 27.
U.S. Bureau of the Census. 1992. U.S. Department of Commerce. Statistical Abstract of the United
States. Washington, DC.
U.S. Bureau of the Census. 1996. Personal communication between Jean Dee and Abt Associates.
March 1.
U.S. Department of Agriculture (USDA). 1988. Machine-Readable Data Sets on Composition of
Foods and Results from Food Consumption Surveys. Hyattsville, MD: U.S. Department of
Agriculture Survey Statistics Branch, Nutrition Monitoring Division, Human Nutrition
Information Service. Administrative Report No. 378 (Cited in Ershow and Cantor, 1989).
U.S. Department of Commerce (USDOC). 1992. Statistical Abstract of the United States, 1992.
Washington, DC: U.S. Bureau of the Census.
USEPA. 1979. Identification and Evaluation of Waterborne Routes of Exposure from Other Than
Food and Drinking Water, Washington, DC: Office of Water Planning and Standards. EPA
440/4-79-016.
163
-------
USEPA. 1989a. Exposure Factors Handbook. Washington, DC. Office of Health and Environmental
Assessment, Exposure Assessment Group. EPA/600/8-89/043.
USEPA. 1989b. Risk Assessment Guidance for Superfund, Volume 1, Human Health Evaluation
Manual (Part A) - Interim Final. Washington, DC: Office of Emergency and Remedial
Response. EPA/540/1-89/002.
USEPA. 1992. Tribes at Risk. The Wisconsin Tribes Comparative Risk Project. Office of Policy
Planning and Evaluation and Region 5. EPA230-R-92-017. October.
USEPA. 1994a. Guidance For Assessing Chemical Contaminant Data for Use in Fish Advisories.
Volume II: Risk Assessment and Fish Consumption Limits. Washington, DC: Office of
Water. EPA 823-B-94-004.
USEPA. 1994b. Integrated Risk Information System (IRIS). Reference Dose (RfD) for Oral
Exposure for Cadmium. Online. (Verification date 5/25/88.) Cincinnati, OH: Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office.
USEPA. 1995a. Water Quality Guidance for the Great Lakes System: Supplementary Information
Document (SID). Washington, DC: Office of Water. EPA-820-B-95-001. March.
USEPA. 1995b. Mercury Study Report to Congress. Volume II: Inventory of Anthropogenic
Mercury Emissions in the United States.
USEPA. 1996. Asian and Pacific Island Seafood Consumption Study. Exposure Information
Obtained Through a Community-Centered Approach. Risk Evaluation Unit, Office of
Environmental Assessment, USEPA Region 10. EPA 910/R-96-007. August.
USEPA. 1997a. Exposure Factors Handbook. National Center for Environmental Assessment,
Office of Research and Development. Washington, DC. EPA/600/P-95/002Fa. August.
USEPA. 1997b. Guidance for Conducting Fish and Wildlife Consumption Surveys. September.
USEPA. 1998. Daily Average Per Capita Fish Consumption Estimates Based on the Combined
USDA 1989, 1990, 1991 Continuing Survey of Food Intakes by Individuals (CSFII).
Volume I: Uncooked Fish ConsumptionNational Estimates; Volume II: As Consumed Fish
Consumption National Estimates. Prepared by SAIC under Contract #68-C4-0046. March.
Wells, P. 1990. Michigan Department of Natural Resources (MDNR), Recreation Division,
Personal Communication to Gary Hurlburt, Surface Water Quality Division, MDNR.
West, P.C., et al. 1989. Michigan Sport Anglers Fish Consumption Survey: A Report to the
Michigan Toxic Substance Control Committee, University of Michigan Natural Resource
164
-------
Sociology Research Lab. Technical Report # 1, Ann Arbor, Michigan, MDMB Contract #87-
20141.
West, P., J.M. Fly, R. Marans, F. Larkin and D. Rosenblatt. 1993. The 1991-92 Michigan Sport
Anglers Fish Consumption Study: Final Report to the Michigan Great Lakes Protection
Fund, Michigan Dept. of Natural Resources. University of Michigan, School of Natural
Resources, the Natural Resource Sociology Research Lab. Technical Report #6. May.
Wheatley,B. 1996. Environment Canada. Personal communication with Abt Associates. March
26.
Wolfe, R.J. and R.J. Walker. 1987. Subsistence Economics in Alaska: Productivity, Geography,
and Development Impacts. Arctic Anthropol. 24(2):56-81.
Young, P. 1994. USEPA Region 9. Personal communication with Abt Associates. July 28.
Zein-Eldin, Z.P. and M.L. Renaud. 1986. Inshore Environmental Effects on Brown Shrimp,
Penaeus aztecus, and White Shrimp, P. setiferus, Populations in Coastal Waters, Particularly
Texas. Marine Fisheries Review. 48(3): 9-16.
2.4. Use of BAFs in the Derivation of AWQC
2.4.1 Introduction
Aquatic organisms are known to accumulate certain types of chemicals in their bodies.
Uptake of these chemicals may occur from exposure to contaminated water, consumption of
contaminated food, and exposure to other sources such as contaminated bottom sediment. This
chemical uptake process is called bioaccumulation. For some chemicals, such as certain highly
hydrophobic chemicals, uptake through the food chain can be the most important route of exposure.
As organisms in higher trophic levels feed on organisms in lower trophic levels, tissue
concentrations of these chemicals increase through the trophic levels so that the concentrations in
the highest trophic level organisms may be many orders of magnitude higher than levels in the
environment. The trophic-level increase in contaminated concentrations is called biomagnification,
and may result in serious adverse health effects for consumers of the highest trophic levels offish.
To protect humans from harmful exposures to bioaccumulative chemicals, EPA proposes to
use bioaccumulation factors (BAFs) in deriving AWQC. These BAFs are ratios of the contaminant
concentration in tissue to the concentration in water, taking into account uptake through
contaminated food, sediment, and water. Chemicals with larger BAFs reflect greater accumulation
in fish tissues compared to chemicals with lower BAFs. The BAFs may be of such large magnitude
that the resulting ambient water quality criterion will be strongly influenced by the BAF.
165
-------
In contrast to the current guidelines, the 1980 AWQC National Guidelines for deriving
human health criteria relied on an alternate type of ratio, the biconcentrationfactor (BCF), to derive
AWQC.22 In contrast to the BAF, the BCF measures uptake of chemicals into fish that have been
exposed only through water (not through food or sediment). Because B AFs account for uptake from
all sources of waterbome exposure of a chemical to an organism (through food, water, and
sediment), EPA believes the use of the BAF to be superior to the BCF for deriving human health
AWQC.
2.4.1.1 Bioaccumulation and Bioconcentration Concepts
Bioaccumulation reflects the uptake and retention of a chemical by an aquatic organism from
all surrounding media (e.g., water, food, sediment). Bioconcentration refers to the uptake and
retention of a chemical by an aquatic organism from water only. Both bioaccumulation and
bioconcentration can be viewed simply as the result of competing rates of chemical uptake and
depuration (chemical loss) by an aquatic organism. However, the rates of uptake and depuration can
be affected by numerous factors including the physical and chemical properties of the chemical, the
physiology and biology of the organism, environmental conditions, ecological factors such as food
web structure, and the amount and source of the chemical. When the rates of chemical uptake and
depuration are equal, the distribution of the chemical between the organism and its source(s) is said
to be at equilibrium or at steady state. For a constant chemical exposure, the time required to
achieve steady state conditions varies according to the properties of the chemical and other factors.
For example, some chemicals require a long time to reach steady state conditions between
environmental compartments (e.g., many months for certain highly hydfophobic chemicals) while
others reach steady state relatively quickly (e.g., hours to days for certain hydrophilic chemicals).
The concept of steady state or equilibrium conditions is very important when assessing or
evaluating bioaccumulation or bioconcentration and applying these principles in real world
situations, such as the derivation of ambient water quality criteria. For some chemicals and
organisms that require relatively long time periods to reach steady state, changes in water column
chemical concentrations may occur on a much more rapid time scale compared to the corresponding
changes in an organism's tissue concentrations. Thus, if the system departs substantially from steady
state conditions, the ratio of the tissue concentration to the water concentration may have little
resemblance to the steady-state ratio and have little predictive value of long-term bioaccumulation
potential.
For highly bioaccumulativepollutants in dynamic systems, reliable B AFs can be determined
only if, among other factors, water column concentrations are averaged over a sufficient period of
time (e.g., a duration approximating the amount of time predicted for the pollutant to reach steady-
"According to the 1980 AWQC National Guidelines, laboratory-measured or predicted bioconcentration factors were used when field-
measured bioconcentration factors (equivalent to what are now called field-measured BAFs) were not available.
166
-------
state). In addition, adequate spatial averaging of both tissue and water column concentrations is
required to develop reliable BAFs for use in deriving human health ambient water quality criteria.
For this reason, a bioaccumulation factor is defined in this guidance as representing the ratio
(in L/kg) of a concentration of a substance in tissue to its concentration in the surrounding water in
situations where the organism and its food are exposed and the ratio does not change substantially
over time. A bioconcentration factor is considered to represent the uptake and retention of a
substance by an aquatic organism from the surrounding water only, through gill membranes or other
external body surfaces, in situations where the tissue-to-water ratio does not change substantially
over time.
This chapter provides the technical basis and rationale for EPA's proposed procedures for
determining BAFs for toxic substances. Section 2.4.2 lists pertinent definitions used throughout the
chapter, Sections 2.4.3 through 2.4.5 describe issues and procedures relevant to estimating BAFs for
nonpolar organic chemicals, Section 2.4.6 describes procedures relevant to the derivation of BAFs
for inorganic chemicals, and Section 2.4.7 discusses the derivation of BAFs for two example
chemicals. Issues associated with applying fish consumption rate information to trophic level-
specific BAFs are discussed in Section 2.4.8.
2.4.2 Definitions
Baseline BAF (BAFf). For organic chemicals, a BAF (in L/kg-lipid) that is based on the
concentration of freely dissolved chemical in the ambient water and the lipid normalized
concentration in tissue; for inorganic chemicals, a BAF that is based on the wet weight of the tissue.
Baseline BCF (BCF[d). For organic chemicals, a BCF (in L/kg-lipid) that is based on the
concentration of freely dissolved chemical in the ambient water and the lipid normalized
concentration in tissue; for inorganic chemicals, a BCF that is based on the wet weight of the tissue.
Bioaccumulation. The net accumulation of a substance by an organism as a result of uptake
.from all environmental sources.
Bioaccumulation Factor (BAF). The ratio (in L/kg-tissue) of the concentration of a
substance in tissue to its concentration in the ambient water, in situations where both the organism
and its food are exposed and the ratio does not change substantially over time. The BAF is
calculated as:
BAF =
(Equation 2.4.1)
167
-------
where:
Ct = Concentration of the chemical in the wet tissue (either whole organism or
specified tissue)
Cw = Concentration of chemical in water
Bioconcentration. The net accumulation of a substance by an aquatic organism as a result
of uptake directly from the ambient water, through gill membranes or other external body surfaces.
Bioconcentration Factor (BCF). The ratio (in L/kg-tissue) of the concentration of a
substance in tissue of an aquatic organism to its concentration in the ambient water, in situations
where the organism is exposed through the water only and the ratio does not change substantially
overtime. The BCF is calculated as:
BCF =
(Equation 2.4.2)
where:
C, = Concentration of the chemical in the wet tissue (either whole organism or
specified tissue)
Cw = Concentration of chemical in water
Biota-Sediment Accumulation Factor (BSAF). The ratio (kg of sediment organic carbon
per kg of lipid) of the lipid-normalized concentration of a substance in tissue of an aquatic organism
to its organic carbon-normalizedconcentrationin surface sediment, in situations where the ratio does
not change substantially over time, both the organism and its food are exposed, and the surface
sediment is representative of average surface sediment in the vicinity of the organism. The BSAF
is defined as:
BSAF =
where:
(Equation 2.4.3)
The lipid-normalized concentration of the chemical in tissues of the biota
Cug/g lipid)
The organic carbon-normalized concentration of the chemical in the surface
sediment (/ug/g sediment organic carbon)
168
-------
Biomagnification. The increase in tissue concentration of poorly depurated materials in
organisms along a series of predator-prey associations, primarily through the mechanism of dietary
accumulation.
Biomagnification Factor (BMP). The ratio (unitless) of the tissue concentration of a
predator organism at a particular trophic level to the tissue concentration hi its prey organism at the
next lowest trophic level, for a given waterbody and chemical exposure. For organic chemicals, a
BMP can be calculated using lipid-normalized concentrations in the tissue of organisms at two
successive trophic levels as:
BMP
(TL, n)
t (TL, n)
1
"t (TL, n-1)
where:
Q(TL,n> = Lipid-normalized concentration in appropriate tissue of
predator organism at trophic level "n"
Qca,n-i) = Lipid-normalized concentration in appropriate tissue of prey
organism at the next lowest trophic level from the predator.
For inorganic chemicals, a BMP can be calculated using chemical concentrations in the tissue of
organisms at two successive trophic levels as:
BMP
(TL, n)
t (TL, n)
1
't (TL, n-1)
where:
Q(TL,n) = Concentration in appropriate tissue of predator organism at trophic
level "n" (may be either wet weight or dry weight concentration so
long as both the predator and prey concentrations are expressed in the
same manner)
Q(TL,n-i) = Concentration in appropriate tissue of prey organism at the next
lowest trophic level from the predator (may be either wet weight or
dry weight concentration so long as both the predator and prey
concentrations are expressed in the same manner)
As explained in the TSD, BMFs can also be related to (and calculated from) FCMs and baseline
BAFs.
169
-------
Depuration. The loss of a substance from an organism as a result of any active or passive
process.
Food-Chain Multiplier (FCM). The ratio of a baseline BAF for an organism of a particular
trophic level to the baseline BCF (usually determined for organisms in trophic level one).
Freely Dissolved Concentration. For hydrophobic organic chemicals, the concentration of
the chemical that is dissolved in ambient water, excluding the portion sorbed onto paniculate or
dissolved organic carbon. The freely dissolved concentration is considered to represent the most
bioavailable form of an organic chemical in water and, thus, is the form that best predicts
bioaccumulation. The freely dissolved concentration can be determined as:
where:
pfd _
v/,u
ffd
(Equation 2.4.4)
Freely dissolved concentration of the organic chemical in ambient water
Total concentration of the organic chemical in ambient water
Fraction of the total chemical in ambient water that is freely dissolved
Lipid-normalized Bioaccumulation Factor (BAF,). The ratio (in L/kg-lipid) of a
substance's lipid-normalized concentration in tissue to its concentration in the ambient water, in
situations where both the organism and its food are exposed and the ratio does not change
substantially over time. The lipid-normalized BAF is calculated as:
BAP =
(Equation 2.4.5)
where:
C{ = Lipid-normalized concentration of the chemical in whole organism or
specified tissue
Cw = Concentration of chemical in water
Lipid-normalized Bioconcentration Factor (BCFt). The ratio (in L/kg-lipid) of a
substance's lipid-normalized concentration in tissue of an aquatic organism to its concentration in
the ambient water, in situations where the organism is exposed through the water only and the ratio
does not change substantially over time. The lipid-normalized BCF is calculated as:
170
-------
where:
c,
BCF •=
(Equation 2.4.6)
Lipid-normalized concentration of the chemical in whole organism or
specified tissue
Concentration of chemical in water
Lipid-normalized Concentration (€5). The total concentration of a contaminant in a tissue
or whole organism divided by the lipid fraction in that tissue or whole organism. The lipid-
normalized concentration can be calculated as:
(Equation 2.4.7)
where:
ct - Concentration of the chemical in the wet tissue (either whole organism or
specified tissue)
f? ~ Fraction lipid content in the organism or specified tissue
Octanol-water Partition Coefficient (Kow). The ratio of the concentration of a substance
in the n-octanol phase to its concentration in the aqueous phase in an equilibrated two-phase octanol-
water system. For log K^,, the log of the octanol-water partition coefficient is a base 10 logarithm.
Organic Carbon-normalized Concentration (Csoc). For sediments, the total concentration
of a contaminant in sediment divided by the fraction of organic carbon in sediment. The organic
carbon-normalized concentration can be calculated as:
where:
C
C s
SOC f
oc
(Equation 2.4.8)
Concentration of chemical in sediment
Fraction organic carbon in sediment
171
-------
Uptake. Acquisition by an organism of a substance from the environment as a result of any
active or passive process.
2.4.3 Determining BAFs for Nonpolar Organics
The calculation of a BAF for a nonpolar organic chemical (chemicals that do not readily
dissolve in water) used in the derivation of AWQC is a two-step process. The first step is to
calculate a baseline BAF for the chemical of interest using information from the field site or
laboratory where the original data were collected (i.e., the lipid content of the species collected and
the freely dissolved fraction of the chemical in water at the site where the data were collected). If
information used to estimate fish consumption rates indicates that organisms are being consumed
from different trophic levels, then baseline BAFs need to be determined for each of the relevant
trophic levels.
The second step is to calculate a BAF (or BAFs) for the chemical that will be used in the
derivation of AWQC, using information from the location where the aquatic species of interest are
consumed (i.e., the lipid content of the aquatic species consumed by humans and the freely dissolved
fraction of the chemical in water at the site where the aquatic species are being consumed). The
difference between a baseline BAF and a BAF used in the derivation of a AWQC is that baseline
BAFs can be used for extrapolating from one species to another and from one water body to another.
This is the case because baseline BAFs are lipid-normalized, enabling extrapolation for organic
chemicals from one species to another; and because they are based on the freely dissolved
concentration of organic chemicals, enabling extrapolation from one water body to another (the
importance of these concepts is discussed below). Baseline BAFs, however, cannot be used directly
in the derivation of AWQC because they may not reflect the conditions in the area of interest (e.g.,
the lipid content of the aquatic species consumed in the area of interest and the freely dissolved
fraction of the chemical in the area of concern).
Depending on the type of information available for a given chemical, different procedures
may be used to determine the baseline BAF. The most preferred BAFs are those derived using
appropriate field data. Field-measuredBAFs, however, have not been determined for all chemicals.
Thus, EPA proposes a hierarchy of procedures to determine BAF values. The data preference for
derivation of baseline BAFs for nonpolar organic substances is as follows (in order of priority):
1. A field-measured baseline BAF derived from a field study of acceptable quality.
2. A predicted baseline BAF derived from a field-measured BSAFs of acceptable
quality.
3. A predicted baseline BAF derived from a laboratory-measured BCF of acceptable
quality and a food-chain multiplier (FCM).
172
-------
4. A predicted baseline BAF derived from an acceptable
multiplier.
. and a food-chain
While EPA recommends the above hierarchy for determining final baseline BAF values, for
comparative purposes, baseline BAFs should be determined for each chemical by as many of the
four methods as available data allow. Comparing baseline BAFs derived using the different methods
recommended above can provide insight for identifying and evaluating any discrepancies in the BAF
determinations that might occur. The information needed to derive a baseline BAF using each of
the four methods is discussed in Section 2.4.4. Section 2.4.5 discusses the information needed to
derive a BAF for use in the calculation of AWQC.
2.4.4 Estimating Baseline BAFs for Nonpolar Organics
All the baseline BAFs for nonpolar organic chemicals should be expressed on a freely-
dissolved and lipid-normalized basis. The procedures for adjusting a field-measured BAF, field-
measured BSAF, or laboratory-measured BCF to a freely-dissolved and lipid-normalized basis are
discussed below.
2.4.4.1 Field-Measured Baseline BAF
EPA's first preference for deriving a BAF for nonpolar organic substances is the use of a
valid field-measured BAF. Field-measured BAFs are preferred to other procedures because they
inherently account for the effects of metabolism, biomagnification, and other factors affecting
bioaccumulation.
The calculation of a field-measured baseline BAF requires information on: (1) a field-
measured BAF based on the total concentration of a chemical in the tissue of the aquatic organism
sampled and the total concentration of the chemical in the ambient water; (2) the fraction of tissue
that is lipid in the aquatic organism of interest; and (3) either the measured or estimated freely
dissolved fraction of the total chemical in the ambient water where the aquatic species were collected
(estimating the freely dissolved fraction for a chemical requires information on the particulate and
dissolved organic carbon content in the ambient water and the K^ of the chemical of interest). The
equation for deriving a field-measured baseline BAF expressed on a freely-dissolved and lipid-
normalized basis is:
Baseline BAF
fd
Measured
fd
(Equation 2.4.9)
- 1
173
-------
where:
Baseline BAFf = BAF expressed on a freely-dissolved and lipid-normalized
basis
Measured B AF-f = BAF based on total concentration in tissue (wet weight basis)
and water
fc = Fraction of the tissue that is lipid
ffd = Fraction of the total chemical that is freely dissolved in the
ambient water
For the derivation of Equation 2.4.9, see Appendix C.
For each trophic level, a species mean baseline BAF is calculated as the geometric mean if
more than one acceptable, measured baseline BAF is available for a given species. For each trophic
level, a trophic level-specificBAF is calculated as the geometric mean of the species mean measured
baseline BAFs. Each of the three components for deriving the baseline BAF are described in further
detail below.
Measured BAF'r
To estimate a measured BAFj, information is needed on the total concentration of the
pollutant in the tissue of the organism and the total concentration of the chemical in ambient water
at the site of sampling. The equation to derive a measured BAFj is:
Measured BAF-f =
Total concentration of chemical in tissue (ug/Ka wet weight)
Total concentration of chemical in the ambient water (ug/L)
(Equation 2.4.10)
Guidance for Measuring Field-Based BAFs
Application of data quality assurance procedures when measuring, estimating, and applying
BAFs is of primary importance. The following procedural and quality assurance requirements
should be met for field-measured BAFs:
• The field studies used should be limited to those that include fish at or near the top
of the aquatic food chain (i.e., in trophic levels 3 and/or 4). In situations where
consumption of lower trophic level organisms represents an important exposure
route, such as certain types of shellfish at trophic level 2, the field study should also
include appropriate target species at this trophic level.
174
-------
The trophic level of the fish species should be determined, taking into account the
life stage(s) consumed and food web structure at the location(s) of interest.
Collection of bioaccumulation field data at a specific site for which criteria are to be
applied and with the species of concern are preferred.
If data cannot be collected from every site for which criteria are to be derived, the site
of the field study should not be so unique that the B AF cannot be extrapolated to
other locations where the criteria and values will be applied.
Samples of the appropriate resident species and the water in which they reside should
be collected and analyzed using appropriate, sensitive, accurate, and precise methods
to determine the concentrations of bioaccumulative chemicals present.
For organic chemicals, the percent lipid should be either measured or reliably;
estimated for the tissue used in the determination of the BAF to permit the measured-
concentration of chemical in the organism's edible tissues to be lipid-normalized.
The concentration of the chemical in the water should be measured in a way that can
be related to particulate organic carbon (POC) and dissolved organic carbon (DOC),
as further described in the forthcoming section on POC and DOC concentrations
(page 14).
For organic chemicals with log Kow greater than four, the concentrations of POC and
DOC in the ambient water should be either measured or reliably estimated.
For inorganic chemicals where lipid normalization does not apply, BAFs should be
used only if they are expressed on a wet weight basis; BAFs reported on a dry weight
basis can be used only if they are converted to a wet weight basis using a conversion
factor that is measured or reliably estimated for the tissue used in the determination
of the BAF.
EPA recommends the use of field-measured BAFs as the first preferred method for
determining BAFs because they incorporate numerous site-specific factors that can affect
bioaccumulation (food web structure, temporal and spatial variation in contaminant levels, and
metabolism of the contaminant). However, in order to ensure that the resulting BAFs accurately
reflect contaminant bioaccumulation and subsequent exposure to the target human population, the
measurement of field-based BAFs must be performed carefully and should consider several factors
that can lead to variability and uncertainty in BAF estimates. Several of these factors are
summarized below. Further discussion of these and other factors is provided in USEPA (1995a;
1995b). EPA is developing additional guidance on performing field studies for determining BAFs
and will provide this guidance for review upon its completion.
175
-------
Selection of Target Species. The choice of the target species for contaminant analysis is one
critical aspect in determining a valid and representative field-measured BAF for establishing A WQC
designed to protect human health. Selection of the target species should be made with knowledge
of the key exposure route(s) involved in bioaccumulationof the contaminant of interest (e.g., uptake
from water, food, sediment/pore water). Several important factors to consider when selecting target
species for contaminant monitoring have been summarized by EPA in their document: Fish
Sampling and Analysis: A Guidance Document for Issuing Fish Consumption Advisories (USEPA,
1993), and are recommended for consideration when identifying target species in BAF studies.
While the objectives offish consumption advisory studies and BAF studies are not entirely identical,
many of the principles described in USEPA (1993) also apply to the determination of BAFs from
field studies.
It is of primary importance that the target species selected be among those species commonly
consumed in the study area and those of commercial, recreational or sustenance fishing value. In
addition, the potential for bioaccumulation of the contaminant(s) of interest should be considered.
Knowledge of the food web structure, likely exposure routes, and contaminant properties (e.g., K^,
for organics) is important for evaluating a species' bioaccumulation potential. Species occupying
trophic level three (e.g., forage fish) or four (e.g., predator fish) are recommended for selection in
BAF studies because they have consistently been among the highest bioaccumulators in the aquatic
food web, particularly for highly hydrophobic chemicals. If possible, the target finfish species
should include at least one species of bottom feeding fish species (trophic level three) and one top
predator species (trophic level four). Including species with different dietary preferences will help
account for the effect of food web structure on bioaccumulation, the effect of which can vary with
the properties of the chemical (i.e., in some cases, bottom feeders can have higher BAFs and in other
cases lower BAFs compared to top predator (piscivorous) species). Organisms occupying trophic
level two (e.g., clams, oysters) should also be sampled if information indicates that consumption of
such organisms is likely to be an important exposure route to contaminants. In addition, the
geographic distribution of the species should be considered in relation to the target human population
intended for protection. Further information pertaining to the selection of target aquatic species for
contaminant analysis for fish advisories is provided in USEPA (1993).
Choice of Sampling Sites. Selection of sampling sites and the frequency at which they are
sampled should take into account numerous factors, many of which relate to the spatial and temporal
variability in the contaminant concentrations in the target aquatic species and environmental media.
If the proper temporal and spatial intervals are not selected, such measurements can lead to erroneous
estimates of bioaccumulation. The sites should be representative of those from which the target
human population are expected to be exposed. In addition, the sampling sites need to be
representative of the area of movement of the target species. This is particularly important for
migratory species which may only spend a portion of the time in the study area of interest.
Temporal and spatial variability can be particularly high for water concentrations of
contaminants. Thus, individual water samples taken at one point in time may not adequately reflect
average exposure to the target species. Water concentrations should be averaged over the
176
-------
approximate time it takes for the target species to reach steady state, which varies depending on the
toxicokineticsof the contaminants in relation to the target organism. For example, chemicals with
high K™, values are expected to reach steady-state in top trophic level organisms much slower than
chemicals with low Kow values, and thus, require greater temporal averaging of water column
concentrations for estimating B AFs. Other factors to consider when determining the frequency of
sampling include the home range of the target species, its life history, and the pattern of contaminant
release (episodic vs. continuous releases). Selection of sampling sites should also consider temporal
and spatial variations in food web structure that may occur across the study area. The desired level
of statistical power should also be considered when determining the number of sampling sites and
replicates.
Biological Considerations. When sampling target species for BAF determinations used in
deriving human health criteria, several biological attributes of the target species should also be
considered. For example, the size/age of the organism can affect the extent of bioaccumulation in
the organism. Young fish can exhibit lower accumulation of some contaminants due to growth
dilution. In addition, the reproductive status (e.g., pre/post spawning) can alter the body burden of
contaminants, with significant contaminant loss observed due to maturation and release of sperm or
eggs. Seasonal variations in lipid content can also lead to differences hi accumulation of
contaminants. In general, the size of the target species should be representative of the size being
consumed by the target human population. If this size range is broad, stratifying sampling strategies
by size class is necessary, particularly when taking composite samples. The timing of sampling
should include the period of most frequent harvesting of the species. Additional discussion of these
and other attributes to consider when sampling finfish and shellfish for contaminant monitoring is
provided in USEPA (1993).
Measurement of Other Important Parameters. For nonpolar organic chemicals, lipid content
of the target species should be measured in the same tissue in which the contaminant was measured
to permit lipid normalization. This will usually be fillet for finfish and edible tissue for shellfish.
In addition, POC and DOC should be measured in the water samples in order to estimate the freely
dissolved fraction. For inorganic chemicals, the bioavailability of various forms of the chemical
should be considered when deciding upon the analyte being measured for the BAF determination.
Where appropriate, BAFs should be expressed for specific forms of the contaminant. For example,
methylmercury is known to be more bioavailablethan inorganic forms of mercury, and the relative
proportions of each can vary significantly over space and time. Thus, BAFs determined for total
mercury without knowledge of the relative proportion of various organic and inorganic forms of
mercury are more uncertain in their applicability to other sites and times than BAFs measured for
specific forms of mercury. Other parameters such as temperature, pH, dissolved oxygen,
conductivity/salinity,total suspended sediments, and sediment grain size should also be measured,
as they may alter the bioavailability and subsequent bioaccumulation of contaminants by aquatic
organisms. EPA will provide additional guidance on the design and conduct of field BAF studies
in its forthcoming guidance document, which is expected to undergo external review in the fall of
1998.
177
-------
Freely Dissolved Fraction of the Chemical in Water (ffd)
Nonpolar organic chemicals can exist in water in several different forms, including freely
dissolved chemicals in the water column, chemicals bound to paniculate matter, and chemicals
bound to dissolved organic matter in the water. The form of the chemical has been shown to affect
bioaccumulation, with the freely dissolved form of a chemical considered to be the best expression
of the bioavailable form to aquatic organisms. Because the amount of chemical that is freely
dissolved may differ among water bodies due to differences in the total organic carbon in the water,
BAFs based on the concentration of freely dissolved chemical will provide the most universal BAF
for organic chemicals when averaging BAFs from different studies. However, BAFs based on the
total concentration of the chemical in water (i.e., the freely dissolved plus that sorbed to particulate
organic carbon and dissolved organic carbon) can often be measured more accurately than BAFs
based on freely dissolved concentrations in water. Therefore, if BAFs based on total water
concentrations are reported in a BAF study, information on the organic carbon content of water (at
the site from which the BAF was measured, if available) is required to predict freely dissolved
concentrations used to determine the baseline BAF. Specifically, the fraction freely dissolved (ffd)
in Equation 2.4.9 must be estimated, using information on the chemical's Kow and both dissolved
and particulate organic carbon contents (DOC and POC) of the water. The equation used to estimate
ffd is:
K
[1 + (POC • K) + (DOC •
where:
POC =
DOC =
(Equation 2.4.11)
Concentration of particulate organic carbon (kg/L) in the ambient water
Concentration of dissolved organic carbon (kg/L) in the ambient water
N-octanol water partition coefficient for the chemical
In this equation, the terms "K^" and "K^/10" are used to estimate the partition coefficients
to POC and DOC, respectively, which have units of L/kg. The scientific basis supporting the
derivation of this equation for estimating the freely dissolved fraction is provided in Appendix D.
POC and DOC Concentrations. As noted above, when converting from the total
concentration of a chemical to a freely dissolved concentration, the POC and DOC should be
obtained from the original study that reports BAFs based on total concentrations of a chemical in
water. However, if the POC and DOC concentrations are not reported in the BAF study, then
reliable estimates of POC and DOC might be available from other studies of the same site used in
the BAF study or closely related sites within the same water body. When using POC/DOC data from
other studies of the same water body for the same or very similar sites, care must be taken to ensure
178
-------
that environmental conditions that may affect POC or DOC concentrations are similar to those in
the BAF study. Information on the spatial and temporal variability of POC and DOC at the site of
interest (and factors influencing this variability) should be used in evaluating the applicability of any
surrogate POC or DOC data for estimating the freely dissolved fraction. For example, differences
in hydrological conditions between the BAF study and the surrogate study (e.g., high vs. low flow
events, mixed vs. stratified water column, tidal cycle differences) and the degree to which such
conditions influence POC and DOC concentrations should be evaluated in deciding whether
surrogate data provide reliable estimates of POC and DOC for the BAF study. Similarly, differences
in other factors which may influence POC and DOC concentrations, such as the sampling season,
sampling depth, proximity to areas of high DOC inputs including wetlands, should be evaluated in
determining the reliability of surrogate data. Additional factors besides the examples listed here may
also be important in determining the reliability of surrogate POC and DOC data for estimating the
baseline BAF.
Guidance on Selecting Appropriate Kow Values
The conversion of total chemical concentrations in water to freely dissolved chemical
concentrations, as well as other procedures discussed in this chapter (including the BSAF method
and use of the food chain model) rely on the K^, for chemicals. A variety of techniques are available
to estimate or predict Kow values, some of which are more or less reliable depending on the K,,w of
the chemical.
As discussed in USEPA 1998a, EPA is proposing and taking comment on two options on
how to select a reliable K^, value. The first option is EPA' s existing guidance published in the Great
Lakes Water Quality Initiative (60 FR 15366, March 23, 1995). A second option is more detailed,
draft guidance on selecting Kow values which EPA has developed and is currently undergoing
external scientific peer review. The salient features of both the GLWQI Kow selection guidance
(option 1) and EPA's new, draft guidance (option 2) are presented below. Additional details of the
new draft KOW selection guidance (option 2) are provided in Appendix F.
Guidance on selecting reliable values of K^ based on the GLWQI approach (option 1) is as
follows:
For chemicals with log K,,w < 4:
Priority Technique
1
Slow-stir
Shake-flask
Generator-column
Measured value from the CLOGP program
179
-------
3 Reverse-phase liquid chromatography on CI8 with extrapolation to zero
percent solvent
4 Reverse-phase liquid chromatography on Clg without extrapolation to zero
percent solvent
5 Calculated by the CLOGP program
For chemicals with log Kow > 4:
Priority Technique
1 Slow-stir
Generator-column
2 Reverse-phase liquid chromatography on Clg with extrapolation to zero
percent solvent
3 Reverse-phase liquid chromatography on C18 without extrapolation to zero
percent solvent
4 Shake-flask
5 Measured value from the CLOGP program
6 Calculated by the CLOGP program
If no measured K^ is available, the K^ must be estimated using the CLOGP program.
Several general points should be kept in mind when using Kow values. Values should be used only
if they were obtained from the original authors or from a critical review that supplied sufficient
information. If more than one "best" K^, value is available for a chemical (i.e., the highest priority
value available), the arithmetic mean of the available log K^s or the geometric mean of the available
K^yS may be used. Because of potential interference due to radioactivity associated with impurities,
values determined by measuring radioactivity in water and/or octanol should be considered less
reliable than values determined by a Kow method of the same priority that employ non-radioactive
techniques except when measurements of parent chemical are done. The values determined using
radioactive methods should be moved down one step in the priority below the values determined
using the non-radioactive technique except when measurements of parent chemical are done.
Because the Kow is an intermediate value in the derivation of a BAF, the value used for the Kow of
a chemical should not be rounded to less than three significant digits. K^, values that are outliers
compared with other values for a chemical should not be used.
180
-------
The salient features of EPA's new draft methodology (option 2) for selecting reliable values
of Kow is described below.
I. Assemble/evaluate experimental and calculated data (e.g.,CLOGP, LOGKOW,
SPARC)
II. If calculated log K,,w's> 8,
A. Develop independent estimates of Kow using:
1. Liquid Chromatography (LC) methods with "appropriate" standards.
(See Appendix F for guidelines for LC application).
2. Structure Activity Relationship (SAR) estimates extrapolated from
similar chemicals where "high quality" measurements are available.
"High quality" SARs are defined in Appendix F of the TSD.
3. Property Reactivity Correlation (PRC) estimates based on other
measured properties (solubility, etc.).
B. If calculated data are in reasonable agreement and are supported by
independent estimates described above, report the average calculated value.
Guidance on determining whether BCOW values are in "reasonable agreement"
are presented in Appendix F of the TSD.
C. If calculated/estimated data do not agree, use professional judgement to
evaluate/blend/weight the calculatedand estimated data to assign a K^ value.
D. Document rationale including relevant statistics.
III. If calculated log Kow's range from 6 - 8,
A. Look for "high quality" measurements. These will generally be slow stir
measurements, the exception being certain classes of compounds where
micro emulsions tend to be less of a problem (i.e., PNA's, shake flask
measurements are good to log K,,w of 6.5).
B. If measured data are available and are in reasonable agreement (both
measurements and calculations), report average measured value.
C. If measured data are in reasonable agreement, but differ from calculated
values, develop independent estimates and apply professional judgement to
evaluate/blend/weightthe measured, calculated and estimated data to assign
K<,w value.
D. If measured data are not in reasonable agreement (or if only one measurement
is available), use II A, B, and C to produce a 'best estimate;' use this value
to evaluate/screen the measured Kow data. Report the average value of
screened data. If no measurements reasonably agree with 'best estimate,'
apply professional judgement to evaluate/blend/weight the measured,
calculated and estimated data to assign Kow.
E. If measured data are unavailable, proceed through II A, B, C and report the
'best estimate.'
F. Document rationale including relevant statistics.
181
-------
IV. If calculated log Kow's < 6,
A. Proceed as in III. Slow stir is the preferred method but shake flask data can
be considered for all chemicals if sufficient attention has been given to
emulsion problems in the measurement.
The general operational guidelines for EPA's new draft methodology for selecting K^ values
are as follows:
1.
2.
3.
4.
For chemicals with log K^ > 5, it is highly unlikely to find multiple "high quality"
measurements. (Note: "high quality" is data judged to be reliable based on the
guidelines presented in Appendix F of the TSD).
"High Quality" measured data are preferred over estimates, but due to the scarcity of
'high quality' data, the use of estimates is important in assigning ]£<,„, 's.
KOW measurements by slow stir are extendable to 10s. Shake flask K,,w measurements
are extendable to 106 with sufficient attention to micro emulsion effects; for classes
of chemicals that are not highly sensitive to emulsion effects (i.e., PNA's) this range
may extend to 106-5.
What is to be considered reasonable agreement in log Kow data (measured or
estimated) depends primarily on the log K^ magnitude. The following standards for
data agreement have been set for this guidance: 0.5 for log Kow > 7; 0.4 for 6 <, log
K^?; 0.3 for log K^ <6.
5.
Statistical methods should be applied to data as appropriate but application is limited
due to the scarcity of data, and the determinate/methodicnature of most measurement
error(s).
The various techniques are summarized as follows:
• The slow-stir method requires adding the test chemical to a reaction flask which
contains a water and octanol phase. The chemical partitions to these two phases
under conditions of slow stirring the flask. After the phases are allowed to separate,
the concentration of the test chemical in each phase is determined (Brooke et al.,
1986). This method is easy to use and can be replicated with a high degree of
confidence. Emulsions, which can contaminate the aqueous phase and influence the
observed K,,w values, can be prevented, and high Kow values can be obtained easily
(de Bruijn et al., 1989). In general, there is reasonable agreement between the slow
stir method and literature data obtained using the generator-column method. For log
KOW values less than 4.5, data agree well with Kows determined based on the shake-
flask method (de Bruijn et al., 1989).
182
-------
• The shake-flask method also involves adding the chemical to a reaction flask with a
mixture of octanol and water. In this method, however, the flask is shaken to obtain
partitioning of the chemical between the octanol and water phases (OECD, 1981).
Several researchers have found that the shake-flask technique is acceptable only for
chemicals with log K,,ws less than certain values. Some researchers found that the
shake-flask technique has been reported to be acceptable only for chemicals which
have log K^ less than 4 (Karickhoff et al., 1979; Konemann et al., 1979; Braumann
and Grimme, 1981; Harnisch et al., 1983; Brooke et al., 1990). Others have found
that the technique is acceptable for chemicals with slightly higher log K,,w values.
Brooke et al. (1986) compared techniques and decided that the shake-flask technique
is acceptable for chemicals with Jog K,,^ up to 5, whereas Chessells et al. (1991)
stated that this technique is acceptable for log K^, values up to about 5.5.
• The generator-column method involves filling a column with an inert material
(silanized Chromosorb W or glass beads) that is coated with water-saturated octanol
and contains the test chemical. Pumping water through the column results in an
aqueous solution in equilibrium with the octanol phase. The water that leaves the
column is extracted with specifically either an organic solvent or a C18 column that
is then eluted with hexane or methanol (DeVoe et al., 1981; Woodburn et al., 1984;
Miller etal., 1984).
• The reverse-phase liquid chromatography method involves adding the test chemical
in a polar mobile phase (such as water or water-methanol) to a hydrophobic porous
stationary phase (the C,g w-alkanes covalently bound to a silica support). The
chemical partitions between the column and the polar aqueous phase. Revalues are
estimated from linear equations between the Kow and retention indices that are
derived for reference chemicals (Konemann et al., 1979; Veith et al., 1979;
McDuffie, 1981; Garst and Wilson, 1984).
• The CLOGPprogram is a computer program that contains measured Kow values for
some chemicals and can calculate K,,w values for additional chemicals based on
similarities in chemical structure between chemicals with measured Kow values and
chemicals for which Kows are to be determined. The method used to calculate the K^
values is described in Hansch and Leo (1979).
• SPARC (SPARC Performs Automated Reasoning in Chemistry) is a mechanistic
model developed at the Ecosystems Research Division of the National Exposure
Research Laboratory of the Office of Research and Development of the U.S.
Environmental Protection Agency by Sam Karickhoff, Lionel Carreira, and co-
workers.
In some situations, available data may require determination of a single K^, value for a class
of chemicals or a or mixture of closely related chemicals (e.g., when toxicity data are class- or
183
-------
mixture-specific). However, it is not possible to determine experimentally a valid K^, for a
substance that is a mixture of chemicals (e.g., PCBs, toxaphene, chlordane). For calculating the
composite freely dissolved fraction used to adjust a composite field-measured BAF to a composite
baseline BAF, a composite K^ value of the mixture can be calculated based on the sum of the total
concentration of the mixture components in water (e.g., individual congeners for PCBs), the sum of
the dissolved mixture components in water, and the DOC and POC from the site for which the BAF
was measured. The equation used to derive the composite K<,w for use in determining a composite
baseline BAF is:
Composite K
DOC
10
+ POC
EC'
*— ' w
Ec
fd
- 1
where:
C" =
1, 2, ... n individual mixture components (e.g., congeners for
PCBs).
total concentration of the mixture component in water.
freely dissolved concentration of the mixture component in water.
Notably, calculation of a composite K,,w is just one of a series of steps involved in deriving
composite baseline B AFs and AWQC B AFs for chemical mixtures for which single toxicity values
apply. These steps include: ( 1 ) first determining a composite initial total BAF (analogous to the total
field-measured BAF for individual chemicals), (2) determining a composite baseline BAF, using the
composite freely dissolved fraction and composite Kow derived using the equation above, (3)
calculating a composite AWQC BAF from the composite baseline BAF based on the composite
freely dissolved fraction at the AWQC site(s). This last step requires calculation of a second
composite K^, which is used to determine the composite freely dissolved fraction at the AWQC
site(s) using POC and DOC data from the AWQC site(s). Additional details of the steps required
in determining a composite K,>w for deriving composite baseline B AFs and composite AWQC B AFs
(including example calculations for PCBs) are provided in 62 FR 117250 (March 12, 1997).
Lipid Normalization of Data
Partitioning of organic chemicals into aquatic organisms has been shown to be a function of
the lipid content of the organism (Mackay, 1982; Connolly and Pederson, 1988; Thomann, 1989).
For this reason, EPA assumes that BAFs and BCFs for lipophilic organic chemicals are directly
proportional to the percent lipid in the tissue or whole body of the organism of interest. For
example, an organism with two percent lipid content would accumulate twice the amount of a
184
-------
chemical as an organism with one percent lipid content, all else being equal. This assumption has
been extensively evaluated in the literature and is generally accepted. To account for the influence
of the lipid content on the BAF or BCF, EPA recommends normalizing the BAF or BCF to the
percent lipid in the fish. This procedure is consistent with other EPA guidance on bioaccumulation
(Stephan et al., 1985; USEPA, 1991).
