_ ~Q, United Slates Office of Publication 9285.7-15-1
J-J»i Environmental Protection Solid Waste and EPA/540/R-93/081
Agency Emergency Rcisponse PB93-963510
February 1994
Superfund
v>EPA Guidance Manual for the
Integrated Exposure Uptake
Biokinetic Model for Lead in
Children
-------
February 1994
Publication Number 9285.7-15-1
EPA 540-R-93-081
PB93-963510
GUIDANCE MANUAL FOR
THE INTEGRATED EXPOSURE UPTAKE BIOKINETIC
MODEL FOR LEAD IN CHILDREN
Prepared by
THE TECHNICAL REVIEW WORKGROUP FOR LEAD
for
THE OFFICE OF EMERGENCY AND REMEDIAL RESPONSE
U.S. ENVIRONMENTAL PROTECTION AGENCY
with Document Production Assistance from
THE ENVIRONMENTAL CRITERIA AND ASSESSMENT OFFICE
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial
products does not constitute endorsement or recommendation for use.
11
-------
PREFACE
The Guidance Manual has been developed to assist the user in providing appropriate
input to the Integrated Exposure Uptake Biokinetic (IEUBK) Model for Lead. The IEUBK
Model is designed to model exposure from lead in air, water, soil, dust, diet, and paint and
other sources with pharaiacokinetic modeling to predict blood lead levels in children
6 months to 7 years old. This manual emphasizes the use of the IEUBK Model for
estimating risks from childhood lead exposure to soil and household dust that might be
encountered at CERCLA/RCRA sites, although other applications of the model are possible.
The manual provides background information on environmental exposure parameters and
recommends some useful approaches that allow flexibility for site-specific risk assessments,
where possible. Default parameters are recommended unless there is sufficient data to
characterize site-specific conditions. A separate Appendix on sampling is being developed
and will be issued later. A Technical Support Document details the basis for the biokinetic
parameters and equations in the IEUBK Model. In addition, EPA is continuing to compare
the results of field studies with model predictions and will release these findings in a later
document.
One of the proposed uses of this model will be support for the implementation of an
Interim Directive of the Office of Solid Waste and Emergency Response (OSWER). This
Interim Directive explains how the IEUBK Model results can be a tool for the determination
of site-specific cleanup levels. In this context, the model is viewed as a predictive tool for
estimating changes in blood concentrations as exposures are modified. The model is also
viewed as a useful tool that should aid the Agency in making more informed choices about
the concentrations of lead that might be expected to impact human health.
The development of the model has included the cooperative efforts of several EPA
programs over nearly a decade. For the last three years, these efforts have been coordinated
by the Technical Review Workgroup for Lead. During its development, the model has
undergone review by outside scientists, and its usefulness has been evaluated by EPA staff,
contractors, and other reviewers assessing site-specific risk. The current version of the
IEUBK model and the Guidance Manual incorporates many of their recommendations.
The use of mathematical and statistical models for environmental risk assessment has
become increasingly widespread because of the many practical difficulties encountered in
controlling human exposure to toxicants with subtle and long-lasting effects. Exposure to
lead during infancy and childhood increases the risk of irreversible neurobehavioral deficits
ill
-------
at levels of internal exposure as low as 10 to 15 jug Pb per 100 mL of blood (10 to
15 jug/dL). Lead has many known sources, and many pathways from its environmental
sources into the child's body (U.S. Environmental Protection Agency, 1986). The
Environmental Protection Agency has long been interested in methods for relating
environmental lead concentrations to blood lead concentrations in children. Earlier
approaches based on statistical correlations provided essential information on the existence
and magnitude of childhood lead uptake from persistent exposure to different environmental
sources, including lead in air, diet, drinking water, soil, dust, and lead-based paint.
Unfortunately, these statistical relationships are limited in their ability to estimate the effects
of alternative lead abatement methods that change pathways as well as sources.
In 1985 the EPA Office of Air Quality Planning and Standards began to develop an
alternative approach for estimating the effectiveness of alternative National Ambient Air
Quality Standards for lead, particularly around point sources of air lead emissions such as
smelters. This was a computer simulation model with two components: (1) a model of the
biokinetics of lead distribution and elimination whose parameters vary with the child's age,
and (2) a multi-source and multi-media lead exposure model in which air lead concentrations
change over time. The biokinetic model was based on studies at New York University by
Naomi Harley, Theodore Kneip, and Peter Mallon. The U.S. Environmental Protection
Agency Clean Air Science Advisory Committee (CASAC) reviewed and found acceptable the
OAQPS staff report documenting the model in 1989. A subsequent OAQPS staff paper
reviewing the National Ambient Air Quality Standard for Lead, which included results of
applying the model to point sources of air lead such as smelters and battery plants, was also
evaluated by CASAC in 1990 (U.S. Environmental Protection Agency, 1990B).