To compare BAFs and BCFs that have been measured in fish that have different lipid
contents, EPA recommends that BAFs and BCFs be normalized by dividing by the mean lipid
fraction of the aquatic organism. Whole body and edible tissue BAFs and BCFs are normalized
using the respective whole body and edible tissue fraction lipid values. Since lipid content is known
to vary from one tissue to another and from one aquatic species to another, EPA recommends the
percent lipid used to normalize the BAF or BCF (whole body or edible tissue) be obtained from the
BAF or BCF study. Unless comparability can be determined across organisms, the fraction lipid
should be determined for the test organism. Lipid content of the fish tissue is affected by the age,
sex and diet of the fish, by the season the fish are sampled, and by differing environmental
conditions. Therefore, it is generally necessary to determine an average percent lipid value for the
test organisms.
EPA recommends using a gravimetric method for determining the percent lipid value
(USEPA, 1995a). The method is easy to use and is employed by many laboratories. It should be
noted that the solvent used to determine lipid content has been shown to affect the percent lipid
values measured in some studies because different solvent systems extract differing fractions of total
lipids (Lapin and Chernova, 1969; Randall et al., 1991; and Cabrini et al., 1992). These authors note
that percent lipid values can vary by as much as a factor of four depending on the solvent system
used. To ensure consistency among States, EPA recommends for lipid analyses that the method of
Bligh and Dyer (1959) which uses chloroform/methanolas an extraction solvent or the lower toxicity
solvent modification of this the method by Kara and Radin (1978) which uses hexane/isopropanol
as an extraction solvent. Other extraction solvents for lipid analyses, e.g., hexane/acetone and
dichloromethane, might provide equivalent results when used with appropriate sample sizes and
extraction times (Honeycutt et al., 1995, de Boer, J., 1988).
In addition to the effect of the solvent on lipid analysis, additional factors may affect
variability of results if they are not adequately controlled (USEPA, 1995a). Use of alcohol as a
solvent may overestimate total lipids because non-lipid material may also be extracted. Several
factors, including solvent contaminants, lipid decomposition from exposure to oxygen or light, and
lipid degradation from changes in pH during cleanup can lead to underestimation of total lipids.
Finally, high temperature may decompose lipid material. Laboratories should consider these sources
of error when conducting and evaluating results of lipid analyses (USEPA, 1995a).
185
-------
2.4.4.2 Baseline BAF Derived from Biota-Sediment Accumulation Factors
(BSAFs)
When acceptable field-measured values of the BAF are not available for a nonpolar organic
chemical, EPA recommends the use of a BSAF to predict the BAF as the second procedure in the
BAF data preference hierarchy. Although BSAFs may be used for measuring and predicting
bioaccumulation directly from concentrations of chemicals in surface sediment, they also can be
useful in estimating a BAF as noted by Cook et al. (1993). Because BSAFs are based on field data
and incorporate effects of metabolism, biomagnification, growth, and other factors, BAFs estimated
from BSAFs will incorporate the net effect of all these factors. The BSAF approach is particularly
beneficial for developing water quality criteria for chemicals such as polychlorinated dibenzo-p-
dioxins, dibenzofurans, and certain biphenyl congeners. These chemicals are detectable in fish
tissues and sediments but are difficultto measure in the water column and are subject to metabolism.
Predicting BAFs from BSAFs requires several steps. First, BSAFs must be measured for the
chemical of interest and for one or more reference chemicals for which measured BAFs are also
available. Second, the relationship between the BSAFs for the chemical of interest and the reference
chemical and the relationship between the chemicals' K^, values must be determined. Finally,
information on the BSAF and K^ relationships and the BAF of the reference chemical(s) should be
used to determine the BAF for the chemical of interest. The following sections describe the
methodology for determining BAFs from BSAFs, the data requirements, and the application and
validation of this procedure for estimating BAFs using data from Lake Ontario.
Determination of BSAF Values
As shown hi the following equation, the BSAF is determined by relating lipid-normalized
concentrations of chemicals in an organism to organic carbon-normalized concentrations of the
chemicals in surface sediment samples associated with the average exposure environment of the
organism.
BSAF =
(Equation 2.4.12)
where:
Lipid-normalized concentration of the chemical in tissues of biota Cug/g
lipid)
Organic carbon-normalized concentration of the chemical in the surface
sediment (/ug/g sediment organic carbon)
186
-------
The lipid-normalized concentration of a chemical in an organism (Q) is determined by:
c -fl
where:
C,
(Equation 2.4. 13)
Concentration of the chemical in the wet tissue (either whole organism or
specified tissue) (/ug/g)
f{ = Fraction lipid content in the organism
The organic carbon-normalized concentration of a chemical in sediment (Csoc) is determined by:
sed
where:
•"sed
(Equation 2.4.14)
= Concentration of chemical in sediment (yUg/g sediment)
foc = Fraction organic carbon in sediment
BSAFs are most useful when measured under steady state or near steady-state conditions in
which chemical concentrations in water are linked to slowly changing concentrations in sediment.
However, because BSAFs are rarely measured for ecosystems which are at equilibrium, the BSAF
inherently includes a measure of the "disequilibrium" of the ecosystem. This disequilibrium can be
assessed for chemicals with log Kow > 3 with the following relationship:
BSAF =
Cfd.K
= DK *
bs
« DK • 2
bs
(Equation 2.4.15)
187
-------
K,
where:
= Concentration of freely dissolved chemical (associated with water) in the
tissues of biota (ug/g wet tissue)
Concentration of freely dissolved chemical (associated with pore water) in
the sediment Cug/g sediment organic carbon)
= Lipid-water equilibrium partition coefficient (C/C^)
KJOC = Sediment organic carbon-water equilibrium partition coefficient (Csoc/Cf)
Dbs = Disequilibrium (fugacity) ratio between biota and sediment (C^/Cf)
Measured BSAFs may range widely for different chemicals depending on Kc, K^, and the
actual ratio of Cbd to Cf. However, at equilibrium, the ratio between the freely dissolved chemical
in the tissue water to sediment pore water (Dbs) is one. Thus, the BSAF under equilibrium conditions
is equal to the ratio K^/K^ (which is thought to range from 1-4)23. When chemical equilibrium
between sediment and biota does not exist, the BSAF will equal the disequilibrium (fugacity) ratio
between biota and sediment (Dbs = C^/Cf) times the ratio of the equilibrium partition coefficients
(approximately 2).
The deviation of Dbs from the equilibrium value of 1.0 is determined by the net effect of all
factors which contribute to the disequilibrium between sediment and aquatic organisms. A
disequilibrium ratio (Dbs) greater than one can occur due to biomagnification or when surface
sediment has not reached steady-state with water. A disequilibrium ratio (Dbs) less than one can
occur as a result of kinetic limitations for chemical transfer from sediment to water, water to food
chain, and biological processes (such as growth or biotransformation of the chemical in the animal
and its food chain). BSAFs are most useful when measured under steady-state conditions. BSAFs
measured for systems with new chemical loadings or rapid increases in loadings may be unreliable
due to underestimation of steady-state Csocs.
Relationship ofBAFs to BSAFs
Differences between BSAFs for different organic chemicals are good measures of the relative
bioaccumulationpotentials of the chemicals. When calculated from a common organism-sediment
sample set, chemical-specific differences in BSAFs primarily reflect the net effect of
biomagnification, metabolism, bioenergetics, and bioavailability factors on each chemical's
disequilibrium ratio between biota and sediment. Thus, the relationship between the BSAF for the
test chemical / and the BSAF for the reference chemical r can be used with additional information
"Because K, and K^ are of similar magnitude and vary in proportion to one another, the BSAF at equilibrium is expected to be at or near
unity.
188
-------
(KoW values for all chemicals and the B AF for the reference chemicals) to predict a B AF for chemical
/'. This approach is consistent with previously proposed guidance, in which ratios of BSAFs for
PCDDs and PCDFs to TCDD were proposed for evaluation of TCDD toxic equivalency associated
with complex mixtures of the dioxin and furan congeners (see 60 FR 15366 for discussion of
bioequivalency factors).
The calculation of the BAF from the BSAF is as follows:
(BSAF). • (K ).
(Baseline BAF{fd). = (Baseline BAF{fd)r • ' ow '
(BSAF) • (K )
7r ow'r
(Equation 2.4.16)
where:
(Baseline BAFJd)i
BAF expressed on a freely-dissolved and lipid-normalized
basis for chemical of interest "i"
(Baseline BAFf)r
(BSAF);
(BSAF)r
(Kow)r
BAF expressed on a freely-dissolved and lipid-normalized
basis for reference chemical "r"
BSAF for chemical"!"
BSAF for the reference chemical "r"
octanol-water partition coefficient for chemical "i"
octanol-water partition coefficient for the reference chemical
66 M
Appendix E presents the derivation of this equation using the general BAF equation relating
concentration in tissue to concentration in water, a relationship between concentrations in sediment
organic carbon and water, and assumptions about equilibrium between water and sediment.
Note that BAF[ds calculated from BSAFs will incorporate any errors associated with
measurement of the BAF[d for the reference chemical and the Kows for both the reference and
unknown chemicals. Such errors can be minimized by comparing results from several reference
chemicals and assuring consistent use of freely dissolved water concentration (C ™) values which are
adjusted for dissolved organic carbon binding effects on the fraction of each chemical that is freely
dissolved (ffd) in unfiltered, filtered, or centrifuged water samples. Other errors may be introduced
by using values based on non-steady state external loading rates or chemicals with strongly reduced
C ™ due to rapid volatilization from water. When selecting K^, values for use in estimating BAFs
from BSAFs, consideration should be given to the similarity of Kow measurement techniques
189
-------
between the reference and target chemicals, in addition to the guidance previously described for
selecting representative K,,w values.
The trophic level to which the baseline BAF applies is the same as the trophic level of the
organisms used in the determination of the BSAF. For each trophic level, a species mean baseline
BAF is calculated as the geometric mean if more than one acceptable baseline BAF is predicted from
BSAFs for a given species. For each trophic level, a trophic level-specific BAF is calculated as the
geometric mean of the acceptable species mean baseline BAFs derived using BSAFs.
Procedural and Quality Assurance Requirements
EPA recommends certain requirements for measuring the data needed for this procedure.
These requirements, described below, apply to BAF values for the reference chemicals, the measured
BSAFs for all chemicals, and the K,,w values used in this procedure.
The data requirements for measuring BAF values that were noted in Section 2.4.4.1 (Field-
Measured BAFs) are also applicable to the measurement of BAFf values to assure reliable values
for the reference chemicals. Data on several reference chemicals should be obtained for use in the
analysis to ensure that predictions are more robust than those that would be obtained using only one
reference chemical. The water sample analyses should approximate the average exposure of the
organism and its food chain over a time period that is most appropriate for the chemical, organism,
and ecosystem. It is preferable to choose at least some reference chemicals that have similar log
KowS and chemical class characteristics as the test chemicals for which the BAF is to be determined.
In addition, for consistency among reference chemicals, each freely dissolved water concentration
used to calculate a BAFf should be based on a consistent adjustment of the concentration of total
chemical in water for DOC and POC using the relationship described in the section titled "Freely
Dissolved Fraction of Chemical in Water."
For measured BSAFs, chemical concentrations in surface sediment and in biota and data on
the percent organic carbon in surface sediment samples are needed. The following procedural and
quality assurance requirements should be met for determining the field-measured BSAFs:
• The field studies used should be limited to those conducted with fish at or near the
top of the aquatic food chain (i.e., in trophic levels 3 and/or 4). In situations where
consumption of lower trophic level organisms represents an important exposure
route, such as certain types of shellfish at trophic level 2, the field study should also
include appropriate target species at this trophic level.
• Samples of surface sediments (0-1 cm is ideal) should be from locations in which
sediment is regularly deposited and is representative of average surface sediment in
the vicinity of the organism.
• The KowS used should be of acceptable quality as described in Section 2.4.4.1 above.
190
-------
The site of the field study should not be so unique that the resulting BAF cannot be
extrapolated to other locations where the criteria and values will be applied.
The percent lipid should be either measured or reliably estimated for the tissue used
in the determination of the BAF.
Application of BSAF Procedure for Predicting Lake Ontario and Green Bay
BAFfs
To demonstrate the use of the BSAF procedure to predict B AFs, EPA has calculated BAFfds
from BSAFs using two independent data sets from Lake Ontario and one from Green Bay. These
data sets come from Oliver and Niimi (1988), the EPA Lake Ontario TCDD Bioaccumulation Study
(USEPA, 1990), and the EPA Green Bay/Fox River Mass Balance Study.
The first data set (Oliver and Niimi, 1988) has been used extensively for construction of food
chain models of bioaccumulationand calculation of food chain multipliers, biomagnificationfactors
and BAFfs from chemical concentrations determined in organisms and water. Oliver and Niimi
(1988) also collected surface sediment data which allows calculation of lakewide average BSAFs.
These data were collected from 1981 to 1984 for PCB congeners and other chlorinated organics.
The second data set (from the TCDD Bioaccumulation Study) includes extensive, samples
of fish and sediment collected in 1987 from Lake Ontario. Samples from this study were later
analyzed for PCDD, PCDF, PCB congeners, and some organochlorinepesticides at EPA. Although
data from the TCDD Bioaccumulation Study have not been published, they are useful to show a
comparison with BAFfs calculated from Oliver and Niimi samples and to provide BAF,fds for
additional organic chemicals not measured by Oliver and Niimi (1988).
Four reference chemicals (the PCB congeners 52, 105 and 118 and DDT) were used for
evaluating chemicals from both Oliver and Niimi (1988) and the TCDD Bioaccumulation Study in
order to examine the variability introduced by the choice of reference chemical.
The third study, the Green Bay/Fox River Mass Balance Study, involved extensive sampling
of water, sediment, and fish in Green Bay in 1989. Brown trout BAF[ds were calculated from PCB
BSAFs measured in the mid-bay region using PCB congeners 52 and 118 as reference chemicals.
The reference chemical BAFfs were determined using water and brown trout data from the same
region.
Tables 2.4.1a and 2.4.1b present the predicted BAFfs from all three data sets as well as
measured BAFfs from Oliver and Niimi (1988) and the TCDD Bioaccumulation Study. The
geometric means of the BAFf predicted using the Lake Ontario data (Oliver and Niimi, 1988; TCDD
Bioaccumulation Study) are reported in Table 2.4.2.
191
-------
There are several assumptions and additional data used for these evaluations. First, the water
analyses of Oliver and Niimi (1988) were adjusted for an estimated 2 mg/L residual dissolved
organic carbon concentration in the centrifuged water (assuming no residual paniculate organic
carbon after centrifuging) and an estimated Kdoc = Kow/10 in order to calculate a freely dissolved
water concentration from ffd (see Section 2.4.4.1 on total vs. freely dissolved concentrations and
Appendix D for calculation of water concentrations). Concentrations of freely dissolved PCBs from
Green Bay were also calculated on the basis of dissolved organic carbon in the water samples and
an assumed K^ = K^/l 0. Log K,,w values were taken from a variety of sources. Log K^s for PCBs
are those reported by Hawker and Connell (1988). Log K<,ws for PCDDs and PCDFs are those
estimated by Burkhard and Kuehl (1986) except for the penta-, hexa-, and hepta-chlorinated
dibenzofurans which were estimated on the basis of assumed similarity to the trends reported for the
PCDDs by Burkhard and Kuehl (1986).
Evaluation ofBAF^s Calculated from Lake Ontario and Green Bay BSAFs
The validity of the BSAF method for predicting BAFs is evaluated in this section using
several approaches: (1) correlating measured vs. predictedlog BAFfs from the same lake and same
study (i.e., Lake Ontario, Oliver and Niimi, 1988), (2) correlating measured vs. predicted log BAFfs
from the same lake (Ontario) but separate studies (Oliver and Niimi, 1988, U.S. EPA, 1990), (3)
comparisons of predicted BAFfs with Kow values, and (4) comparisons of predicted and measured
BAFfs from different lakes (i.e., Lake Ontario and Green Bay, Lake Michigan). These comparisons
were based on data presented in Tables 2.4.la and 2.4.Ib.
Exhibit 2.4.1 illustrates that measured log BAFfs calculated using water data from Oliver
and Niimi (1988) generally agree with log BAFfs predicted from BSAFs determined using sediment
data from the same study. The correlation coefficient (r) for the correlation of BAFs using data from
Tables 2.4.1a and 2.4.1b is 0.92 and indicates that the data are well correlated. One deviation of
predicted BAFs from measured BAFs should be noted, however. For chlorinated benzenes and
toluenes, BAFfs predicted from BSAFs are underestimated compared with measured BAF[ds. This
underestimation may be due to altered water-sediment fugacity gradient in response to rapid
volatilizationfrom water. The better agreement between measured and predicted BAFfs for PCBs,
on the other hand, is facilitated by the lower volatilization of PCBs from water. In addition to the
correlation shown in Exhibit 2.4.1, the ratios between the BAFs (which indicates the magnitude of
difference between the values) were plotted as a frequency distribution, as shown in Exhibit 2.4.2.
The magnitude of difference between these two BAFs is within a factor of four in the majority of
cases.
Exhibit 2.4.3 demonstrates the predictability of BAFs for the same chemical but based on
BSAF from different studies. The predicted log BAFfs using data from the EPA TCDD
Bioaccumulation Study (U.S. EPA, 1990) (collected several years after the Oliver and Niimi samples
were collected) correlate equally well with the predicted log BAFfs calculated from Oliver and
Niimi (1988) data. An r value of 0.94 was obtained for the correlation of BAFs using data from
Tables 2.4.1a and 2.4.1b. Exhibit 2.4.4 shows that for the majority of chemicals, predicted BAFs
192
-------
from the TCDD Bioaccumulation Study are within a factor of two of predicted BAFs from Oliver
and Niimi (1988); measured and predicted BAFs for all chemicals are within a factor often of each
other.
Exhibit 2.4.5 shows the relationship of log BAFfs calculated from EPA BSAFs using lake
trout data from the TCDD Bioaccumulation Study (Cook et al., 1994) to log K^s. The
bioaccumulativePCDDsand PCDFs (2,3,7,8-chlorinated) have BAFfs 10- to 1,000-fold less than
PCBs with similar K^s, which is expected due to PCDD and PCDF metabolism in fish. It should
be noted, however, that some of the chlordane and nonachlor B AFf s do not have the expected
correlations with K,^. This is shown in Exhibit 2.4.5 by the BAFf s for five of six chlordanes and
nonachlors that are much greater than those for PCBs with the same estimated log K^. This finding
is unexpected because PCBs are not metabolized in fish and would be expected to have higher BAFs
than other chemicals with the same KOW values. Therefore, the log Kow values chosen here for the
chlordanes and nonachlors may be significantly underestimated.
All of the above correlations were based on the BSAF procedure using the Oliver and Niimi
(1988) Lake Ontario salmonid B AFf for PCB congener 52 as a reference chemical. As noted earlier,
the BSAF procedure is strengthened through use of several reference chemicals with both a range
of KOWS and ability to be accurately measured in water. Using additional reference chemicals (PCB
congeners 105 and 118 and DDT) results in correlations with other measured and predicted BAF[ds
from Tables 2.4. la and 2.4. Ib that are very similar to comparisons seen using PCB congener 52 as
a reference chemical.
A good test for robustness of the BSAF procedure for predicting B AFf s is comparison of two
independent data sets based on different ecosystems and conditions. Such a comparison can be made
for bioaccumulationof PCBs in Lake Ontario fish and Green Bay fish. Although both ecosystems
are specific to the Great Lakes area, Green Bay is a shallower, smaller, and more eutrophic body of
water than Lake Ontario. The correlation between the PCB log BAFf s for brown trout predicted
from BSAFs using Green Bay data and measured log B AFf s based on Oliver and Niimi (1988) data
is shown in Exhibit 2.4.6. An r value of 0.91 was obtained for the correlation of these log BAFs
(using data from Tables 2.4.1 a and 2.4.1 b), showing that the values are well correlated. Exhibit 2.4.7
shows that, most frequently, the BAFs differ from each other by less than a factor of two, and all
chemical BAFs are within a factor of ten of each other. The correlation between predicted log
B AFf s from Green Bay (for brown trout) and predicted lake trout log BAFs using Oliver and Niimi
(1988) salmonid and water measurements and lake trout BSAFs from the EPA TCDD
Bioaccumulation Study is shown in Exhibit 2.4.8. The r value for the relationship between these log
BAFs using data from Tables 2.4. la and 2.4. Ib is 0.90. As shown in Exhibit 2.4.9, the most frequent
difference in these BAFs is less than or equal to a factor of two. However, for PCB congener 198,
the difference between BAFs is 85, as indicated by the value farthest to the right in the exhibit.
Despite the complex exposures of Green Bay fish (which result from movement and
interaction of biota through gradients of decreasing PCBs, nutrients and suspended organic carbon
extending from the Fox River to the outer bay and Lake Michigan), good agreement exists between
193
-------
Green Bay brown trout predicted log BAFs and both field-measured log BAFfs and predicted log
BAFf s from BSAFs from Lake Ontario, using PCB 52 as a reference chemical. Although not shown,
good agreement also exists for the predicted BAFs using PCB 118 as the reference chemical. In
addition to the above comparisons, correlations of predicted log BAFs with log K^. values from
Green Bay show relationships that are similar to the log BAF -log Kow relationship for predicted
BAF[ds from Lake Ontario data.
Based on the above correlations and ratios, the BSAF method appears to work well not only
for predicting BAFs using data from the same system (Lake Ontario) but also for predicting BAFs
between systems (Green Bay vs. Lake Ontario). These evaluations support the use of the BSAF
method for predicting BAFs.
194
-------
Table 2.4.1a: Great Lakes Trout BAFjds Calculated from Measured BSAFs/BAFs
Chemical
dieldrin
ddt
dde
ddd
mirex
photomirex
g-chlordane
t-chlordane
c-chlordane
t-nonachlor
c-nonachlor
alpha-hch
gamma-hch
hcbd
ocs
hcb
pcb
1235tcb
1245tcb
1234tcb
13Stcb
124tcb
123tcb
245tct
236tct
pet
Total-PCB
PCB 5
PCB 6
Log
K»
5.3
6.45
6.76
6.06
6.89
6.89
6.0
6.0
6.0
6.0
6.0
3.78
3.67
4.84
6.29
5.6
5.11
4.56
4.56
4.59
4.17
3.99
4.1
4.93
4.93
6.36
6.14
4.97
5.06
Measured Values
BSAF
Ol. & Niimi
(1988)
1.09
4.14
0.28
1.43
5.48
2.22
2.45
0.69
0.98
0.09
0.04
0.01
1.85
log BAF
OI. & Niimi
(1988)
7.78
8.35
7.00
8.13
8.07
6.79
4.69
4.93
8.07
6.40
5.81
5.07
7.81
BSAF
EPA
(1990)
6.65
1.67
7.7
1.31
2.00
4.77
10.5
0.51
0.36
Predicted BAF™
log BAF
Ol. & Niimi
ref PCB 52
7.87
8.76
6.90
8.43
9.01
7.73
5.55
4.89
7.67
5.95
5.07
4.11
7.79
log BAF
EPA
ref PCB 52
7.67
8.22
9.19
8.55
7.85
8.23
8.57
7.25
6.16
log BAF
OI. & Niimi
ref PCB 105
7.54
8.43
6.56
8.09
8.68
7.40
5.22
4.56
7.33
5.62
4.73
3.78
7,46
log BAF
EPA
ref PCB 105
6.95
7.5
8.47
7.84
7.13
7.51
7.85
6.54
5.44
195
-------
Table 2.4.Ia: Great Lakes Trout BAF?d s Calculated from Measured BSAFs/BAFs
Chemical
PCB8
PCB12
PCB13
PCB16
PCS 17
PCB1S
PCS 22
PCB25
PCB26
PCB32
PCB33
PCB40
PCB 42
PCB 44
PCS 45
PCB 46
PCB 49
PCB 52
PCB 53
PCB 63
PCB 64
PCB 66
PCB 74
PCB 77
PCB SI
PCB 82
PCB 83
PCB 84
PCB 85
PCB 87
Log
1C.
5.07
S.22
5.29
5.16
5.25
5.24
5.58
5.67
5.66
5.44
5.60
5.66
5.76
5.75
5.53
5.53
5.85
5.84
5.62
6.17
5.95
6.20
6.20
6.36
6.36
6.20
6.26
6.04
6.30
6.29
Measured Values
BSAF
Ol. & Niimi
(1988)
0.15
0.26
0.21
0.25
1.72
0.18
0.15
0.10
0.52
0.48
0.57
0.69
0.61
1.84
0.73
0.85
3.45
2.45
3.04
1.45
log BAF
Ol. & Niimi
(1988)
5.92
5.52
5.75
6.39
6.76
5.32
6.55
7.49
6.96
7.13
7.01
6.51
7.51
7.79
7.66
8.13
8.28
7.89
BSAF
EPA
(1990)
0.44
0.99
0.1
0.27
0.33
0.44
0.49
0.18
0.4
0.22
0.02
0.42
0.82
0.61
0.29
0.67
0.18
1.33
1.29
1.37
Predicted BApf
log BAF
Ol. & Niimi
refPCB52
5.80
6.05
6.28
6.44
7.28
6.09
6.15
6.06
6.86
6.82
6.67
7.07
7.01
7.27
7.20
7.52
8.12
7.97
7.91
7.85
log BAF
EPA
refPCBS2
6.41
6.79
5.79
6.56
6.74
6.85
6.84
6.46
6.90
6.42
5.38
7.01
7.63
7.53
7.37
. 7.73
7.00
7.93
7.96
7.97
log BAF
Ol. & Niimi
refPCBlOS
5.47
5.71
5.95
6.11
6.94
5.75
5.82
5.72
6.53
6.48
6.34
6.74
6.67
6.93
6.86
7.18
7.79
7.64
7.57
7.51
log BAF
EPA
refPCBlOS
5.69
6.07
5.07
5.84
6.02
6.13
6.12
5.74
6.18
5.70
4.66
6.29
6.91
6.81
6.65
7.01
6.28
7.21
7.24
7.25
196
-------
Table 2.4.1a: Great Lakes Trout BAF^s Calculated from Measured BSAFs/BAFs
Chemical
PCB91
PCB92
PCS 95
PCB97
PCB99
PCB 100
PCB 101
PCB 105
PCB 110
PCB 118
PCB 119
PCB 126
PCB 128
PCB 129
PCB 130
PCB 132
PCB 136
PCB 138
PCB 141
PCB 146
PCB 149
PCB 151
PCB 153
PCB 156
PCB 158
PCB 167
PCB 171
PCB 172
PCB 174
PCB 177
Log
K~
6.13
6.35
6.13
6.29
6.39
6.23
6.38
6.65
6.48
6.74
6.58
6.89
6.74
6.73
6.8
6.58
6.22
6.83
6.82
6.89
6.67
6.64
6.92
7.18
7.02
7.27
7.11
7.33
7.11
7.08
Measured Values
BSAF
Ol. & Niimi
(1988)
1.25
1.43
1.40
0.68
2.45
2.70
1.53
4.09
3.61
1.75
0.87
10.87
4.25
2.75
3.22
2.33
3.38
4.22
3.97
2.71
1.54
3.53
log BAF
Ol. & Niimi
(1988)
6.92
8.11
7.25
7.39
7.45
8.13
7.79
8.15
7.56
7.37
8.30
8.32
8.73
7.99
8.51
8.32
8.74
9.01
BSAF
EPA
(1990)
0.64
0.28
1.51
1.78
1.06
4.49
0.82
1.72
3.83
3.21
2.78
1.13
2.15
1.74
1.25
0.93
1.65
1.91
1.52
0.69
1.36
1.25
1.91
Predicted BAF™
log BAF
Ol. & Niimi
refPCBS2
7.61
7.89
7.66
7.61
8.15
8.47
8.05
8.74
8.68
8.36
7.90
8.64
8.84
8.64
8.78
8.42
8.55
8.93
9.16
8.93
8.68
9.01
log BAF
EPA
refPCB52
7.48
7.28
8.12
8.03
7.95
8.85
7.94
8.52
S.71
8.94
8.73
8.33
8.68
8.61
8.53
8.19
8.40
8.75
8.75
8.66
9.01
8.75
8.91
log BAF
Ol. & Niimi
ref PCB 105
7.28
7.55
7.33
7.27
7.82
8.13
7.71
8.40
8.35
8.02
7.57
8.30
8.51
8.31
8.45
8.09
8.22
8.59
8.83
8.59
8.35
8.68
log BAF
EPA
ref PCB 105
6.76
6.56
7.40
7.31
7.23
8.13
7.22
7.80
7.99
8.22
8.01
7.61
7.96
7.89
7.81
7.47
7.69
8.03
8.03
7.94
8.29
8.03
8.19
197
-------
Table 2.4.1a: Great Lakes Trout BAFJ's Calculated from Measured BSAFs/BAFs
Chemical
PCB178
PCS 180
PCBI83
PCB185
PCS 189
PCB194
PCB195
PCS 197
PCB 198
PCB201
PCB20S
PCB 206
PCB207
PCB209
PCB 24+27
PCB 28*31
PCB 37+42
PCB 47+48
PCB 41+64+71
PCB 56+60
PCB 70+76
PCB 66+95
PCB 56+60+81
PCB 84+92
PCB 87+97
PCB 137+176
PCB 138+163
PCB 156+171+202
PCB 182+187
PCB 157+200
Log
K«
7.14
7.36
7.20
7.11
7.71
7.80
7.56
7.3
7.62
7.62
8.00
8.09
7.74
8.18
5.40
5.67
5.8
5.82
5.87
6.11
6.17
6.17
6.19
62
6.29
6.8
6.91
7.18
7.19
7.23
Measured Values
BSAF
Ol. & Niimi
(1988)
4.48
3.78
5.62
1.55
1.53
1.90
1.53
0.34
0.47
0.66
0.14
0.25
0.52
1.23
1.49
0.55
2.45
3.80
logBAF
Ol. & Niimi
(1988)
8.58
9.03
8.56
6.89
7.18
7.56
7.96
8.08
8.43
BSAF
EPA
(1990)
2.76
3.26
2.68
2.24
0.71
2.47
1.1
6.55
1.13
0.48
0.34
0.89
0.03
0.12
0.19
0.62
0.65
0.46
0.31
0.61
0.53
1.22
1.16
2.23
1.25
1.56
Predicted BAF™
log BAF
Ol. & Niimi
refPCBS2
9.18
9.32
9.33
8.68
9.37
9.22
9.19
8.91
9.15
8.95
8.70
6.17
6.77
7.29
7.72
7.32
8.06
9.15
log BAF
EPA
refPCBS2
9.13
9.42
9.17
9.01
9.11
9.74
8.89
9.98
9.22
9.23
9.17
9.24
8.20
6.02
6.50
7.14
7.17
7.08
7.15
7.50
7.44
7.83
8.41
8.81
8.82
8.97
log BAF
Ol. & Niimi
refPCBlOS
8.84
8.99
9.00
8.35
9.03
8.89
8.85
8.58
8.81
8.61
8.36
5.83
6.43
6.95
7.39
6.98
7.73
8.81
log BAF
EPA
ref PCB IDS
8.41
8.70
8.46
8.29
8.39
9.02
8.17
9.26
8.50
8.51
8.45
8.52
7.48
5.30
5.78
6.42
6.46
6.36
6.43
6.78
6.72
7.11
7.69
8.09
8.10
8.25
198
-------
Table 2.4.1a: Great Lakes Trout BARfs Calculated from Measured BSAFs/BAFs
Chemical
PCB 170+190
PCB 195+208
PCB 196+203
2378-TCDD
12378-PeCDD
123478-HxCDD
123678-HxCDD
123789-HxCDD
1234678-HpCDD
OCDD
2378-TCDF
12378-PeCDF
23478-PeCDF
123478-HxCDF
123678-HxCDF
123789-HxCDF
234678-HxCDF
1234678-HpCDD
1234789-HpCDD
OCDF
Log
K«
7.37
7.64
7.65
7.02
7.5
7.8
7.8
7.8
8.2
8.6
6.5
7.0
7.0
7.5
7.5
7.5
7.5
8.0
8.0
8.8
Measured Values
USAF
Ol. & Niimi
(1988)
2.06
1.56
log BAF
Ol. & Niimi
(1988)
9.20
9.26
BSAF
EPA
(1990)
4.17
0.72
1.12
0.059
0.054
0.018
0.0073
0.0081
0.0031
0.00074
0.047
0.013
0.095
0.0045
0.011
0.037
0.04
0.00065
0.023
0.00099
Predicted BAF™
log BAF
Ol. & Niimi
ref PCB 52
9.06
9.23
log BAF
EPA
ref PCB 52
9.53
9.04
9.25
7.34
7.78
7.60
7.21
7.26
7.24
7.02
6.72
6.66
7.52
6.70
7.09
7.61
7.65
6.36
7.91
7.34
log BAF
Ol. & Niimi
ref PCB 105
8.73
8.89
log BAF
EPA
ref PCB 105
8.81
8.33
8.53
6.62
7.06
6.88
6.49
6.54
6.52
6.30
6.00
5.94
6.81
5.98
6.37
6.90
6.93
5.64
7.19
6.62
a. Oliver and Niimi (1988). BAF c calculated from measured BAFs using freely dissolved equation 2.4.1 1, DOC=2.0 rag/L, POC=0.0 mg/L, K doc=Kcw/10,
K^K^,. Predicted BAFs based on equation 2.4.16.
b. U.S. EPA 1990. (TCDD Bioaccumulation Study). Predicted BAFs based on equation 2.4.16.
c. Green Bay/Fox River Mass Balance Study (As described in Great Lakes Water Quality Initiative Technical Support Document for the Procedure to
Determine Bioaccumulation Factors. EPA-820-B-95-005. March 1995.)
199
-------
Table 2.4.1b: Great Lakes Trout BAFfs Calculated from Measured BSAFs/BAFs
Chemical
dicldrin
ddt
Ate
ddd
mircx
pholomircx
g-chlordane
t-chlordanc
c-chlofdane
t-nonachlor
c-nonachlor
alpha-hch
gamma-hob
hcbd
OCX
hcb
pcb
1235Kb
I245lcb
1234tcb
I35icb
124tcb
123lcb
245tct
236tet
p«
ToWl-PCB
PCB 5
PCB 6
Log
K..
5.3
6.45
6.76
6.06
6.89
6.89
6
6
6
6
6
3.78
3.67
4.84
6.29
5.6
5.11
4.56
4.56
4.59
4.17
3.99
4.1
4.93
4.93
6.36
6.14
4.97
5.06
Predicted Values
log BAF
OI.&
Niimi'
refDDT
7.78
8.67
6.80
8.33
8.92
7.64
5.46
4.80
7.57
5.86
4.97
4.02
7.70
log BAF
EPA*
refDDT
7.23
7.78
8.75
8.11
7.41
7.78
8.13
6.81
5.72
log BAF
OI.&
Niimi*
refPCB
118
7.29
8.18
6.31
7.84
8.43
7.14
4.97
4.31
7.08
5.37
4.48
3.53
7.21
log BAF
EPAb
refPCB
118
7.30
7.85
8.82
8.18
7.48
7.85
8.20
6.88
5.79
BT-BSAF
EPA-G
Bay"
,
0.14
1.7 '
log BAF
EPA-G .
Bay'
refPCB 52
4.88
6.05
log BAF
EPA-G
Bay«
refPCB
118
5.12
6.29
200
-------
Table 2.4.1 b: Great Lakes Trout BAFfs Calculated from Measured BSAFs/BAFs
Chemical
PCB8
PCB12
PCB13
PCB16
PCB17
PCB 18
PCB22
PCB 25
PCB 26
PCB 32
PCB 33
PCB 40
PCB 42
PCB 44
PCB 45
PCB 46
PCB 49
PCB 52
PCB 53
PCB 63
PCB 64
PCB 66
PCB 74
PCB 77
PCB 81
PCB 82
PCB 83
PCB 84
PCB 85
PCB 87
Log
K»
5.07
5.22
5.29
5.16
5.25
5.24
5.58
5.67
5.66
5.44
5.60
5.66
5.76
5.75
5.53
5.53
5.85
5.84
5.62
6.17
5.95
6.20
6.20
6.36
6.36
6.20
6.26
6.04
6.30
6.29
Predicted Values
log BAF
OI.&
Niimi*
refDDT
5.71
5.95
6.19
6.35
7.18
5.99
6.06
5.96
6.77
6.72
6.58
6.98
6.91
7.17
7.10
7.42
8.03
7.88
7.81
7.75
log BAF
EPA"
refDDT
5.97
6.35
5.34
6.12
6.29
6.41
6.39
6.02
6.46
5.98
4.94
6.57
7.19
7.09
6.93
7.29
6.56
7.49
7.51
7.53
log BAF
01. &
Niimi*
refPCB
118
5.22
5.46
5.70
5.86
6.69
5.50
5.57
5.47
6.28
6.23
6.09
6.49
6.42
6.68
6.61
6.93
7154
7.39
7.32
7.26
log BAF
EPAb
refPCB
118
6.04
6.42
5.41
6.19
6.36
6.48
6.46
6.09
6.53
6.05
5.01
6.64
7.26
7.16
7.00
7.36
6.63
7.56
7.59
7.60
BT-BSAF
EPA-G
Bay1
0.14
0.75
0.64
0.39
0.73
0.95
0.29
0.69
1.16
0.61
3.34
4.74
2.12
4.37
3.1
2.46
4.12
11.6
4.05
5.67
7.2
7.25
6.13
log BAF
EPA-G
Bay'
refPCBSZ
4.98
5.89
5.81
5.94
6.30
6.40
5.83
6.26
6.36
6.08
7.14
7.28
6.71
7.57
7.46
7.36 •
7.74
8.19
7.57
7.78
7.66
7.92
7.84
log BAF
EPA-G
Bay"
refPCB
118
5.22
6.13
6.05
6.18
6.54
6.64
6.07
6.50
6.60
6.32
7.38
7.52
6.95
7.81
7.70
7.60
' 7.98
8.43
7.81
8.02
7.90
8.16
8.08
201
-------
Table 2.4.1 b: Great Lakes Trout BAFfs Calculated from Measured BSAFs/BAFs
Chemical
PCB91
PCB92
PCS 95
PCB97
PCS 99
PCBIOO
PCB101
PCBI05
PCB110
PCB118
PCBI19
PCBI26
PCB128
PCBI29
PCB130
PCB132
PCB136
PCB138
PCS 141
PCS 146
PCB149
PCBIS1
PCS 153
PCB156
PCB15S
PCS 167
PCB171
PCS 172
PCBI74
PCS 177
Log
K..
6.13
6.35
6.13
6.29
6.39
6.23
6.3S
6.65
6.48
6.74
6.5S
6.89
6.74
6.73
6.8
6.58
6.22
6.83
6.82
6.89
6.67
6.64
6.92
7.18
7.02
7.27
7.11
7.33
7.11
7.08
Predicted Values
log BAF
OI.&
Niimi'
refDDT
7.52
7.79
7.57
7.51
8.06
8.37
7.95
8.64
8.59
8.26
7.81
8.55
8.75
8.55
8.69
8.33
8.46
8.84
9.07
8.83
8.59
8.92
log BAF
EPA"
refDDT
7.04
6.84
7.67
7.58
7.51
8.41
7.50
8.08
8.27
8.50
8.29
7.89
8.24
8.16
8.09
7.74
7.96
8.31
8.31
8.21
8.57
8.31
8.47
log BAF
OI.&
Niimi*
refPCB
118
7.02
7.30
7.08
7.02
7.57
7.88
7.46
8.15
8.10
7.77
7.32
8.05
8.26
8.06
8.20
7.84
7.97
8.34
8.58
8.34
8.10
8.43
log BAF
EPAb
refPCB
118
7.11
6.91
7.74
7.65
7.58
8.48
7.57
8.15
8.34
8.57
8.36
7.96
8.31
8.24
8.16
7.81
8.03
8.38
8.38
8.28
8.64
8.38
8.54
BT-BSAF
EPA-G
Bay*
8.44
6.42
7.18
1.71
10.01
5.35
4.15
4.96
3.03
10.21
11.21
9.30
10.0
8.7
9.7
5.35
16.0
4.46
8.04
log BAF
EPA-G
Bay*
refPCB 52
7.82
7.86
8.01
7.23
8.14
8.14
7.86
8.20
7.83
8.51
8.61
8.55
8.66
8.37
8.39
8.41
9.24
8.52
8.75
log BAF
EPA-G
Bay"
refPCB
118
8.06
8.10
8.25
7.47
8.38
8.38
8.10
8.44
8.07
8.75
8.85
8.79
8.90
8.61
8.63
8.65
9.48
8.76
8.99
202
-------
Table 2.4.1b: Great Lakes Trout BAFfs Calculated from Measured BSAFs/BAFs
Chemical
PCB 178
PCB180
PCB 183
PCB 185
PCB 189
PCB 194
PCB 195
PCB 197
PCB 198
PCB 201
PCB 205
PCB 206
PCB 207
PCB 209
PCB 24+27
PCB 28+31
PCB 37+42
PCB 47+48
PCB 41+64+71
PCB 56+60
PCB 70+76
PCB 66+95
PCB 56+60+81
PCB 84+92
PCB 87+97
PCB 137+176
PCB 138+163
PCB 156+171+202
PCB 182+187
PCB 157+200
Log
K..
7.14
7.36
7.20
7.11
7.71
7.80
7.56
7.3
7.62
7.62
8.00
8.09
7.74
8.18
5.40
5.67
5.8
5.82
5.87
6.11
6.17
6.17
6.19
6.2
6.29
6.8
6.91
7.18
7.19
7.23
Predicted Values
log BAF
O1.&
Niimi1
refDDT
9.08
9.23
9.24
8.59
9.27
9.13
9.10
8.82
9.05
8.85
8.60
6.07
6.68
7.19
7.63
7.22
7.97
9.05
log BAF
EPAb
refDDT
8.69
8.98
8.73
8.56
8.67
9.30
8.45
9.54
8.78
8.79
8.73
8.79
7.76
5.58
6.05
6.70
6.73
6.64
6.71
7.05
7.00
7.39
7.97
8.36 "
8.38
8.53
log BAF
OI.& '
Niimi*
refPCB
118
8.59
8.74
8.75
8.10
8.78
8.64
8.60
8.33
8.56
8.36
8.11
5.58
6.18
6.70
7.14
6.73
7.48
8.56
log BAF
EPAfc
refPCB
118
8.76
9.05
8.80
8.63
8.74
9.37
8.52
9.61
8.85
8.86
8.80
8.86
7.83
5.65
6.12
6.77
6.80
6.71
6.78
7.12
7.07
7.46
8.04
8.43
8.45
8.60
BT-BSAF
EPA-G
Bay*
10.96
6.5
3.23
3.45
3.29
0.46
4.79
3.09
0.95
1.3
' 0.19
1.55
0.67
6.75
7.86
2.55
1.14
1.2
3.1
1.15
7.25
6.3
1.43
11.94
10.70
9.38
8.66
log BAF
EPA-G
Bay1
refPCB 52
9.16
8.78
8.38
9.01
9.08
8.05
9.06
9.25
8.83
8.62
8.22
6.35
6.26
7.39
7.47
7.04
6.93
7.01
7.43
7.02
7.82
7.85
7.72
8.75
8.97
8.92
8.93
log BAF
EPA-G
Bay
refPCB
118
9.40
9.02
8.62
9.25
9.32
8.29
9.30
9.49
9.07
8.86
8.46
6.59
6.50
7.63
7.71
7.28
7.17
7.25
7.67
7.26
8.06
8.09
7.96
8.99
9.21
9.16
9.17
203
-------
Table 2.4.1 b: Great Lakes Trout BAFfs Calculated from Measured BSAFs/BAFs
Chemical
PCB 170+190
PCB 195+208
PCB 196+203
2378-TCDD
12378-PeCDD
123478-HxCDD
123678-HxCDD
123789-HxCDD
123334678-Hp
OCDD
2378-TCDF
12378-PeCDF
23478-PcCDF
12347S-HxCDF
123678-HxCDF
123789-HxCDF
234678-HxCDF
Log
K..
737
7.64
7.65
7.02
7.5
7.8
7.8
7.8
8.2
8.6
6.5
7.0
7.0
7.5
7.5
7.5
7.5
Predicted Values
log BAF
O1.&
Niimi'
refDDT
8.97
9.13
log BAF
EPAb
refDDT
9.09
8.60
8.80
6.90
7.34
7.16
6.77
6.81
6.80
6.57
6.28
6.22
7.08
6.26
6.65
7.17
7.21
log BAF
OI.&
Niimi*
refPCB
118
8.48
8.64
log BAF
EPA*
refPCB
118
9.16
8.67
8.87
6.97
7.41
7.23
6.84
6.88
6.87
6.64
6.35
6.29
7.15
6.33
6.72
7.24
7.28
BT-BSAF
EPA-G
Bay"
4.10
1.01
4.24
log BAF
EPA-G
Bay*
refPCB 52
8.74
8.41
9.04
log BAF
EPA-G
Bay*
refPCB
118
8.98
8.65
9.28
204
-------
Table 2.4.1 b: Great Lakes Trout BAFfs Calculated from Measured BSAFs/BAFs
Chemical
1234678-HpCD
1234789-HpCD
OCDF
Log
K..
8.0
8.0
8.8
Predicted Values
log BAF
OI.&
Niimi'
refDDT
log BAF
EPAb
refDDT
5.92
7.47
6.90
log BAF
Ol. &
Niimi'
refPCB
118
log BAF
EPA"
refPCB
118
5.99
7.54
6.97
BT-BSAF
EPA-G
Bay1
log BAF
EPA-G
Bay*
refPCB 52
log BAF
EPA-G
Bay"
refPCB
118
a. Oliver and Niimi (1988). BAF{ calculated from measured BAFs using freely dissolved equation 2.4.1 1, DOC=2.0 mg/L, POC=0.0 mg/L, K Joc=KouyiO,
Kp^Kw. Predicted BAFs based on equation 2.4.16.
b. U.S. EPA 1990. (TCDD Bioaccumulation Study). Predicted BAFs based on equation 2.4.16.
c. Green Bay/Fox River Mass Balance Study (As described in Great Lakes Water Quality Initiative Technical Support Document for the Procedure to
Determine Bioaccumulation Factors. EPA-820-B-95-005. March 1995.)
205
-------
Table 2.4.2: Mean BAFjds from Lake Ontario BSAFs for Salmonids
Chemical
dieldrin
ddt
dde
ddd
mirex
photomirex
g-chlordane
t-chlordane
c-chlordane
t-nonachlor
c-nonachlor
alpha-hch
gamma-hch
hcbd
ocs
hcb
pcb
1235tcb
1245tcb
1234tcb
135tcb
124tcb
123tcb
245tct
236tct
pet
PCBs
5
6
8
12
13
16
17
18
22
25
26
logK™,
5.30
6.45
6.76
6.06
6.89
6.89
6.00
6.00
6.00
6.00
6.00
3.78
3.67
4.84
6.29
5.60
5.11
4.56
4.50
4.59
4.17
3.99
4.10
4.93
4.93
6.36
4.97
5.06
5.07
5.22
5.29
5.16
5.25
5.24
5.58
5.67
5.66
Number
BAFs
4
8
8
4
8
4
4
4
4
4
4
4
4
4
4
4
4
4
4
8
8
8
8
8
Mean
logBAF?
7.29
7.73
8.66
6.64
8.17
8.76
7.48
7.46
7.84
8.18
6.87
5.30
4.64
7.41
5.70
4.81
3.86
5.78
6.03
5.98
5.60
6.10
6.27
6.75
Mean
BAF?
1.93e+07
5.33e+07
4.56e+08
4.39e+06
1.49e+08
5.74e+08
3.00e+07
2.91e+07
6.95e+07
1.53e+08
7.43e+06
2.00e+05
4.34e+04
2.58e+07
5.01e+05
6.47e+04
7.25e+03
6.02e+05
1.06e+06
9.52e+05
3.96e+05
1.27e+06
1.87e+06
5.57e+06
206
-------
Table 2.4.2: Mean BAF[ds from Lake Ontario BSAFs for Salmonids (continued)
Chemical
PCBs
32
33
40
42
44
45
46
49
52
53
63
64
66
74
77
81
82
83
84
85
87
91
92
95
97
99
100
101
105
110
118
119
126
128
129
130
132
,08K.
5.44
5.60
5.66
5.76
5.75
5.53
5.53
5.85
5.84
5.62
6.17
5.95
6.20
6.20
6.36
6.36
6.20
6.26
6.04
6.30
6.29
6.13
6.35
6.13
6.29
6.39
6.23
6.38
6.65
6.48
6.74
6.58
6.89
6.74
6.73
6.80
6.58
Number
BAFs
4
8
8
4
8
4
8
4
8
4
4
4
4
8
4
4
8
4
4
8
4
8
4
4
4
8
4
8
8
8
8
4
4
8
8
4
4
Mean
log BAF"
5.84
6.18
5.94
6.61
6.54
6.04
5.71
6.82
6.69
7.02
7.25
6.94
7.26
7.51
6.99
7.35
7.17
7.55
7.65
7.58
7.59
7.23
7.64
7.41
6.90
7.54
7.64
7.73
8.34
7.68
8.31
8.33
8.56
8.39
8.03
8.30
7.65
Mean
BAFcfd
6.84e+05
1.50e+06
8.72e+05
4.06e+06
3.46e+06
1.09e+06
. 5.08e+05
6.61e+06
4.90e+06
1.04e+07
1.77e+07
8.80e+06
1.83e+07
3.23e+07
9.68e+06
2.24e+07
1.48e+07
3.53e+07
4.50e+07
3.83e+07
3.89e+07
1.69e+07
4.32e+07
2.55e+07
7.95e+06
3.49e+07
4.40e+07
5.43e+07
2.18e+08
4.74e+07
2.04e+08
2.12e+08
3.63e+08
2.44e+08
1.06e+07
1.98e+08
4.47e+07
207
-------
Table 2.4.2: Mean BAFfc from Lake Ontario BSAFs for Salmonids (continued)
Chemical
PCBs
136
138
141
146
149
151
153
156
158
167
171
172
174
177
178
180
183
185
189
194
195
197
198
201
205
206
207
209
24+27
28+31
37+42
47+48
41+64+71
56+60
70+76
66+95
56+60+81
84+92
logK™
6.22
6.83
6.82
6.89
6.67
6.64
6.92
7.18
. 7.02
7.27
7.11
7.33
7.11
7.08
7.14
7.36
7.20
7.11
7.71
7.80
7.56
7.30
7.62
7.62
8.00
8.09
7.74
8.18
5.40
5.67
5.80
5.82
5.87
6.11
6.17
6.17
6.19
6.20
Number
BAFs
4
4
8
8
8
8
8
4
4
4
4
4
8
8
8
8
8
8
4
8
4
4
4
8
8
8
8
8
8
8
4
8
4
4
8
4
4
4
Mean
log BAF?
8.39
8.59
8.31
8.34
7.98
8.16
8.52
8.91
8.37
8.27
8.67
8.63
8.40
8.64
8.83
9.05
8.94
8.53
8.72
9.23
8.97
8.50
9.60
8.89
8.75
8.84
8.77
8.13
5.78
6.31
6.76
6.91
6.70
6.76
7.29
7.06
7.06
7.45
Mean
BAF?
2.44e+08
3.88e+08
2.03e+08
2.18e+08
9.66e+07
1.45e+08
3.31e+08
8.12e+08
2.32e+08
1.87e+08
4.72e+08
4.24e+08
2.51e+08
4.38e+08
6.80e+08
1.13e+09
8.63e+08
3.36e+08
5.30e+08
1.72e+09
9.32e+08
3.20e+08
3.98e+09
7.70e+08
5.64e+08
6.90e+08
5.92e+08
1.35e+08
5.98e+07
2.06e+06
5.70e+06
8.18e+06
4.97e+06
5.82e+06
1.96e+07
1.14e+07
1.16e+07
2.82e+07
208
-------
Table 2.4.2: Mean BAFfc from Lake Ontario BSAFs for Salmonids (continued)
Chemical
PCBs
87+97
137+176
138+163
156+171+202
182+187
157+200
170+190
195+208
196+203
PCDDs
2378-TCDD
12378-PeCDD
123478-HxCDD
123678-HxCDD
123789-HxCDD
1234678-HpCDD
OCDD
PCDFs
2378-TCDF
12378-PeCDF
23478-PeCDF
123478-HxCDF
123678-HxCDF
123789-HxCDF
234678-HxCDF
1234678-HpCDF
1234789-HpCDF
OCDF
,ogKm
6.29
6.80
6.91
7.18
7.19
7.23
7.37
7.64
7.65
7.02
7.50
7.80
7.80
7.80
8.20
8.60
6.50
7.00
7.00
7.50
7.50
7.50
7.50
8.00
8.00
8.80
Number
BAFs
4
4
4
4
4
4
8
4
8
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
Mean
logBAFf
7.81
8.03
8.42
8.44
8.89
8.59
8.98
8.66
8.92
6.95
7.40
7.22
6.83
6.87
6.85
6.63
6.34
6.28
7.14
6.32
6.70
7.23
7.27
5.98
7.53
6.96
Mean
BAF?
6.46e+07
1.07e+08
2.64e+08
2.76e+08
7.85e+08
3.86e+08
9.53e+08
4.58e+08
8.27e+08
9.00e+06
2.49e+07
1.65e+07
6.71e+06
7.44e+06
7.16e+06
4.29e+06
2.16e+06
1.89e+06
1.38e+07
2.07e+06
5.07e+06
1.70e+07
1.84e+07
9.47e+05
3.35e+07
9.10e+06
209
-------
LJL
00
TJ
"E
o
E
CO
CO
c
O
CD
CM
!5
X
LU
0)
a!
LL
CO
m
o
CO
LL.
CD
•o
£
3
(0
CO
0)
o
O
CO
LL
5 co
.o
13
z
i
O
E
CO
LL
CD
O)
O
- co
- in
co
CD
CO
Ll_
<
DQ
_o
CD
13
CO
03
CD
E
CM
in
O
Q_
O £
u_
CD
-------
00
(A C
8 •-
11
00 C
00 (0
at ,_
rl
«* .1 o
*Z S
CM -o O
.-e c £
5 w -o
X
111
w
O
c
'•&
2
•M
Q.
to
U.
2
T3
C
(0
w
U.
<
CD
(0
o
V
n
a:
M—
O
0)
O)
(0
a:
saouejjnooo
211
-------
LJL
O
CD
DO S>
•of
s o
!5
•= CQ
x f
LJJ
- CO
CO
CD
CO
LL
CO
DQ
08
fe
5
o
CO
co, LL
DQ
CD
O
CM
in
O
P cvi
® ii
0) o
en Q
212
-------
^ -o
a. ° CD
rsi Q. o ^j.
±! « =0 §
.2 LL 0) O)
1 S a^
T- w 'c
S£i
i 2 z
O
JQ
C
(0
§£
IS
CO
CO
A
CO
CD
A
CO
4
A
3
A
CO
_0
+3
as
tf
M—
O
0)
D)
C
to
a:
seouejjnooo |o
213
-------
LL
<
CQ
3
O
OS
C3
<+•»
c
O
0)
J*:
cc
"D
O
2
Q.
X
LJLJ
O)
O
CVJ
m
m
o
Is
p
CD It
II
0) O
tr Q
214
-------
CQ
05
CD
c
o
0)
CD
08
0)
o
O
• •
CO
CM
15
IE
X
LJL1
CO
LL
CQ
CD
D
03
CO
CD
cvj
CD
O 2
w
u.
<
m
m
5
< o
CO T-
§ I
0) i£
CD II
m E
O «?
•
•=J o
CC o
UJ Q
215
-------
TJ
CD
-------
CO
CO
E
o
(0
LL
CD
a.
LLJ
JS
CD
o
O
• •
oo
c\
x
LLJ
3
O
H
CD
CO
00
0)
3
0)
•*•*
I
o
CO
CO
*•»
o
CO
C/3
UL
CD
D)
.O
CD
03
CO
CD
— o>
CO
U_
00 CO
D)
D
— LO
13
O
h=
CD
03
03
CD I
'
-------
s
1
0)
D)
i
seouajjnooo
218
-------
2.4.4.3 Baseline BAF Derived from a Laboratory-Measured BCF and Food-
Chain Multiplier
For the third tier in the data preference hierarchy for nonpolar organic chemicals, EPA
recommends the use of a predicted BAF derived from a technically defensible, laboratory
measurement of the BCF and an appropriate food chain multiplier (FCM). A FCM is determined
as the ratio of the baseline BAF (BAF[d) of an organism at a particular trophic level to the baseline
BCF (usually determined for trophic level one).24 FCMs with values greater than 1.0 indicate
biomagnification and typically apply to organic chemicals with K,,w values between 4.0 and 9.0.
Laboratory-measuredBCFs are preferred over predicted BCFs because laboratory-measured BCFs
inherently account for effects of chemical metabolism on the BCF during its measurement.
The equation for calculating a baseline BAF from a laboratory-measured BCF is:
Baseline BAF fd = (FCM)
Measured BCF *
- 1
fd
where:
Baseline BAFf
FCM
(Equation 2.4.17)
BAF expressed on a freely-dissolved and lipid-normalized
basis
BCF based on total concentration in tissue and water
Fraction of the tissue that is lipid
Fraction of the total chemical in the test water that is freely
dissolved
Food-chain multiplier obtained from Tables 2.4.4, 2.4.5, or
2.4.6 by linear interpolation for the appropriate trophic level
as necessary (or from appropriate field data)
For each trophic level, the species mean baseline BAF is calculated as the geometric mean
if more than one acceptable baseline BAF is predicted from laboratory-measured BCFs for a given
"Note: Equilibrium partitioning theory would predict BCFcfd approximately equal to K.,, thus the BCFcfd would not be trophic level
dependent.
219
-------
species. For each trophic level, the trophic level-specific BAF is calculated as the geometric mean
of the species mean baseline BAFs based on laboratory-measured BCFs.
Procedural and Quality Assurance Requirements for Measured BCFs
A measured BCF derived from results of a laboratory exposure study is acceptable if the
study has met certain specific technical criteria. These criteria include, but are not limited to:
1. The test organism should not be diseased, unhealthy, or adversely affected by the
concentration of the chemical because these attributes may alter accumulation of
chemicals by otherwise healthy organisms.
2. The total concentration of the chemical in the water should be measured and should
be relatively constant during the steady-state time period.
3. The organisms should be exposed to the chemical using a flow-through or renewal
procedure.
4. For organic chemicals, the percent lipid should be either measured Or reliably
estimated for the tissue used in the determination of the BCF.
5. For organic chemicals with log K^, greater than four, the concentrations of POC and
DOC in the test solution should be either measured or reliably estimated. For organic
chemicals with log J^ less than four, virtually all of the chemical is predicted to be
freely dissolved, except in water with extremely high DOC and POC concentrations,
which is not characteristic of laboratory dilution water used in BCF determinations.
6. Laboratory-measured BCFs should be determined using fish species, but BCFs
determined with molluscs and other invertebrates may be used with caution. For
example, because invertebrates metabolize some chemicals less efficiently than
vertebrates, a baseline BCF determined using invertebrates is expected to be higher
than a comparable baseline BCF determined using fish.
7. If laboratory-measured BCFs increase or decrease as the concentration of the
chemical increases in the test solutions in a bioconcentrationtest, the BCF measured
at the lowest test concentration above control concentrations should be used (i.e., a
BCF should not be calculated from a control treatment). The concentrations of an
inorganic chemical in a bioconcentration test should be greater than normal
background levels and greater than levels required for normal nutrition of the test
species if the chemical is a micronutrient, but below levels that adversely affect the
species. Bioaccumulation of an inorganic chemical might be overestimated if
concentrations are at or below normal background levels due to, for example,
nutritional requirements of the test organisms.
220
-------
8. For inorganic chemicals, BCFs should be used only if they are expressed on a wet
weight basis. BCFs reported on a dry weight basis cannot be converted to wet
weight unless a conversion factor is measured or reliably estimated for the tissue
used in the determination of the BAF.
9. BCFs for organic chemicals may be based on measurement of radioactivity only
when the BCF is intended to include metabolites, when there is confidence that there
is no interference due to metabolites, or when studies are conducted to determine the
extent of metabolism, thus allowing for a proper correction.
10. The calculation of the BCF must appropriately address growth dilution.
11. Other aspects of the methodology used should be similar to those described by
ASTM(1990).
In addition, the magnitude of the octanol-water partition coefficient (K^) and the availability
of corroborating BCF data should be considered. For example, some chemicals with high log K,,ws
may require longer than 28 days to reach steady state conditions between the organism and the water
column. As with B AFs, the BCFs should be divided by the mean lipid fraction to express the value
on a lipid-normalized basis.
Food-Chain Multipliers
The food-chain multiplier represents a measure of a chemical's tendency to biomagnify in
aquatic food webs. For non-polar organic chemicals, FCMs can be determined from
bioaccumulation models or directly from field data (tissue residues).
For model-derivedFCMs, EPA recommends using the food web model by Gobas (1993) to
determine FCMs for nonpolar organic chemicals. There are several advantages to using the Gobas
(1993) model. First, uptake into both benthic and pelagic food chains is measured, incorporating
exposure of organisms to chemicals from both the sediments and the water column. Second, the
input data needed to run the model can be readily defined. Third, the model-predicted B AFs (which
are used to derive the FCMs) are in agreement with field-measured B AFs for chemicals, even those
with very high log K^. Finally, the model predicts chemical residues in benthic organisms using
equilibrium partitioning theory, which is consistent with EPA's sediment quality criteria effort.
The Gobas (1993) model predicts the chemical residues in the organisms, which are then
used to estimate BAFs for each species in the food chain:
221
-------
BAF
fd_
fd
(Equation 2.4.18)
where:
BAF fd= Lipid-normalizedBAF using the freely dissolved concentration in the water
fd =
Freely dissolved concentration of the chemical in the water column
C = Lipid-normalized concentration in appropriate tissue
Food-chain multipliers are then calculated from the predicted BAFfs using the following equation:
BAF
fd
FCM =
K
where:
K,,
(Equation 2.4.19)
n-octanol/water partition coefficient
Data Requirements to Predict the Food-Chain Multiplier. The food chain model by Gobas
(1993) requires specific data on the structure of the food chain and the water quality characteristics
of the water body of interest including:
• Feeding preferences, weights, and lipid contents for each species in the food chain.
• Water temperature.
• Organic carbon content of the sediment and the water column.
• Concentrations of the chemical in the sediment and freely dissolved concentration
of the chemical in the water column.
• Densities of lipid and organic carbon.
222
-------
• Metabolic transformation rate constant.
• Kow values, estimated using the methods described in Section 2.4.4.1 (subsection
entitled: Guidance on Selecting Appropriate Kow Values).
It should be noted that the model of Gobas (1993) does not include solubility controls or
limitations; thus, the concentration of the chemical in the water used with the model is arbitrary for
determining the BAFs, i.e., the ratio of the concentration of the chemical in the tissue to the
concentration of the chemical in the water column (BAF) obtained using a 1 ng/L concentration of
the chemical in the water will be equal to that obtained using a 150 Aig/L concentration of the
chemical for a specified K^,.
It should be noted that the model of Gobas (1993) takes the total concentration of the
chemical in the water and, before doing any predictions, calculates the freely dissolved concentration
of the chemical in the water. The freely dissolved concentration of the chemical in the water is then
used in all subsequent calculations by the model. By setting the concentration of the DOC and POC
to 0 mg/L, the total concentration of the chemical input to the model becomes equal to the freely
dissolved concentration of the chemical in the water. This allows the fixing of the chemical
concentration relationship between sediment and water phases in the model. BAFs were deteriiilued
by dividing the chemical residues predicted by the model of Gobas (1993) by the freely dissolved
concentration of the chemical in the water; therefore, they are not influenced by the concentration
of DOC input to the model.
Measured chemical residues in fishes assigned to trophic level 3 can be higher than those in
piscivorous fishes (trophic level 4) from the same food chain. Potential causes of the higher
concentrations (on a lipid basis) in the trophic level 3 fish include 1) growth rates which are much
slower than rates for predator fishes; 2) slower rates of metabolism than the predator fishes for the
chemicals of interest; and 3) feeding preferences for trophic level 3 fish, including predation on other
fish. In the development of FCMs, the feeding preferences for smelt (see Gobas 1993) consisted of
a mixture of trophic level 2 and 3 organisms, i.e., mysids, Diporeia sp., and sculpin. This mixture
of different trophic levels combined with bioenergetic factors for the smelt caused the predicted
concentrations of the chemicals and subsequently, the derived FCMs, to be slightly larger than those
for the piscivorous fishes (trophic level 4).
223
-------
Table 2.4.3: Environmental Parameters and Species Characteristics Used
With the Model of Gobas (1993) for Deriving Chain Multipliers
Environmental parameters:
Mean water temperature: 8°C
Organic carbon content of the sediment: 2.7%
Organic carbon content of the water column: 0.0 kg/L
Density of lipids: 0.9 kg/L
Density of organic carbon: 0.9 kg/L
Metabolic transformation rate constant: 0.0 day1
= 25*logKow.
Species characteristics;
Phytoplankton
Lipid content: 0.5%
Zooplankton: Mysids (Mysis relicta)
Lipid content: 5.0%
Diporeia sp.
Lipid content: 3.0%
Sculpin (Cottus cognatus)
Lipid content: 8.0%
Weight: 5.4 g
Diet: 18% zooplankton, 82% Diporeia sp. (pelagic/benthic food web)
100% Diporeia sp. (all-benthic food web)
100% zooplankton (all-pelagic food web)
Alewives (Alosa pseudoharengus)
Lipid content: 7.0%
Weight: 32 g
Diet: 60% zooplankton, 40% Diporeia sp. (pelagic/benthic food web)
100% Diporeia sp. (all-benthic food web)
100% zooplankton (all-pelagic food web)
Smelt (Osmerus mordax)
Lipid content: 4.0%
Weight: 16 g
Diet: 54% zooplankton, 21% Diporeia sp., 25% sculpin
54% Diporeia sp., 46% sculpin (all-benthic food web)
68% zooplankton, 32% sculpin (all-pelagic food web)
Salmonids (Salvelinus namaycush, Oncorhynchus mykiss, Oncorhynchus velinus
namaycush)
Lipid content: 11.0%
Weight: 241 Og
Diet: 10% sculpin, 50% alewives, 40% smelt
224
-------
Table 2.4.4:
Food-Chain Multipliers for Trophic Levels 2, 3, & 4
(Pelagic and Benthic Structure)
LogK,,w
<2.0
2.0
2.5
3.0
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
4.0
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
5.0
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
6.0
6.1
6.2
6.3
6.4
6.5
6.6
6.7
Trophic Level 2
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
Trophic8 Level 3
1.000
1.005
1.010
1.028
1.034
1.042
1.053
1.067
1.083
1.103
1.128
1.161
1.202
1.253
1.315
1.380
1.491
1.614
1.766
1.950
2.175
2.452
2.780
3.181
3.643
4.188
4.803
5.502
6.266
7.096
7.962
8.841
9.716
10.556
11.337
12.064
12.691
13.228
13.662
13.980
14.223
Trophic Level 4
1.000
1.000
1.002
1.007
1.007
1.009
1.012
1.014
1.019
1.023
1.033
1.042
1.054
1.072
1.096
1.130
1.178
1.242
1.334
1.459
1.633
1.871
2.193
2.612
3.162
3.873
4.742
5.821
7.079
8.551
10.209
12.050
13.964
15.996
17.783
19.907
21.677
23.281
24.604
25.645
26.363
225
-------
Table 2.4.4:
Food-Chain Multipliers for Trophic Levels 2, 3, &
(Pelagic and Benthic Structure)
LogK™
6.8
6.9
7.0
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
8.0
8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8
8.9
9.0
Trophic Level 2
1.000
1.000
1,000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
Trophic" Level 3
14.355
14.388
14.305
14.142
13.852
13.474
12.987
12.517
11.708
10.914
10.069
9.162
8.222
7.278
6.361
5.489
4.683
3.949
3.296
2.732
2.246
1.837
1.493
4
Trophic Level 4
26.669
26.669
26.242
25.468
24.322
22.856
21.038
18.967
16.749
14.388
12.050
9.840
7.798
6.012
4.519
3.311
2.371
1.663
1.146
0.778
0.521
0.345
0.226
The FCMs for trophic level 3 are the geometric mean of the FCMs for sculpin and alewife.
226
-------
Table 2.4.5:
Food-Chain Multipliers for Trophic Levels 2,3, & 4
(All-Pelagic Structure)
Trophic Level 2
Trophic' Level 3
Trophic Level 4
<2.0
2.0
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
4.0
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
5.0
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
LOGO
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
.000
.000
.000
.000
.000
1.000
1.001
1.001
1.001
1.001
1.001
1.002
1.002
1.002
1.003
1.004
1.005
1.006
1.007
1.009
1.011
1.014
1.018
1.022
1.028
1.034
1.043
1.053
1.066
1.081
1.099
1.121
1.147
1.176
1.210
1.248
1.289
1.333
1.379
1.425
1.471
1.000
1.001
1.001
1.001
1.002
1.002
1.002
1.003
1.003
1.004
1.005
1.006
1.007
1.009
1.011
1.013
1.016
1.021
1.026
1.032
1.040
1.050
1.063
1.078
1.097
1.121
1.150
1.185
1.228
1.280
1.342
1.415
1.502
1.603
1.719
1.851
1.999
2.162
2.337
2.521
2.711
227
-------
Table 2.4.5:
Food-Chain Multipliers for Trophic Levels 2, 3, & 4
(All-Pelagic Structure)
LogK,,,,
6.0
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
7.0
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
8.0
8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8
8.9
9.0
Trophic Level 2
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
Trophic" Level 3
1.514
1.554
1.589
1.619
1.643
1.660
1.671
1.674
1.669
1.657
1.636
1.606
1.567
1.518
1.458
1.389
1.308
1.219
1.122
1.020
0.915
0.810
0.707
0.610
0.520
0.438
0.366
0.303
0.249
0.204
0.166
Trophic Level 4
2.900
3.083
3.254
3.407
3.536
3.637
3.705
3.738
3.733
3.688
3.602
3.474
3.305
3.094
2.848
2.570
2.270
1.958
1.647
1.349
1.076
0.835
0.631
0.466
0.336
0.237
0.164
0.112
0.075
0.050
0.033
The FCMs for trophic level 3 are the geometric mean of the FCMs for sculpin and alewife.
228
-------
Table 2.4.6:
Food-Chain Multipliers for Trophic Levels 2, 3, & 4
(AII-Benthic Structure)
LogK™
<2.0
2.0
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
4.0
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
5.0
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
Trophic Level 2
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
Trophic8 Level 3
1.000
1.009
1.010
1.011
1.013
1.015
1.018
1.022
1.026
1.032
1.039
1.048
1.060
1.074
1.092
1.114
1.142
1.177
1.222
1.277
1.347
1.433
1.541
1.676
1.843
2.050
2.306
2.620
3.004
3.470
4.032
4.702
5.492
6.411
7.462
8.643
9.942
11.337
12.800
14.293
15.774
Trophic Level 4
1.000
1.001
1.001
1.001
1.002
1.002
1.002
1.003
1.003
1.004
1.005
1.006
1.008
1.010
1.013
1.017
1.022
1.029
1.039
1.053
1.072
1.099
1.138
1.195
1.276
1.392
1.559
1.796
2.131
2.595
3.232
4.087
5.215
6.668
8.501
10.754
13.457
16.617
20.213
24.192
28.468
229
-------
Table 2.4.6:
Food-Chain Multipliers for Trophic Levels 2, 3, & 4
(AH-Benthic Structure)
LogK,,*
6.0
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
7.0
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
8.0
8.1
8.2
8.3
8.4
•8.5
8.6
8.7
8.8
8.9
9.0
Trophic Level 2
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
Trophic" Level 3
17.202
18.539
19.753
20.822
21.730
22.469
23.037
23.433
23.659
23.717
23.606
23.326
22.873
22.246
21.443
20.467
19.327
18.040
16.629
15.129
13.580
12.026
10.510
9.068
7.732
6.522
5.448
4.513
3.711
3.032
2.465
Trophic Level 4
32.920
37.405
41.764
45.836
49.472
52.544
54.949
56.610
57.472
57.501
56.679
55.007
52.507
49.227
45.254
40.714
35.780
30.657
25.572
20.744
16.359
12.547
9.368
6.822
4.856
3.387
2.321
1.567
1.045
0.689
0.451
The FCMs for trophic level 3 are the geometric mean of the FCMs for sculpin and alewife.
230
-------
The freely dissolved concentrations of the chemicals in the water column were calculated
from the data of Oliver and Niimi (1988) using the equations of Gschwend and Wu (1985) and Cook
et al. (1993)25 for freely dissolved fraction:
1
fd 1 + DOC • KJ + POC • K
doc
(Equation 2.4.21)
and freely dissolved concentration:
poc
fd _
= c' • f.
w fd
(Equation 2.4.22)
where:
ffd = Fraction of the chemical which is freely dissolved in the water
DOC = Concentration of dissolved organic carbon
POC = Concentration of particulate organic carbon
Kdoc = Partition coefficient for the chemical between the DOC and the freely
dissolved phase in the water
= Partition coefficient for the chemical between the POC phase and the freely
dissolved phase in the water
= Total concentration of the chemical in the water
= Freely dissolved concentration of the chemical in the water
The concentrations in the water reported by Oliver and Niimi (1988) were obtained by liquid-
liquid extraction of aliquots of Lake Ontario water which had passed through a continuous-flow
centrifuge to remove POC. Therefore, the concentrations in the water reported by Oliver and Niimi
(1988) include both the freely dissolved chemical and the chemical associated with the DOC in the
water sample. The above equations were used to derive the freely dissolved concentrations of the
"These equations were used to derive the equation for fa presented in Section 2.4.4.1, and are discussed in Appendix D.
231
-------
chemicals in the water by setting the POC = 0 mg/L and DOC = 2 mg/L. K^s used to derive the
freely dissolved concentrations have been reported elsewhere (USEPA, 1995c).
In Exhibit 2.4. 1 0, n^ is plotted against K^ for each chemical reported by Oliver and Niimi
(1988). Because the chemical residues by Oliver and Niimi (1988) for the foraging and piscivorous
fishes were almost entirely for the PCBs and pesticides, a regression equation of the form log II =
A x log KOU, + B was determined using this set of chemicals. Using the geometric mean regression
technique, the slope (standard deviation) of this equation was 1.07 (0.078). This slope was not
significantly different from 1.0, and thus, the relationship between the nsocw/K<)W of the individual
PCB and pesticide compounds. The average (standard deviation) ratio was 24.7 (25.7). The
following relationship was therefore selected to define IISOCW in this investigation: IISOCW = 25 * log
In addition to determining FCMs for organic substances using the Gobas (1993) model, EPA
also recommends the use of FCMs derived from field data where data are sufficient to enable
scientifically valid and reliable determinations to be made. Currently, field-measured FCMs are the
only method recommended for estimating FCMs for inorganic substances because appropriate
model-derived estimates are not yet available. Similarly, field-measured FCMs can also be
determined for organic substances. Compared to the model-based FCMs described previously,
properly derived field-based FCMs may offer some advantages in some situations. For example,
field-measured FCMs rely on measured contaminant concentrations in tissues of biota and therefore
inherently account for any contaminant metabolism which may occur. Field-measured FCMs may
also be useful for estimating BAFs for some highly hydrophobic contaminants whose water column
concentrations are very difficult to determine with accuracy and precision. Furthermore, field-
measured FCMs may better reflect local conditions that can influence bioaccumulation, such as
differences in food web structure, exposure pathways, water body type, and target species. Finally,
use of field-measured FCMs in estimating BAFs may enable existing data on contaminant
concentrations in aquatic organisms to be used in situations where companion water column data are
unavailable or are judged to be unreliable for calculating a BAF.
232
-------
0
CO
1
•»J
1_
0
«*-
4-1
C
'.^
0
cr
C
o
"5 •— *
£0 CO
^3 CO
C O5
0 ^
o
SI
o :=
ll
c w
CD *—
o .>
1°
C M_
5 0
O «j
*- "n
o ^
II
§ O
E
"•5
o
CO
• •
o
V-
•
CS
•4-»
co
CD
C
CD
•n CO
— ' i—
"o OJ
CO *^ r-
CD -L Q CD
2 O CD 0>
O '1H ( "^<^1 C ^^
*=• co ifi^J ow
w Q i Q
Q. S ^ CD
N 0
§ Q,
JD
i
O
•
V
<3
'd
< ^o
. xl
O
1 111 1 i I I T 1 , ,
5 oooo
jr oooo
X 0 0 CO r-
UJ CO T-
0
^
• •
0 O
-
V^CL
^^B ^^
* ^
_ "S
C •
c u
CM o
CO J 5
o ^*
°1
J
to
h-
o
CD ^
o>
o
10
rf
O O CO t—
CO T-
233
-------
As discussed below and in Appendix C, FCMs are related to and can be determined from
biomagnification factors (BMP). For example:
1.
2.
3.
FCM
^
) (BMF^.3) (BMF
where:
FCM = Food chain multiplier for designated trophic level (TL2, TL3, or TL4)
BMP = Biomagnification factor for designated trophic level (TL2, TL3, or TL4)
The basic difference between FGMs and BMFs is that FCMs relate back to trophic level one
(or trophic level two as assumed by the Gobas (1 993) model), whereas BMFs always relate back to
the next lowest trophic level. For nonpolar organic chemicals, biomagnification factors can be
calculated from lipid-normalizedtissue residue concentrations determined in biota at a site according
to the following equation.
BMP - = C./(C.)
where:
BMP ™ =
Cf= Lipid-normalized concentration of chemical in tissue of appropriate biota that occupy
the specified trophic level (TL2, TL3, or TL4).
For inorganic chemicals, BMFs are determined as shown above, except that tissue
concentrations expressed on a wet-weight basis and are not lipid normalized. In calculating field-
derived BMFs for determining FCMs, care must be taken to ensure that the biota upon which they
are based actually represent functional predator-prey relationships at the study site, and therefore,
would accurately reflect any biomagnification that may occur at the site.
As with field-measured BAFs, the potential advantages of using field data for estimating
bioaccumulationcan be offset by improper collection and use of information. In calculating field-
based FCMs, steps similar to those recommended for determining field-measured BAFs need to be
taken to ensure that the resulting FCMs accurately represent potential exposures to the target
population at the site(s) of interest. Some of the general procedural and quality assurance
requirements that are important for determining field-measured FCMs include:
234
-------
1. A food web analysis should be conducted for the site from which the tissue concentration
data are to be determined (or have been already been determined) to identify the appropriate trophic
levels for the aquatic organisms and appropriate predator-prey relationships. To assist in trophic
level determinations, EPA is in the process of finalizing its draft trophic level and exposure analysis
documents (USEPA 1995d; 1995e; 1995f) which include trophic level analyses of numerous species
in the aquatic-based food web.
2. The aquatic organisms sampled from each trophic level should reflect the most important
exposure pathways leading to human exposure via consumption of aquatic organisms. For higher
trophic levels (e.g., 3 and 4), aquatic species should also reflect those that are commonly consumed
by humans.
3. Collection of tissue concentration field data for a specific site for which criteria are to be
derived and with the specific species of concern are preferred.
4. If data cannot be collected from every site for which criteria are to be derived, the site of
the field study should not be so unique that the FCM values cannot be extrapolated to other locations
where the criteria and values will apply.
5. Samples of the appropriate resident species and the water in which they reside should be
collected and analyzed using appropriate, sensitive, accurate, and precise methods to determine the
concentrations of bioaccumulative chemicals present in the tissues.
6. For organic chemicals, the percent lipid should be either measured or reliably estimated
for the tissue used in the determination of the lipid normalized concentration in the organism's edible
tissues.
7. The tissue concentrations should reflect average exposure conditions over the time period
required to achieve steady-state conditions for the contaminant in the target species (usually trophic
level three or four organisms).
2.4.4.4 Baseline BAF from Predicted BCF and Food-Chain Multiplier
As the fourth tier in the data preference hierarchy for nonpolar organics (e.g., when
acceptable field-measured BAFs, BSAFs, or laboratory-measured BCFs are unavailable), EPA
recommends multiplying the FCM by the Kow for the chemical for estimating the baseline BAF.
This is equivalent to the direct use of the food chain bioaccumulationmodel for estimating the BAF
when model-derived FCMs are used. For each trophic level, the equation for calculating this
baseline BAF is:
235
-------
Baseline BAF = (FCM)(predicted baseline BCF) = (FCM)(KQw)
(Equation 2.4.23)
where:
Baseline BAF =
FCM
BAF expressed on a lipid-normalizedbasis using the freely dissolved
concentration of the chemical in water
Food-chain multiplier obtained from Table 2.4.3,2.4.4, or 2.4.5 by
linear interpolation (or from appropriate field data)
KOW = Octanol-water partition coefficient
Use of the K,,w in place of the baseline BCF is supported by equilibrium partitioning theory.
The linear relationship between the BCF and K^ is also based on the underlying assumption that
the bioeoncentrationprocess can be viewed as a partitioning of a chemical between the lipids of the
aquatic organisms and water and that the K^ is a useful surrogate for this partitioning process
(Mackay, 1982). These authors presented a thermodynamic basis for the partitioning process for
bioconcentration and, in essence, the BCF on a lipid-normalized basis (and freely dissolved
concentration of the chemical in the water) should be similar if not equal to the K^, for organic
chemicals.
In addition, empirical data support the use of the Kow in place of the BCF. As indicated by
Isnard and Lambert (1988), numerous studies have demonstrated a linear relationship between the
logarithm of the BCF and the logarithm of the octanol-water partition coefficient (Kow) for organic
chemicals for fish and other aquatic organisms. In addition, when the regression equations are
constructed using BCFs reported on a lipid-normalized basis, the slopes and intercepts are not
significantly different from one and zero, respectively. For example, de Wolf et al. (1992) adjusted
a relationship reported by Mackay (1982) to a 100 percent lipid basis (lipid normalized basis) and
obtained the following relationship:
log BCF = 1.00 log Kow - 0.08
(Equation 2.4.24)
For chemicals with large log K,,w s (i.e., greater than 6.0), reported BCFs are often not equal to the
KOW for non-metabolizable chemicals. BCFs for non-metabolizable chemicals are equal to the K,,w
when the BCFs are reported on lipid-normalized basis, determined using the freely dissolved
concentration of the chemical in the exposure water, corrected for growth dilution, determined from
steady-state conditions or determined from accurate measurements of the chemical's uptake (k|) and
elimination (k2) rate constants from and to the water, respectively, and determined using no solvent
236
-------
carriers in the exposure. Therefore, EPA recommends that the K,,w can be used as an approximation
oftheBCF.
It is important to recognize the BAF estimated using this method is based on non-
metabolizable chemicals. Thus, predicted BAFs will be larger than laboratory-measured BCFs for
chemicals that undergo some metabolism. For some chemicals, such as PAHs, the predicted BAF
can be higher than the measured BAF.
2.4.4.5
Metabolism
One factor affecting bioaccumulation is metabolism of the chemical by aquatic organisms.
Many organic chemicals that are taken up by aquatic organisms are transformed to some extent by
the organism's metabolic processes, but the rate of metabolism varies widely across chemicals and
species. For most organic chemicals, metabolism increases the depuration rate and reduces the BAF
of the parent compound.
The procedures to measure or predict BAFs differ in the extent to which they account for
metabolism. Field-measured BAFs and BSAFs inherently account for any metabolism of the
chemical. Predicted BAFs that are obtained by multiplying a laboratory-measuredBCF by a model-
derived FCM take into account the effect of metabolism on the BCF, but not on the FCM. Use of
field-derived FCMs takes into account metabolism. A food chain model prediction of the BAF (the
fourth data preference for nonpolar organics) makes no allowance for chemical metabolism. Despite
the differential effects of metabolism on predicted BAFs, information is not available for predicting
the effect of metabolism on predicted BAFs that rely on the food-chainmultiplier or predicted BCFs.
2.4.4.6 Mixtures
For chemical classes where sufficient data on the relative toxicities of individual members
of the class is available, toxicity equivalency factors (TEF) can be used to assess the total toxicity
risk of the mixture (for further discussion of TEFs, see Appendix I, Section A.3.f and Appendix III,
Section F.4 of the Federal Register notice). To date, adequate data to support use of TEFs has been
found in only one class of compounds (dioxins) (USEPA, 1989). Because individual chemicals of
a class (e.g., PCDDs and PCDFs) not only differ in their relative toxicities, but also their relative
bioaccumulation potentials, bioaccumulation equivalency factors (BEF) can also be used to account
such differences. Bioaccumulationequivalency factors have been developed for PCDDs and PCDFs
based on bioaccumulation data for the Great Lakes (60 FR 15366). As adequate data become
available to establish TEFs for other chemical classes, the BEF methodology described in 60 FR
15366 and U.S. EPA 1995c should be considered for assessing combined risk from chemical
mixtures.
237
-------
2.4.5 BAFs Used in Deriving AWQC
As discussed above for nonpolar organic chemicals, after the baseline B AF has been derived
for a chemical using one of the four methods, the next step is to calculate a BAF that will be used
in the derivation of AWQC. This requires information on: (1) the baseline BAF for the chemical of
interest using one of the four methods described above; (2) the percent lipid of the aquatic organisms
consumed by humans at the site of interest; and (3) the freely dissolved ;fraction of the chemical in
the ambient water of interest.
2.4.5.1 General Equation for an AWQC BAF
For each trophic level, the equation for deriving a BAF to be used in deriving AWQC is
applicable to all four methods and is:
BAF for
d
= [(Baseline BAF )TL
(Equation 2.4.25)
(f£)TL _ - 1] • (ffd)
where:
BAF for AWQC (TLn) =
BAF at trophic level "n" used to derive AWQC based on site
conditions for lipid content of consumed aquatic organisms
for trophic level "n" and the freely dissolved fraction in the
site water
Baseline BAFf (TL n) = BAF expressed on a freely dissolved and lipid-normalized
basis for trophic level "n"
n)
ffd
Fraction lipid of aquatic species consumed at trophic level "n"
Fraction of the total chemical in water that is freely dissolved
2.4.5.2
Baseline BAF
The baseline BAFs used in this equation are those derived from the equations presented in
Section 2.4.3 above.
238
-------
2.4.5.3 Lipid Content of Aquatic Species Eaten by Humans
The lipid content of the aquatic species consumed by humans is required when deriving
BAFs for a nonpolar organic chemical that will be used for deriving ambient water quality criteria
(AWQC). Information on lipid content is needed because it affects the extent of bioaccumulation
of nonpolar organic chemicals in aquatic organisms (Mackay, 1982; Connolly and Pederson, 1988;
Thomann, 1989) and therefore, is important in characterizing the potential contaminant exposure to
the target population (e.g., general population, sport anglers, subsistence anglers).
The lipid content of aquatic organisms can vary considerably across different species, across
different locations for a given species across seasons, and across different age classes (life stages)
of a species at a given location. In addition to lipid content, the types and quantity of aquatic species
eaten by individuals differ substantially throughout the United States. In order to account for some
of this variability in determining a representative lipid content of consumed aquatic species, EPA
recommends that the lipid fraction of aquatic organism be weighted by the consumption rate of those
aquatic species consumed by the target population based on information^/ro/w the local or regional
survey. Information on consumptionrates and lipid content is most accurately determined on a local
or regional basis and is recommended as the first choice for estimating lipid content of consumed
species. Since baseline BAFs are determined for each trophic level and must be adjusted to reflect
the lipid content of consumed aquatic species, EPA recommends that the consumption-weighted
lipid content of consumed aquatic organisms also be determined for each trophic level. If sufficient
information is not available to derive trophic level specific lipid contents, then States and Tribes may
choose to calculate an overall consumption-weighted lipid content that combines data from the
relevant trophic levels.
EPA recognizes that local or regional fish consumption data are not always available.
Therefore, EPA has derived default, national estimates of consumption-weighted lipid content for
use in deriving national AWQC, when local or regional information is unavailable. If local data on
both aquatic species consumptionrates and lipid contents are not available, States may wish to use
national default lipid values calculated by EPA. Using the general relationship in Equation 2.4.26
and information on national finfish and shellfish consumption rates at various trophic levels, EPA
has developed a national default consumption-weighted mean lipid values of 2.3% at trophic level
2, 1.5% at trophic level 3, and 3.1% at trophic level 4 (expressed to two significant digits for
convenience). The data sources, calculations and assumptions supporting of these national default
lipid values are described below.
Data Sources
To estimate a national default consumption-weighted percent lipid value for humans,
information is needed at a national level on: (1) the type and quantity of aquatic biota consumed by
humans; (2) the percent lipid of the aquatic biota consumed by humans; and (3) the trophic level of
the consumed species. These data are described below.
239
-------
Fish Consumption Data. Information on the types and quantity of aquatic organisms
consumed in the U.S. were obtained from USDA's Continuing Survey of Food Intake by Individuals
(CSFI1) (USEPA, 1998b). This survey provided daily average per capita estimates of fish
consumption for the U.S. population for categories of estuarine, freshwater, and marine fish and
shellfish. Although other regional or local surveys were available, the CSFH was selected because
it provided consumption information on a national basis and was the most recent data available. In
this survey, consumption rates were divided into 16 categories representing mostly estuarine species
and 5 categories representing mostly freshwater species. Mean per capita consumption rates were
characterized for individuals 18 years and older. For a detailed discussion of the use of the CSFII
data see the chapter on exposure in this TSD. Table 2.4.7 displays the habitat classification, CSFII
consumption categories, mean per capita consumption rates, and the fraction of total estuarine and
freshwater consumption represented by each category in the survey.
Lipid Content of Consumed Species. The second type of information required in deriving
national default values for lipid content includes data on the lipid content of consumed aquatic
species. Six primary data sources were used to estimate lipid content. These include: EPA's
STORET data base, EPA's National Study of Chemical Residues in Fish (USEPA, 1992), two
reviews from National Marine Fisheries Service of the National Oceanic and Atmospheric
Administration (Sidwell, 1981; Kryznowek and Murphy, 1987), data from the California Toxic
Substances Monitoring Program (TSMP), Green Bay Mass Balance Study (USEPA, 1992b, 1995g),
and a study of the Hudson River conducted by the New York Department of Environmental
Conservation (HydroQual, Inc., portions published in Armstrong and Sloan, 1988). Each of these
data sources are discussed in more detail below.
STORET. Data from EPA's STORET (STOrage and RETrieval of U.S. Waterways
Parametric data) data base, a waterway-relatedmonitoring data base, were retrieved by downloading
tissue and sediment chemistry data from ambient non-land-basedmonitoring stations, including lipid
content, from 1980 through October, 1993. Data are stored in STORET by many government
agencies, both federal and state. Most of the data used in this analysis were collected in the
Midwest, in particular in Illinois, Minnesota, Michigan, Iowa, Kansas, Missouri, and Nebraska. The
number of individual organisms per sample is not known. Information on the common and Latin
name for the species sampled, the tissue sampled, the percent lipid content, the collection date and
location, and the agency responsible for the data were retrieved from STORET. Data from the other
data bases used in this analysis were confirmed as not being present in STORET, which could have
caused double-counting of some samples.
240
-------
Table 2.4.7: Aquatic Organism Categories and Average Consumption Rates from CSFII
Estimated Mean Percentage of
USDA CSFII Consumption Rate Total
Habitat Category (g/person/day) Consumption
Rate
Estuarine shrimp
perch
estuarine flatfish
crab
flounder
oyster
mullet
croaker
herring
smelt
clam
scallop
anchovy
scup
sturgeon
Freshwater catfish
trout
carp
pike
freshwater
salmon
1.72959
0.60368
0.52735
0.37126
0.29941
0.22555
0.08756
0.06749
0.03925
0.03753
0.03146
0.00322
0.00292
0.00068
0.00054
1.18227
0.44946
0.05727
0.02337
0.01096
30.08%
10.50%
9.17%
6.46%
5.21%
3.92%
1.52%
1.17%
0.68%
0.65%
0.55%
0.06%
0.05%
0.01%
0.01%
20.56%
7.82%
1.00%
0.41%
0.19%
Source: USDA combined 1989, 1990, and 1991 Continuing Survey of Food Intakes by Individuals (CSFII).
Notes: Estimates are projected from a sample of 8,478 individuals 18 years of age and older in the U.S. population
of 177,807,000 individuals 18 years of age or older using 3-year combined survey weights.
The population for this survey consisted of individuals in the 48 conterminous states.
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Database for individual food intake surveys.
The number of digits does not imply their significance.
241
-------
National Study of Chemical Residues in Fish. This EPA study was a one-time screening
investigation to determine the prevalence of selected bioaccumulative pollutants in fish (USEPA,
1992a). Three to five fish collected from one location were used for each composite sample. For
each composite sample, two measurements of the percent lipid content were obtained, one from the
test for dioxins/furans and one from the test for other xenobiotics. The average of the two lipid
values was used to represent each sample data point. Location and sampling date information were
available as was the common name of the species collected and the tissue type sampled (whole body,
fillet).
NOAA Data. Data on the lipid content of estuarine species from the National Marine
Fisheries Service of the National Oceanic and Atmospheric Administration were available from two
reviews (Sidwell, 1981; Kryznowek and Murphy, 1987). These reviews consist of compilations of
data from primary literature sources. Information on the specific location and number of individuals
per value were not available in these reviews. Data not collected in North America were excluded,
where information was available to make this distinction. Information was available on common and
Latin name, tissue type, and method of preparation (e.g., raw, cooked). Only samples that were
indicated as being fresh or raw, or for which no preparation information was available, were used
in the analysis. Other information such as the number of individuals in a sample, age, weight, and
sex were not available. This data source was used to augment the data from the other data sources
which were very limited in quantity. For those categories (catfish, trout, carp, pike, and perch) for
which we had extensive data from the other sources, data from NOAA were not used.
California Toxic Substances Monitoring Program (TSMP). This program is run by the
California Environmental Protection Agency, California State Water Resources Control Board,
Division of Water Quality, Monitoring and Assessment Unit. The TSMP was organized to provide
a uniform statewide approach to the detection and evaluation of the occurrence of toxic substances
in fresh water, and to a limited extent, in estuarine and marine waters, through the analysis offish
and other aquatic life. Samples are collected annually and composite samples of six fish are
collected when possible. The data base provides information on age, sample collection date,
location, number of organisms per sample, weight, and length. Most samples for the species of
interest are fillet samples, although some whole organism samples are also present. Data were
obtained via the Internet at World Wide Web site: http://www.swrcb.ca.gov.
Green Bay Mass Balance Study. This study includes lipid content from aquatic species from
Green Bay, Lake Michigan. All data are whole body samples. Included in the data base are
information on collection date, zone of bay in which the sample was collected, age, and number of
fish in average or composite. Data were obtained from HydroQual, Inc. (U.S. EPA 1992b;U.S. EPA
1995g). These data are also available on the World Wide Web at www.epa.gov/ghipo/monitor.html.
Hudson River Data. The New York Department of Environmental Conservation conducted
sampling from which percent lipid content for fish species from the Hudson River were available.
All data are for muscle fillet samples, and there is one individual per sample. In this data base,
242
-------
collection date, length, weight, and sex are available, as well as location in the form of river mile.
Data were obtained from HydroQual, Inc. (U.S. EPA 1995h; 1995i and Armstrong and Sloan, 1988).
Data Analysis
Data Screening/Treatment. Several steps were required to prepare the data for the calculation
of average, consumption-weightedlipid content. As described further in the following section, each
record was assigned to a CSFII species category based information on the common and Latin names
given in the various data bases and information on whether they could reasonably be expected to be
consumed. Data for those species that could not be assigned to a CSFII species category were
omitted. For the STORET data, only those data that contained the common name, Latin name, and
species code were used so as to maintain consistency of species names within and across data bases.
In addition, several steps were taken to correct species codes where known mistakes occurred. An
upper bound lipid content was also set at 35 percent to exclude extreme values that were considered
outliers. Very few values occurred above 35 percent.
Data were screened by tissue types that corresponded to those considered most commonly
consumed by humans. For all finfish species, with the exception of herring, anchovy, and smelt
species, data were considered only for selected tissues that include: fillet, fillet/skin, muscle, meat,
and flesh samples. The great majority of samples used for the finfish species were fillet or fillet/skin
samples. For crab species, data considered in the analysis included tissues designated as edible flesh,
edible portion, or edible skinned, in addition to those described for finfish. For the remaining
shellfish species (clam, oyster, scallop, and shrimp) and for herring, anchovy and smelt, whole body
samples were considered in addition to the tissue types used for fish and crab species. These criteria
were established based on the portions of a species are believed to be consumed and constraints on
the availability of data.
Selection of Species for Inclusion in Lipid Analysis. Given information from the CSFII
survey on the types of aquatic organisms consumed in the U.S., the next step in calculating the
consumption-weighted average lipid content involved assigning species to the general CSFII
categories. In most cases, information was not available from the CSFII survey to identify which
species were included for determining the consumption rates listed in Table 2.4.7. Therefore,
inclusion of species and accompanying lipid data into a CSFII category was based on: (1) their
taxonomicandpublicly-perceivedlinkageto a CSFII category, (2) consideration of their likelihood
for being caught (either recreationally or commercially) and consumed in the U.S., and (3) their
occurrence in either fresh or estuarine waters for at least some portion of their life cycle. The species
included in the lipid analysis and their relationship to the CSFII consumption categories is provided
in Table 2.4.8. Information from numerous published sources were used to help determine whether
a species met these criteria. Because several of the CSFII species categories were broad in terms of
the types of species that could be included, some species were assigned to multiple CSFII survey
categories. For example, flounder species fit into both estuarine flatfish and flounder categories.
In such cases, appropriate records were included in both CSFII categories. Notably, some species
243
-------
that are commonly caught and consumed but did not fit into a CSFII category, such as bass and
walleye, were not included in this analysis.
Lipid Content of Species in CSFII Categories. Based on lipid data from the aforementioned
sources, an average lipid content was determined for each of the species in the CSFII consumption
categories (Table 2.4.8). Next, the overall average lipid content of each CSFII category was
determined as the average of the corresponding individual species mean lipid values. Ideally, if
sufficient national consumption data were available at the species level, the overall average lipid
value for each CSFII category would be determined on a consumption-weighted basis. However,
sufficient national information was not available below the CSFII category level and thus, equal
weights were assigned to each species mean lipid value. For example, lipid contents were available
for several species of trout (e.g., rainbowtrout, brown trout, and others), whereas consumption rates
were available from the CSFII only for trout as a group. Thus, mean lipid values for all trout species
were averaged and combined with the consumption rate for trout from the CSFII.
Trophic Level Assignments to CSFII Categories. National fish and shellfish consumption
data from the CSFII (see Table 2.4.7) indicate that on average, individuals consume aquatic
organisms from a variety of trophic levels (e.g., oysters and clams in trophic level two, flounder and
shrimp in trophic level three, perch and certain catfish species in trophic level four). Therefore, for
the purposes of calculating national AWQC using the CSFII consumption data, BAFs need to be
derived that are applicable to each of these trophic levels and should be adjusted to reflect the
average lipid content of organisms consumed in each of the trophic levels. To estimate the
consumption-weighted average lipid content in each of the three trophic levels (and to estimate
consumption rates of aquatic organisms within each of the trophic levels—see Section 2.4.8), a
trophic level designation must be assigned to each of the consumption rate categories of the CSFII
shown in Table 2.4.7.
In order to estimate the trophic level status of consumed aquatic species, one should ideally
rely on information concerning the identity, size, age, and diets of individual aquatic species
consumed. This information is useful because not only can individual species differ in their trophic
status, but trophic level status can also differ for different sizes (ages) of individuals within a species
because diets often change with age/size of the organism. Information on the identity, and size (age)
of consumed aquatic species should be obtained from the fish consumption survey, if available.
Information on the diet of consumed aquatic species might be available on a local or regional basis,
but more often is scattered about in the scientific literature based on studies of various sites around
the United States. If local or regional information is not available, then EPA recommends the use
of the most recent version of the document: Trophic Level and Exposure Analysis for Selected
Piscivorous Birds andMammals (USEPA, 1995d, 1995e, 1995f), which contains information on the
dietary composition of numerous aquatic species. This draft document is currently being revised
based on peer review comments and is expected to be made final in 1998.
244
-------
Table 2.4.8: Lipid Data for Aquatic Species Included in the Derivation of a National Default
Consumption-Weighted Lipid Value
CSFH
Consumption
Category
Anchovy
Carp
Catfish
Clam
Crab
Croaker
Estuarine
Flatfish
Estuarine
Salmon®
Flounder
Common Name
Northern anchovy
Striped anchovy
Carp
Black bullhead
Brown bullhead
Channel catfish
White catfish
Yellow bullhead
Butter clam
Geoduck clam
Hard clam
Littleneck clam
Soft shell clam
Venus clam
Blue crab
Dungeness crab
King crab
Snow crab
Atlantic croaker
Spot
White croaker
Yellowfin croaker
Gulf flounder
Rainbow smelt(1)
Southern flounder
Starry flounder
Summer flounder
Winter flounder
Atlantic salmon
Chinook salmon
Chum salmon
Coho salmon
Pink salmon
Sockeye salmon
Gulf flounder
Southern flounder
Starry flounder
Scientific Name
Engraulis mordax
Anchoa hepsetus
Cyprinm carpio
Ameiurus melas
Ameiurus nebulosus
Ictalurus punctatus
Ameiurus catus
Ameiurus natalis
Saxidomus nuttalli
Panope generosa
Mercenaria mercenaria
Protothaca staminea
My a arenaria
Tapes philippinarum
Callinectes sapidus
Cancer magister
Paralithodes camtschatica
Chionoectes bairdi
Micropogonias undulatus
Leiostomus xanthurus
Genyonemus lineatus
Umbrina roncador
Paralichthys albigutta
Osmerus mordax
Paralichthys lethostigma
Platichthys stellatus
Paralichthys dentatus
Pseudopleuronectes
americanus
Salmo salar
Oncorhynchus tshawytscha
Oncorhynchus keta
Oncorhynchus kistuch ,
Oncorhynchus gorbuscha
Oncorhynchus nerka
Paralichthys albigutta
Paralichthys lethostigma
Platichthys stellatus
Species Mean
Lipid (%)
11.70
2.80
4.45
1.12
2.79
5.00
2.15
0.75
1.22
3.20
1.04
0.75
1.29
2.60
2.33
1.15
0.80
1.30
3.30
10.35
2.33
3.11
0.80
4.50
1.20
0.97
0.40
1.00
5.65
2.09
4.25
2.17
5.00
8.12
0.80
1.20
0.97
Sample CSFII
Size Mean
Lipid (%)
2 7.25
1
1433 4.45
1 1 2.36
704
1108
46
29
2 1.68
1
2
2
5
1
7 1.40.
2
1
1
3 4.77
2
7
1
1 1.48
88
1
3
1
2
2 4.55
271
2
308
2
4
1 0.87
1
3
245
-------
Table 2.4.8: Lipid Data for Aquatic Species Included in the Derivation of a National Default
Consumption-Weighted Lipid Value
CSFEE
Consumption
Category
Freshwater
Salmon
Herring
Mullet
Oyster
Perch
Pike
Scallop
Scup
Shrimp
Smelt
Sturgeon
Trout
Common Name
Summer flounder
Winter flounder
Atlantic salmon
Kokanee salmon(3)
Chinook salmon
Chum salmon
Coho salmon
Pink salmon
Sockeye salmon
Atlantic herring
Blueback herring
Pacific herring
Striped mullet
Eastern oyster
European oyster
Olympia oyster
Pacific oyster
White perch
Yellow perch
Chain pickerel
Northern pike
Atlantic bay scallop
Sea scallop
Scup
Brown shrimp
Northern pink shrimp
Pink shrimp
White shrimp
Rainbow smelt
White sturgeon
Brook trout
Brown trout
Cutthroat trout
Lake trout
Rainbow trout
Scientific Name
Paralichthys dentatus
Pseudoplewrnectes americanus
Salmo solar
Oncorhynchus nerka
Oncorhynchus tshawytscha
Oncorhynchus keta
Oncorhynchus kistuch
Oncorhynchus gorbuscha
Oncorhynchus nerka
Clupea harengus
Alosa aestivalis
Clupea pallasi
Mugil cephalus
Crassostrea virginica
Ostrea edulis
Ostrea lurida
Crassostrea gigas
Morons americana
Perca flavescens
Esox niger
Esox lucius
Aequipecten irradians
Placopectens magellanicus
Stenotomus chrysops
Penaeus aztecus
Pandalus borealis
Penaeus duorarum
Penaeus setiferus
Osmerus mordax
Acipenser transmontanus
Salvelinus fontinalis
Salmo trutta
Salmo clarki
Salvelinus namaycush
Oncorhynchus mykiss
Species Mean
Lipid (%)
0.40 ,
1.00
5.65
2.28
2.09
4.25
2.17
5.00
8.12
13.04
8.63
9.34
4.49
1.94
1.65
0.50
2.40
5.34
0.66
1.21
0.47
0.60
0.80
3.70
0.93
0.78
0.78
0.52
4.50
1.09
1.51
3.81
1.23
10.90
4.00
Sample
Size
1
2
2
3
271
2
308
2
4
5
3
8
9
8
2
1
8
296
220
5
356
1
2
1
3
2
2
2
88
6
7
142
16
380
123
CSFII
Mean
Lipid (%)
4.22
10.34
4.49
1.62
3.00
0.84
0.70
3.70
0.75
4.50
1.09
4.29
01 Information from the CSFII survey indicated that rainbow smelt were included in the calculation offish consumption rates for estuarine flatfish.
Because these species are anadromous, data on lipid content were also included for freshwater salmon category.
°»Information from the American Fisheries Society publication: Common and Scientific Names of Fishes from the United States and Canada (AFS,
1991) indicates that freshwater stocks of sockeye salmon are commonly referred to as Kokanee.
246
-------
For the national CSFII survey, very limited data were available to further delineate the
identity and size of species consumed within each of the CSFII categories in Table 2.4.7. For most
of the CSFII categories, this lack of information was not viewed as problematic, because rather
unambiguous assignments of trophic status could be made to these categories (e.g., all oysters are
considered to be trophic level two). However, for other CSFII categories, assignment of trophic
status required some reasonable assumptions to be made and therefore reflect greater uncertainty.
The folio wing procedures were used hi assigning trophic status to the CSFII consumption categories.
1. DatafromEPA'sdrafttrophicleveldocument(USEPA, 1995d, 1995e, 1995f)and
other sources were used to estimate the trophic level of species that could
reasonably be classified in each of the CSFII consumption categories. Species
level trophic assignments were performed as follows.
a. For game fish that correspond to the CSFII categories, data were used for
edible size ranges (about 20 cm [8 inches] or larger).
b. For species where multiple size ranges were available, preferences was given
to the larger specimens in determining the species trophic level.
c. Trophic level 2 was assigned to a species if appropriate trophic level data
ranged between 1.6 and 2.4; trophic level 3 if trophic level data ranged from
2.5 to 3.4; trophic level 4 if trophic level data were 3.5 or higher. This is
consistent with the approach taken in the Great Lakes Water Quality
Initiative guidance (USEPA, 1995c).
2. Once the species level trophic assignments were completed, this information was
used to assign a trophic level to each CSFII consumption rate category as follows.
a.
b.
In situations where a CSFII category was represented by the vast majority of
species within a single trophic level, that trophic level was assigned to the
CSFII category (e.g., trout, estuarine flatfish, smelt).
In one situation (catfish), the CSFII consumption rate was equally divided
between trophic level 3 and 4 because about half the species were determined
to be trophic level 3 and about half trophic level 4.
c. For shrimp, trophic level 3 was assigned based on data for the "general
shrimp category" in USEPA (1995f) because other data (grass shrimp,
mysids) were for species that are not consumed by humans.
The results of the trophic level assignments are shown in Table 2.4.9.
247
-------
Cfl
•S
1
03
CSFII Consumption C
esponding to
h
>•
0
U
.§
"5
a
CO
.—
3
CT
•s
"S
o>
S
1
CO
3
1
H-l
."
15
B.
O
r*"*
• ft
0\
^f
.2
*
^
"O cj •?
E Sj'i.f
3 2 Si
U < E- -J
"g 0 -
.§ !*1
13 1 5
5
o
Z
•H *** 'S*
-B — 6C
Q. a> c
o g w
•Is -s-
g-lf 8
~
«*
»
„
ll
S^
B
0
B U
E E
O CJ
O Z
e
.S
<••*
n. ^
E %
S S w>
CA o 'el
U U U
co
CO
c
o
2
n cs n
o "E. S
^ O ^i
C C) C
CO N CS
Feeds on zoopl
Feeds on large
Feeds on zoopl
! ! i
CO £. CO
JB
— '1
i 31
H
1
5
!
s
E >>
O >
Z g
^»
o
•s
B
CO
"
Cf"5.
IB i M
ti 1 -sll
§1 1 *al
1| ^ 111
«^ ii^^
- e E ra ^ >o
•5. « 'g. E 5 ^?
O es "• o to |« ^
oofeo.> cs-£gv
>• E j= D.E u 0.^2-0
^M ^M V^
ro ro fN
1 1 1
CN 00 CS
CS CS CS
i _ 'S-
! ^ CN
1 E
ff\ ^
CS CO
1 1-^1
! o V;1
•2
1
s
s.
§
§•
U
f~
o
E
o
U
fr
cs
U
^^
1 1
co 'S'
CO CO CO T(-
i| oil |& I
Hi • if! f!!H
Hi Jit Mill,
lisl d S-SscS^^I IS^
8.1 1{| l^lil ^-S-|tl
o ., E § 0.55 e^-^is «8c3e E • -e
•"•-co.E'o go°_ Mg)§ o^^
CS CO CS j.
rf? « -f? ^ .
1 ill OO '
O\ ' t-; f-; ^ '
CS CS CS
M w i S i ^
E § e
, , 2 S ^
i i i A co >A
§ §
a § 2
J 1 -I i
^3 P -g s
c ^52 %
5 «5, « .ex,
a t^ a
-------
09
(M
o
b£
•S*
03
g
0
• M
•*•*
a
I Consum
ta
XI
U
3
•esponding
*••
1U
a
.1
"3
V
a,
QC
«
es
S
cr
"^1
o
**
e
i
g>
•P«
V5
52
j
0,
2
M
^B
*
•
hj
M
- "g-as
|tfl
w <» i
.•8.8s
£ E js ~
•5 M a. ~
u w o ;
w < H -
W3
"o
.a * "a?1
•S3 ^. bjn
0. « B
314
.a s c*
•ffi *"
0."3 g
H J S.
*
Q>
.E3
t»
Scientific
Name
Common
Name
B
•2
i* *•*
S B a*
n o es
U U U
' ^^ ^3" CO
M JO
o 3 2 o o <« g 0
^ 'S o. IgEiu g § «"§
-"cl -CoVo So •> nc
cs>«3 ^^^-^.^ ? '5™
o*1 •» C O1^ ®® ^* *^P >-jf| _S -C *" **
CSCa^O ^^^O^"^ i^ WP
||| iiflfl fi Ii
MSoiJ ^N55 o 't* <& .2 -
W CQ ^-^ ^5 WS^2 £ hn"3"3 o£2 ^"2
ir» °5 ^
• cn Tf ""^ i i
S £ 5 - ! !
£ £• •* ^ °« ^
en m m en (N
_
o 5 E °
U-l 10 0 >o
en -«4- v-i TJ-
O ^5 ~
•C ^ CO fli
O E E >•
•o
c
u
_c
U
fN
tN (
B
4*J
«
•a
o
"c c
CO O
D.S
if
N CN fN CN
v- O t- i_ t- l_
O U U U
•a N -a -a -o t:
u m o u a> a>
<2J S 18 <& .S «
*-. ^ U b« u* L^
Q> . fl> O Q> Cv
ii T) — = £ J=
E £ E iZ tZ cZ
ts rji CN
U < So < CQ Q
U
en
m
en
p
i^^
« 0 g"
^«N 2-
111
E^l:
§E^
^>.¥-
Feed primari
organic debr
crustaceans (
(2%)
i
CN
> =
"*"• CO
15-
ll!
**• «
|18
III
0 " ..
Opportunist!
polychaetes,
clams (From
1
1
1
!
Micropogonias
undulatus
L_
^
«1
Atlantic cro;
JH
e
o
-------
.i
o
jr
u
a
0
••G
CL
esponding to CSFII Consum]
fa
o
U
S
"3
OJ
c.
C/J
u
•«•»
s
O"
«*H
o
"S
«
i
I
"^5
a>
•j
P*H
u
S
2
H
-%
ON
TT
M
^
-•g|2
b Ji B. "3
3 ||e
u 'Si O >
Q. to u cu
W < H -J
'o
Z
.a 'S' /aJv
a. ~ B
H J S
O*~* ^— ^
*•• ** ^^
cL ^* ^
S w S
H j e
cio
u
.S
*^ ^
.B ^
*3 J5
co Z
o
E o
i i
B
.2
-|£
E *" 4»
C/3 O ^8
O U U
en
|
d.
£ u
Si *
Feeds on benthic organisms (e.g., s
urchins, sand dollars, brittle stars, s
molluscs, & worms) most of which
detritivores.
s
en
C*3
en"
j
i
•S "S
i s
o o
n •§ o" c
3 0 2 3
i2 ^ m =
^ ^wX UJ (^
en
.^
Assumed equal to general flounder
^.
en
i
p
en
fN
en
i
i
Righteye
flounders
en en
'•B '•£
Assumed same as general flounder
Assumed same as general flounder
TT •*
en" en
en en
f-4 (N
en en
i i
52 Qi
•§ ^
•w O
&2 £3
-^ "5.
••5 C*
'~ "§
ft! £
"- U
•n "°
^5 Gl
C "^
Starry flou
Winter floi
en
0.
U
u
X
<1)
Assumed same TL as other smelt, <
for adult smelt in Great Lakes
1
CO
1
Jo
s:
•^
S>
i i
"2
u
Smelt (gen
,
s
CO
.s
u
1
on
i
"E.
o
S
§
CO
-o
0>
1
*«.
CO
1
8
&^
o
S
g
^)
s
a
•4-t
"o
CO
American
(Rainbow)
<5
"ca
§ ..
In Great Lakes, > lyr old feed on s
fish and on Mysis (TL3); and othei
inverts & zooplankton (TL2)
j
^"i
en
,
i
en en
During first year, feed on zoopl.
Assumed same as other smelt
i i
*"*• en
en
i !
JK.
s
o3
»
a
,&
^x.
fl*
S
.SP
"E
"5 *-
11
^^ CO
en en
Assumption
Assumed same TL as other smelt.
i i
en en
| |
to
|
t
&
J;
t*
i
o
S
i £
SX /o1
3 tO 4-
| 6 1
Illl
-------
1
o
1
fi
.2
•esponding to CSFII Consumpl
M
O
U
.§
*3
!&
.g
«
O"
O
1
1
.£?
'oj
"«5
15
2
^•f
CS
Slie
'3 .S? n. "S
en
"o
•C 'Z* bfi
a. « c
i j B!
o ~
a. "<5 S
O ^ m
«. « s
H
cu
.H
Scientific
Name
Common
Name
s
.0
CSFII
Consumpt
Category
*"
' m m m en m
£
- •§ 2 .1 S 1 ^ «S
SJSCS T3 H5 "5 "5 T3
en en .c i- i_ i_ fc. i_
..O O Q> U O O
?? § -i 1 "c 1 "c"c
O *-S ;S 3 3 3 33
-— ' m £, 0 O O O O
en _o o C rtrf rA ^^ 03 CT C3 QJ
^ra^go « " c
CS o Q
.£ -s §
CO IX N
CS. •*
O —
CS CS
— ; CS
CS CS
i i
i i
O "r; en TS
a 2 -3 2
J5 o en O
— C to C
O 3 0
!
*
-^f
.^
Ml
•5. S "3
ill
11|
E ?> j=
3 -en
c t ti=
O CJ ""?
0 > 0
E -s 2
0 ." ~
o jr ^>
en *= t.
I C CS
O 1) O>
CS J3 C
OO
2
en
^
rn
g
0
1
CS
J2
1
}
J=
1
1
CD
1
-------
S
*n
o
•4*^
63
U
B
o
«
1
esponding to CSFII Cons
t
a
tn
.«
"3
B,
CO
.J-
•*^
«
cr
(M
O
8
a>
S
.&
1
1
o<
o
H
* •
0\
^
r*
jO
"S
«
fr"
•O o IT
*• .1? a >
•o -, _
u, u .s 3
jrj OD Q. "3
U *M O >
co < H J
o
.a «• -s1
J= _ OB
Q. es
^"* T 0£
^ *J \^f
.H S ^
"o."o S"
?* i
«
00
u
'S u
•i 1
C3
O
E 10cm; diet primarily all fish.
Larger specimens consume sm
I j
en
c
E ^ i
o a> "j;
A — a
I
.1 1
J 1
1 1
u
'5u 0£
Northern
Pickerel
(redfin&
Jd
CU
CN
(S
Bivalves feed on plankton
(S
(N
0
f**J
^-
04
!
i
w [
11
a.
"3
CO
CO
•a
c
ra
11
"o ^
Bottom feeder, primarily on m
worms, and small crustaceans I
Evermann, 1969)
i
!
E,
1
§
I
1
53
n.
3
CO
Cu
3
CO
c*^ CS ^f
^
__ w
11
1 1
Filter feeding on zoopl.
Detritivore, primarily on plant
Cold water forms; during warr
restricted to hypolimnion
• ! "1"
o — -22
"5 § >»
co O 2
n.
.E
CO
i ^_ ^^ ^_ ff*
i rr ^*
"^ to
$ s § I
1 2 "I 5
S «i> -S 3
5 | -S g.
f 1 1 i, o. >?
g c /? e ^ "5 -2 ~ ^
^eo 8° 8°~ ^'Cj
1
co
-------
1
§D
•w
es
8
.0
•*•»
Q.
&
S
'esponding to CSFII Cons
mm
o
U
.i
'3
V
O,
'•S
s
V
^^
«*.<
o
s->
e
i
8
DJD
*i
•^^
"«
a>
.«
JS
a.
0
.^
ON
^
S
- "g .as
1||1
slis
'S .SP o.'S
U in O >
Q. t« V. «
c« < H J
'S
u 'S' ^
•S « M
Q* V tS
P S ^
^M M^ ?_*^
ej —
||I
H J§-
o>
,N
CO
c
** «
S S
y |TO
fi f~i
e
o
E «
s e
5^
•*^
- 1 e
2 s SB
S S i
0 O 03
U U U
^
•* ^r
w e
. . .IM C ^ ^ ' * "^
SwT^g o So|° —
•S ~ "s « -g 'E _g ^ .S -- «? 'C .S 85
& O C *PT » &0 ^ ^*^ QJ ^^ "O £!
jliif?li^llillli
!lll?t§'ll|slla"-f
||'|.|v|||||^|>||^|l
"T T
i 1
i i
§
"a
i
GO •<£;
*S oQ
•1 *
S g fe
II I
^2 u
c
0 C
U 0
op rr
4-»
3
Xi en
.¥ s -i
^ J> 03 m
10-40cm; at most, 7-8% fish in
remainder primarily benthic in\
also some zoopl. in some popul
< 40cm; consumes invertebrate:
> 40cm; becomes piscivorous
: : :
i t i
3
s e
! 1
§ i
1 5
3
e
-------
.i
u
1
u
c
.S
M
Q.
S
S
CO
O
U
ft
«
2
BJD
^B
T3
B
O
C,
en
E
o
U
£
'S
o
0,
o
•«
9
O
•*•»
B
u
S
6JD
's
•— «
>
GJ
.—
IS
a,
o
H
ON
•
o
A
H
_ "S .y s
- e -C —
8||5
•3 ti ~
tn ** ••• -^
8 e .e 71
5 .£? Q.TS
S in O >
S. S b U
cn < E- J
Z
u - ,-.
iS — OJJ
o, y a
0 > 63
H -2 &
ti ^^ —
•5 "^ ^
o" > ^
H J ^
en
u
tf^
.§ s
CA Z
e
o
E o
E E
O S3
U Z
e
«
c E-"
— 3 §)
E S «
C/3 O 68
U U U
-
.
M
c
u
r oldest specim
thage
e 1
o w
~ en
o. 5
E 5
3 "
Ul J
< H
' r-i
^* *
^ i
old adults
J
1
g
5*
1
e
rt> JT^ .-• >^ ^r? ^T* *~* r* +* ""
3? ^*^ ?•* p^ OT sO •& • gj f^.,
c« *^ x? w ^5 ^^ ®^ f*** 5 ^S- ^ o
"77s? 2^ r- f*^ e ^ en «o^O
•J^ 00^§C,«=8§n c&ij3cx-— g°jrS —
•St> "nu'S'o'S-o'§'J"o
^^o •-'2^t^-0{ut<",^^"£
1 1 « E i -J, § 1 E.S>'1 s w
°^f|^l^l|i^ll
o — «•>
ii i •*" TT TT i
' ' ' ' .J J.
II 1 10 t^; O *
cn m ^
10 t*^* O\ C'i _-.
co r*" ^- • • * m
§ § E 1 § s E
00 5 0 0 g 0
c<-> ^r o m •* ,5 o
0 0 ^ 0 0 f f?
— m A CN m A V
•§ .g
S* H,^
£• 5^ S
i l i
§ § i
S .K ^
01 w O
•S -=t s;
C§ CO O
c 3
(U O
i i ;
> s i
3 — 'S Js •£
is a is j B:
«2
o
en
g|
i "C
*o ^fl
tN U
xf S j=
en 4— en
if i
? S o
VI —
-*-» i2 O
^ > g
§ .a E
i i
VO ^.
en
§ E
O 0
w> O
1 I/-J
m A
T3
§
a
•8
1
O
c/3
C
.1
0.
o
•a
1
*O
1
1
•o
•§
O
c-
_J
oc
o
1
a
•S3
•o
'i
en
1
&
CB
O
•a
u
J>
en
§
S
s
^
•«3
«
— u
i: e
— • tn
•g.3
§ s
2 u
El
0 3
* «
'- "-
fc> 3
M.
ca ob
— ~S
E -3
C/5 >
g.-
•o -J
a
op-a
-------
Calculation of Consumption-WeightedLipid Content, by Trophic Level. Using consumption
rate data from CSFII (daily average per capita estimates of individuals 18 years and older- Table
2.4.7), the mean lipid content estimated for organisms assigned to each CSFII category (Table 2.4.8),
and the trophic level assignments of each CSFII consumption category (Table 2.4.9), consumption
weighted mean lipid content determined within each trophic level according to the following
equation.
' CR.
CR
tot
where:
CR,-
CR™
(Equation 2.4.26)
Lipid fraction representative of aquatic species eaten by the target population
that correspond to a given trophic level
Consumption rate of species "i" of a given trophic level eaten by the target
population
Consumption rate of all species at that same trophic level eaten by the target
population
Lipid fraction of species "i" eaten by the target population
Calculation of the consumption-weightedlipid content values is shown in Table 2.4.10. The mean.
consumption weighted percent lipid values were calculatedas (expressed here to 2 significant figures
for convenience):
Trophic Level Two: 2.3%
Trophic Level Three: 1.5%
Trophic Level Four: 3.1%
Because of limitations in the availability and precision of the used to estimate consumption
rates, lipid content, and trophic level status, uncertainty exists in the estimation of national default,
consumption-weighted lipid content. To illustrate some of this uncertainty, "high" and "low"
estimates of the consumption weighted lipid content values were determined using the species with
the highest and lowest species mean lipid value, respectively, within each CSFII category. "High"
and "low" estimates of percent lipid content values within each trophic level are:
"High" Estimate for Trophic Level Two: 3.0%
"High" Estimate for Trophic Level Three: 2.2%
"High" Estimate for Trophic Level Four: 6.2%
255
-------
"Low" Estimate for Trophic Level Two: 1.5%
"Low" Estimate for Trophic Level Three: 0.77%
"Low" Estimate for Trophic Level Four: 0.91 %
The reason that there is not a greater difference between the mean lipid content values (where
each species within a CSFII category was given equal weighting) and the "high" and "low" estimates
is likely because the mean consumption rates in the CSFII survey are weighted heavily by relatively
lean aquatic organisms such as shrimp, crab, perch, and flounder. Therefore, because the
consumption of aquatic organisms may differ on a local or regional basis from that reflected in the
CSFII survey, EPA recommends that States and Tribes give preference to using local and regional
data on consumption patterns over national default estimates, when available.
2.4.5.4 Freely Dissolved Fraction
The next step in calculating a BAF used in deriving an AWQC involves adjusting the
baseline BAF to account for the freely fraction of the chemical at the site(s) to which the AWQC will
apply. The same equation used to estimate the freely dissolved fraction for determining a baseline
BAF (Equation 2.4.11) is used to estimate the freely dissolved fraction for determining the AWQC
BAF. However, hi this case, however, the POC and DOC values should be based on the site(s)
where the BAF and the criterion will be applied and not where the samples were collected for
determining the BAF. If POC and DOC data are not available for the site(s) to which the AWQC
will apply, then data from sites closely related to those to which the AWQC sites can be used. Care
should be taken to ensure that conditions affecting the POC and DOC concentrations at the surrogate
sites are representative of conditions at the AWQC sites. States and tribes are encouraged to use
local or regional data when appropriate and scientifically defensible. If such data are unavailable,
then the default values for POC and DOC can be used. EPA has developed national default values
of 0.48 mg/L (4.8 x 10'7 kg/L) for POC and 2.9 mg/L (2.9x 10"6 kg/L) for DOC. Both of these values
are 50th percentile values (medians) based on an analysis of over 132,000 DOC values and 81,000
POC values contained in EPA's STORET data base. These default values reflect the combination
of values for streams, lakes and estuaries across the United States. Further delineation of the POC
and DOC concentrations in different water body types is provided in Table 2.4.11. These default
values, which are derived at a more disaggregated level may provide more appropriate estimates of
POC and DOC concentrations associated with the field BAF study compared to the national default
medians listed above. The K,,w value for the chemical is the same as used for deriving the baseline
BAF for the chemical.
256
-------
Table 2.4.10: Calculation of National Default Consumption-Weighted Mean Lipid
Content of Consumed Aquatic Organisms
Habitat
Estuarine
Estuarine
Estuarine
Estuarine
Estuarine
Freshwater
Freshwater
Estuarine
Estuarine
Estuarine
Estuarine
Estuarine
Estuarine
Estuarine
Estuarine
Freshwater
Freshwater
Estuarine
'reshwater
istuarine
7reshwater
CSFH
Category (1)
clam
mullet
oyster
scallop
anchovy
carp
catfish
crab
croaker
estuarine
flatfish
flounder
herring
scup
shrimp
smelt
catfish
freshwater
salmon
perch
pike
sturgeon
trout
Assigned
Trophic
Level12'
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
Trophic
Level
Weighting
Factor0'
1.0
1.0
1.0
1.0
1.0
1.0
0.5
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
0.5
1.0
1.0
1.0
1.0
1.0
Average
Percent
Lipids (4)
1.68
4.49
1.62
0.70
7.25
4.45
2.36
1.40
4.77
1.48
0.87
10.34
3.70
0.75
4.50
2.36
2.28
3.00
0.84
1.09
4.29
Mean
Consumption
Rate
(g/person/dav)
0.03146
0.08756
0.22555
0.00322
0.00292
0.05727
1.18227
0.37126
0.06749
0.52735
0.29941
0.03925
0.00068
1.72959
0.03753
1.18227
0.01096
0.60368
0.02337
0.00054
0.44946
Trophic Level
Trophic
Trophic
Trophic
Level 2
Level 3
Level 4
Total:
Trophic Level
Weighted Mean
Consumption rate
(g/person/dav)
0.03146
0.08756
0.22555
0.00322
0.00292
0.05727
0.59114
0.37126
0.06749
0.52735
0.29941
0.03925
0.00068
1.72959
0.03753
0.59114
0.01096
0.60368
0.02337
0.00054
0.44946
Sum of
Consumption
Rates
(g/pers./day)
0.34779
3.72389
1.67915
5.75082
CSFII
Category
Weights
0.09046
0.25176
0.64852
0.00926
0.00078
0.01538
0.15874
0.09970
0.01812
0.14161
0.08040
0.01054
0.00018
0.46446
0.01008
0.35205
0.00653
0.35952
0.01392
0.00032
0.26767
Sum of
Weights
1.00
1.00
1.00
Consumption-
Weighted
Percent Lipid
Values
0.15219
1.13069
1.05162
0.00648
0.00568
0.06842
0.37486
0.13918
0.08648
0.20927
0.07022
0.10897
0.00068
0.34821
0.04535
0.83133
0.01486
1.07956
0.01168
0.00035
1.14837
Consumption-
Weighted
Mean Percent
Lipid
2.34%
1.46%
3.09%
individuals 18 years and older (USEPA, 1998b and Table 2.4.7).
m Trophic level designation of organisms corresponding to CSFII consumption categories, as described in the text and Table 2.4.9.
P) Trophic level weighing factor used to apportion consumption rates to multiple trophic levels for catfish only (see text).
w Mean lipid content for species assigned to each CSFII category as described in the text and Table 2.4.8.
257
-------
Table 2.4.11: National Default Values for POC and DOC in U.S. Water Bodies
WATER
BODY TYPE
Stream/River
Lake
Estuary
All Types
DOC (mg/L)
50th%
(Median)
4.0
2.1
2.7
2.9
Mean
6.2
3.0
3.4
4.9
POC (mg/L)
50th%
(Median)
0.70
0.31
0.90
0.48
Mean
1.3
0.43
1.1
0.83
Source: USEPA STORET data base, data retrieval February, 1996.
Sample sizes for DOC are: 77,637 (stream/river); 40,472 (lake); 14,376 (estuary); 132,485 (all water bodies)
Sample sizes for POC are: 30,236 (stream/river); 39,931 (lake); 10,920 (estuary); 81,087 (all water bodies)
2.4.6 Determining BAFs for Inorganic Substances
Unlike organic chemicals, the lipid-BAF relationship does not generally apply to the
determination of BAFs for inorganic chemicals. Thus, BAFs and BCFs for inorganics should not
be expressed on a lipid-normalized basis, and are not as transferable from one species to another, or
one tissue to another, as with organic chemicals. Bioaccumulation of some trace metals is
substantially greater in internal organs than muscle tissue. For example, BCFs for various tissues
of the rainbow trout after exposure to cadmium for 178 days are as follows (Giles, 1988):
liver 325
kidney 75
gut and skin 7
muscle 1
Merlini and Pozzi (1977) reported that lead bioconcentrated 30 times more in bluegill liver than in
bluegill muscle tissue after eight days. Because of the differential uptake to different tissues and
species, the BAFs should be measured in edible tissues.
BAFs or BCFs measured in plants or invertebrate animals may be available. However, these
factors might be one or more orders of magnitude greater than BAFs or BCFs for the edible tissue
offish as noted in Table 5 in each of the EPA criteria documents for cadmium, copper, lead and
nickel (USEPA, 1985a; USEPA, 1985b; USEPA, 1985c; and USEPA, 1986). For this reason,
invertebrate BAFs and BCFs should only be used in the derivation of human health criteria when
they are considered to be a significant component of the diet of the target consumers.
258
-------
Although bioaccumulationof many inorganic chemicals is similar to their bioconcentration,
mercury and certain other metals are subject to methylation through microbial action in nature, and
may biomagnify through the food chain. For example, research demonstrates that methyl mercury
is bioaccumulativein fish and biomagnifles in aquatic food webs (Grieb et al. 1990; Gardner, 1978).
The following two procedures, in order of priority, should be used to estimate BAFs for
inorganic chemicals. EPA is not aware of any other generic procedures for predicting the BAF for
these substances.
Field-measured B AFs are the most preferred BAFs for inorganic chemicals. Section 2.4.4.1
(Field-MeasuredBAFs) describes data requirements for measuring BAF values using field data for
nonpolar organic chemicals. These requirements are applicable to field-measured BAFs for
inorganic chemicals as well, except that inorganic BAFs should not be lipid-normalized because
bioaccumulation of inorganics is not proportional to lipid content. However, as noted above,
inorganic bioaccumulation can differ dramatically between tissues. Thus, BAFs based on uptake
into edible tissue should be used to calculate human health criteria.
If field-measuredBAFs are not available for inorganic chemicals, a laboratory BCF may be
used to estimate bioaccumulation of inorganic substances from water. The BCF may be used
because for most inorganic substances, bioaccumulation and bioconcentration are similar. Section
2.4.4.3 (Predicted BAF Based on Laboratory-Measured BCF and a Food-Chain Multiplier)
describes acceptable data for measuring BCFs for organic chemicals in the laboratory, which are
applicable to BCFs measured for inorganic chemicals. For inorganic chemicals where dietary
exposure forms a significant portion of the exposure to target organisms, (e.g. mercury, selenium)
BCFs should be used in conjunction with field-derived food chain multipliers.
2.4.7 Example Calculations
The two examples below illustrate how BAFs are developed using two of the four methods
recommended for deriving BAFs for nonpolar organic chemicals for use in establishing AWQC for
human health. The first example illustrates the development of a BAF when field-measured BAFs
are available. The second example illustrates the development of BAFs using a laboratory-measured
BCF and a food-chain multiplier.
2.4.7.1 Example 1: Field-Measured BAF for Chemical M
The calculation of a BAF used in the derivation of a human health criteria is a two-step
process. The first step is to derive baseline BAFs for appropriate trophic levels. The second step is
to derive BAFs that can be used in deriving human health AWQC. Each of these steps are illustrated
below.
259
-------
Baseline BAFfor Chemical M
This example illustrates the development of a baseline B AF appropriate for trophic level four
for a lipophilic chemical M. Data are available from Lake Washington (a hypothetical lake) on the
total concentrationof chemical M in fish tissue in lake trout and the water column. A review of the
dietary preferences of lake trout indicates that this organism is at trophic level four for larger size
ranges commonly consumed by the target population (USEPA, 1995d, 1995e, 1995f). The
development of a baseline BAF for a given trophic level requires information on a field-measured
BAF (Measured BAF^), the fraction of the chemical that is freely dissolved in the ambient water (ffd),
and the fraction lipid content of the species sampled (ft). The equation for calculating a baseline BAF
using a field-measured BAF is:
Baseline BAF
fd _
Measured
- 1
fd
(Equation 2.4.27)
To determine a field-measured BAF, information is needed on the total concentration of chemical
M in fish tissue and in ambient water at the site of sampling. For this example, the mean total
concentration for chemical M in fish tissue is 100 ng/g and the mean total water column
concentration 160 pg/L. Data from the field studies indicates that the mean water column
concentration reflects adequate temporal and spatial averaging based on the K^ of this chemical and
is representative of the average exposure offish to chemical M. The field-measured BAFj for
chemical M is 625,000 L/kg, as demonstrated below.
Field-measured BAF-i- =
Total concentration of chemical M in fish tissue
Total concentration of chemical M in the water column
(Equation 2.4.28)
Field-measured BAF| =
ng/gVl-OQQ pg/ngVl.OOO g/kg> - 625,000 L/kg-tissue
160 pg/L
(Equation 2.4.29)
To determine the fraction of chemical M that is freely dissolved in the ambient water requires
information on the paniculate organic carbon (POC) and dissolved organic carbon (DOC) in the
ambient water where the samples were collected and the Kow of chemical M. For this example, the
median POC concentrationfrom Lake Washington, where the samples were collected, is 0.6 mg/L
260
-------
(6.0 x 10-7kg/L) and the median DOC concentration is 8.0 mg/L (8.0 x lO^kg/L). Importantly, the
POC and DOC concentrations used in calculating the freely dissolved fraction for baseline BAFs is
from the waterbody used in the BAF study. Use of default POC and DOC concentrations in the
derivation of baseline BAFs is not appropriate. The K^ for chemical M is 100,000 or a log K^ of
5.0. The fraction freely dissolved for chemical M is 0.8722, as shown below.
K
[1 + (POC • Kow) + (DOC •
(Equation 2.4.30)
[1 -i- (6.0 x 10"7kg/L • 100,000 L/kg) + (8.0 x 10"6 kg/L • 100'000 L/kg)]
= 0.8772
10
(Equation 2.4.31)
The freely dissolved fraction has been expressed to four significant digits for convenience. The
scientific basis supporting this equation for estimating the freely dissolved fraction is described in
Section 2.4.4.1 and Appendix D.
Finally, the mean fraction lipid content of the fish species sampled in Lake Washington was
8 percent. Using the baseline BAF equation and information on the field-measured BAF, the
fraction freely dissolved, and the fraction lipid content provides a baseline BAF for lake trout of
8,906,166, which is illustrated below.
Baseline BAF
fd _
625,000
0.8772
- 1
0.08
= 8,906,166 L/kg-lipid
(Equation 2.4.32)
For the purposes of this example, it has been assumed that only one acceptable BAF value
is available for trophic level four organisms. Thus, the baseline BAF for trophic level four is equal
to this baseline BAF. Had other acceptable field-measured BAFs been available for trophic-level
261
-------
four organisms, then the baseline BAF for trophic level four would have been calculated as the
geometric mean of the acceptable baseline BAFs at trophic level four.
BAF for Chemical M to Be Used in Deriving A WQC
After the derivation of trophic level-specific baseline BAFs for chemical M (described in the
previous section), the next step is to calculate BAFs that will be used in the derivation of AWQC.
This step is necessary to adjust the baseline BAFs to conditions that are expected to affect the
bioavailability of chemical M at the sites applicable to the AWQC. Derivation of AWQC BAFs
requires information on: (1) the baseline BAF at appropriate trophic levels, (2) the percent lipid of
the aquatic organisms consumed by humans at the site(s) of interest (trophic level specific), and (3)
the freely dissolved fraction of the chemical in ambient water at the site(s) of interest. For each
trophic level, the equation for deriving a BAF to used in deriving AWQC is:
BAF for AWQC(TLn) = [(Baseline BAF/" )TL n • (f,)TL n + 1] • (fffl)
(Equation 2.4.33)
where:
BAF for AWQC (7Ln) =
Baseline BAF[d(TLn) =
n)
fd
BAF at trophic level "n" used to derive AWQC based on site
conditions for lipid content of consumed aquatic organisms
for trophic level "n" and the freely dissolved fraction in the
site water
BAF expressed on a freely dissolved and lipid-normalized
basis for trophic level "n"
Fraction lipid of aquatic species consumed at trophic level "n"
Fraction of the total chemical in water that is freely dissolved
For the purposes of this example, an AWQC BAF is being calculated only for aquatic organisms at
one trophic level (trophic level four). If fish consumption data indicates the target population
consumes significant quantities from multiple trophic levels, then AWQC BAFs should be derived
for each of the appropriate trophic levels.
For chemical M, the baseline BAF at trophic level four is calculated to be 8,906,166 L/kg-
Hpid as described above. The fraction lipid content of aquatic species consumed by the target
population at trophic level four is assumed to be 3.1 % based on the national default lipid content for
trophic level four derived in Section 2.4.5.3. The freely dissolved fraction of chemical M that is
expected in the sites applicable to the AWQC was determined to be 0.9615. This value is calculated
as shown below using hypothetical expected POC and DOC concentrations at the sites applicable
262
-------
to the AWQC of 0.3 mg/L (3.0 x 10'7 kg/L) and 1.0 mg/L (1.0 x 10-6kg/L), respectively, and the
same K^ of 100,000 for chemical M.
K
[1 + (POC • Kow) + (DOC
(Equation 2.4.34)
10
[1 + (3.0 x 10~7 kg/L • 100,000 L/kg) + (1.0 x 10"6 kg/L • 10°'000 L/kg)]
(Equation 2.4.35)
10
ffd = 0.9615
Using the AWQC BAF equation described previously, the AWQC BAF for trophic level four
organisms is calculated to be 265,463 L/kg as shown below.
AWOC BAF for Trophic Level Four
= [(8,906,166 L/kg-lipid>(0.031) +1] • (0.9615)
265,463 L/kg-tissue
This AWQC BAF relates the total concentration in water to the total concentration in tissue of
trophic level four organisms, based on the expected conditions that would affect the bioavailability
of chemical M (i.e., freely dissolved fraction at AWQC sites and lipid content of consumed aquatic
organisms).
2.4.7.2 Example 2: Laboratory-Measured BCF for Chemical R
When a field-measured BAFj or field-measured BSAF are not available, a laboratory-
measured B AFj along with a food-chain multiplier should be used to derive a baseline BAF and then
an AWQC BAF for use in deriving human health criteria.
Baseline BAF for Chemical R
The development of a baseline BAFf for chemical R specific to a given trophic level requires
information on a laboratory measured BCF (measured BCFj), the fraction of the chemical that is
freely dissolved in the test water (ffd), the fraction lipid content of the species sampled (Q, and the
food-chain multiplier for the chemical (FCM). For a given trophic level, the equation for calculating
a baseline BAFf using a laboratory BCFj and food-chain multiplier is:
263
-------
fd _
Baseline BAF™ = (FCM)
Measured BCFT*
fd
- 1
(Equation 2.4.36)
The basis of this equation is described in Section 2.4.4.3.
The laboratory-measured BCF requires information on the total concentration of chemical
R in fish tissue and the total concentration of chemical R in the test water. For this example, the
mean total fish tissue concentration for chemical R is 10 ng/g and the mean total test water
concentration is 3 ng/L. The laboratory-measured BCF is 3333 L/kg.
Laboratory measured BCF-r =
Total concentration of chemical R in fish tissue
Total concentration of chemical R in test water
(Equation 2.4.37)
Laboratory measured BCF-f =
ng/g)(1.000 g/Kg1) = 3333 L/kg-tissue
3 ng/L
(Equation 2.4.38)
To determine the fraction of chemical R that is freely dissolved in the test water requires
information on the particulate organic carbon (POC) and dissolved organic carbon (DOC) in the test
water and the KOW of chemical R. For this example, the median POC concentration in the test water
is 0.6 mg/L (6.0 x 10'7 kg/L) and the median DOC concentration is 8.0 mg/L (8.0 x 10'6 kg/L). The
Kow for chemical R is 10,000 or a log Kow of 4.0. The fraction freely dissolved for chemical R is
0.9860, as shown below.
K
[1 + (POC • K ) + (DOC • ——)]
L ^ ow' v 10
(Equation 2.4.39)
264
-------
[1 + (6.0 x 10"7 kg/L • 10,000 L/kg) + (8.0 x 10"6 kg/L 10>000
= 0.9862
10
L/kg)]
(Equation 2.4.40)
The freely dissolved fraction has been expressed to four significant digits for convenience. The
scientific basis supporting this equation is explained in Section 2.4.4.1 and Appendix D.
The fraction lipid content of the fish species sampled in the laboratory is 8 percent. The
food-chain multiplier based on a log K,,w of 4 is 1.072, as indicated in Table 2.4.4 (assuming mixed
benthic and pelagic food web structure and trophic level four for the tested species). Using the
baseline BAF[d equation and the information on the laboratory-measured BCFf, the fraction freely
dissolved, the fraction lipid content, and the FCM provides a baseline BAF[d of 45,274 L/kg - lipid,
which is used in the derivation of the BAF as described in the next section.
Baseline BAF{fd = (1.072)
3333
0.9862
0.08
= 45,274 L/kg-lipid
(Equation 2.4.41)
For the purposes of this example, it has been assumed that only one acceptable baseline BAF
value could be derived for trophic level four organisms. Thus, the baseline BAF for trophic level
four is equal to this baseline BAF. Had other acceptable BCFs been available for trophic-level four
organisms, then the trophic level four baseline BAF would have been calculated as the geometric
mean of the acceptable BCF-predicted baseline BAFs for trophic level four.
A WQCBAFfor Chemical R
After the derivation of trophic level-specific baseline BAFs for chemical R (described in the
previous section), the next step is to calculate BAFs that will be used in the derivation of AWQC.
This step is necessary to adjust the baseline BAFs to conditions that are expected to affect the
bioavailability of chemical R at the sites applicable to the AWQC. Derivation of AWQC BAFs
requires information on: (1) the baseline BAF at appropriate trophic levels, (2) the percent lipid of
the aquatic organisms consumed by humans at the site(s) of interest (trophic level specific), and (3)
the freely dissolved fraction of the chemical in ambient water at the site(s) of interest. For each
trophic level, the equation for deriving a BAF to used in deriving AWQC is:
BAF for AWQC(TL n) = [(Baseline BAF{fd )TL n
(Equation 2.4.42)
1]
265
-------
where:
BAFforAWQC(TLn) =
Baseline BAFf(TLn) =
fd
BAF at trophic level "n" used to derive AWQC based on site
conditions for lipid content of consumed aquatic organisms
for trophic level "n" and the freely dissolved fraction in the
site water
BAF expressed on a freely dissolved and lipid-normalized
basis for trophic level "n"
Fraction lipid of aquatic species consumed at trophic level "n"
Fraction of the total chemical in water that is freely dissolved
For the purposes of this example, an AWQC BAF for chemical R is being calculated only for aquatic
organisms at one trophic level (trophic level four). If fish consumption data indicates the target
population consumes significant quantities from multiple trophic levels, then AWQC BAFs should
be derived for each of the appropriate trophic levels.
For chemical R, the baseline BAF at trophic level four is calculated to be 45,274 L/kg-lipid
as described above. The fraction lipid content of aquatic species consumed at trophic level four is
assumed to be 3.1% based on the national default lipid content for trophic level four derived in
Section 2.4.5.3. The freely dissolved fraction of chemical R that is expected in the sites applicable
to the AWQC was determined to be 0.9924. This value is calculated as shown below using
hypothetical expected POC and DOC concentrations at the sites applicable to the AWQC of 0.48
mg/L (4.8 x lO'7 kg/L) and 2.9 mg/L (2.9 x 10"6 kg/L), respectively, and the same Kow of 10,000 for
chemical R.
f,, =
fd
K
[1 + (POC • KQW) + (DOC
(Equation 2.4.43)
1
10
[1 -i- (4.8 x 10"7 kg/L • 10,000 L/kg) + (2.9 x 10"6 kg/L •
(Equation 2.4.44)
10,000
10
L/kg)
f,. = 0.9924
to
266
-------
Using the AWQC BAF equation described previously, the AWQC BAF for trophic level four
organisms is calculated to be 1,394 L/kg as shown below.
AWOC BAF for Trophic Level Four
= [45,274 L/kg-lipid>(0.03 !)+!]• (0.9924)
1,394 L/kg-tissue
This AWQC BAF relates the total concentration in water to the total concentration in tissue of
trophic level four organisms, based on the expected conditions that would affect the bioavailability
of chemical R (i.e., freely dissolved fraction at AWQC sites and lipid content of consumed aquatic
organisms).
2.4.8 Trophic Level-Specific Fish Consumption Rates
When local or regional data are unavailable for calculating fish and shellfish consumption
rates, EPA has derived national default consumption rates of 17.8 g/person/d, 39 g/person/d, and
86.3 g/person/d based on the 90th, 95th and 99th percentile of average per capita fish consumption
from the adult U.S. population (see Section 2.3 on exposure). These default consumption figures
reflect total consumption of aquatic organisms across all trophic levels. However, as described in
above, EPA recommends that BAFs be determined separately for specific trophic levels because
accumulation of chemicals is often related to the trophic position of the aquatic organism,
particularly for highly persistent, lipophilic organic chemicals. The question then becomes how to
best relate available fish consumption rates to the trophic level-specific BAFs in the calculation of
AWQC.
When calculating AWQC, EPA recommends that if possible, fish consumption rates be
determined for individual trophic levels for which BAFs have been derived. For example, if
available fish and shellfish consumption survey data indicate that the target population is consuming
significant portions of aquatic organisms at trophic levels two, three, and four, then both BAFs and
fish consumption rates should be determined for each of these trophic levels to provide the most
accurate representation of contaminant exposure via the consumption of aquatic organisms. In this
example, applying the total consumption rate from all three trophic levels to a BAF that is derived
for a single trophic level may not accurately reflect likely exposure to the target population, if BAFs
differ greatly by trophic level.
Calculating fish consumption rates for individual trophic levels requires information on the
trophic status of consumed species from the consumption survey. Determination of trophic status
of aquatic organisms is best determined on a local or regional basis and should involve consideration
of the size (age) of aquatic organisms in addition to their dietary preferences. If local or regional
information is not available, then EPA recommends the use of the most recent version of the
document: Trophic Level and Exposure Analysis for Selected Piscivorous Birds and Mammals
(USEPA 1995d, 1995e, 1995f), which contains information on the dietary composition of numerous
aquatic species. This draft document is currently being revised based on peer review comments and
267
-------
is expected to be made final in 1998. Described below is the derivation of trophic level-specific fish
consumption rates for EPA's CSFII-based, national default fish consumption rates listed above.
In estimating trophic level-specific consumption rates appropriate to the national default
consumption rates from the CSFII consumption survey (USEPA, 1998b), EPA first estimated the
trophic level of aquatic organisms corresponding to each of the CSFII consumption categories for
mean per capita consumption of adults (see Table 2.4.9 and associated text for trophic level
determinations of the various CSFII consumption categories; see Table 2.4.7 for mean per capita
consumption rates from the CSFII survey). Since the national default consumption rates of 17.8,
39.0, and 86.3 g/person/day reflect consumption at the 90th, 95th and 99th percentiles, respectively,
trophic level assignments would have ideally been made according to consumption patterns
corresponding to these higher percentiles because consumption patterns might differ at higher
percentiles. However, inherent limitations in the data from the CSFII survey prevented a meaningful
assessment of consumption patterns at these upper percentiles. Therefore, the consumption pattern
reflective of mean per capita consumption rates was assumed to adequately reflect consumption at
the higher consumption rate percentiles.
The second step involved calculating the fraction of total fish and shellfish consumption at
trophic level two, three and four using the mean per capita consumption rates of adults from CSFII
survey (Table 2.4.7). The fractions of total fish consumption at specific trophic levels is shown
below:
Fl
(TLn)
•(FI, TL n) pj
(.11 TL)
where:
Ma TLn)
FI(TLn)
FI
(TLall)
Fraction of total mean per capita fish consumption at trophic level "n"
Mean per capita fish consumption rate at trophic level "n"
(g/person/day)
Total mean per capita fish consumption rate for all trophic levels
(g/person/day)
Fraction total fish consumption at:
Trophic Level Two: 0.06048
Trophic Level Three: 0.64754
Trophic Level Four: 0.29198
= 0.34779 (g/pers./day) / 5.75082 (g/pers./day)
= 3.72389 (g/pers./day) / 5.75082 (g/pers./day)
= 1.67915 (g/pers./day) / 5.75082 (g/pers./day)
268
-------
These fractions indicate that on a national, per capita average basis, the majority of fish and
shellfish consumption is occurring at trophic level three, folio wed by trophic level trophic level four.
This is corroborated by the comparatively greater consumption of shrimp, catfish, perch, estuarine
flatfish, trout, crab, and flounder (Table 2.4.7). Finally, the trophic level-specific consumption rates
applicable to 90th, 95th, and 99th percentile national default consumption values were calculated using
the following equation:
j i"1 percentile _
(TL n)
» i " percentile
(all TL)
(FI, TL n)
where:
ith percentile
ith percentile
(FI, TL n)
(TLn)
(TL all)
Estimated fish consumption rate at trophic level "n" for the i*
percentile
Total fish consumption rate for all trophic levels for the ith
percentile
Fraction of total mean per capita fish consumption at trophic
level "n" (determined above)
Trophic level-specific consumption rates ( in g/person/day) corresponding to the national
default consumptionrates of 17.8g/d, 39.0g/d, and 86.3g/d (i.e., the 90*, 95th and 99th percentile of
mean per capita consumption of adults from the CSFII survey) are shown below.
TL2
TL3
TL4
90th
1.1
11.5
5.2
All TL 17.8
95th
2.4
25.2
11.4
39.0
99th
5.2
55.9
25.2
86.3
EPA recognizes that in some situations, States and Tribes may lack sufficient data to
determine trophic level-specific fish consumption rates that are applicable to their target populations
and site(s) of concern. In these situations, EPA recommends that States and Tribes assign the total
fish consumption rates to the highest BAF determined across the relevant trophic levels. In most
cases, this will be the BAF corresponding to trophic level four, but in some cases, it may be trophic
level three. This approach reflects a assumption that may be conservative, the degree to which will
depend on the actual consumption pattern of the target population.
269
-------
r
2.4.9 References
American Fisheries Society. 1991. Common and Scientific Names of Fishes from the United States
and Canada. Fifth Edition. Bethesda, MD: American Fisheries Society. AFS Spec. Publ. No.
20.
Armstrong, R.W. and RJ. Sloan. 1988. PCB Patterns in Hudson River Fish: I. Resident Freshwater
Species. In: C.L Smith (ed). Fisheries Research in the Hudson River: The Hudson River
Environmental Society. Albany, NY: State University of New York Press, pp. 304-324.
Baumann, T. and L.H. Grimme. 1981. Determination of Hydrophobic Parameters for Pyridazinone
Herbicides by Liquid-Liquid Partition and Reversed-Phase High-Performance Liquid
Chromatography. J. Chromatog. 206: 7-15.
Bligh, E.G. and WJ. Dyner. 1959. A Rapid Method of Total Lipid Extraction and Purification.
Canadian J. Biochem. and Physiol. 37:911-917.
Brooke, D.N., A.J. Dobbs and N. Williams. 1986. Octanol:Water Partition Coefficients (P):
Measurement, Estimation, and Interpretation, Particularly for Chemicals with P > 10s.
Ecotoxicol. Environ. Safety. 11:251-260.
Brooke, D., I. Nielsen, J. de Bruijn and J. Hermens. 1990. An Interlaboratory Evaluation of the
Stir-Flask Method for the Determination of Octanol-Water Partition Coefficient (log Pow)-
Chemosphere. 21:119-133.
Burkhard, L. 1994. Food Chain Multipliers for the GLWQI. ERL-Duluth.
Burkhard, L.P. and D.W. Kuehl. 1986. N-Octanol Water Partition Coefficients by Reverse Phase
Liquid Chromatography/Mass Spectrometry for Eight Tetrachlorinated Planar Molecules.
Chemosphere. 15:163-167.
Burkhard, L.P., B.R. Sheedy, D.J. McCauley, and G.M. DeGraeve. 1997. Bioaccumulation Factors
for Chlorinated Benzene, Chlorinated Butadienes and Hexachloroethane. Environ. Toxicol.
Chem. 16(8): 1677-1686.
Cabrini, L., L. Landi, C. Stefanelli, V. Barzanti, and A.M. Sechi. 1992. Extraction of Lipids and
Lipophilic Antioxidantsfrom Fish Tissue: A Comparison Among Different Methods. Comp.
Biochem. Physiol. 101B(3):383-386.
Chessells, M., D.W. Hawker and D.W. Connell. 1991. Critical Evaluation of the Measurement of
the 1-Octanol/WaterPartition Coefficient of Hydrophobic Compounds. Chemosphere. 22:
1175-1190.
270
-------
Chiou, G.T. 1985. Partition Coefficients of Organic Compounds in Lipid-Water Systems and
Correlation with Fish Bioconcentration Factors. Environ. Sci. Technol. 19: 57-62.
Connolly, J. and C. Pedersen. 1988. A Thermodynamic-based Evaluation of Organic Chemical
Accumulation in Aquatic Organisms. Environ. Sci. Technol. 22: 99-103.
Cook, P.M. 1994. Prediction of Bioaccumulation Factors (BAFs) from Biota-Sediment
Accumulation Factor (BSAF) Measurements. ERL-Duluth.
Cook, P.M., R.J. Erickson. R.L. Spehar, S.P. Bradbury and G.T. Ankley. 1993. Interim Report on
Data and Methods for Assessment of 2,3,7,8-tetrachlorodibenzo-p-dioxin Risks to Aquatic
Life and Associated Wildlife. Duluth, MM: USEPA, Environmental Research Laboratory.
EPA/600/R-93/055.
de Boer, J. 1988. Chlorobiphenyls in Bound and Non-bound Lipids of Fishes: Comparison of
Different Extraction Methods. Chemosphere 17:1803-1810.
de Bruijn, J., F. Busser, W. Seinen and J. Hermans. 1989. Determination of Octanol/water Partition
Coefficients of Hydrophobic Organic Chemicals with the "Slow-Stirring" Method. Environ.
Toxicol. Chem. 8:499-512.
De Wolf, W., J.H.M. de Bruijn, W. Seinen, and J. L.M. Hermens. 1992. Influence of
Biotransformation on the RelationshipBetween BioconcentrationFactors and Octanol-Water
Partition Coefficients. Environ. Sci. Technol. 26: 1197-1201.
DeVoe, H., M.M. Miller and S.P. Wasik. 1981. Generator Columns and High Pressure Liquid
Chromatography for Determining Aqueous Solubites and Octanol-Water Partition
Coefficients of Hydrophobic Substances. J. Res. Natl. Bur. Stand. 86: 361-366.
Eadie, B.J., N.R. Morehead and P.P. Landrum. 1990. Three-Phase Partitioning of Hydrophobic
Organic Compounds in Great Lake Waters. Chemosphere. 20:161-178.
Eadie, B. J., N.R. Morehead, J. Val Klump and P.P. Landrum. 1992. Distribution of Hydrophobic
Organic Compounds between Dissolved and Particulate Organic Matter in Green Bay
Waters. J. Great Lakes Res. 18:91-97.
Gardner, D. 1978. Mercury in Fish and Waters of the Irish Sea and Other United Kingdom Fishing
Grounds. Nature. 272:49-51.
Garst, I.E. and W.C. Wilson. 1984. Accurate, Wide-Range, Automated, High-Performance Liquid
Chromatographic Method for the Estimation of Octanol/Water Partition Coefficients. I:
Effect of Chromatographic Conditions and Procedure Variables on Accuracy and
Reproducibility of the Method. J. Pharm. Sci. 73:1616-1623.
271
-------
Giles, M.A. 1988. Accumulation of Cadmium by Rainbow Trout, Salmo gairdneri, During
Extended Exposure. Can. J. Aquat. Sci. 45: 1045-1053.
Gobas, F.A.P.C. 1993. A Model for Predicting the Bioaccumulation of Hydrophobic Organic
' Chemicals in Aquatic Food-Webs: Application to Lake Ontario. Ecological Modeling. 69:
1-17.
Gobas, F.A.P.C., K.E. Clark, W.Y. Shiu, and D. Mackay. 1989. Bioconcentration of
PolybrominatedBenzenes and Biphenyls and Related Superhydrophobic Chemicals in Fish:
Role of Bioavailability and Elimination into Feces. Chemosphere 8: 231-245.
Grieb, T.M., C.T. Driscoll, S.P. Gloss, C.L. Schofield, G.L. Bowie and D.B. Porcella. 1990.
Factors Affecting Mercury Accumulationin Fish in the Upper Michigan Peninsula. Environ.
Toxicol. Chem. 9: 919-930.
Gschwend. P.M. and S. Wu. 1985. On the Constancy of Sediment-Water Partition Coefficients of
Hydrophobic Organic Pollutants. Environ. Sci. Technol. 19: 90-96.
Hansch, C. and A. J. Leo. 1979. Substituent Constituents for Correlation Analysis in Chemistry and
Biology. New York: John Wiley and Sons.
Hara, A. and N.S. Radin. 1978. Lipid Extraction of Tissues with a Low-toxicity Solvent. Anal.
Biochem. 90:420-426.
Harnish, M., HJ. Mockel and G. Schulze. 1983. Relationship Between log Pow Shake-Flask Values
and Capacity Factors Derived from Reverse-Phase High-Performance Liquid
Chromatography for n-Alkylbenzenes and Some OECD Reference Substances. J.
Chromatog. 282: 315-332.
Hawker, D.W. and D.W. Connell. 1988. Octanol-Water Partition Coefficients of Polychlorinated
Biphenyl Congeners. Environ. Sci. Technol. 22: 382-387.
Hemond, H.F. and EJ. Fechner. 1994. Chemical Fate and Transport in the Environment. San
Diego: Academic Press.
Herbert, B.E., P.M. Bertsch and J.M. Novak. 1993. Pyrene Sorption by Water-Soluble Organic
Carbon. Environ. Sci. Technol. 27:398-403.
Honeycutt, M.E., V. A. McFarland, and D.D. McCant. 1995. Comparison of Three Lipid Extraction
Methods for Fish. Bull. Env. Contain, and Toxicol. 55:469-472.
Isnard, P., and S. Lambert. 1988. Estimating Bioconcentration Factors from Octanol-Water
Partition Coefficients and Aqueous Solubility. Chemosphere 17: 21-34.
272
-------
Jordan, D.S. and B.W. Evermann. 1969. American Food and Game Fishes. New York: Dover
Publications, Inc. 574pp.
Karickhoff, S.W., D.S. Brown and T.A. Scott. 1979. Sorption of Hydrophobic Pollutants on
Natural Sediments. Water Research 13: 241-248.
Konemann, H., R. Zelle, F. Busser and W.E. Hammers. 1979. Determination of log P^, Values
Chloro-Substituted Benzenes, Toluenes and Anilines by High Performance Liquid
Chromatography on ODS-silica. J. Chromatogr. 178:559-565.
Krzynowek, J., and J. Murphy. 1987. Proximate Composition, Energy, Fatty Acid, Sodium, and
Cholesterol Content of Finfish, Shellfish, and their Products. Department of Commerce,
National Oceanic and Atmospheric Administration, National Marine Fisheries Service.
NOAA Technical Report NMFS 55. July.
Landrum, P.P., S.R. Nihart, B.J. Eadie and W.S. Gardner. 1984. Reverse-Phase Separation Method
for Determining Pollutant Binding to Aldrich Humic and Dissolved Organic Carbon of
Natural Waters. Environ. Sci. Technol. 18:187-192.
Lapin, V.I. and Y.G. Chernova. 1969. Procedure for Lipid Extraction from Crude Fish Tissues.
J. Ichthyol. 10(4):563-566.
Leo, A.J. 1988. Unified Medchem Software, Version 3.53, Pomona Medicinal Chemistry Project.
Mackay,D. 1982. Correlation of Bioconcentration Factors. Environ. Sci. Technol. 16:274-278.
McKim, J., P. Schmeider and G. Veith. 1985. Absorption Dynamics of Organic Chemical
Transport Across Trout Gills as Related to Octanol-Water Partition Coefficient. Toxicol.
Applied Pharmacol. 77:1-10.
Mercer, L. 1989. Fishery Management Plan for Atlantic Croaker. Atlantic States Marine Fisheries
Commission. Fisheries Management Report No. 10.
Merlini, M. and G. Pozzi. 1977. Lead and Freshwater Fishes: Part I-Lead Accumulation and Water
pH. Environ. Pollut. 12:167-172.
Miller, M.M., S. Ghodbane, S.P. Wasik, Y.D. Terwari and D.E. Marthre. 1984. Aqueous
Solubilities, Octanol/Water Partition Coefficients and Entropies of Melting of Chlorinated
Benzenes and Biphenyls. J. Chem. Eng. Data 29: 184-190.
McDuffie, B. 1981. Estimation of Octanol/Water Partition Coefficients for Organic Pollutants
Using Reversed-Phase HPLC. Chemosphere 10: 73-83
273
-------
Oliver, E.G., and A.J. Niimi. 1988. Trophodynamic Analysis of Polychlorinated Biphenyl
Congeners and Other Chlorinated Hydrocarbons in the Lake Ontario Ecosystem. Environ.
Sci. Technol. 22: 388-397.
Pereira, W.E., C.E. Ostad, C.T. Chiou, T.I. Brinton, L.B. Barber II, D.K. Demcheck and C.R.
Demas. 1988. Contamination of Estuarine Water, Biota, and Sediment by Halogenated
Organic Compounds: A Field Study. Environ. Sci. Technol. 22: 772-778.
Randall, R.C., H. Lee H, R.J. Ozretich, J.L. Lake and R.J. Pruell. 1991. Evaluation of Selected
Lipid Methods for Normalizing Pollutant Bioaccumulation. Environ. Toxicol. Chem. 10:
1431-1436.
Sidwell, V.D. 1981. Chemical and Nutritional Composition of Finfishes, Whales, Crustaceans,
Mollusks, and Their Products. NOAA Technical Memorandum NMFS F/SEC-11.
Department of Commerce, National Oceanic and Atmospheric Administration, National
Marine Fisheries Service. January.
Stephan, C.E. 1993. Derivation of Proposed Human Health and Wildlife Bioaccumulation Factors
for the Great Lakes Initiative. ERL-Duluth.
Stephen, C.R., D.I. Mount, D.J. Hansen, J.H. Gentile, G.A. Chapman and W.A. Brungs. 1985.
Guidelines for Deriving Numerical National Water Quality Criteria for the Protection of
Aquatic Organisms and Their Uses. USEPA, Office of Research and Development,
Environmental Research Labs, Duluth, MN; Narragansett, RI; Corvallis, OR.
Thomann,R.V. 1989. BioaccumulationModel of Organic Chemical Distribution in Aquatic Food
Chains. Environ. Sci. Technol. 23: 699-707.
USEPA. Large Lakes Research Station, Grosse He, Michigan. (Technical Report prepared by J.P.
Connolly et al., Manhattan College under cooperative agreement No. CR-815396).
USEPA. 1985a. Ambient Water Quality Criteria for Cadmium-1984. Washington, DC: Office of
Water. EPA 440/5-84-032.
USEPA. 1985b. Ambient Water Quality Criteria for Copper-1984. Washington, DC: Office of
Water. EPA 440/5-84-031.
USEPA. 1985c. Ambient Water Quality Criteria for Lead-1984. Washington, DC: Office of
Water. EPA 440/5-84-027.
USEPA. 1986. Ambient Water Quality Criteria for Lead-1986. Washington, DC: Office of Water.
EPA 440/5-86-004.
274
-------
USEPA. 1990. Lake Ontario TCDD Bioaccumulation Study-Final Report. New York: USEPA,
Region II.
USEPA. 1991. Assessment and Control of Bioconcentratable Contaminants in Surface Waters.
Draft. Washington, DC: Office of Water.
USEPA. 1992a. National Study of Chemical Residues in Fish. Washington, DC: Office of Water,
Office of Science and Technology, Standards and Applied Science Division. EPA-823-R-
92-008a and EPA-823-R-92-008-b.
USEPA. 1992b. Development and Application of a Model of PCBs in the Green Bay, Lake
Michigan Walleye and Brown Trout and Their Food Webs. Riverdale, NY: Manhattan
College.
USEPA. 1993. Fish Sampling and Analysis: A Guidance Document for Issuing Fish Consumption
Advisories. Washington, DC: Office of Water, Office of Science and Technology.
USEPA. 1995a. Guidance for Assessing Chemical Contaminant Data for Use in Fish Advisories.
Volume 1: Fish Sampling and Analysis Second Edition. USEPA, Office of Water. EPA
823-R-95-007. September.
USEPA. 1995b. Final Water Quality Guidance for the Great Lakes System. Federal Register
60(56): 15366+. March 23.
USEPA. 1995c. Great Lakes Water Quality Initiative Technical Support Document for the
Procedure to Determine Bioaccumulation Factors. EPA-820-B-95-005. March.
USEPA. 1995d. Trophic Level and Exposure Analyses for Selected Piscivorous Birds and
Mammals. Volume I: Analyses of Species for the Great Lakes. Draft. Washington, DC:
Office of Water (Available in the GLWQI docket).
USEPA. 1995e. Trophic Level and Exposure Analyses for Selected Piscivorous Birds and
Mammals. Volume II: Analyses of Species in the Conterminous United States. Draft.
Washington, DC: Office of Water (Available in the GLWQI docket).
USEPA. 1995f. Trophic Level and Exposure Analyses for Selected Piscivorous Birds and
Mammals. Volume III: Appendices. Draft. Washington, DC: Office of Water (Available in
the GLWQI docket).
USEPA. 1995g. Addendum to Green Bay Final Report Food Chain Model Projections. U.S. EPA
Large Lakes Research Station, Grosse He, Michigan. (Technical Report prepared by
HydroQual Inc., Mahwah, NJ., under Contract No. 68-C3-0332).
275
-------
USEPA. 1995h. Phase 2 Report. Further Site Characterization and Analysis Database Report:
Hudson River PCBs ReassessmentRI/FS. Volume 2A. U.S.EPA Region II, New York, NY.
(Technical Report prepared by TAMS Consultants, Inc and Gradient Corporation, under
Contract No. 68-S9-2001).
USEPA. 1995i. Bioaccumulation of Superlipophilic Organic Chemicals: Data Compilation and
Analysis. Washington, DC: Office of Water, Office of Science and Technology. (Prepared
by HydroQual, Inc. under Contract # 68-C3-0332).
USEPA. 1998a. Notice of Draft Revisions to the Methodology for Deriving Ambient Water
Quality Criteria for the Protection of Human Health. Federal Register Notice.
USEPA. 1998b. Daily Average Per Capita Fish Consumption Estimates Based on the Combined
USDA 1989, 1990, 1991 Continuing Survey of Food Intakes by Individuals (CSFII).
Volume I: Uncooked Fish ConsumptionNational Estimates; Volume II: As Consumed Fish
Consumption National Estimates. Prepared by SAIC under Contract #68-C4-0046. March.
Veith, G.D., N.M. Austin, and R.T. Morris. 1979. A Rapid Method for Estimating log P for
Organic Chemicals. Water Res. 13:43-47.
Woodburn, K.B., WJ. Doucette and A.W. Andren. 1984. Generator Column Determination of
Octanol /Water Partition Coefficients for Selected Polychlorinated Biphenyl Congeners.
Environ. Sci. Technol. 18:457-459.
Yin, C. and J.P. Hassett. 1986. Gas-Partitioning Approach for Laboratory and Field Studies of
Mirex Fugacity in Water. Environ. Sci. Technol. 20:1213-1217.
Yin, C. and J.P. Hassett. 1989. Fugacity and Phase Distribution of Mirex in Oswego River and
Lake Ontario Waters. Chemosphere. 19: 1289-1296.
3. MINIMUM DATA CONSIDERATIONS
3.1 Background
The 1980 AWQC National Guidelines did not present specific minimum data requirements.
However, the following minimum data requirements were implied from the text:
3.1.1 Threshold Effects Guidelines
Animal dose-response toxicity data were used in developing guidelines for deriving criteria
based on noncarcinogenic responses. The following guidelines for deriving criteria were adopted:
276
-------
• A free-standing Frank Effect Level (PEL) is unsuitable for the derivation of criteria.
• A free-standingNo Observed Effect Level (NOEL) is unsuitable for the derivation of criteria.
If multiple NOELs are available, with or without additional data on Lowest Observed Effect
Levels (LOELs), No Observed Adverse Effect Levels data (NOAELs), or Lowest Observed
Adverse Effect Levels (LOAELs). The highest NOEL should be used to derive a criterion.
• A NOAEL, LOEL, or LOAEL can be suitable for criteria derivation. A well-defined
NOAEL from a chronic study (90-day study was considered minimum) may be used directly,
applying the appropriate uncertainty factor. For a LOEL, a judgment needs to be made as
to whether it actually corresponds to a NOAEL or a LOAEL. In the case of a LOAEL, an
additional uncertainty factor is applied; the magnitude of the additional uncertainty factor is
judgmental and should lie in the range of 1 to 10. Caution must be exercised not to substitute
FELs for LOAELs.
3.1.2 Non-Threshold Effects
This section discusses lifetime animal studies or human studies where excess cancer risk has
been associated with exposure to the agent.
3.1.2.1 Animal Studies
For some chemicals, several studies conducted at several doses and different routes
of exposure are available for different animal species, strains, and sexes. A choice
must be made as to which of the data sets from several studies are to be used in the
model. The procedures listed below, used in evaluating these data, are consistent
with the estimate of a maximum-likely-risk.
The data (i.e., dose and tumor incidence) used in the model are data sets where the
incidence is statistically significantly higher than the control for at least one test dose
level and/or where the tumor incidence rate shows a statistically significantly trend
with respect to dose level. The data set which gives the highest estimate of lifetime
carcinogenic risk, ql*, estimated from each of these data sets, is used for risk
assessment.
• If sufficient data exist for two or more significant tumor sites in the same study, the
number of animals with at least one of the specific tumor sites under consideration
was used as incidence data in the model.
• Since to a close approximation, the surface area is proportional to the 2/3 power of
the weight as would be the case for a perfect sphere, the exposure in mg/(body
weights/day is similarly considered to be an equivalent exposure.
277
-------
• Use data from organ sites which are statistically higher than the control data.
3.1.3 Exposure Assumptions
The three exposure-related parameters listed below were provided in the 1980 AWQC
National Guidelines. Although the concept of accounting for dietary and inhalation exposures was
included, no parameters were provided.
• 2 L/day drinking water consumption;
• 6.5 g/day consumption of fish; and
• Lipid-normalized bioconcentration factor (BCF).
3.2 Minimum Data Considerations in the Federal Register Notice
Many sections of the Federal Register notice which accompanies this Technical Support
Document (TSD) include discussions of data quality. While many of these discussions are
qualitative in nature, they may help direct the reader to the kinds of data which meet a minimally
successful risk assessment. For example, in the exposure section, there is a discussion of what
constitutes acceptable data for conducting a relative source contribution assessment; in addition,
there is a discussion regarding minimally acceptable fish consumption surveys and data collection.
In developing bioaccumulation factors, there is a discussion in the Federal Register of what is
regarded as a minimally acceptable BCF and Kow. That section also cites a field guidance document
which will contain minimum data requirements for assessing field-measured BAFs and field-
measured lipid levels and POC/DOC. Once this document is finalized in 1998, the results will be
cited and incorporated into the final TSD.
On the lexicological side of the human health methodology, the following minimum data is
suggested for RfD development:
3.2.1 Noncancer - Data Suggestions
3.2.1.1 RfD Development (Minimal Data)
• One well-conductedsubchronic (90 days) mammalian bioassay by the oral route of
exposure hi which a NOAEL or LOAEL can be derived.
• If the most critical endpoint is an acute effect, which occurs short-term, it should be
used as the basis of the RfD.
• One short-term developmental study, if it can be shown that the developmental
toxicity endpoint is the critical effect given other subchronic or chronic studies.
278
-------
• One developmental study cannot be used as the basis of an RfD on its own unless
other studies exist to support its use.
Of course, a more ideal data set is preferred but not always available. The following data set
is considered complete and likely to have much less uncertainty associated with the resulting RfD:
3.2.1.2 RfD Development (Ideal Situation)
• One well-conducted epidemiological study; or
• Two or more adequate chronic studies in two animal species, one of which must be
with rodents, by the oral route of exposure in which one can identify a NOAEL and
LOAEL;and
• One adequate mammalian multi-generation reproductive toxicity study by the oral
route of exposure; and
• Two adequate mammalian developmental toxicity studies by the oral route of
exposure in different species; and
• Mechanistic, pharmacokinetic and target organ toxicity data; and
• If the most critical endpoint is an acute effect, which occurs short-term, it should be
used as the basis of the RfD.
• The species most biologically relevant to humans is known; in the absence of the
most biologically relevant species, the most sensitive species is chosen as the basis
for RfD development. For example, study results from an animal whose
pharmacokinetics and toxicokinetics match those of a human would be considered
the most biologically relevant.
Minimum data suggestions for benchmark and categorical regression analyses are currently
evolving. However, the examples and text provided in this TSD under the noncancer section do
provide some information of the data needs of each of these analyses.
3.2.2 Cancer - Data Suggestions
3.2.2.1 Minimum Data
A minimally acceptable data base for cancer assessment is one similar to the weight of
evidence established by the 1986 Guidelines for Carcinogen Risk Assessment (e.g., A, B, and C
classifications) fully described at 51 FR 33992 and in the Federal Register notice which
279
-------
accompanies this TSD. However, such a data base may be lacking information on mode of action,
which is important for making judgments using the new Cancer Guidelines of 1996 (61 FR 17960).
At a minimum, some information is needed to determine the mode of action; otherwise, the chemical
must be treated as a linear compound.
3.2.2.2 Ideal Situation
The goal is to establish a complete data base which includes not only adequate tumor data
from chronic studies, as described above, but data on mode of action, metabolism, pharmacokinetics,
and target toxicity. The ultimate goal in any cancer assessment is to establish the mechanism by
which the cancer develops. Since the number of studies for a weight-of-evidence is yet to be
established, there is no quantitative guidance being presented today. These minimum data
requirements may be established in time to incorporate them into the final TSD. However, as with
all determinations based on a weight-of-evidence, the number of studies which demonstrate (1) a
carcinogenic effect in a number of animal species and sexes, and (2) a particular mode of action,
determines the confidence in the overall weight of evidence.
3.2.3 Exposure - Data Suggestions
Numerous suggestions are made at the beginning of the exposure analyses section of this
TSD regarding the factors used in the AWQC derivation. These factors include (1) body weight of
the individuals exposed; (2) drinking water ingestion rates; (3) fish consumption rates; (4) incidental
ingestion of water; and (5) the relative source contribution factor to account for other exposures.
Body weights and fish intake assumptions are used in each criterion. The suggestions made are to
help the exposure assessor locate sources of information for conducting exposure analyses and do
not prescribe minimum data considerations. However, the specific approach to estimating non-water
sources of exposure when setting AWQC (i.e., the Exposure Decision Tree Approach) discusses data
adequacy considerations. This discussion is not repeated here; the reader is referred to the data
adequacy subsection in Section 23 A.I of this TSD.
3.3 Site-Specific Criterion Calculation
The 1980 AWQC National Guidelines allowed for site-specific modifications to reflect local
environmental conditions and human exposure patterns. The methodology stated that "local" may
refer to any appropriate geographic area where common aquatic environmental or exposure patterns
exist. Thus "local" may signify a Statewide, regional, river reach or entire river.
In today's proposal, site-specific criteria may be developed as long as the site-specific data,
either toxicologicalor exposure-related, is justifiable. For example, a State should use a site-specific
fish consumption rate that represents at least the central tendency (median or mean) of the
population surveyed (either sport or subsistence, or both). If a site-specific fish consumption rate
for sport anglers or subsistence anglers is lower than an EPA default value, it may be used in
calculating AWQC. To justify such a level (either higher or lower than EPA defaults) the State
280
-------
should present survey data it used in arriving at the site-specific fish consumption rate. The same
conditions apply to site-specific calculations of BAF, percent fish lipid, or the RSC. In the case of
deviations from toxicological values (IRIS values: verified noncancer and cancer assessments), EPA
recommends that the data upon which the deviation is based be presented to and approved by the
Agency before a criterion is developed.
3.4 Organoleptic Criteria
The 1980 A WQC National Guidelines provided for the development of organoleptic criteria
if organoleptic data were available for a specific contaminant. The methodology also made a clear
distinction that organoleptic criteria and toxicity-based criteria are derived from completely different
endpoints and that organoleptic criteria have no demonstrated relationship to potential adverse
human health effects. The 1992 National Experts Workshop participants and the Great Lakes
Committees of the Initiative both recommended that EPA place highest priority on setting toxicity-
based criteria, rather then using limited resources to set organoleptic criteria. Both efforts, the GLI
and the National Experts Workshop, concluded that organoleptic effects, while significant from an
aesthetic standpoint, were not a significant health concern and did not merit significant expenditures
of time and effort. While it can be argued that organoleptic properties indirectly affect human health
(people may drink less water or eat less fish due to objectionable taste or odor), they have not been
demonstrated to result in direct adverse effects, such as cancer or other types of toxicity.
3.5 Criteria for Chemical Classes
The 1980 AWQC National Guidelines allowed for the development of criteria for chemical
classes. A chemical class was defined as any group of chemical compounds which were reviewed
in a single risk assessment document. The Guidelines also stated that in criterion development,
isomers should be regarded as part of a chemical class rather than as a single compound. A class
criterion, therefore, was an estimate of risk/safety which applied to more than one member of a class.
It involved the use of available data on one or more chemicals of a class to derive criteria for other
compounds of the same class in the event that insufficient data were available to derive compound-
specific criteria. The criterion applied to each member of the class, rather than to the sum of the
compounds within the class. The 1980 methodology also acknowledged that, since relatively minor
structural changes within the class of compounds can have pronounced effects on their biological
activities, reliance on class criteria should be minimized.
The 1980 methodology prescribed the following analysis when developing a class criterion:
A detailedreviewof the chemical and physical properties of the chemicals within the
group should be made. A close relationship within the class with respect to chemical
activity would suggest a similar potential to reach common biological sites within
tissues. Likewise, similar lipid solubilities would suggest the possibility of
comparable absorption and distribution.
281
-------
Qualitative and quantitative data for chemicals within the group are examined.
Adequate toxicologicaldata on a number of compounds with a group provide a more
reasonable basis for extrapolation to other chemicals of the same class than minimal
data on one chemical or a few chemicals within the group.
Similarities in the nature of the toxicological response to chemicals in the class*
provide additional support for the prediction that the response to other members of
the class may be similar. In contrast, where the biological response has been shown
to differ markedly on a qualitative and quantitative basis for chemicals within a class,
the extrapolation of a criterion to other members is not appropriate.
• Additional support for the validity of extrapolation of a criterion to other members
of a class could be provided by evidence of similar metabolic and pharmacokinetic
data for some members of the class.
The proposal in the Federal Register allows for the development of a criterion for classes of
chemicals, as long as the 1980 methodology guidance is followed and a justification is provided
through the analysis of mechanistic data, pharmacokinetic data, structure-activity relationship data,
and limited acute and chronic toxicity data. When potency differences between members of a class
are great (such as in the case of chlorinated dioxins and furans), toxicity equivalency factors (TEFs)
may be more appropriately developed than one class criterion.
3.6 Criteria for Essential Elements
The 1980 AWQC National Guidelines acknowledged that developing criteria for essential
elements, particularly metals, must be a balancing act between toxicity and essentiality. The 1980
guidelines state:
that the criteria must consider essentiality and cannot be established at levels which
would result in deficiency of the element in the human population. The difference
between the RDA and the daily doses causing a specified risk level for carcinogens
or the ADIs (now RfDs) for noncarcinogens defines the spread of daily doses from
which the criterion may be derived. Because errors are inherent in defining both
essential and maximum-tolerable levels, the criterion is derived from the dose levels
near the center of such dose ranges.
In the current proposal, EPA endorses the guidance from the 1980 methodology and adds that
the process for developing criteria for essential elements should be similar to that used for any other
chemical with minor modifications. The RfD represents concern for one end of the exposure
spectrum (toxicity), whereas the RDA represents the other end (minimum essentiality). Where the
RDA and RfD values might occasionally appear to be similar in magnitude to one another, it does
not imply incompatibility of the two methodological approaches, nor does it imply inaccuracy or
error in either calculation.
282
-------
Appendices
-------
-------
Appendix A
TABLE A.1
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals 18 Years of Age or Older in the U.S. Population - Finfish and Shellfish
Grams/person/day
90% Interval*
Habitat
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
Marine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
5.59
0.00
17.80
39.04
86.30
12.42
0.00
45.98
64.08
111.38
18.01
0.00
60.64
86.25
142.96
4.91
0.00
14.89
36.13
81.99
11.55
0.00
44.48
61.61
101.94
16.85
0.00
57.06
80.29
134.23
6.28
0.00
20.63
42.16
96.67
13.29
0.00
48.34
68.05
120.49
19.17
0.00
64.63
91.00
154.15
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 8,478 individuals to the U.S. population of 177,807,000 using 3-year
combined survey weights.
ource of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food Intakes
by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the USDA's
Nutrient Data Base for Individual Food Intake Surveys.
A-l
-------
TABLE A.2
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals 18 Years of Age or Older in the U.S. Population - Finfish and Shellfish
Milligrams/kilogram/person/day
90% Interval*
Habitat
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
Marine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
75.56
0.00
242.49
547.61
1,171.84
172,86
0.00
624.83
911.05
1,573.20
248.42
0.00
829.02
1,197.36
2,014.67
66.37
0.00
205.05
493.47
1,123.52
160.73
0.00
598.84
877.29
1,468.43
232.19
0.00
791.06
1,133.18
1,839.55
84.75
0.00
277.26
587.37
1,252.78
184.99
0.00
670.34
952.66
1,713.17
264.64
0.00
872.61
1,264.74
2,180.87
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 8,478 individuals to the U.S. population of 177,807,000 using 3-year
combined survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-2
-------
TABLE A.3
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals of Age 14 and Younger in the Acute Consumers* - Finfish and Shellfish
Grams/person/day
Habitat
Statistic
Estimate
Fresh/Estuarine
n = 295
N= 6,267,000
Marine
n= 663
N= 13,190,000
All Fish
n = 807
N= 16,159,000
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
45.73
28.35
108.36
136.24
214.6
73.62
56.00
153.20
176.90
337.24
74.80
56.49
153.70
178.08
337.46
Note: Acute consumer = Individual who consumed fish at least once during the 3-day reporting period.
n = sample size
N = population size
Estimates are projected from a sample of acute consumers of age 14 and younger to the population of acute consumers
'f age 14 and younger using 3-year combined survey weights. The population for this survey consisted of individuals
n the 48 conterminous states.
ource of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
ntakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-3
-------
TABLE A.4
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals of Age 14 and Younger in the Acute Consumers* - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat
Statistic
Estimate
Fresh/Estuarine
n = 295
N = 6,267,000
Marine
= 663
N= 13,190,000
All Fish
= 807
= 16,159,000
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
1,721.99
1,271.12
3,760.67
4,208.18
9,789.49
2,532.95
2,107.05
5,068.69
6,376.47
8,749.02
2,624.35
2,172.61
5,020.14
6,904.83
10,384.82
+ Note: Acute consumer = Individual who consumed fish at least once during the 3-day reporting period.
n = sample size
N = population size
Estimates are projected from a sample of acute consumers of age 14 and younger to the population of acute consumers
of age 14 and younger using 3-year combined survey weights. The population for this survey consisted of individuals
in the 48 conterminous states.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-4
-------
TABLE A.5
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Females of Age 15 to 44 in the Acute Consumers* - Finfish and Shellfish
Grams/person/day
Habitat
Statistic
Estimate
Fresh/Estuarine
n = 445
N= 10,853,000
Marine
n = 774
N= 17,967,000
All Fish
n= 952
N = 21,924,000
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
61.40
35.22
148.83
185.44
363.56
76.53
62.96
149.78
178.74
271.06
88.80
69.95
170.01
212.56
361.04
+ Note: Acute consumer = Individual who consumed fish at least once during the 3-day reporting period.
n = sample size
N = population size
Estimates are projected from a sample of female acute consumers of age 15 to 44 to the population of female acute
consumers of age 15 to 44 using 3-year combined survey weights. The population for this survey consisted of
individuals in the 48 conterminous states.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food Intakes
by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
JSDA's Nutrient Data Base for Individual Food Intake Surveys.
A-5
-------
TABLE A.6
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Females of Age 15 to 44 in the Acute Consumers* - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat
Statistic
Estimate
Fresh/Estuarine
= 445
N = 10,853,000
Marine
n = 774
N= 17,967,000
AH Fish
n= 952
N = 21,924,000
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
961.58
533.18
2,578.81
3,403.75
6,167.24
1,227.41
986.25
2,469.67
3,007.98
4,800.68
1,414.54
1,100.44
2,726.46
3,740.83
6,703.25
Note: Acute consumer = Individual who consumed fish at least once during the 3-day reporting period.
n = sample size
N = population size
Estimates are projected from a sample acute consumers of females of age 14 and younger to the population of acute
consumers of females of age 14 and younger using 3-year combined survey weights. The population for this survey
consisted of individuals in the 48 conterminous states.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-6
-------
TABLE A.7
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals 18 Years of Age or Older in the Acute Consumers* - Finfish and Shellfish
Grams/person/day
90% Interval*
Habitat Statistic
Fresh/Estuarine Mean
n = 1,541 50th %
N = 37, 166,000 90th %
95th %
99th %
Marine Mean
n = 2,432 50th %
N = 57,830,000 90th %
95th %
99th %
AH Fish Mean
n = 3,007 50th %
N = 70,949,000 90th %
95th %
99th %
Estimate Lower Bound Upper
70.91
42.45
176.58
230.41
402.56
91.49
77.56
172.29
215.62
313.05
106.39
85.36
206.76
258.22
399.26
* Percentile intervals were estimated using the percentile bootstrap method with
+ Note: Acute consumer = Individual who
n = sample size
N = population size
64.16
37.24
165.08
224.00
358.58
87.35
74.89
168.00
201.99
292.80
102.37
84.00
197.84
241.00
336.50
1,000 bootstrap replications.
Bound
77.65
46.91
193.26
255.55
518.41
95.64
78.52
182.00
225.63
324.81
110.41
87.36
213.00
266.86
423.56
consumed fish at least once during the 3-day reporting period.
Estimates are projected from a sample of acute consumers 18 years of age or older to the population of acute
consumers 1 8 years of age or older using 3-year combined survey weights. The population for this survey consisted of
individuals in the 48 conterminous states.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
[Tie fish component of foods containing fish was calculated using data from the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
recipe file for release 7 of the
A-7
-------
TABLE A.8
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals 18 Years of Age or Older in the Acute Consumers* - Finfish and Shellfish
Milligrams/kilogram/person/day
90% Interval*
Habitat
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
n= 1,541
= 37,166,000
Marine
n = 2,432
N = 57,830,000
All Fish
n = 3,007
N = 70,949,000
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
959.15
601.88
2,442.97
3,116.28
5,151.98
1,270.78
1,062.93
2,467.68
3,116.74
4,250.22
1,461.71
1,189.29
2,802.28
3,588.11
5,355.90
867.58
532.31
2,233.16
2,839.90
4,432.30
1,214.65
1,019.60
2,331.88
2,906.16
4,037.74
1,406.34
1,156.77
2,685.81
3,308.93
5,095.58
1,050.72
656.86
2,606.66
3,303.96
6,931.61
1,326.90
1,087.06
2,585.09
3,264.98
4,387.96
1,517.09
1,225.43
2,868.73
3,798.54
5,766.99
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
* Note: Acute consumer = Individual who consumed fish at least once during the 3-day reporting period.
n = sample size
N = population size
Estimates are projected from a sample of acute consumers 18 years of age or older to the population of acute
consumers 18 years of age or older using 3-year combined survey weights. The population for this survey consisted of
individuals in the 48 conterminous states.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-8
-------
TABLE A.9
DAILY AVERAGE PER CAPITA ESTIMATES
As Consumed Fish
U.S. Population - Finfish and
OF FISH CONSUMPTION
Shellfish
Grams/person/day
90% Interval*
Habitat
Fresh/Estuarine
Marine
All Fish
Statistic Estimate
Mean 4.71
50th % 0.00
90th % 12.62
95th % 32.16
99th % 82.45
Mean 10.94
50th % 0.00
90th % , 39.51
95th % 59.62
99th % 106.84
Mean 15.65
50th % 0.00
90th % 55.02
95th % 78.34
99th % 133.46
Lower Bound Upper
4.17
0.00
10.91
29.81
77.17
10.14
0.00
37.29
57.03
104.59
14.67
0.00
51.38
75.21
125.27
Bound
5.25
0.00
13.98
35.15
86.40
11.73
0.00
42.91
61.84
114.55
16.63
0.00
56.00
80.56
140.21
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 1 1,912 individuals to the
combined survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991
Intakes by Individuals (CSFII).
U.S. population of 242,707,000 using 3-year
USD A Continuing Survey of Food
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-9
-------
TABLE A.10
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
U.S. Population - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat
90% Interval*
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
Marine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
74.16
0.00
204.00
547.64
1,274.55
186.06
0.00
663.00
991.96
1,942.17
260.22
0.00
880.47
1,308.54
2,356.54
65.74
0.00
177.97
505.10
1,197.29
170.81
0.00
627.39
960.40
1,815.48
242.60
0.00
844.35
1,267.15
2,224.54
82.57
0.00
225.16
565.37
1,324.90
201.31
0.00
717.18
1,044.69
2,042.99
277.83
0.00
918.79
1,346.71
2,556.68
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 11,912 individuals to the U.S. population of 242,707,000 using 3-year
combined survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-10
-------
TABLE A.11
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
U.S. Population, Acute Consumers* - Finfish and Shellfish
Grams/person/day
90% Interval*
Habitat
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
n= 1,892
N = 44,946,000
Marine
n = 3,184
N = 73,100,000
All Fish
n = 3,927
N = 89,800,000
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
68.00
39.52
170.84
224.78
374.74
87.77
71.77
169.39
209.50
320.41
100.63
80.79
197.44
253.38
371.59
61.92
36.16
158.74
212.91
336.50
83.74
69.73
167.00
198.11
292.80
96.66
79.29
188.74
231.51
359.29
74.07
44.68
181.79
245.98
431.34
91.80
74.23
173.65
221.73
341.88
104.60
83.90
205.12
264.45
401.61
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Acute consumer = Individual who consumed fish at least once during the 3-day reporting period.
n = sample size
N = population size
estimates are projected from the sample to the population of acute consumers in the 48 conterminous states, using
3-year combined survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
JSDA's Nutrient Data Base for Individual Food Intake Surveys.
A-ll
-------
TABLE A.12
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
U.S. Population, Acute Consumers* - Finfish and Shellfish
Milligrams/kilogram/person/day
90% Interval*
Habitat
Statistic
Estimate
Lower Bound
Upper Bound
Fresh/Estuarine
n = 1,892
N = 44,946,000
Mean
50th %
90th %
95th %
99th %
1,076.80
656.62
2,695.81
3,399.46
6,526.10
980.00
588.84
2,546.77
3,132.65
5,270.61
1,173.61
709.37
2,819.33
3,839.47
6,931.61
Marine
n = 3,184
N- 73,100,000
Mean
50th %
90th %
95th %
99th %
1,495.37
1,151.58
2,956.38
3,887.52
6,510.73
1,422.63
1,120.00
2,838.46
3,770.65
5,772.57
1,568.12
1,181.14
3,083.70
4,113.22
6,852.01
All Fish
n = 3,927
N = 89,800,000
Mean
50th %
90th %
95th %
99th %
1,674.31
1,307.30
3,299.54
4,258.69
7,126.90
1,606.79
1,267.12
3,133.69
4,065.32
6,644.11
1,741.83
1,339.46
3,462.35
4,483.83
7,794.41
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
* Note: Acute consumer = Individual who consumed fish at least once during the 3-day reporting period.
n = sample size
N = population size
Estimates are projected from the sample to the population of acute consumers in the 48 conterminous states, using
3-year combined survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food Intakes
by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the USDA's
Nutrient Data Base for Individual Food Intake Surveys.
A-12
-------
DAILY
TABLE A.13
AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
New England Region - Finfish and Shellfish
Grams/person/day
Habitat
Fresh/Estuarine
Marine
All Fish
90
Statistic Estimate
Mean 4.94
50th % 0.00
90th % 17.58
95th % 30.11
99th % 72.04
Mean 16.96
50th % 0.00
90th % 56.95
95th % 74.59
99th % 124.64
Mean 21.90
50th % 0.00
90th % 73.56
95th % 91.33
99th % 145.86
* Percentile intervals were estimated using the percentile bootstrap method
>Jote: Estimates are
weights.
% Interval*
Lower Bound Upper
3.87
0.00
6.92
25.76
64.61
15.57
0.00
49.00
73.91
99.13
19.95
0.00
65.07
83.81
138.94
with 1,000 bootstrap replications.
projected from a sample of 595 individuals to the regional population of 12,769,000 using
Bound
6.01
0.00
25.76
57.86
74.67
18.35
0.00
66.28
83.81
178.05
23.85
0.00
74.68
96.30
178.05
survey
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the USDA's
Nutrient Data Base for Individual Food Intake Surveys.
A-13
-------
TABLE A.14
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Middle Atlantic Region - Finfish and Shellfish
Grams/person/day
Habitat
90% Interval*
Statistic
Estimate
Lower Bound Upper Bound
Fresfa/Estuarioe
Marine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
3.79
0.00
12.13
25.29
61.72
14.89
0.00
52.01
67.21
148.91
18.68
0.00
61.26
80.54
153.23
2.62
0.00
9.28
19.32
48.00
12.76
0.00
46.27
60.67
113.21
15.59
0.00
55.67
69.72
149.28
4.96
0.00
13.71
31.42
84.21
17.02
0.00
55.67
74.63
154.15
21.77
. 0.00
64.74
94.85
171.37
1 Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 1,585 individuals to the regional population of 37,330,000 using
survey weights.
Source of individual consumption data: Combined 1989,1990, and 1991 USDA Continuing Survey of Food Intakes
by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the USDA's
Nutrient Data Base for Individual Food Intake Surveys.
A-14
-------
TABLE A.15
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
South Atlantic Region - Finfish and Shellfish
Grams/person/day
Habitat
Fresh/Estuarine
Marine
All Fish
Statistic
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
90% Interval*
Estimate Lower Bound Upper
4.92
0.00
16.72
30.45
77.54
11.41
0.00
44.56
63.37
102.78
16.33
0.00
57.62
83.39
130.78
* Percentile intervals were estimated using the percentile bootstrap method with 1,000
Note: Estimates are projected from a sample
using survey weights.
3.93
0.00
8.63
28.10
72.35
9.72
0.00
36.43
57.88
91.51
15.10
0.00
54.65
77.30
122.02
bootstrap replications.
Bound
5.91
0.00
20.77
39.25
86.41
13.11
0.00
49.55
65.33
107.98
17.57
0.00
65.74
92.88
139.45
of 2,245 individuals to the regional population of 42,307,000
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe
USDA's Nutrient Data Base for Individual Food Intake Surveys.
file for release 7 of the
A-15
-------
TABLE A.16
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
East North Central Region - Finfish and Shellfish
Grams/person/day
Habitat
90% Interval*
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
Marine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
2.88
0.00
5.10
18.24
58.24
10.33
0.00
38.82
56.88
113.83
13.21
0.00
47.50
72.05
114.31
1.99
0.00
1.71
14.49
56.50
8.33
0.00
37.33
52.63
104.59
11.07
0.00
43.87
65.87
106.60
3.77
0.00
5.94
26.02
66.00
12.33
0.00
42.96
65.03
132.39
15.35
0.00
51.55
80.51
132.39
Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 2,222 individuals to the regional population of 41,565,000
using survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-16
-------
TABLE A.17
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
East South Central Region - Finflsh and Shellfish
Grams/person/day
Habitat
Fresh/Estuarine
Marine
All Fish
Statistic
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Estimate
10.66
0.00
37.64
58.41
165.12
5.97
0.00
23.87
37.29
61.16
16.63
0.00
52.26
69.94
165.12
* Percentile intervals were estimated using the percentile bootstrap method
Note: Estimates are projected from a sample of 671
weights.
Source of individual consumption data: Combined
Intakes by Individuals (CSFII).
90% Interval*
Lower Bound Upper
5.84
0.00
28.11
51.53
86.40
4.41
0.00
18.54
32.27
56.00
12.03
0.00
45.44
66.77
119.53
with 1,000 bootstrap replications.
Bound
15.49
0.00
49.93
72.36
224.33
7.54
0.00
32.27
47.13
76.45
21.24
0.00
60.67
86.40
224.33
individuals to the regional population of 15,1 13,000 using survey
1989, 1990, and 1991 USDA Continuing Survey of Food
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-17
-------
TABLE A.18
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
West North Central Region - Finfish and Shellfish
Grams/person/day
Habitat
Fresh/Estuarine
Marine
All Fish
Statistic
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Estimate
4.48
0.00
4.41
25.84
104.32
8.37
0.00
31.50
49.00
99.62
12.85
0.00
42.99
63.05
141.07
* Percentile intervals were estimated using the percentile bootstrap method
Note: Estimates are projected from a sample of 785
using survey weights.
Source of individual consumption data: Combined
Intakes by Individuals (CSFII).
90% Interval*
Lower Bound Upper
2.22
0.00
0.31
14.42
58.05
5.71
0.00
27.83
40.36
66.20
8.06
0.00
36.39
55.03
112.00
with 1,000 bootstrap replications.
Bound
6.74
0.00
9.19
47.73
178.33
11.03
0.00
39.88
56.00
103.24
17.64
0.00
54.78
86.40
178.33
individuals to the regional population of 17,720,000
1989, 1990, and 1991 USDA Continuing Survey of Food
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-18
-------
TABLE A.19
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
West South Central Region - Finfish and Shellfish
Grams/person/day
Habitat
Fresh/Estuarine
Marine
All Fish
Statistic
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
* Percentile intervals were estimated using the
Estimate Lower
7.04
0.00
23.85
55.46
112.68
6.02
0.00
22.23
33.75
92.12
13.06
0.00
49.69
74.77
114.22
percentile bootstrap method with 1,000
90% Interval*
Bound Upper
4.82
0.00
17.18
47.71
.77.17
4.71
0.00
18.72
28.38
55.67
10.04
0.00
37.48
57.88
100.74
bootstrap replications.
Bound
9.26
0.00
37.48
74.95
115.75
7.33
0.00
27.67
41.99
100.20
16.08
0.00
55.92
85.33
115.75
Note: Estimates are projected from a sample of 1,287 individuals to the regional population of 26,321,000
using survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe
USDA's Nutrient Data Base for Individual Food Intake Surveys.
file for release 7 of the .
A-19
-------
TABLE A.20
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
Mountain
As Consumed Fish
Region - Finfish and
Shellfish
Grams/person/day
Habitat
Fresh/Estuarine
Vlarine
All Fish
Statistic
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Estimate
3.23
0.00
0.48
20.90
78.60
7.97
0.00
30.96
52.68
89.62
11.20
0.00
39.32
58.55
95.84
90% Interval*
Lower Bound Upper
1.86
0.00
0.00
10.51
55.50
5.64
0.00
27.83
37.80
63.71
8.26
0.00
36.89
55.67
91.00
Bound
4.60
0.00
5.25
36.00
96.67
10.29
0.00
34.86
58.73
91.00
14.13
0.00
48.20
68.37
136.75
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 889 individuals to the regional population of 13,385,000
using survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991
Intakes by Individuals
(CSFII).
USDA Continuing Survey of Food
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data
Base for Individual Food Intake Surveys.
A-20
-------
TABLE A. 21
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH
Pacific
As Consumed Fish
Region - Finfish and Shellfish
CONSUMPTION
Grams/person/day
Habitat
Fresh/Estuarine
Marine
AH Fish
Statistic
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Estimate Lower
3.93
0.00
10.16
26.46
68.74
12.88
0.00
50.78
69.74
111.49
16.81
0.00
55.87
83.44
122.64
* Percentile intervals were estimated using the percentile bootstrap method with 1,000
Note: Estimates are projected from a sample
using survey weights.
90% Interval*
Bound Upper
2.78
0.00
8.53
22.72
51.53
10.18
0.00
43.68
60.67
106.93
14.32
0.00
51.53
71.74
116.93
bootstrap replications.
Bound
5.07
0.00
12.08
29.79
91.02
15.58
0.00
54.65
79.57
121.52
19.29
0.00
59.76
95.95
169.54
of 1,633 individuals to the regional population of 36,197,000
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe
USDA's Nutrient Data Base for Individual Food Intake Surveys.
file for release 7 of the
A-21
-------
TABLE A. 22
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
New England Region - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat
90% Interval*
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
Marine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
72.50
0.00
271.02
503.29
1,031.43
283.20
0.00
964.92
1,321.16
2,083.17
355.70
0.00
1,059.55
1,536.32
2,362.63
59.10
0.00
103.78
366.45
755.19
251.27
0.00
852.34
1,163.41
1,789.78
315.28
0.00
1,019.19
1,323.53
2,120.35
85.89
0.00
363.88
723.60
1,240.52
315.14
0.00
1,041.69
1,547.33
2,631.23
396.12
0.00
1,269.56
1,568.72
3,014.33
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 595 individuals to the regional population of 12,769,000 using survey
weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-22
-------
TABLE A. 23
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Middle Atlantic Region - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat
90% Interval*
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
Marine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
63.05
0.00
204.00
413.18
1,223.45
239.72
0.00
768.83
1,128.64
2,047.28
302.77
0.00
941.32
1,416.61
2,510.55
45.79
0.00
130.89
320.99
759.78
211.16
0.00
706.68
930.83
1,697.69
262.82
0.00
803.12
1,209.29
2,302.98
80.31
0.00
260.42
597.44
1,324.90
268.29
0.00
834.11
1,365.60
2,352.51
342.72
0.00
1,160.53
1,561.31
2,673.55
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 1,585 individuals to the regional population of 37,330,000
using survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
JSDA's Nutrient Data Base for Individual Food Intake Surveys.
A-23
-------
TABLE A. 24
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
South Atlantic Region - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat
90% Interval"
Statistic
Estimate
Lower Bound
Upper Bound
Tresh/Estuarine
Vlarine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
73.39
0.00
256.50
509.78
1,216.10
178.44
0.00
641.43
977.52
1,690.98
251.83
0.00
858.43
1,297.53
2,175.84
60.04
0.00
150.60
451.81
1,092.18
154.11
0.00
604.63
913.81
1,331.69
234.98
0.00
795.11
1,192.41
1,943.82
86.73
0.00
290.43
596.27
1,353.36
202.77
0.00
690.25
1,071.87
1,943.8
268.67
0.00
943.87
1,411.80
2,323.75
: Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 2,245 individuals to the regional population of 42,307,000 using
survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-24
-------
TABLE A. 25
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
East North Central Region - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat
90% Interval*
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
Marine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
44.34
0.00
75.18
316.64
858.02
197.24
0.00
682.31
1,114.31
2,128.19
241.58
0.00
827.02
1,388.72
2,376.91
29.23
0.00
25.97
222.48
808.03
155.33
0.00
624.84
943.16
1,923.34
198.48
0.00
731.33
1,241.40
1,930.85
59.46
0.00
100.90
408.50
1,001.91
239.14
0.00
747.86
1,333.60
3,430.60
284.68
0.00
943.16
1,513.97
3,430.60
Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 2,222 individuals to the regional population of 41,565,000 using
survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-25
-------
TABLE A. 26
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
East South Central Region - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat
90% Interval*
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
Marine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
199.53
0.00
807.10
1,086.94
2,552.93
94.04
0.00
340.21
569.69
1,370.45
293.57
0.00
977.86
1,342.17
2,786.16
106.27
0.00
573.22
1,021.39
2,257.50
70.81
0.00
258.49
435.77
880.38
209.10
0.00
898.05
1,035.12
2,257.50
292.79
0.00
1,021.39
1,355.11
3,044.97
117.28
0.00
425.71
690.12
1,644.42
378.05
0.00
1,023.86
1,585.92
3,044.97
1 Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 671 individuals to the regional population of 15,113,000 using survey
weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-26
-------
TABLE A. 27
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
West North Central Region - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat
90% Interval*
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
Marine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
68.68
0.00
44.73
428.01
1,819.78
165.41
0.00
546.17
839.85
2,354.93
234.09
0.00
756.81
1,222.04
2,988.90
31.08
0.00
5.23
129.96
929.24
101.17
0.00
438.28
762.30
1,809.91
137.32
0.00
578.06
94.1.54
2,224.38
106.29
0.00
99.14
904.27
2,310.54
229.64
0.00
711.48
981.18
2,890.40
330.86
0.00
829.11
1,653.86
3,733.22
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 785 individuals to the regional population of 17,720,000 using survey
weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-27
-------
TABLE A. 28
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
West South Central Region - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat Statistic
Fresh/Estuarine Mean
50th % -
90th %
95th %
99th %
Marine Mean
50th %
90th %
95th %
99th %
AH Fish Mean
50th %
90th %
95th %
99th %
90% Interval*
Estimate Lower Bound Upper Bound
106.38
0.00
428.38
863.50
1,398.61
96.30
0.00
354.59
655.36
1,189.90
. 202.68
0.00
840.80
1,085.60
1,725.49
* Percentile intervals were estimated using the percentile bootstrap method with
Note: Estimates are projected from a sample of 1
survey weights.
74.14 138.63
0.00 0.00
297.04 565.37
717.62 1,038.97
1,112.73 1,416.47
71.85 120.75
0.00 0.00
291.60 396.92
548.14 747.56
1,088.33 1,697.71
154.13 251.23
0.00 0.00
698.95 887.76
920.62 1,133.18
1,416.47 1,755.07
1,000 bootstrap replications.
,287 individuals to the regional population of 26,321,000 using
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
recipe file for release 7 of the
A-28
-------
TABLE A. 29
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Mountain Region - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat
90% Interval*
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
Marine
All Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
43.34
0.00
7.32
295.74
1,028.60
132.52
0.00
458.07
751.69
1,535.18
175.86
0.00
641.64
957.82
1,702.77
27.12
0.00
0.00
145.18
750.63
88.63
0.00
421.93
680.90
1,344.97
128.71
0.00
542.17
817.21
1,421.80
59.55
0.00
64.26
458.98
1,115.58
176.42
0.00
583.88
902.64
1,713.17
223.01
0.00
729.72
1,075.34
1,786.53
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 889 individuals to the regional population of 13,385,000 using survey
weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-29
-------
TABLE A. 30
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Pacific Region - Finfish and Shellfish
Milligrams/kilogram/person/day
Habitat
90% Interval*
Statistic
Estimate
Lower Bound Upper Bound
Fresh/Estuarine
Marine
AH Fish
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
Mean
50th %
90th %
95th %
99th %
59.58
0.00
182.83
416.82
1,018.21
226.16
0.00
799.42
1,159.26
2,173.89
285.75
0.00
946.55
1,413.34
2,589.08
44.67
0.00
125.87
345.98
861.13
162.61
0.00
760.71
1,096.99
1,995.09
222.88
0.00
873.95
1,295.67
2,181.37
74.50
0.00
221.07
472.91
1,114.52
289.71
0.00
945.97
1,369.73
2,598.14
348.61
0.00
1,054.94
1,489.91
2,661.40
* Percentile intervals were estimated using the percentile bootstrap method with 1,000 bootstrap replications.
Note: Estimates are projected from a sample of 1,633 individuals to the regional population of 36,197,000
using survey weights.
Source of individual consumption data: Combined 1989, 1990, and 1991 USDA Continuing Survey of Food
Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-30
-------
TABLE A.31
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals IS Years of Age or Older in the U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day'
Estuarine
Freshwater
Marine
Shrimp
Perch
Flatfish (Estuarine)
Crab (Estuarine)
Flounder
Oyster
Mullet
Croaker
Herring
Smelts
Clam (Estuarine)
Scallop (Estuarine)
Anchovy
Scup
Sturgeon
Catfish
Trout
Carp
Pike
Salmon (Freshwater)
Tuna
Flatfish (Marine)
Cod
Salmon (Marine)
Haddock
1.72959
0.60368
0.52735
0.37126
0.29941
0.22555
0.08756
0.06749
0.03925
0.03753
0.03146
0.00322
0.00292
0.00068
0.00054
1.18227
0.44946
0.05727
0.02337
0.01096
4.71788
1.28921
1.26813
0.91786
0.61729
Notes: Estimates are projected from a sample of 8,478 individuals 18 years of age or older to the population of 177,807,000 individuals
18 years of age or older using 3-year combined survey weights. The population for this survey consisted of individuals in the 48
conterminous states. .
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of Food Intakes by
Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-31
-------
TABLE A.31 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals 18 Years of Age or Older in the U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
Marine (Con't.)
Unknown
Crab (Marine)
Pollock
Clam (Marine)
Ocean Perch
Porgy
Scallop (Marine)
Sea Bass
Lobster
Swordfish
Sardine
Squid
Pompano
Sole
Mackerel
Whiting
Shark
Halibut
Mussels
Whitefish
Snapper
Octopus
Barracuda
Abalone
Seafood
Fish
0.43234
0.37254
0.35788
0.30679
0.30502
0.28389
0.25467
0.25446
0.17743
0.13812
0.12760
0.10485
0.10096
0.07188
0.06481
0.02596
0.02396
0.01911
0.00888
0.00735
0.00512
0.00151
0.00103
0.00057
0.00077
Notes: Estimates are projected from a sample of 8,478 individuals 18 years of age or older to the population of 177,807,000 individuals
18 years of age or older using 3-year combined survey weights. The population for this survey consisted of individuals in the 48
conterminous states.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of Food Intakes by
Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-32
-------
TABLE A.31 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals 18 Years of Age or Older in the U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day
All Species
Tuna
Shrimp
Flatfish (Marine)
Cod
Catfish
Salmon (Marine)
Haddock
Perch
Flatfish (Estuarine)
Trout
Crab (Marine)
Pollock
Crab (Estuarine)
Clam (Marine)
Ocean Perch
Porgy
Flounder
Scallop (Marine)
Sea Bass
Lobster
Oyster
Swordfish
Sardine
Squid
Pompano
Sole
4.71788
1.72959
1.28921
1.26813
1.18227
0.91786
0.61729
0.60368
0.52735
0.44946
0.43234
0.37254
0.37126
0.35788
0.30679
0.30502
0.29941
0.28389
0.25467
0.25446
0.22555
0.17743
0.13812
0.12760
0.10485
0.10096
-------
TABLE A.31 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals 18 Years of Age or Older in the U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
AH Species (Con't.)
Mullet
Mackerel
Croaker
Whiting
Carp
Herring
Smelts
Clam (Estuarine)
Shark
Halibut
Pike
Mussels
Salmon (Freshwater)
Whitefish
Snapper
Octopus
Scallop (Estuarine)
Anchovy
Barracuda
Abalone
Fish
Scup
Seafood
Sturgeon
0.08756
0.07188
0.06749
0.06481
0.05727
0.03925
0.03753
0.03146
0.02596
0.02396
0.02337
0.01911
0.01096
0.00888
0.00735
0.00512
0.00322
0.00292
0.00151
0.00103
0.00077
0.00068
0.00057
0.00054
Notes: Estimates are projected from a sample of 8,478 individuals 18 years of age or older to the population of 177,807,000 individuals
18 years of age or older using 3-year combined survey weights. The population for this survey consisted of individuals in the 48
conterminous states.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of Food Intakes by
Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys. .
A-34
-------
TABLE A.32
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals 14 Years of Age and Younger in the U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
Estuarine
Freshwater
Marine
Shrimp
Perch
Flatfish (Estuarine)
Flounder
Oyster
Crab (Estuarine)
Mullet
Clam (Estuarine)
Croaker
Smelts
Anchovy
Scollop (Estuarine)
Catfish
Trout
Carp
Pike
Salmon (Freshwater)
Tuna
Cod
Pollock
Flatfish (Marine)
Ocean Perch
Porgy
Salmon (Marine)
0.36872
0.28899
0.17291
0.08613
0.03172
0.02880
0.02396
0.01273
0.00335
0.00080
0.00061
0.00034
0.47501
0.34732
0.02862
0.01170
0.00321
2.82208
1.13423
0.64386
0.42271
0.38098
0.32992
0.26894
Notes: Estimates are projected from a sample of 2,977 individuals of age 14 and younger to the population of 55,163,000 individuals of
age 14 and younger using 3 years combined survey weights.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of Food Intakes by
'ndividuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
JSDA's Nutrient Data Base for Individual Food Intake Surveys.
A-35
-------
TABLE A.32 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals 14 Years of Age and Younger in the U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
Marine (Con't.)
Unknown
All Species
Haddock
Clam (Marine)
Squid
Sea Bass
Pompano
Lobster
Mackerel
Mussels
Crab (Marine)
Scallop (Marine)
Whiting
Whitefish
Halibut
Sardine
Seafood
Fish
Tuna
Cod
Pollock
Catfish
Flatfish (Marine)
Ocean Perch
Shrimp
0.24788
0.14478
0.12532
0.07716
0.06221
0.05980
0.04899
0.03597
0.03354
0.02981
0.02808
0.01170
0.00995
0.00765
0.00005
0.00568
2.82208
1.13423
0.64386
0.47501
0.42271
0.38098
0.36872
Notes: Estimates are projected from a sample of 2,977 individuals of age 14 and younger to the population of 55,163,000 individuals of
age 14 and younger using 3 years combined survey weights.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of Food Intakes by
Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-36
-------
TABLE A.32 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals 14 Years of Age and Younger in the U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
All Species (Con't.)
Haddock
Trout
Porgy
Perch
Salmon (Marine)
Flatfish (Estuarine)
Clam (Marine)
Squid
Flounder
Sea Bass
Pompano
Lobster
Mackerel
Mussels
Crab (Marine)
Oyster
Scallop (Marine)
Crab (Estuarine)
Carp
Whiting
Mullet
Clam (Estuarine)
Pike
Whitefish
Halibut
Sardine
Fish
Croaker
0.24788
0.34732
0.32992
0.28899
0.26894
0.17291
0.14478
0.12532
0.08613
0.07716
0.06221
0.05980
0.04899
0.03597
0.03354
0.03172
0.02981
0.02880
0.02862
0.02808
0.02396
0.01273
0.01170
0.01170
0.00995
0.00765
0.00568
0.00335
Notes: Estimates are projected from a sample of 2,977 individuals of age 14 and younger to the population of 55,163,000
individuals of age 14 and younger using 3 years combined survey weights.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of Food Intakes by
Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-37
-------
TABLE A.32 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Individuals 14 Years of Age and Younger in the U.S. Population - Mean Consumption by Species within Habitat
Estimated Mean
Habitat
Species
Salmon (Freshwater)
Smelts
Anchovy
Scallop (Estuarine)
Seafood
(grams/person/aayj
0.00321
0.00080
0.00061
0.00034
0.00005
Notes: Estimates are projected from a sample of'2,977 individuals of age 14 and younger to the population of 55,163,000
individuals of age 14 and younger using 3 years combined survey weights.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of Food Intakes by
Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-38
-------
TABLE A.33
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Females of Age 15 to 44 in the U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
Estuarine
Freshwater
Marine
Shrimp
Perch
Flatfish (Estuarine)
Flounder
Crab (Estuarine)
Mullet
Oyster
Croaker
Clam (Estuarine)
Herring
Scallop (Estuarine)
Anchovy
Catfish
Trout
Carp
Pike
Salmon (Freshwater)
Tuna
Flatfish (Marine)
Cod
Pollock
Haddock
Salmon (Marine)
Crab (Marine)
1.52145
0.55348
0.45313
0.23224
0.22766
0.05635
0.02736
0.02672
0.01925
0.01112
0.00225
0.00018
0.81492
0.34703
0.07291
0.00756
0.00479
4.41949
1.10778
1.04468
0.48699
0.43548
0.40098
0.26511
Notes: Estimates are projected from a sample of 2,891 females of age 15 to 44 to the population of
58,750,000 females of age 15 to 44 using 3-year combined survey weights.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of
Food Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7
of the USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-39
-------
TABLE A.33 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Females of Age 15 to 44 in the U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
Marine (Con't.)
All Species
Porgy
Clam (Marine)
Lobster
Scallop (Marine)
Sea Bass
Ocean Perch
Swordfish
Squid
Pompano
Mackerel
Sole
Halibut
Sardine
Whiting
Whitefish
Octopus
Abalone
Mussels
Seafood
Shark
Tuna
Shrimp
Flatfish (Marine)
Cod
Catfish
Perch
Pollock
0.23300
0.21896
0.20638
0.19840
0.18125
0.17925
0.17758
0.06367
0.06076
0.04620
0.03611
0.03495
0.03436
0.02711
0.00756
0.00318
0.00312
0.00280
0.00046
0.00034
4.41949
1.52145
1.10778
1.04468
0.81492
0.55348
0.48699
Notes: Estimates are projected from a sample of 2,891 females of age 15 to 44 to the population of 58,750,000
females of age 15 to 44 using 3-year combined survey weights.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of Food Intakes
by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-40
-------
TABLE A.33 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Females of Age 15 to 44 in the U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
All Species (Con't.)
Flatfish (Estuarine)
Haddock
Salmon (Marine)
Trout
Crab (Marine)
Porgy
Flounder
Crab (Estuarine)
Clam (Marine)
Lobster
Scallop (Marine),
Sea Bass
Ocean Perch
Swordfish
Carp
Squid
Pompano
Mullet
Mackerel
Sole
Halibut
Sardine
Oyster
Whiting
Croaker
Clam (Estuarine)
Herring
0.45313
0.43548
0.40098
0.34703
0.26511
0.23300
0.23224
0.22766
0.21896
0.20638
0.19840
0.18125
0.17925
0.17758
0.07291
0.06367
0.06076
0.05635
0.04620
0.03611
0.03495
0.03436
0.02736
0.02711
0.02672
0.01925
0.01112
Notes: Estimates are projected from a sample of 2,891 females of age 15 to 44 to the population of
58,750,000 females of age 15 to 44 using 3-year combined survey weights.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of
Food Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7
of the USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-41
-------
TABLE A.33 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
Females of Age 15 to 44 in the U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
AH Species (Con't.)
Pike
Whitefish
Salmon (Freshwater)
Octopus
Abalone
Mussels
Scallop (Estuarine)
Seafood
Shark
Anchovy
0.00756
0.00756
0.00479
0.00318
0.00312
0.00280
0.00225
0.00046
0.00034
0.00018
Notes: Estimates are projected from a sample of 2,891 females of age 15 to 44 to the population of
58,750,000 females of age 15 to 44 using 3-year combined survey weights.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of
Food Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7
of the USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-42
-------
TABLE A.34
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
Estuarine
Freshwater
Marine
Shrimp
Perch
Flatfish (Estuarine)
Crab (Estuarine)
Flounder
Oyster
Mullet
Croaker
Herring
Smelts
Clam (Estuarine)
Scallop (Estuarine)
Anchovy
Scup
Sturgeon
Catfish
Trout
Carp
Pike
Salmon (Freshwater)
Tuna
Cod
Flatfish (Marine)
Salmon (Marine)
1.37241
0.52580
0.43485
0.29086
0.24590
0.17419
0.07089
0.05021
0.02937
0.02768
0.02691
0.00247
0.00228
0.00050
0.00040
1.06776
0.43050
0.04846
0.01978
0.00881
4.19998
1.22827
1.06307
0.73778
Notes: Estimates are projected from a sample of 8,478 individuals 18 years of age or older to the population of
177,807,000 individuals 18 years of age or older using 3-year combined survey weights. The population for this
survey consisted of individuals in the 48 conterminous states.
Source of individual consumption data: USDA Combined 1989,1990, and 1991 Continuing Survey of Food Intakes
by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of the
LJSDA's Nutrient Data Base for Individual Food Intake Surveys.
A-43
-------
TABLE A.34 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
Marine (Con't.)
Haddock
Pollock
Crab (Marine)
Ocean Perch
Clam (Marine)
Porgy
Scallop (Marine)
Sea Bass
Lobster
Swordfish
Squid
Sardine
Pompano
Sole
Mackerel
Whiting
Halibut
Mussels
Shark
Whitefish
Snapper
Octopus
Barracuda
Abalone
Seafood
0.51533
0.44970
0.33870
0.31878
0.30617
0.29844
0.21805
0.20794
0.20001
0.13879
0.12196
0.10313
0.09131
0.07396
0.06379
0.05498
0.02463
0.02217
0.01901
0.00916
0.00539
0.00375
0.00111
0.00075
0.00043
Notes: Estimates are projected from a sample of 8,478 individuals 18 years of age or older to the
population of 177,807,000 individuals 18 years of age or older using 3-year combined survey weights.
The population for this survey consisted of individuals in the 48 conterminous states.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of
Food Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of
the USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-44
-------
TABLE A.34 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
Unknown
All Species
Fish
Tuna
Shrimp
Cod
Catfish
Flatfish (Marine)
Salmon (Marine)
Perch
Haddock
Pollock
Flatfish (Estuarine)
Trout
Crab (Marine)
Ocean Perch
Clam (Marine)
Porgy
Crab (Estuarine)
Flounder
Scallop (Marine)
Sea Bass
Lobster
Oyster
Swordfish
Squid
Sardine
0.00186
4.19998
1.37241
1.22827
1.06776
1.06307
0.73778
0.52580
0.51533
0.44970
0.43485
0.43050
0.33870
0.31878
0.30617
0.29844
0.29086
0.24590
0.21805
0.20794
0.20001
0.17419
0.13879
0.12196
0.10313
Notes: Estimates are projected from a sample of 8,478 individuals 18 years of age or older to the
population of 177,807,000 individuals 18 years of age or older using 3-year combined survey weights.
The population for this survey consisted of individuals in the 48 conterminous states.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of
Food Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of
the USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-45
-------
TABLE A.34 (continued)
DAILY AVERAGE PER CAPITA ESTIMATES OF FISH CONSUMPTION
As Consumed Fish
U.S. Population - Mean Consumption by Species within Habitat
Habitat
Species
Estimated Mean
(grams/person/day)
Pompano
Sole
Mullet
Mackerel
Whiting
Croaker
Carp
Herring
Smelts
Clam (Esruarine)
Halibut
Mussels
Pike
Shark
Whiteflsh
Salmon (Freshwater)
Snapper
Octopus
Scallop (Estuarine)
Anchovy
Fish
Barracuda
Abalone
Scup
Seafood
Sturgeon
0.09131
0.07396
0.07089
0.06379
0.05498
0.05021
0.04846
0.02937
0.02768
0.02691
0.02463
0.02217
0.01978
0.01901
0.00916
0.00881
0.00539
0.00375
0.00247
0.00228
0.00186
0.00111
0.00075
0.00050
0.00043
0.00040
Notes: Estimates are projected from a sample of 8,478 individuals 18 years of age or older to the
population of 177,807,000 individuals 18 years of age or older using 3-year combined survey weights.
The population for this survey consisted of individuals in the 48 conterminous states.
Source of individual consumption data: USDA Combined 1989, 1990, and 1991 Continuing Survey of
Food Intakes by Individuals (CSFII).
The fish component of foods containing fish was calculated using data from the recipe file for release 7 of
the USDA's Nutrient Data Base for Individual Food Intake Surveys.
A-46
-------
Appendix B
Evaluation of the Quality of Data Set(s) for Use in Deriving an RfD
The derivation of RfDs begins with a thorough review and assessment of the toxicological
data base to identify the type and magnitude of possible adverse health effects associated with a
chemical. This evaluation should include an examination of the full range of possible health effects,
including acute, short-term (14 to 28 days), subchronic, reproductive/developmental, and chronic
effects.
To be useful for supporting the derivation of an RfD, a study must meet certain standards
with regard to experimental design, conduct and data reporting. This appendix provides general
guidance on criteria for appropriate study design for a variety of types of toxicity studies. These
guidelines provide the assessor with a means to evaluate the quality and adequacy of data.
Appropriate studies are used both for the evaluation of potential hazard of the chemical and for the
derivation of the RfD.
Acute Toxicity Determination
Studies of acute exposure (one dose or multiple dose exposure occurring within a short time
(e.g. less than 24 hours)) are widely available for many chemicals. Acute toxicity [often expressed
in terms of the lethal dose (or concentration) to 50 percent of the population (LD50 or LC50)] is
usually the initial step in experimental assessment and evaluation of a chemical's toxic
characteristics. Such studies are used in establishing a dosage regimen in subchronic and other
studies and may provide initial information on the mode of toxic action of a substance. Because
LD50 or LC50 studies are of short duration, inexpensive and easy to conduct, they are commonly used
in hazard classification systems.
Acute lethality studies are of limited use, however, in the derivation of chronic criteria, since
the establishment of chronic criteria should never be based on exposures that approach acutely lethal
levels. However, the data from such studies do provide information on health hazards likely to arise
from individual short-term exposures. Such studies provide high dose effects data from which to
evaluate potential effects from exposures which may temporarily exceed the acceptable chronic
exposure level. An evaluation of the data should include the incidence and severity of all
abnormalities, the reversibility of abnormalities observed other than lethality, gross lesions, body
weight changes, effects on mortality, and any other toxic effects.
In recent years guidelines have been established to improve quality and provide uniformity
in test conditions. Unfortunately, many published LD50 or LC50 tests were not conducted in
accordance with current EPA or OECD guidelines (USEPA, 1985; OECD, 1987) since they were
conducted prior to establishment of those guidelines. For this reason, it becomes necessary to
examine each test or study to determine if the study was conducted in an adequate manner.
The following is a list of ideal conditions compiled from various testing guidelines which
may be used for determination of adequacy of acute toxicity data. Many published studies do not
B-l
-------
report details of test conditions making such determinations difficult. However, test conditions
guidelines that might be considered ideal may include:
General:
• Animal age and species identified.
• Minimum of 5 animals per sex per dose group (both sexes should be used).
• 14-day or longer observation period following dosing.
• Minimum of 3 dose levels appropriately spaced (most statistical methods
require at least 3 dose levels).
• Identification of purity or grade of test material used (particularly important
in older studies).
• If a vehicle used, the selected vehicle is known to be non-toxic.
• Gross necropsy results for test animals.
• Acclimation period for test animals before initiating study.
Specific conditions for oral LDSO:
• Dosing by gavage or capsule.
• Total volume of vehicle plus test material remain constant for all dose levels.
• Animals were fasted before dosing.
Specific conditions for dermal LDSO:
• Exposure on intact, clipped skin and involve approximately 10 percent of
body surface.
• Animals prevented from oral access to test material by restraining or covering
test site.
B-2
-------
Specific conditions for inhalation LC50:
• Duration of exposure at least 4 hours.
« If an aerosol (mist or paniculate), the particle size (median diameter and
deviation) should be reported.
Although the above listed conditions would be included in an ideally conducted study, not
all of these conditions need to be included in an adequately conducted study. Therefore, some
discretion is required on the part of the individual reviewing these studies (USEPA, 1985; OECD,
1987).
Short-Term Toxicity Studies (14-Day or 28-Day Repeated Dose Toxicity)
Short-term exposure generally refers to multiple or continuous exposure usually occurring
over a 14-day to 28-day time period. The purpose of short-term repeated dose studies is to provide
information on possible adverse health effects from repeated exposures over a limited time period.
The following guidelines were derived using the OECD Guidelines for Testing of Chemicals
(OECD, 1987) for determining the design and quality of a repeated dose short-term toxicity study:
• Minimum of 3 dose levels administered and an adequate control group used.
• Minimum of 10 animals per sex, per dose group (both sexes should be used).
• The highest dose level should ideally elicit some signs of toxicity without inducing
excessive lethality and the lowest dose should ideally produce no signs of toxicity.
• Ideal dosing regimes include 7 days per week for a period of 14 days or 28 days.
• All animals should be dosed by the same method during the entire experiment period.
• Animals should be observed daily for signs of toxicity during the treatment period
(i.e., 14 or 28 days). Animals that die during the study are necropsied and all
survivors in the treatment groups are sacrificed and necropsied at the end of the study
period.
• All observed results, quantitative and incidental, should be evaluated by an
appropriate statistical method.
• Clinical examinations should include hematology and clinical biochemistry,
urinalysis may be required when expected to provide an indication of toxicity.
Pathological examination should include gross necropsy and histopathology.
B-3
-------
The findings of short-term repeated dose toxicity studies should be considered in terms of
the observed toxic effects and the necropsy and histopathological findings. The evaluation will
include the incidence and severity of abnormalities, gross lesions, body weight changes, effects on
mortality, and other general or specific toxic effects (OECD, 1987).
These guidelines represent ideal conditions and studies will not be expected to meet all
standards in order to be considered to be adequate. For example, the National Toxicology Program's
cancer bioassay program has generated a substantial data base of short-term repeated dose studies.
The study periods for these range from 14 days to 20 days with 12 to 15 doses administered
generally for 5 dose levels and a control. Since the quality of this data is good, it is desirable to
consider these study results even though they do not always identically follow the protocol.
Subchronic and Chronic Toxicity
Studies involving subchronic exposure (occurring usually over 3 months) and chronic
exposure (those involving an extended period of time, or a significant fraction of the subject's
lifetime) are designed to permit a determination of no-observed-effect levels (NOEL) and toxic
effects associated with continuous or repeated exposure to a chemical. Subchronic studies provide
information on health hazards likely to arise from repeated exposure over a limited period of time.
They provide information on target organs, the possibilities of accumulation, and, with the
appropriate uncertainty factors, may be used in establishing water quality criteria for human health.
Chronic studies provide information on potential effects following prolonged and repeated exposure.
Such effects might require a long latency period or are cumulative in nature before manifesting
disease. The design and conduct of such tests should allow for detection of general toxic effects
including neurological, physiological, biochemical, and hematological effects and exposure-related
pathological effects.
The following guidelines were derived using the EPA Health Effects Testing Guidelines
(USEPA, 1985), for determining the quality of a subchronic or chronic (long term) study. Additional
detailed guidance may be found in that document. These guidelines represent ideal conditions and
studies will not be expected to meet all standards in order to be considered for use as the basis for
RfD derivation. Ideally, a subchronic/chronic study should include:
• Minimum of 3 dose levels administered and an adequate control group used.
• Minimum of 10 animals for subchronic, 20 animals for chronic studies per sex, per
dose group (both sexes should be used).
• The highest dose level should elicit some signs of toxicity without inducing
excessive lethality and the lowest dose should ideally produce no signs of toxicity.
• Ideal dosing regimes include dosing for 5-7 days per week for 13 weeks or greater
(90 days or greater) for subchronic, and at least 12 months or greater for chronic
studies in rodents. For other species, repeated dosing should ideally occur over 10
B-4
-------
percent or greater of animal's lifespan for subchronic studies and 50 percent or
greater of the animal's lifespan for chronic studies.
• All animals should be dosed by the same method during the entire experimental
period.
• Animals should be observed daily during the treatment period (i.e., 90 days or
greater).
• Animals that die during the study are necropsied and, at the conclusion of the study,
surviving animals are sacrificed and necropsied and appropriate histopathological
examinations carried out.
• Results should be evaluated by an appropriate statistical method selected during
experimental design.
• Such toxicity tests should evaluate the relationship between the dose of the test
substance and the presence, incidence and severity of abnormalities (including
behavioral and clinical abnormalities), gross lesions, identified target organs, body
weight changes, effects on mortality, and any other toxic effects noted in USEPA
(1985).
Developmental Toxicity
Guidelines for reproductive and developmental toxicity studies have been developedby EPA
(USEPA, 1985 and OECD, 1987). Developmental toxicity can be evaluated via a relatively short-
term study in which the compound is administered during the period of organogenesis. Based on
the EPA Health Effects Testing Guidelines (USEPA, 1985), ideal studies should include:
• Minimum of 20 young, adult, pregnant rats, mice, or hamsters or 12 young, adult,
pregnant rabbits recommended per dose group.
• Minimum of 3 dose levels with an adequate control group used.
• The highest dose should induce some slight maternal toxicity but no more than 10
percent mortality. The lowest dose should not produce grossly observable effects in
dams or fetuses. The middle dose level, in an ideal situation, will produce minimal
observable toxic effects.
• Dose period should cover the major period of organogenesis (days 6 to 15 gestation
for rat and mouse, 6 to 14 for hamster, and 6 to 18 for rabbit).
• Dams should be observed daily; weekly food consumption and body weight
measurements should be taken.
B-5
-------
• Necropsy should include both gross and microscopic examination of the dams; the
uterus should be examined so that the number of embryonic or fetal deaths and the
number of viable fetuses can be counted; fetuses should be weighted.
• One-third to one-half of each litter should be prepared and examined for skeletal
anomalies and the remaining animals prepared and examined for soft tissue
anomalies.
As with any other type of study, the appropriate statistical analyses must be performed on
the data for a study to qualify as a good quality study. In addition, developmental studies are unique
in the sense that they yield two potential experimental units for statistical analysis, the litter and the
individual fetus. The EPA testing guidelines do not provide any recommendation on which unit to
use, but the Guidelines for the Developmental Toxicity Risk Assessment (USEPA, 1991) states that
"since the litter is generally considered the experimental unit in most developmental toxicity
studies . . ., the statistical analyses should be designed to analyze the relevant data based on
incidence per litter or on the number of litters with a particular endpoint." Others have also
identified the litter as the preferred experimental unit (Palmer, 1981 and Madson et al, 1982).
Information on maternal toxicity is very important when evaluating developmental effects
because it helps determine if differential susceptibility exists for the offspring and mothers. Since
the conceptus relies on its mother for certain physiological processes, interruption of maternal
homeostasis could result in abnormal prenatal development. Substances which affect prenatal
development without compromising the dam are considered to be a greater developmental hazard
than chemicals which cause developmental effects at maternally toxic doses. Unfortunately,
maternal toxicity information has not been routinely presented in earlier studies and has become a
standard practice in studies only recently. In an attempt to use whatever data are available, maternal
toxicity information may not be required if developmental effects are serious enough to warrant
consideration regardless of the presence of maternal toxicity.
Reproductive Toxicity
The EPA Health Effects Testing Guidelines (USEPA, 1985) include guidelines for both
reproduction and fertility studies and developmental studies. These EPA guidelines can serve as the
ideal experimental situation with which to compare study quality. Studies being evaluated do not
need to match precisely but rather should be similar enough that one can be assured that the chemical
was adequately tested and that the results are a reliable estimate of the true reproductive or
developmental toxicity of the chemical.
These guidelines also recommend a two-generation reproduction study to provide
information on the ability of a chemical to impact gonadal function, conception, parturition and the
growth and development of the offspring. Additional information concerning the effects of a test
compound on neonatal morbidity, mortality, and developmental toxicity may also be provided. The
recommendationsfor reproductive testing are lengthy and quite detailed and may be reviewed further
in the EPA Health Effects Testing Guidelines. In general, the test compound is administered to the
B-6
-------
parental (P) animals (at least 20 males and enough females to yield 20 pregnant females) at least 10
weeks before mating, through the resulting pregnancies and through weaning of their offspring (Fl
or first generation). The compound is then administered to the Fl generation similarly through the
production of the second generation (or F2) offspring until weaning. Recommendations for numbers
of dose groups and dose levels are similar to those reported for developmental studies. Details
should also be provided on mating procedures, standardization of litter sizes (if possible, 4 males and
4 females from each litter are randomly selected), observation, gross necropsy and histopathology.
Full histopathology is recommended on the following organs of all high dose and control P and Fl
animals used in mating: vagina, uterus, testes, epididymides, seminal vesicles, prostate, pituitary
gland, and target organs. Organs of animals from other dose groups should be examined when
pathology has been demonstrated in high dose animals (USEPA, 1985).
References
Interagency Regulatory Liaison Group (IRLG)
Madson, J.M. et al. 1982. Teratology Test Methods for Laboratory Animals. In: Principles and
Methods of Toxicology. Hayes, A.W. (ed). New York: Raven Press.
Organization for Economic Cooperation and Development (OECD), 1987. Guidelines for Testing
of Chemicals. Paris, France.
Palmer, A.K., 1981. Regulatory Requirements for Reproductive Toxicology: Theory and Practice.
In: Developmental Toxicology. Kimmel,C.A.andJ.Buelke-Sam (eds). New York: Raven
Press.
USEPA. 1985. Health Effects Testing Guidelines. 40 CFR Part 798. Federal Register Vol. 50.
September 27.
USEPA. 1991. Final Guidelines for Development Toxicity Risk Assessment. Federal Register
56: 63798-63826. December 5.
B-7
-------
-------
Appendix C
Derivation of Basic Equations Concerning Bioconcentration and Bioaccumulation of
Organic Chemicals
Introduction
Most work dealing with the bioconcentration and bioaccumulation of organic chemicals has
concerned chemicals whose log K,,ws are greater than 3. The purpose of this appendix is to explain
why modifications of the equations generally used with such chemicals are necessary so that the
equations also are appropriate for chemicals whose K^s, BCFs, or BAFs are less than 1,000, and
to derive all of the appropriate equations that are used in the calculation of BAFs for the final
Guidance.
Background
BCFs were originally defined as:
BCFTl =
where:
BCFTl
= Total BCF (i.e., a BCF that is based on the total concentrations of the chemical
in the water and in the aquatic biota)
= Total concentration of the chemical in the aquatic biota, based on the wet weight
of the aquatic biota
= Total concentration of the chemical in the water around the aquatic biota
This is not the nomenclature that was used originally, but it is used here for clarity.
It was subsequently realized that extrapolation of BCFs for organic chemicals from one species
to another would be more accurate if the BCFs were normalized on the basis of the amount of lipid
in the aquatic biota. It was also realized that extrapolation of BCFs for organic chemicals from one
water to another would be more accurate if the BCFs were calculated on the basis of the freely
dissolved concentration of the organic chemical in the water around the aquatic biota. Thus, two
additional BCFs were defined and used:
C
BCF.4 =
C1
Sir
(2)
C-l
-------
BCF,
where:
BCF0l
Q
BCF.
i fd
"W
(3)
Lipid-normalized total BCF (i.e., normalized to 100 percent lipid and based on
the total concentration of the chemical in the water around the biota)
Lipid-normalized concentration of the chemical in the aquatic biota
Lipid-normalized, freely dissolved BCF
t
CIQ = Freely dissolved concentration of chemical in the water around the aquatic
biota
The experimental definition of Cc is:
the total amount of chemical in the aquatic biota
the amount of lipid in the aquatic biota
(f)(B)
(4)
where:
B = Wet weight of the aquatic biota.
L = Weight of the lipid in the aquatic biota.
f{ = Fraction of the aquatic biota that is lipid = L/B
Using Equation 4 to substitute for Q in Equation 2 and then using Equation 1:
BCF/ =
BCFTl
(5)
If frd=the fraction of the chemical in the water around the aquatic biota that is freely dissolved, then:
C-2
-------
fd
(6)
Using Equations 4 and 6 to substitute for Cc and Cd in Equation 3 and then using Equation 1:
BCF/d =
BCFT'
W&
(7)
Equations 1, 5, and 7 show the relationships between the three different BCFs.
Theoretical justification for use of both lipid-normalization and the freely dissolved
concentration of the organic chemical in the ambient water is based on the concept of equilibrium
partitioning, whereas practical justification is provided by the general similarity of the value of
BCF{ra for an organic chemical across both species and waters. It will be demonstrated, however,
that a more complete application of equilibrium partition theory shows that BCF{IQ extrapolates well
only for chemicals whose Kows are greater than 1,000, whereas a different BCF extrapolates well for
organic chemicals whose K^s are greater than 1,000 as well as for chemicals whose Kows are less
than 1,000.
Partition Theory and Bioconcentration
Equilibrium partition theory provides the understandingnecessary to ensure proper use of Kows,
BCFs, and BAFs in the derivation of water quality criteria for organic chemicals. For the purpose
of applying partition theory, aquatic biota can be modeled as consisting of water, lipid, and non-lipid
organic matter (Barber et al., 1991). In this model, an organic chemical in aquatic biota exists in
three forms:
1. Chemical that is freely dissolved in the water that is in the biota.
2. Chemical that is partitioned to the lipid that is in the biota.
3. Chemical that is partitioned to non-lipid organic matter in the biota. The total
concentration of chemical in the water inside the biota includes chemical that is
partitioned to lipid and non-lipid organic matter in the water.
According to this model:
C-3
-------
where:
fw
CL
CN
(8)
= Fraction of the aquatic biota that is water
= Freely dissolved concentration of the organic chemical in the water in the
aquatic biota
= Fraction of the aquatic biota that is lipid
= Concentration of the organic chemical in the lipid
= Fraction of the aquatic biota that is non-lipid organic matter
= Concentration of the organic chemical in the non-lipid organic matter in the
aquatic biota
The most important partitioning of the organic chemical within the aquatic biota is between the lipid
and the water, which is described by the following equation:
c
c
L
ftf
WB
(9)
where:
KLW = the lipid-water partition coefficient.
"KLw" (Gobas 1993) is used herein because it is more descriptive than "KL," which is used by
DiToro et al., (1991). This partition coefficient is central to the equilibrium partition approach that
is used to derive sediment quality criteria (DiToro et al., 1991), the Gobas model that is used to
derive Food-Chain Multipliers for the final Guidance, and the equations given here that are used to
derive BCFs and BAFs for the final Guidance.
In order for Equations 8 and 9 to be correct, partition theory requires that the concentration of
the organic chemical in the lipid, CL, be defined as:
_ _ the amount of chemical partitioned to lipid in aquatic biota
L the amount of lipid in the aquatic biota
It is difficult to determine CL experimentally because it is not easy to measure only the
chemical that is partitioned to the lipid (i.e., it is not easy to separate the three different kinds of
C-4
-------
chemical that, according to the model, exist in aquatic biota). Because all of the organic chemical
in the biota is measured when C, is determined, C, can be determined easily, and Cc is higher than
It is useful to define another BCF as:
BCF fd =
fd
(10)
Because CL is lower than C,, BCF,10 < BCF/°.
Li t
The only difference between KLW and BCFLra is that the denominator in KLW is C^, whereas
the denominator in BCFLra is C^0. When partition theory applies, however, all phases are in
equilibrium and so:
£ fd _ £ fd
(11)
Therefore, when the organic chemical is not metabolized by the aquatic biota and when growth
dilution is negligible:
fd - v
L ~ *xw
Because octanol is a useful surrogate for lipid, a reasonable approximation is that:
K = K
(12)
where:
(13)
K,,w = the octanol-water partition coefficient.
Thus:
predicted BCF_M = Krw - K
LJ J_i W (
(14)
By using Equations 9 and 1 1 to substitute for CL and C in Equation 8:
(fN)(CN)
(15)
C-5
-------
By using Equation 6 to substitute for C^d in Equation 15:
CB = (fwXffdXCw) + (fc)(BCFLfd)(ffd)(C;) + (fN)(CN)
Dividing by C' gives:
C l
'w
fd
= (fw)(ffd) - (fc)(BCFL)(ffd)
(f
w
Using Equation 1 and rearranging gives:
- (fa)[ t, + (f,)(BCF«)
Using Equation 6:
w
(ff)(BCFLfd)
fd
Substituting x =
w
and rearranging gives:
BCFT£ = (ffd)[ x - (f()(BCFLfd)
(16)
(17)
(18)
(19)
(20)
The term " (f()(BCFLfd) " accounts for the amount of organic chemical that is partitioned to the lipid
in the biota, whereas in "x," the term "fw" accounts for the amount of organic chemical that is freely
CN
dissolved in the water in the biota and the term "(£,)( ——)" accounts for the amount of organic
C
chemical that is partitioned to non-lipid organic matter in the biota. The relative magnitudes of these
three terms depend on the following:
C-6
-------
Because of bones and other inorganic matter, the sum of fw + f, + fN must be less than 1.
fw is usually about 0,7 to 0.9.
Because f( must be measured if the BAF or BCF is to be useful, fc is known for the aquatic
biota; it is usually between 0.03 and 0.15.
• The term "(—^-)" is similar to BCFLfd (see Equation 10) and is therefore probably
Cw
related to Kow (see Equation 14), although the affinity of the chemical for non-lipid
organic matter is probably much less than its affinity for lipid.
Although such considerations aid in understanding "x," the magnitude of "x" in Equation 20
is important only for chemicals whose log Kows are in the range of 1 to 3. For organic chemicals
whose log KOWS are about 1, ffd is about 1. In addition, such chemicals distribute themselves so as
to have similar concentrations in water and in the different organic phases in the aquatic biota, which
means that BCFTl will be approximately 1 if both metabolism and growth dilution are negligible.
An organic chemical whose log K<,w is less than 1 will also have a BCFT* on the order of 1 because
water is the predominant component in aquatic biota. Setting "x" equal to 1 is about right in the
range of log K,>ws in which it is not negligible (see also McCarty et al., 1992).
Substituting x = 1 into Equation 20:
BCFT' -
(f{)(BCFLfd)
(21)
Rearranging gives:
BCFT
BCFfd = ( —I
fd
(22)
fd
BCFLW can be called the "baseline BCF" because it is the most useful BCF for extrapolating from
one species to another and from one water to another for organic chemicals with both high and low
KOWS. The baseline BCF is intended to reference bioconcentration of organic chemicals to
partitioning between lipid and water.
Equations 12,13, and 22 demonstrate that both Kow and
C-7
-------
BCF' !
( -JT^ - 1 )(£)
rfd rf
are useful approximations of the baseline BCFs. It will probably be possible to improve both
approximations within a few years, but such improvements might not affect the BCFs substantially
and probably will not require changes in the rest of the equations or the terminology.
When BCF^ is greater than 1,000, the "-1" in Equation 22 is negligible and so this equation
becomes equivalent to Equation 7 (i.e., when BCFTl is large, BCF{fd is a useful approximation of
the baseline BCF).
Bioaccumulation
By analogy with Equations 21 and 22:
BAFT< = (ffd)[ 1 + (ff)(BAFLfd) ]
(23)
BAFLfd =
BAF
fd
(24)
BAFLfd can be called the "baseline BAF" because it is the most useful BAF for extrapolating from
one species to another and from one water to another for chemicals with both high and low Kows.
It is convenient to define a food-chain multiplier (FCM) as:
baseline BAF
FCM =
baseline BCF BCF
fd
(25)
Some of the consequences of Equation 25 are:
1. Substituting Equations 22 and 24 into Equation 25:
FCM =
BAFJ - f.
fd
BCF; - ffd
(26)
C-8
-------
2.
3.
Therefore, BAFT' - (FCM)(BCFTl) only when ffd is much less than BAFT' and BCF^.
When FCM = 1 (as for trophic level 2 in the Gobas model):
baseline BAF = baseline BGF
(27)
Predicted baseline BAFs can be obtained using FCMs and the following rearrangement of
Equation 25:
predicted baseline BAF = (FCM)(baseline BCF)
a. Using a laboratory-measured BCF in Equation 22:
predicted baseline BAF = (FCM)(measured BCFLfd)
= (FCM)(
BCFT'
fd
(28)
(29)
(30)
b. Using a predicted BCF in Equation 14:
predicted baseline BAF = (FCM)( predicted BCFtfd)
l_i
= (FCM)(KQW)
(31)
(32)
The FCMs used to calculate predicted baseline BAFs must be appropriate for the trophic level
of the aquatic biota for which the predicted baseline BAF is intended to apply.
Although BAFs can be related to BCFs using FCMs, BAFs, and BCFs can also be related using
Biomagnification Factors (BMFs). The two systems are entirely compatible, but confusion can
result if the terms are not used consistently and clearly. Because both systems are used in the final
Guidance and elsewhere, it is appropriate to explain the relation between the two here. The basic
difference is that FCMs always relate back to trophic level one, whereas BMFs always relate back
to the next trophic level. In the FCM system:
BAF
TLl
= BCF
= (FCM112)(BAFTL1)
= (FCM1L3)(BAF1U)
C-9
-------
In the BMP system:
= (FCM11.4)(BAFTLI)
= BCF
BAF^ =(BMF1L3)(BAFn,2)
BAP™ '
Therefore:
BMF-n.2 =FCMTL2
=(FCMTL3)/(FCMTL2)
Both metabolism and growth dilution can cause BMFs to be less than 1 .
Calculation of Criteria
Baseline BCFs and BAFs can be extrapolated between species and waters, but they cannot be
used directly in the calculation of criteria that are based on the total concentration of the chemical
in the water. The BCFs and BAFs that are needed to calculate such criteria can be calculated from
measured and predicted baseline BCFs and BAFs using the following equations, which are derived
from Equations 21 and 23:
BCF< = [ 1 + (baseline BCF)(ff) ](ff.)
BAFT< = [ 1 + (baseline BAF)(f() ](ffd)
(33)
(34)
References
Barber, M.C., L.A. Suarez, and R.R. Lassiter. 1991. Modeling Bioaccumulation of Organic
Pollutants in Fish with an Application to PCBs in Lake Ontario Salmonids. Can. J. Fish.
Aquat. Sci. 48:318-337.
C-10
-------
DiToro, D.M., C.S. Zarba, D.J. Hansen, W.J. Berry, R.C. Swartz, C.E. Cowan, S.P. Pavlou, H.E.
Allen, N.A. Thomas, and P.R. Paquin. 1991. Technical Basis for Establishing Sediment
Quality Criteria for Nonionic Organic Chemicals Using Equilibrium Partitioning. Environ.
Toxicol. Chem. 10:1541-1583.
Gobas, F.A.P.C. 1993. A Model for Predicting the Bioaccumulation of Hydrophobic Organic
Chemicals in Aquatic Food-Webs: Application to Lake Ontario. EcologicalModeling 69:1 -17.
McCarty, L.S., D. Mackay, A.D. Smith, G.W. Ozburn, and D.G. Dixon. 1992. Residue-Based
Interpretation of Toxicity and Bioconcentration QSARs from Aquatic Bioassays: Neutral
Narcotic Organics. Environ. Toxicol. Chem. 11:917-930.
C-ll
-------
-------
Appendix D
Derivation of the Equation Defining ffd
Experimental investigations have shown that hydrophobic organic chemicals exist in water in
three phases, (1) the freely dissolved phase; (2) sorbed to suspended solids (particulate organic
carbon); and (3) sorbed to dissolved organic matter (Hassett and Anderson, 1979; Carter and Suffet,
1982; Landrum et al., 1984; Gschwend and Wu, 1985; McCarthy and Jimenez, 1985; Eadie et al.,
1990,1992). The total concentration of the chemical in water is the sum of the concentrations of the
sorbed chemical and the freely dissolved chemical (Gschwend and Wu, 1985; Cook et al., 1993):
= Cfd + POC • C + DOC • C_,
w poc doc
(1)
where:
Cld =
C,
'poc
•-doc
POC =
Concentration of freely dissolved chemical in the ambient water (kg of
chemical/L of water)
Total concentration of the chemical in the ambient water (kg of chemical/L
of water)
Concentration of chemical sorbed to the particulate organic carbon in the
ambient water (kg of chemical/kg of organic carbon)
Concentration of chemical sorbed to the dissolved organic carbon in the
water (kg of chemical/kg of organic carbon)
Concentration of particulate organic carbon in the ambient water (kg of
organic carbon/L of water)
DOC = Concentration of dissolved organic carbon in the ambient water (kg of
organic carbon/L of water)
The above equation can also be expressed using partitioning relationships as:
K + DOC • K, J
poc doc'
(2)
where:
D-l
-------
K.
•poc
K,
•poc
*-doc
and
equilibrium partition coefficient of the chemical between POC and the freely
dissolved phase in the ambient water
equilibrium partition coefficient of the chemical between DOC and the freely
dissolved phase in the ambient water
From Equation 2, the fraction of the chemical which is freely dissolved in the water can be
calculated using the following equations:
c
w
(3)
POC • K
poc
DOC
(4)
Experimental investigations by Eadie et al. (1990,1992), Landrum et al. (1984), Yin and
Hassett (1986, 1989), Chin and Gschwend (1992), and Herbert et al. (1993) have shown that KdOC
is directly proportional to the K^ of the chemical and is less than the Kow. The KAoc can be estimated
using the following equation:
K
K, ~
doc
10
(5)
The above equation is based upon the results of Yin and Hassett (1986, 1989), Chin and
Gschwend (1992), and Herbert et al. (1993). These investigations were done using unbiased
methods, such as the dynamic headspace gas-partitioning (sparging) and the fluorescence methods,
for determining the
Experimental investigations by Eadie at al. (1990, 1992) and Dean et al. (1993) have shown
that Kpoc is approximately equal to the K^ of the chemical. The K^ can be estimated using the
following equation:
D-2
-------
K * K
poc o
(6)
By substituting Equations 5 and 6 into Equation 4, the following equation is obtained:
DOC • K
(1 + POC • K + (
^ cm v JO
ow>
(7)
The utility in using the freely dissolved equation described above to derive baseline B AFs
applicable to multiple sites has been evaluated recently in a study conducted by Burkhard et al.
(1997). In their study, Burkhard et al. measured BAFs for various chlorinated butadienes,
chlorinated benzenes and hexachloroethane for three species of forage fish and blue crab in Bayou
d'Inde of the Calcasieu River system, Louisiana. Using the freely dissolved equation, Burkhard et
al. adjusted their field-measured BAFs to baseline BAFs (BAFf) and compared these to baseline
BAFs determined for other trophic level three species in two other field studies (Pereria et al., 1988;
Oliver and Niimi, 1988). The field study by Pereria et al. (1988) was conducted in different sites
within the Calcasieu River system and that of Oliver and Niimi (1988) in Lake Ontario. Burkhard
et al. found no significant difference between B AFf determined in their study and those determined
by Pereria et al. (1988) (Tukey's, a = .05). However, for one chemical (HCBD) about an order of
magnitude difference was observed in the measured BAFf between the two studies. Burkhard et al.
further noted their baseline BAFs were not substantially different than those derived for Lake
Ontario, suggesting broader applicability of properly derived baseline BAFs.
References
Carter, C.W. and I.H. Suffet. 1982. Binding of DDT to Dissolved Humic Materials. Environ.
Sci. Technol. 16:735-740.
Chin, Y., and P.M. Gschwend. 1992. Partitioning of Polycyclic Aromatic Hydrocarbons to
Marine Porewater Organic Colloids. Environ. Sci. Technol. 26:1621-1626.
Cook, P.M., R.J. Erickson. R.L. Spehar, S.P. Bradbury and G.T. Ankley. 1993. Interim Report on
Data and Methods for Assessment of 2,3,7,8-tetrachlorodibenzo-p-dioxin Risks to Aquatic
Life and Associated Wildlife. Duluth, MN: USEPA, Environmental Research Laboratory.
EPA/600/R-93/055.
D-3
-------
Dean, K.E., M.M. Shafer, and D.E. Armstrong. 1993. Particle-Mediated Transport and Fate
of a Hydrophobia Organic Contaminant in Southern Lake Michigan: the Role of Major
Water Column Particle Species. J. Great Lakes Res. 19: 480-496.
Eadie, B.J., N.R. Morehead and P.P. Landrum. 1990. Three-Phase Partitioning of Hydrophobic
Organic Compounds in Great Lake Waters. Chemosphere. 20:161-178.
Eadie, B.J., N.R. Morehead, J. Val Klump and P.P. Landrum. 1992. Distribution of Hydrophobic
Organic Compounds between Dissolved and Particulate Organic Matter in Green Bay
Waters. J. Great Lakes Res. 18:91-97.
Gschwend. P.M. and S. Wu. 1985. On the Constancy of Sediment-Water Partition Coefficients
of Hydrophobic Organic Pollutants. Environ. Sci. Technol. 19: 90-96.
Hassett, J.P. and M.A. Anderson. 1979. Association of Hydrophobic Organic Compounds with
Dissolved Organic Matter in Aquatic Systems. Environ. Sci. Technol. 13: 1526-1529.
Herbert, B.E., P.M. Bertsch and J.M. Novak. 1993. Pyrene Sorption by Water-Soluble Organic
Carbon. Environ. Sci. Technol. 27:398-403.
Landrum, P.P., S.R. Nihart, BJ. Eadie and W.S. Gardner. 1984. Reverse-Phase Separation
Method for Determining Pollutant Binding to Aldrich Humic and Dissolved Organic Carbon
of Natural Waters. Environ. Sci. Technol. 18:187-192.
McCarthy, J.F. and B.D. Jimenez. 1985. Interaction Between Polycyclic Aromatic Hydrocarbons
and Dissolved Humic Material: Binding and Dissociation. Environ. Sci. Technol. 19:1072-
1076.
Yin, C. and J.P. Hassett. 1986. Gas-Partitioning Approach for Laboratory and Field Studies of
Mirex Fugacity in Water. Environ. Sci. Technol. 20:1213-1217.
Yin, C. and J.P. Hassett. 1989. Fugacity and Phase Distribution of Mirex in Oswego River and
Lake Ontario Waters. Chemosphere. 19: 1289-1296.
D-4
-------
Appendix E
Derivation of the Equation to Predict BAF from the BSAF
Several steps are involved in the derivation of the equation to predict the BAF for a chemical
from the BSAF. First, in the basic equation for BAF for a given chemical, BSAF and Csoc can be
substituted for Q for a given chemical / as follows:
(BAF*), = (BSAF).
1
(Cfd).
^ w 'i
The chemical concentration quotient between sediment organic carbon and a freely dissolved state
in overlying water may be symbolized by nsoc, as follows:
(2)
Thus the ratio of BAFfs for chemical i and a reference chemical r may be expressed as:
(BAF{\
(3)
If both chemicals have similar fugacity ratios between water and sediment, the following assumption
can be made:
(n J
v soerr
(4)
therefore:
•(BAFf), = (BAF{\
(BSAF)r(KJf
(5)
The assumption of equal or similar fugacity ratios between water and sediment for each chemical
is equivalent to assuming that for all chemicals used in BAFcfd calculations: (1) the concentration
E-l
-------
ratios between sediment and suspended solids in the water, and (2) the degree of equilibrium
between suspended solids and C ™ are the same. Thus, errors could be introduced by inclusion of
chemicals with non-steady-state external loading rates or chemicals with strongly reduced C ™ due
to rapid volatilization from the water.
E-2
-------
Appendix F
EPA New Draft Protocol for Determining Octanol-Water Partition Coefficients (K<,w)
For Compounds with Log KOW Values > 5
1.
Introduction
The octanol- water partition coefficient (K^) is one of the most widely used chemical
parameters. The K^ of a chemical has been found to be representative of a chemical's
propensity to partition into biotic and abiotic components of the environment as well as a
chemical's propensity to accumulate hi living organisms. Because of these associations, the Kow
is widely used to predict a chemical's behavior in the environment and to evaluate a chemical's
impact on human health.
The octanol- water partition coefficient (K^) is a unitless measure and is defined as the
ratio of the equilibrium concentrations, C, of a chemical in the two phases of a system consisting
of n -octanol and water at standard temperature and pressure (STP, 25° C, 1 atm):
where Coct represents the concentration in the n -octanol phase, and Cw represents the
concentration in the water. The concentrations in the respective phases are expressed in the same
volume-referenced units (i.e., mg/ml, mole/L, etc.), therefore, the Kow is a unitless property.
Since the value of the partition coefficient spans orders of magnitude, it is frequently expressed
on a log scale (base ten) such that a given chemical has a log K^ value which may range from 1
to >8. This parameter is also called the log P value.
Some specific applications of the K,,w within the U.S. EPA include: evaluation of a
chemical's potential to bioaccumulate in aquatic life, wildlife and humans; modeling the fate,
transport and distribution of a chemical in the environment; prediction of the distribution of a
contaminant in a living organism; classification of persistent bioaccumulators for regulatory
actions; derivation of soil screening levels; calculation of water quality benchmarks; and
derivation of Sediment Quality Criteria.
Although a seemingly simple experimental determination, Kow measurement is beset with
difficulties. The appropriateness and accuracy of laboratory methods to directly measure a Kow
are influenced by a number of factors which include the magnitude of the value itself. For
chemicals with log K^ values at or exceeding 5, common sources of measurement error include:
(1) failure to achieve equilibrium; (2) incomplete phase separation or interphase mixing during
sampling; (3) emulsion effects derived from "excessive" mixing or induced by contaminants; (4)
propensity of the chemical to self-associate, tautomerize or form hydrates; and (5) the presence
of small quantities of contaminants with a lower K,,w value. These errors tend not to be random,
but to give measured numbers lower than the true value, frequently by an order of magnitude or
more. The likelihood and degree of error increases with increasing Kow and also seems to be
F-l
-------
more prevalent for certain classes of chemicals (such as halogenated compounds or phthalate
esters). As a result, in addition to direct experimental measurement methods, techniques to
indirectly experimentally measure or estimate K^ values have been developed.
Direct experimental measurement techniques include the shake-flask approach, generator
column, and slow-stir methods. The shake-flask method is the classical approach and fairly
straight-forward for chemicals with log K^, values below 5. For chemicals with higher log K^
values, the shake-flask approach requires large volumes of water and formation of emulsions
becomes a significant impediment to accurate measurements. The generator-column approach
was developed to measure the partition coefficients of more hydrophobic chemicals (those with
larger log K,,w values). This is a laborious method which results in more reliable data than the
shake-flask approach for chemicals with higher log K,,w values, but some discontinuities in the
data for higher-chlorinated PCB congeners have been observed. A third direct measurement
technique is the slow-stir method. In this method, careful stirring and close temperature control
can prevent or limit the formation of emulsions and reliable very high partition coefficients can
be obtained relatively easily.
Because of the difficulty of directly and accurately measuring Kow; values, estimation
methods have been developed. These methods can be divided into two types: those requiring a
training set of chemicals with measured Kows and those based upon fundamental chemical
thermodynamics. Those methods requiring a training set of chemicals use Quantitative Structure
Property Relationships (QSPRs) or Quantitative Structure Activity Relationships (QSARs) to
derive K^. In QSPRs, K^ values are correlated with the values for other chemical parameters—
either measured or calculated—using data available from the training set of chemicals. In
QSARs, KOW values are derived from fragment constants obtained from the training set of
chemicals.
One application of QSPRs is estimating K^s indirectly from other experimental
measurements. In this approach, the Kow is correlated with another measured property. These
techniques include the use of reversed-phase high performance liquid chromatography (HPLC)
and reversed-phase thin-layer chromatography (TLC). In applying these approaches, Kows are
estimated from linear equations relating retention times on the reversed-phase column to the K,,w
values. The equations are developed based on a set of reference chemicals for which KOW values
are well established. These are relatively efficient methods because they do not require
quantification of concentrations, but the linear equations can not be extrapolated beyond the Kow
range represented by the reference chemicals from which the equation was derived. In
application, values for the reference chemicals are usually shake-flask values obtained from the
literature, resulting in unreliable K,,w estimates for chemicals with higher log K,,w values.
In addition to direct and indirect measurement methods, QSPRs are also used to establish
correlations between the K^ and calculated properties. For example, Hawker and Connell
(1988) developed a correlative relationship between log K^ and molecular surface area using
approximately two dozen PCBs. They then estimated log K^s for the remaining PCBs by
F-2
-------
inputting the molecular surface area of each PCB. This technique is limited to estimating Kows
for chemicals which are similar to the chemicals used in developing the relationship.
In QSARs, hydrophobic fragment values are derived from a large data base of measured
K,,ws. These fragment constants are used to estimate KOW in two ways. One approach is to
estimate the K,jW by adding up the values for all the fragments composing the chemical, either by
atom or by functional group. The other approach is to start with a measured K^ value for a
structurally similar compound and add or subtract the fragment constants for functional groups or
atoms to estimate the KOW for the specific compound. In both these cases, the calculated Kow
value must also be corrected for proximity effects between structurally close substituent groups,
and the K,,w value derived is only as good as the data associated with the training set of
chemicals. This method is also limited to predicting K^s for chemicals with structures similar to
those within the training set. Computer-based models exist which apply QSAR approaches to
estimate K,jWs. CLOGP1 and LOGKOW2 data bases are both applications of this approach.
Other computer methods are based on fundamental chemical structure theory and are not
limited by nor require a training set of chemicals with measured K^s. For example, the SPARC3
model consists of a set of core models describing intra- and inter-molecular interactions. These
models are linked by appropriate thermodynamic relationships to provide estimates of reactivity
parameters under desired conditions (e.g., temperature, pressure, solvent).
Given the numerous techniques available to determine the Kow and its numerous and
important applications across the Agency, the U.S. EPA has formed an Agency Kow Work Group
to draft the following guidance for selecting reliable values of K^ and ultimately for developing
a data base of reliable K^, values. In determining these recommended KOW values, the preferable
option would be to recommend actual measured values. For chemicals with log Kow values
below 5, the classical shake-flask approach is adequate to obtain these measurements. Although
recent advances in measurement technology (development of the slow stir method and increased
'CLOGP is a molecular fragment-based model developed at Pomona College by Albert Loe, Corwin
Hanch, and co-workers. This model has undergone extensive development and exhaustive testing. (See Hansch
and Leo, 1995, for model description and performance data.)
2LOGKOW is essentially an expanded CLOGP with more recent training data and additional fragment
constants. The developers were Philip Howard, William Meylan and co-workers at Syracuse Research Corporation.
(See Meylan and Howard, 1994, for model details and performance information.)
3SPARC (SPARC Performs Automated Reasoning in Chemistry) is a mechanistic model developed at
the Ecosystems Research Division of the National Exposure Research Laboratory of the Office of Research and
Development of the U.S. Environmental Protection Agency by Sam Karickhoff, Lionel Carreira, and co-workers.
A prototype version was used for which no performance data for KQW estimation is available. The model
complements the aforementioned models because development, training, and testing were done away from
data. (See Hilal, Carreira, and Karickhoff, 1994, for model description.)
F-3
-------
awareness of, and compensation for, determinate errors) have significantly improved the quality
of data available for chemicals with higher log KOW values, there remains a serious shortage of
reliable measured data for compounds with higher log K«,w values (log K^, > 5). Unfortunately, it
is frequently these chemicals that exhibit a propensity to accumulate in living tissues or bind to
soils and sediments.
2. Protocol For Determining Recommended Kow Values
Measured data are preferable for determining recommended K^ values. However, the
absence or scarcity of reliable data necessitates the use of estimation methods in evaluating data
and in assigning K^s. K^ estimates used in this exercise include: (1) QSARs (e.g., CLOGP,
LOGKOW, and fragment additions or subtractions); (2) QSPRs (e.g., HPLC and TLC methods)
and (3) estimation methods based on fundamental chemical structure theory (e.g. SPARC). All
of these approaches except the last one listed (the SPARC model) require measured K^ values
for a training set of chemicals.
Assigning a K,,w from these data will necessarily involve scientific judgement in
evaluating not only the reliability of all data inputs but also the accretion/concretion of evidence
in support of the recommended Kow value. Supporting rationale will be provided for each
recommended value.
2.1 Operational Guidelines for Kow Selection Protocol
For chemicals with log Kow > 5, it is highly unlikely to find multiple "high quality"
measurements. (Note: "high quality" is data judged to be reliable based on the guidelines
presented in Section 3)
• "High quality" measured data are preferred over estimates, but due to the scarcity of
"high quality" data, the use of estimates is important in assigning
KOW measurements by slow stir are extendable to 108. Shake flask K^, measurements are
extendable to 106 with sufficient attention to micro emulsion effects; for classes of
chemicals that are not highly sensitive to emulsion effects (i.e., PNAs) this range may
extend to 106S.
What is to be considered reasonable agreement in log Kow data (measured or estimated)
depends primarily on the log KOW magnitude. The following standards for data agreement
have been set for this guidance: 0.5 for log KQW > 7; 0.4 for 6 < log K^ <. 7 ; 0.3 for log
K™ < 6.
Statistical methods should be applied to data as appropriate but application is limited due
to the scarcity of data, and the determinate/methodic nature of most measurement
error(s).
F-4
-------
2.2 Tiered Procedure for Setting KQW Values
I. Assemble/evaluate experimental and calculated data (e.g.,CLOGP, LOGKOW,
SPARC)
II. If calculated log Kows> 8,
A. Develop independent estimates of K^ using:
1 . Liquid Chromatography (LC) methods with "appropriate" standards. (See
Section 3 for guidelines for LC application.)
2. Structure Activity Relationship (SAR) estimates extrapolated from similar
chemicals where "high quality" measurements are available. "High quality"
S ARs are described in Section 3 .
3. Property Reactivity Correlation (PRC) estimates based on other measured
properties (solubility, etc.)
B. If calculated data are in "reasonable" agreement and are supported by independent
estimates described above, report average calculated value. What is to be
considered reasonable agreement in log Kow data (measured or estimated) depends
primarily on the log KOW magnitude. The following standards for data agreement
have been set for this guidance: 0.5 for log K^ > 7; 0.4 for 6 z log Kow < 7 ; 0.3
forlogKo
ow
C. If calculated/estimated data do not agree, use professional judgement to
evaluate/blend/weight calculated and estimated data to assign a K,,w value.
D. . Document rationale including relevant statistics.
III. If calculated log Kows range from 6-8,
A. Look for "high quality" measurements. These will generally be slow stir
measurements, the exception being certain classes of compounds where micro
emulsions tend to be less of a problem (i.e., PNAs, shake flask measurements are
good to 6.5).
B. If measured data are available and are in reasonable agreement (both
measurements and calculations), report average measured value.
C. If measured data are in reasonable agreement, but differ from calculated values,
develop independent estimates and apply professional judgement to
evaluate/blend/weight measured, calculated and estimated data to assign Kow
value.
F-5
-------
D. If measured data are not in reasonable agreement (or if only one measurement is
available), use II A, B, and C to produce a 'best estimate;' use this value to
evaluate/screen measured data; report average value of screened data. If no
measurements reasonably agree with 'best estimate,' apply professional
judgement to evaluate/blend/weight measured, calculated and estimated data to
assign K^.
E. If measured data are unavailable, proceed through II A, B, C and report the 'best
estimate.'
F. Document rationale including relevant statistics.
IV. If calculated log K^s < 6,
A. Proceed as in III. Slow stir is the preferred method but shake flask data can be
considered for all chemicals if sufficient attention has been given to emulsion
problems in the measurement.
3. Guidelines for Evaluating Measured and Liquid Chromatography-estimated KOW$
3.1 Assessment of Measured K,,w Values
3.1.1 Molecular Speciation. In order to interpret measured data, it is necessary to understand
the molecular species present in both the octanol and water phases including ionization,
self-association, tautomerization, and hydrate formation. For these reasons, it is difficult
to conduct or interpret such measurements for mixtures of unknown composition or for
single molecules of unknown structure. Solutes composed of more than one molecular
species may also show substantial temperature dependence of Kow reflecting relative
change in Speciation in the octanol and water phases.
• Ionization. For weakly ionizable molecules, shake flask measurements are
conducted in solutions of a stable, non-extractable buffer to suppress ionization.
• Self-Association - For molecules able to associate through hydrogen-bonding
(e.g., amines, carboxylic acids, phenols, especially if cyclic dimers can form),
measurement needs to be conducted at a sufficiently low concentration that Kow
reflects only unassociated form of the molecule in both water and octanol phases.
Measurements at several concentrations with no change in Kow provide an
indication that this is the case.
• Tautomerization - If the molecule is likely to exist in more than one tautomeric
form, the ratio of tautomers may be different in the octanol and water phases. If
that is the case, the measured KOW may not be a very meaningful number.
F-6
-------
3.1.2
3.1.3
3.1.4
Sometimes molecules exhibit both ionization and tautomerization, leading to
further complications.
Hydrate Formation - Similar to the case of tautomerization, hydrates may exist to
different degrees in the water and octanol phases thus confounding the
interpretation of the measured value. A comparison of estimated and measured
KOW for chloral hydrate suggests that such hydration may be much less important
in the octanol than water phases, making the compound more lipophilic than
would be expected from the hydrate structure alone.
Shake Flask or Slow-Stirring Considerations. (1) Water and octanol phases should be free
of impurities; (2) mixing should be of sufficient duration (e.g., 7 days for dioctyl
phthalate) to reach steady state equilibrium, particularly for very hydrophobic chemicals;
(3) analytical measurement of both phases is particularly important when using volatile
solutes; (4) Avoid formation of emulsions during mixing and centrifuge before
measuring; (5) experimental protocol should be particularly scrutinized for K,,w
measurements 4-6; (6) ratio of octanol to water should be reduced for high K^ chemicals;
and (7) sorption to glass (e.g., for pyrethroids) during workup can be a problem.
General Considerations. Solute should be stable to hydrolysis during the course of the
experiment. Solutes should be of high purity as the presence of a less lipophilic impurity
exerts a dominant effect in the measured Kow value. Mixtures such as chlorinated
paraffins (containing thousands of isomers, congeners, and degrees of chlorination)
therefore cannot be determined except by chromatographic methods.
Indicators of Potential Concern. Inconsistency with other measured values, with
estimated value, or inconsistency among estimated values. The importance of
professional judgement and knowledge of chemistry cannot be overemphasized in
making the best KOW assignments. For example, inconsistency between measured and
predicted may reflect only problems in the training set used based upon poor
experimental values when better data have since become available.
3.2 Assessment of KOW Values Estimated using Liquid Chromatographic Techniques
An estimated K<,w value would be considered "appropriate" provided the following
experimental conditions existed during its determination:
3 .2. 1 K,,ws used for the reference compounds consist of "high quality"slow stir measurements.
Better estimates for K^s are obtained when reference and test chemicals are
similar.
F-7
-------
3.2.2 A minimum of five chemicals are used in developing the log capacity factor (k1)- log K^
calibration relationship. The K^s of the reference chemicals should be evenly distributed
and should span 3 to 4 orders of magnitude.
3.2.3 The log k' - log K™ calibration curve is linear and has a correlation coefficient greater
than 0.95.
3.2.4 The KOW estimated for the test chemical is within the range of Kows for the reference
compounds or does not exceed the upper end of the range of K^s for the reference
compounds beyond 0.5 log units without adequate justification.
3.2.5 Chemical speciation must be accounted for in performing the measurements. For
example, with ionizable chemicals, measurements must be performed on the unionized
form by using an appropriate buffer with a pH below the pK for an acid and above the pK
for a base.
3.2.6 Reference and test chemicals are of known purity and structure. Independent
confirmation of the identity and purity of the reference and test chemicals is required or
highly desirable.
3.2.7 Chemical mixtures can be used as the source of test chemicals provided accurate
identities can be assigned to individual chromatographic components.
4. Estimation of KQW from Molecular Fragments
For computing thermodynamic properties it is often useful to consider a molecule as a
collection of molecular fragments, each making a distinct contribution to the property of interest,
which is relatively independent of the rest of the molecule. The rationale behind the method is
that a large number of structures can be generated from a relatively small number of fragments,
and thus a large number of estimates can be derived from a small number of experimentally
determined fragment constants. The accuracy of the estimation, however, necessarily improves
as the specificity of the fragment environment increases, which entails an increase in the number
of fragments or corrective factors that must be considered. This approach is applied at different
levels of sophistication. One user may employ a few fragment constants and generate 'first order1
estimates whereas another may make numerous corrections or adjustments reflecting more
fragment specificity for a given molecular environment. For a more complete discussion of group
fragment methods one should consult Hansch and Leo, 1995.
4.1 Addition of Ring Fragments
For condensed ring aromatics, the addition of rings is given by
ft, = log K(QW) (pyrene) log K(QW) (phenanthrene) = 0.50
F-8
-------
iy = log K(ow) (anthracene) log K(QW) (naphthalene) «= 1.20
/(c3H) = °'5 [lQg K(ow)
log K(QW) (naphthalene) « 0.85
where /£ H \ > /L H\ , /I \ are the fragment addition constants for a, p, and y condensation
respectively.
4.2 Addition of Substituents
The addition of a substituent, S (replacing a H atom) is a primary application of this
method. In this case
(R
(R
where R is the base molecule and Hs is a substituent constant, which is experimentally
determined. Tables for common substituents are readily available or can be easily determined
from measured data. One must distinguish (i.e., have different substituent constants for)
attachment to aliphatic, ethylenic, acetylenic, and aromatic carbon atoms in 'R'. Also corrections
must be made for multiple substitution if attachment is to the same or adjacent carbons. The
following are examples of Kow estimation. The fragment constant for Cl attached to aromatic
carbon can be derived from:
= log KQW (chlorobenzene) log K (benzene) » 0.71
With this constant, one can derive
log K(QW) (1,3,5 -trichlorobenzene) * log
+ 3 (0.71) = 4.26
An exhaustive list of substituent constants is included in the aforementioned Hansch and Leo
(1995) reference.
F-9
-------
5. Example Application of EPA Draft K^ Selection Protocol
BENZO(A)ANTHRACENE
CAS
56-55-3
logKow
5.79
5.61
5.79
4.00
5.00
5.66
5.83
5.52
Method
shake flask
shake flask
RP-HPLC
RP-HPLC-E
RP-HPLC
CLOGP
SPARC
LOGKOW
Reference/Comments
Medchem
Steen & Karickhoff (1979)
Wang etal (1986)
Brooke etal( 1986)
Brooke etal (1986)
Calculated values < 6, therefore enter Step IV of protocol.
Go to step III of protocol with inclusion of shake flask data.
Ill A. Two shake-flask measurements in good agreement with one another and avg
= 5.70
B. Calculators: SPARC, LOGKOW and CLOGP in good agreement with range
from 5.52-5.83 and avg = 5.67.
Shake-flask measurements in excellent agreement with calculators, therefore recommended value
is 5.7Q—average of the two shake-flask measurements.
F-10
-------
BENZO(K)FLUORANTHENE
CAS
207-08-9
log KOW
6.12
6.30
6.11
Method
CLOGP
SPARC
LOGKOW
Reference/Comments
Calculated values in range 6-8, therefore enter Step III of protocol.
Ill A. No measured values available.
E. Calculators: SPARC, LOGKOW and CLOGP in excellent agreement with
range from 6.11-6.30 and avg = 6.18.
• Estimate Kow from Molecular Fragment Constants:
Fluoranthene (5.12) + fa(C4H2) (1.20) = 6.32 This is in good agreement with
calculated values.
Recommended value is average of three calculators = 6.18.
BENZO(A)PYRENE
CAS
50-32-8
lOg KOW
5.98
6.34
6.04
6.00
5.99
6.42
6.24
6.04
5.97
6.12
6.25
6.11
Method
gen. col.
shake flask
shake flask
shake flask
shake flask
RP-HPLC
RP-HPLC
RP-HPLC
unknown
CLOGP
SPARC
LOGKOW
Reference/Comments
Miller etal( 1985)
Steen & Karickhoff (1979)
Medchem
Mallon & Harrison (1984)
Mallon & Harrison (1984)
Rappaport & Eisenreich (1984)
Hanal et al(1981) (60% solvent values
Wang etal( 1986)
Hansch & Leo
Calculated values in range 6-8, therefore enter Step III of protocol.
Ill A. There are no slow-stir measurements.
There are four shake flask measurements in range 5.99-6.34 with an average
-6.09
B. Calculators: SPARC, LOGKOW and CLOGP in excellent agreement with
range from 6.11 -6.25 and average = 6.16 .
• Estimate Kow from Molecular Fragment Constants:
pyrene (5.05) +f S(C4H2) (1.20) = 6.25 This is in good agreement
with calculated and measured values.
Three RP-HPLC values (range 6.04 - 6.42) with average 6.26.
This compound is on the border of whether shake flask data acceptable but PNAs are less susceptible
to emulsification using the shake flask approach than other compounds. Therefore shake flask
measurements may be acceptable up to 6.5 rather than 6.0. Therefore, we will assume the shake flask
F-ll
-------
measurements for this compound are accurate and recommend going with the average of the four
shake flask measurements = 6.09
DI-N-OCTYL PHTHALATE
CAS
117-84-0
logKow
8.06
9.49
8.39
8.54
Method
RP-HPLC
CLOGP
SPARC
LOGKOW
Reference/Comments
McDuffie(1981)
Calculated values are >8.0 therefore enter step II of protocol.
II A. No published measured values were available, J. Ellington (EPA, Athens)
measured 8.1 by slow-stir.
D. Calculators: Not in good agreement: SPARC, 8.39, CLOGP, 9.49, &
LOGKOW, 8.54.
SPARC and LOGKOW calculators are close to the Ellington value,
• Can use branching correction and estimate from diethylhexyl phthalate ( 2
slow stir measurements, average 7.3) . A secondary branch contributes
approximately '-0.3 'per which would place the unbranched dioctyl phthalate
at 7.9 (7.3 + 2(0.3)); this is close to the Ellington value.
The RP-HPLC estimate is 8.06.
Recommended value is 8.1 (Ellington, unpublished), supported by two of the calculators and two
other estimates.
F-12
-------
CAS log Kow
129-00-0 5.07
5.18
5.18
5.09
5.09
5.08
5.05
4.88
4.93
4.89
4.76
5.52
5.05
4.97
4.96
5.08
4.89
4.95
5.02
4.93
PYRENE
Method
slow stir
gen. col.
shake flask
shake flask
shake flask
shake flask
shake flask
shake flask
RP-HPLC-E
RP-HPLC-E
RP-HPLC-E
RP-HPLC
RP-HPLC
RP-HPLC
RP-HPLC
RP-HPLC
RP-HPLC
CLOGP
SPARC
LOGKOW
Reference/Comments
Stancil(1994)
Miller etal( 1985)
Karickhoffetal(1979)
Means et al (1980); Hassett et al (1980)
Wang etal( 1986)
Medchem
Ellington & Stand! (1988)
Hansch & Leo (1979); Medchem
Hammers etal( 1982)
Tomlinson & Hafkenschield (1986)
Hafkenschield & Tomlinson (1983)
Rurkhardetal(1985)
McDuffle(1981)
Chin etal (1986)
Rapaport & Eisenreich (1984)
Wang etal (1986)
Hanal et al (1981) (50% acetonitrile)
Calculated values < 6, therefore enter Step IV of protocol.
Go to step III of protocol which permits consideration of shake flask data.
Ill A. Seven of eight measured values are in good agreement (range = 4.88 to 5.18);
shake flask, slow stir, generator column avg 5.08.
B. Calculators: All three calculators are in excellent agreement with avg = 4.94,
in agreement with the measured value.
Recommended value is average of slow stir, shake flask and generator
column values = 5:08
6.
References
Hansch, C. and A. Leo. 1995. Exploring QSAR. American Chemical Society.
Hawker, D. W. and D.W. Connell. 1988. Environ. Sci. Technol. 22: 382-387.
Hilal, S. H., L. A. Carreira, and S. W. Karickhoff. 1994. Quantitative Treatments of Solute/Solvent
Interactions. Theoretical and Computational Chemistry 1: 291-353.
Meylan, W. M. and P. H. Howard. 1995. J. Pharm. Sci. 84: 83-92.
F-13
-------
-------
Appendix G
Amount of Commercial Food Items Consumed and Intake of Chemical X
from Commercial Food Items
Food Item
cheddar cheese
beef roast
steak
beef loin
pork chop
pork roast
lamb chop
veal cutlet
turkey breast
bologna
cod/haddock
fishsticks
corn grits
popcorn
cornbread
biscuits
pancakes
cereal
raisins
prunes
tomato
squash
Amount of
food item
consumed*
(g/day)
7.07
15.06
1.93
28.09
8.07
3.98
1.13
1.11
3.42
7.90
8.58
1.70
4.03
1.18
7.41
4.58
5.51
2.23
0.34
0.34
18.26
1.47
Average
chemical X
concentration
0»g/g)
4.55E-04
4.55E-04
5.00E-04
5.00E-04
8.41E-04
1.18E-03
2.27E-04
2.05E-04
7.50E-04
1.82E-04
5.23E-04
2.05E-03
3.64E-04
2.05E-04
2.50E-04
9.09E-04
5.45E-04
2.45E-03
4.77E-04
2.73E-04
4.55E-04
5.91E-04
Average
intake of
chemical X
(mg/kg-day)
4.59E-08
9.78E-08
1.38E-08
2.01E-07
9.70E-08
4.70E-08
3.66E-09
3.24E-09
3.67E-08
2.05E-08
6.40E-08
5.00E-09
2.09E-08
3.44E-09
2.65E-08
5.95E-08
4.30E-08
7.81E-08
2.34E-09
1.32E-09
1.19E-07
1.24E-08
High-end
chemical X
concentration
(«g/g)
2.00E-02
l.OOE-02
1.50E-02
2.20E-02
2.10E-02
3.40E-02
l.OOE-02
9.00E-03
2.30E-02
8.00E-03
2.30E-02
9.00E-03
1.60E-02
9.00E-03
1.10E-02
2.40E-02
2.40E-02
6.40E-02
2.10E-02
1.20E-02
2.00E-02
2.60E-02
High-end
intake of
chemical X
(mg/kg-day)
2.0E-06
2.2E-06
4.1E-07
8.8E-06
2.4E-06
1.9E-06
1.6E-07
1.4E-07
1.1E-06
9.0E-07
1.72E-06
2.18E-07
9.2E-07
1.5E-07
1.2E-06
1.6E-06
1.9E-06
2.0E-06
l.OE-07
5.8E-08
5.2E-06
5.5E-07
G-l
-------
Food Item
pizza
meatloaf
potpie
margarine
butter
carrael candy
Intake from all
foods containing
chemical X
Amount of
food item
consumed*
(g/day)
8.77
8.04
1.64
4.44
2.84
2.37
161.50
Average
chemical X
concentration
(wg/g)
2.27E-04
6.82E-04
5.68E-04
4.55E-04
5.45E-04
1.36E-04
Average
intake of
chemical X
(mg/kg-day)
2.85E-08
7.83E-08
1.33E-08
2.88E-08
2.21E-08
4.62E-09
1.18E-06
High-end
chemical X
concentration
(ug/g)
l.OOE-02
2.30E-02
2.50E-02
2.00E-02
2.40E-02
6.00E-03
High-end
intake of
chemical X
(mg/kg-day)
1.3E-06
2.6E-06
5.9E-07
1.3E-06
9.7E-07
2.0E-07
4.26E-05
Total daily intake of all foods* = 2,582 grams/day
Intake of Chemical X after subtracting intake from meat
[Calculation = intake from all foods - (fraction of meat that is fish*chemical intake from all
commercial meat)]
general population
sportfisher
subsistence fisher
1.13E-06
1.13E-06
1.06E-06
4.10E-05
4.10E-05
3.86E-05
*These estimates are weighted averages of intakes for males and females in two age groups: 25-30
year olds and 60-65 year olds
G-2
-------
Appendix H
Ambient Water Quality Criteria for the Protection of Human Health:
Acrylonitrile
Summary
nphis criteria document updates the national criteria for
JL Acrylonitrile using new methods and information
described in the Federal Register Notice (USEPA, 1998a)
and Technical Support Document (USEPA, 1998b) to
calculate ambient water quality criteria. These new methods
include approaches to determine dose-response relationships
for both carcinogenic and non-carcinogenic effects, updated
information for determining exposure factors (e.g., values for
fish consumption), exposure assumptions, and procedures to
determine bioaccumulation factors. For more detailed
information please refer to the U.S. EPA Ambient Water
Quality Criteria (AWQC) document for Acrylonitrile
(USEPA, 1998c).
Background Information
The AWQC is being derived for acrylonitrile (CAS No. 107-
13-1). The chemical formula is C3H3N2. Acrylonitrile
occurrence hi environmental media is not well-documented.
Several regional and local drinking water surveys were
found and one limited study analyzed ambient air samples.
Limited information is also available on acrylonitrile
migration into foods from packaging materials.
Acrylonitrile is largely used hi the manufacture of
copolymers for the production of acrylic and modacrylic
fibers. Other major uses include the manufacture of
acrylonitrile-butadiene-styrene (ABS) and styrene
acrylonitrile (SAN) (used hi production of plastics), and
nitrile elastomers and latexes. It is also used in the
synthesis of antioxidants, Pharmaceuticals, dyes, and
surface-active agents.
According to the U.S. Environmental Protection Agency's
(EPA) Toxic Release Inventory, the total release of
acrylonitrile into the environment in 1990 by manufacturers,
was 8,077,470 pounds. The two largest pathways of release
were underground injection, which accounted for 61% (or
4,925,276 pounds) of the total release, and emissions into the
air, which accounted for 39% (or 3,148,049 pounds) of the
total release. Release of acrylonitrile into water bodies was
reported at 3,877 pounds and release onto land was reported
at 268 pounds.
A baseline BAF of 1.5 was calculated for acrylonitrile. The
baseline BAF was calculated using a value of 0.17 for the log
KOW and 1.000 for the food-chain multiplier (FCM) at trophic
level 4. A value of 0.17 was selected as a typical value of the
log Kow for acrylonitrile (USEPA 1998c). A value of 1.000
was selected as the FCM for trophic level 4, reflective of top
predator fish based on a log K,,w of 2.0 from USEPA (1998b).
Using these data, the baseline BAF was calculated as: K^ *
FCM = (10017 )*1.000 = 1.5 (rounded to two significant
digits).
Based upon sufficient evidence from animal studies (multiple
tumor types in several strains of rats by several routes) and
limited evidence from human studies (lung tumors in
workers), positive mutagenicity, acrylonitrile is considered as
a likely human carcinogen by any route. A linear approach
is used for the low dose extrapolation.
AWQC Calculation
For Ambient Waters Used as Drinking Water Sources
The cancer-based AWQC was calculated using the RSD and
other input parameters listed below:
AWQC = RSD x
BW
4
i=2
£ (FI. x BAF)
where:
RSD
BW
DI
FI
BAF
= Risk specific dose (1.6 x lO^6 mg/kg-day at 10"6
lifetime risk)
= Human body weight assumed to be 70 kg
= Drinking water intake assumed to be 2 L/day
= Fish intake at trophic level i, i=2,3, and 4; total
intake assumed to be 0.01780 kg/day
= Bioaccumulation factor at trophic level i (i=2,3,
and 4) equal to 1.03, 1.02, and 1.05 L/kg-tissue for
trophic levels 2,3, and 4, respectively.
This is a preliminary summary of a criteria document being prepared for the derivation of the Ambient Water Quality Criteria (AWQC) for the
protection of human health from exposure to acrylonitrile. The calculated AWQC values presented in this draft are subject to revision pending
inclusion of further information concerning exposure as well as possible changes in the toxicological information used to derive the criterion.
-------
This yields concentrations of 5.5 x 10'5 mg/L (or 0.05 ug/L),
for a 10"* (one in a million) lifetime cancer risk.
For Ambient Waters Not Used as Drinking Water
Sources
When the water body is to be used'for recreational purposes
and not as a source of drinking water, the drinking water
value (DI above) is eliminated from the equation and it is
substituted with an incidental ingestion value (II). The
incidental intake is assumed to occur from swimming and
other activities. The fish intake value is assumed to remain
the same. The default value for incidental ingestion is 0.01
L/day. When the above equation is used to calculate the
AWQC with the substitution of an incidental ingestion of
0.01 L/day an AWQC of 4.0 x W3 mg/L (or 4.0 ug/L) is
obtained for a 10"6 lifetime cancer risk.
Site-Specific or Regional Adjustments to Criteria
Several parameters in the AWQC equation can be adjusted
on a site-specific or regional basis to reflect regional or local
conditions and/or specific populations of concern. These
include fish consumption, incidental water consumption as
related to regional/local recreational activities, BAF
(including factors used to derive BAFs, percent lipid offish
consumed by target population, and species representative of
given trophic levels), and the relative source contribution.
States are encouraged to make adjustments using the
information and instructions provided in the Technical
Support Document (USEPA, 1998b).
REFERENCES
USEPA. 1998a. Federal Register Notice: Proposed Water
Quality Criteria Methodology Revisions; Human
Health. (See Attached Copy).
USEPA. 1998b. Ambient Water Quality Criteria Derivation
Methodology; Human Health. Technical Support
Document. EPA/822/B-98/005. July. (See
Attached Copy).
USEPA. 1998c. Ambient Water Quality Criteria for the
Protection of Human Health: Acrylonitrile.
EPA/822/R-98/006. July.
This is * preliminary summary ofa criteria document being prepared for the derivation of the Ambient Water Quality Criteria (AWQC) for the
protection of human health from exposure to 1,3-dichloropropene. The calculated AWQC values presented in this draft are subject to revision
pending inclusion of further information concerning exposure as well as possible changes in the toxicological information used to derive the
criterion.
-------
Ambient Water Quality Criteria for the Protection of Human Health:
1,3-Dichloropropene (1,3-DCP)
Summary
This criteria document updates the national criteria for 1,3-
DCP using new methods and information described in
the Federal Register Notice (USEPA, 1998a) and Technical
Support Document (USEPA, 1998b) to calculate ambient
water quality criteria. These new methods include
approaches to determine dose-response relationships for both
carcinogenic and non-carcinogenic effects, updated
information for determining exposure factors (e.g., values for
fish consumption), exposure assumptions, and procedures to
determine bioaccumulation factors. For more detailed
information please refer to the U.S. EPA Ambient Water
Quality Criteria (A WQC) document for 1,3-Dichloropropene
(1,3-DCP) (USEPA, 1998c).
Background Information
The A WQC is being derived for 1,3-Dichloropropene (CAS
No. 542-75-6). The chemical formula is C3H4C12 and
molecular weight is 110.98 (pure isomers). At 25°C, the
physical state of 1,3-DCP is a pale yellow to yellow liquid.
Dichloropropene (DCP) is used as soil fumigant in the United
States to control soil nematodes on crops grown in sandy
soils. The EPAs National Toxics Inventory data base
reported air emissions of 18,820,000 pounds/year in the U.S.
(USEPA, 1996a). Numerous studies have sampled for DCP
(and isomers) in drinking water, groundwater and surface
waters across the U.S. (Hall et al., 1987; Miller et al., 1990;
RIDEM, 1990; Rutledge, 1987; STORET, 1992). All of
these studies report concentrations of 1,3-DCP usually at or
below the detection limits (USEPA, 1998c).
The AWQC Bioaccumulation factor (BAF) is 2.2 L/kg of
tissue for 1,3-DCP. This BAF is based on the total
concentration of 1,3-DCP in trophic level four biota divided
by the total concentration in water, assuming default values
for the freely-dissolved fraction and lipid content of
consumed aquatic organisms.
The cancer risk evaluation of 1,3-DCP uses the new methods
in the proposed cancer guidelines (USEPA, 1996), which are
described in the Federal Register Notice (USEPA, 1998a)
and in the Technical Support Document (USEPA, 1998b).
Based upon sufficient evidence from animal studies (multiple
tumor types in several species by oral, inhalation, and dermal
routes), positive mutagenicity, and structural analogues, 1,3-
DCP is considered "likely to be carcinogenic to humans by
all routes of exposure." Based on the mutagenic mode of
action, a linear low dose approach is recommended.
AWQC Calculation
For Ambient Waters Used as Drinking Water Sources
The cancer-based AWQC was calculated using the RSD and
other input parameters listed below:
AWQC = RSD x
BW
•5J (FI x BAF )
;-">
l-i
where:
RSD
BW
DI
FI
BAF
= Risk specific dose 1.0 x 10'5 mg/kg/day (lO"6 risk)
= Human body weight assumed to be 70 kg
= Drinking water intake assumed to be 2 L/day
= Fish intake at trophic level i, i=2,3,and 4 total
intake assumed to be 0.01780 kg/day
= Bioaccumulation factor at trophic level i
(i=2,3,and 4), equal to 2.32, 1.86, and 2.78 L/kg-
tissue for trophic levels 2,3,and 4, respectively.
This yields a value of 3.4 x 10" mg/L, or 0.34 ug/L (rounded
from 0.343 ug/L).
For Ambient Waters Not Used as Drinking Water
Sources
When the water body is used for recreational purposes and
not as a source of drinking water, the drinking water value is
eliminated from the equation and it is substituted with an
incidental ingestion value. The incidental intake is assumed
to occur from swimming and other activities. The fish intake
value is assumed to remain the same. The default value for
incidental ingestion is 0.01 L/day. When the above equation
is used to calculate the AWQC with the substitution of an
incidental ingestion of 0.01 L/day an AWQC of 1.4 xlO'2
mg/L (14 ug/L) is obtained.
This is a preliminary summary of a criteria document being prepared for the derivation of the Ambient Water Quality Criteria (AWQC) for the
protection of human health from exposure to 1,3-dichloropropene. The calculated AWQC values presented in this draft are subject to revision
pending inclusion of further information concerning exposure as well as possible changes in the toxicological information used to derive the
criterion.
-------
Site-Specific or Regional Adjustments to Criteria
Several parameters in the A WQC equation can be adjusted
on a site-specific or regional basis to reflect regional or
local conditions and/or specific populations of concern.
These include fish consumption; incidental water
consumption as related to regional/local recreational
activities; BAF (including factors used to derive BAFs ,
percent lipid offish consumed by the target population, and
species representative of given trophic levels); and the
relative source contribution. States are encouraged to make
adjustments using the information and instructions
provided in the Technical Support Document (USEPA,
1998b).
REFERENCES
USEPA. 1998a. Federal Register Notice: Proposed Water
Quality Criteria Methodology Revisions; Human
Health. (See Attached Copy).
USEPA. 1998b. Ambient Water Quality Criteria
Derivation Methodology; Human Health.
Technical Support Document. EPA/822/B-
98/005. July. (See Attached Copy).
USEPA. 1998c. Ambient Water Quality Criteria for the
Protection of Human Health: 1,3-
Dichloropropene(l,3-DCP). EPA/822/R-98/005.
July.
This Is a preliminary summary of a criteria document being prepared for the derivation of the Ambient Water Quality Criteria (AWQC) for the
protection of human health from exposure to 1,3-dichloropropene. The calculated AWQC values presented in this draft are subject to revision
pending inclusion of further information concerning exposure as well as possible changes in the toxicological information used to derive the
-------
Ambient Water Quality Criteria for the Protection of Human Health:
Hexachlorobutadiene (HCBD)
Summary
'T'his criteria document updates the national criteria for
J. HCBD using new methods and information described in
the Federal Register Notice (USEPA, 1998a) and Technical
Support Document (USEPA, 1998b) to calculate ambient
water quality criteria. These new methods include
approaches to determine dose-response relationships for both
carcinogenic and non-carcinogenic effects, updated
information for determining exposure factors (e.g., values for
fish consumption), exposure assumptions, and procedures to
determine bioaccumulation factors. For more detailed
information please refer to the U.S. EPA Ambient Water
Quality Criteria (A WQC) document for hexachlorobutadiene
(HCBD)(USEPA, 1998c).
Background Information
The AWQC is being derived for hexachlorobutadiene (CAS
No. 87-68-3). The chemical formula is C4C16 and molecular
weight is 260.76. At 25°C, HCBD is a colorless liquid.
HCBD is used as a solvent in chlorine gas production, as an
intermediate in the manufacture of rubber compounds and
lubricants, and as a pesticide. The EPA's National Toxics
Release Inventory data base reported total emissions to the
environment in 1990 of 5,591 pounds/year in the U.S., of
which 4,906 pounds was to air. Numerous studies have
sampled for HCBD in drinking water, groundwater and
surface waters across the U.S. (see USEPA 1998c for a
summary). The vast majority of samples are at trace levels
or below the detection limits (DL = 0.1 ug/L).
The AWQC Bioaccumulation factor (BAF) is 620 L/kg of
tissue for HCBD. This BAF is based on the total
concentration of HCBD in trophic level four biota divided by
the total concentration in water, assuming default values for
the freely-dissolved fraction and lipid content of consumed
aquatic organisms.
The cancer risk evaluation of HCBD uses the new methods
described in the Federal Register Notice (USEPA, 1998a)
and in the Technical Support Document (USEPA, 1998b).
Based on a renal tumor finding in one chronic feeding study
at one high dose in one species (both sexes of Sprague-
Dawley rats), "via oral route, HCBD is considered as
likely to be carcinogenic to humans only at very high
exposure conditions, where significant renal toxicity
occurs." There is some mutagenic activity in the presence of
metabolic activation. Thus, a mutagenic mode of action
cannot be ruled out. As a result, both the cancer-based, linear
low dose approach and the non-linear margin of exposure
approaches are used for deriving the AWQC.
AWQC Calculation
For Ambient Waters Used as Drinking Water Sources
The cancer-based AWQC was calculated using the RSD and
other input parameters listed below:
AWQC = RSD x
BW
DI+2^ (FL x BAF.)
where:
RSD
BW
DI
FI
BAF
= Risk specific dose 2.5 x 1Q'5 mg/kg/day (10'6 risk)
= Human body weight assumed to be 70 kg
= Drinking water intake assumed to be 2 L/day
= Fish intake at trophic level i, i=2,3, and 4; total
intake assumed to be 0.01780 kg/day
= Bioaccumulation factor at trophic level i (i=2,3,
and 4) equal to 1,518,2,389, and 1,294 L/kg-tissue
for trophic levels 2,3, and 4, respectively.
This yields a value of 4.6 x
(rounded from 0.0462 ug/L).
10'5 mg/L, or 0.046 ug/L
The AWQC using the margin of exposure approach was
calculated using the following equation and input parameters
listed below.
AWQC =
SF
- RSC x
BW
22 (FI. x BAF.)
i=2 ' '
This is a preliminary summary of a criteria document being prepared for the derivation of the Ambient Water Quality Criteria (AWQC) for the
protection of human health from exposure to HCBD. The calculated AWQC values presented in this draft are subject to revision pending inclusion
of further information concerning exposure as well as possible changes in the toxicological information used to derive the criterion.
-------
where:
REFERENCES
Pdp = Point of departure (0.054 mg/kg/day)
SF =• Safety factor of 300
RSC = Relative source contribution from air of 1 .2 x 1 0"4
mg/kg-day, subtracted in this case
BW = Human body weight assumed to be 70 kg
DI m Drinking water intake assumed to be 2 L/day
FI - Fish intake at trophic level i, i=2,3, and 4; total
intake assumed to be 0.01780 kg/day
BAF ~ Bioaccumulation factor at trophic level i (i=2,3,
and 4) equal to 1,518, 2,389, and 1,294 L/kg-tissue
for trophic levels 2,3, and 4, respectively.
This yields an AWQC of 1 .1 x 10"4 mg/L (0.1 1
For Ambient Waters Not Used as Drinking Water
Sources
When the waterbody is used for recreational purposes and not
as a source of drinking water, the drinking water value is
eliminated from the equation and it substituted with an
incidental ingestion value. The incidental intake is assumed
to occur from swimming and other activities. The fish intake
value is assumed to remain the same. The default value for
incidental ingestion is 0.01 L/day. When the linear approach
is used to calculate the AWQC with the substitution of an
incidental ingestion of 0.01 L/day a cancer-based AWQC of
4.9 x 10-5 mg/L (or 0.049 ug/L, rounded from 0.0487 ng/L)
is obtained. When the non-linear margin of exposure
approach is used with the substitution of an incidental
ingestion of 0.01 L/day, the AWQC is 1.2 x 10"4 mg/L (or
0.12 ug/L, rounded from 0.117 ug/L).
Site-Specific or Regional Adjustments to Criteria
Several parameters in the AWQC equations can be adjusted
on a site-specific or regional basis to reflect regional or local
conditions and/or specific populations of concern. These
include fish consumption; incidental water consumption as
related to regional/local recreational activities; BAF
(including factors used to derive BAFs, percent lipid offish
consumed by the target population, and speciesrepresentative
of given trophic levels); and the relative source contribution.
States are encouraged to make adjustments using the
information and instructions provided in the Technical
Support Document (USEPA, 1998b).
USEPA. 1998a. Federal Register Notice: Proposed Water
Quality Criteria Methodology Revisions; Human
Health. (See Attached Copy).
USEPA. 1998b. Ambient Water Quality Criteria Derivation
Methodology; Human Health. Technical Support
Document. EPA/822/B-98/005; July. (See Attached
Copy).
USEPA. 1998c. Ambient Water Quality Criteria for the
Protection of Human Health: Hexachlorobutadiene
(HCBD). EPA/822/R-98/004. July.
This is * preliminary summary of a criteria document being prepared for the derivation of the Ambient Water Quality Criteria (AWQC) for the
protection of human health from exposure to HCBD. The calculated AWQC values presented in this draft are subject to revision pending inclusion
of further information concerning exposure as well as possible changes in the toxicological information used to derive the criterion.
------- |