Those who had been involved in developing the lead model then received a large and
growing number of requests on applications of the model in a wide variety of other contexts
not originally intended for model use. The largest number of these requests involved the use
of the model to estimate the effects of soil lead abatement at Superfund sites.
The air model was further developed to include enhancements in absorption and
biokinetics. In November, 1991, the Indoor Air Quality and Total Human Exposure
Committee (IAQTHEC) of EPA's Science Advisory Board (SAB) reviewed the Uptake
Biokinetic Model for Lead (version 0.4) and evaluated its use in assessing total lead
exposures and in aiding in developing soil cleanup levels at residential CERCLA/RCRA
sites. The Committee's Report was transmitted to EPA Administrator William K. Reilly in
March, 1992. The Committee concluded that while refinements in the detailed specifications
IV
-------
of the model would be needed, the approach followed in developing the model is sound. The
Committee stated that the model can effectively be applied for many current needs even as it
continues to undergo refinement for other applications, based upon experience gained in its
use.
The Committee was concerned that the reliability of the results obtained using the
model is very much dependent on the selection of the various coefficients and default values
that were used. In particular, the Committee identified the need for guidance on the
"proper" geometric standard deviation (GSD) and the use of default values for other
parameters. In addition to these general comments, specific comments were included in the
Report. The comments of the SAB and other reviewers have been considered in this revision
of the Guidance Manual.
Since the SAB review, EPA has further refined the model. The four main components
of the current IEUBK model are: (1) an exposure model that relates environmental lead
concentrations to age-dependent intake of lead into the gastrointestinal tract; (2) an absorption
model that relates lead intake into the gastrointestinal tract and lead uptake into the blood;
(3) a biokinetic model that relates lead uptake in the blood to the concentrations of lead in
several organ and tissue compartments; and (4) a model for uncertainty in exposure and for
population variability in absorption and biokinetics. A Technical Support Document that
details the selection of parameters and equations in the model is available.
As with any multicompartmental model, pools in the compartmental analysis can be
identified with specific organs or organ systems only if biological concentrations of the
compartments are known. For some compartments, the biological concentrations have been
measured at a number of time points so that the movement of lead from one compartment to
another can be estimated. The biokinetic and absorption components of the model, however,
are not observed directly but are inferred from accessible data.
In developing the IEUBK Model, EPA has learned much from "real world"
comparisons of blood lead and predicted values—not only that the model works, but also that
it can be made to work better. Guidance on the appropriate use of the model is based on our
experiences, where possible, and on the experiences of many users and reviewers of the
model. Many of the most useful parts of the Guidance Manual have been suggested by these
reviewers.
-------
While the model has been used to support the NAAQS for Lead, the Clean Water Act
national regulations, and several other regulatory and enforcement issues, EPA is continuing
its validation of the IEUBK Model with detailed evaluation of additional data collected from
different types of sites. Comparison of predicted and empirical blood lead concentrations
will be described in the Field Study Data Set Comparisons Document described in
Section 1.2.2.
Although EPA is releasing version 0.99d of the IEUBK Model to ensure consistent
application among users, the Agency will continue to evaluate the results of validation
exercises and different applications of the model. The Environmental Protection Agency will
determine periodically whether refinements to the model are warranted, considering scientific
advancements and the development of alternative approaches.
The Environmental Protection Agency welcomes the suggestions of those using the
IEUBK model. Questions regarding the site-specific application of the IEUBK Model should
be raised with the appropriate Regional Toxics Integration Coordinator. Comments on the
technical content of the manual or suggestions for its improvement may be brought to the
attention of the Technical Review Workgroup for Lead, whose current addresses are listed on
page xxi.
VI
-------
TABLE OF CONTENTS
Page
PREFACE iii
LIST OF TABLES xiii
LIST OF FIGURES xvii
LIST OF SCREENS xix
TECHNICAL REVIEW WORKGROUP FOR LEAD xxi
GLOSSARY OF MODEL TERMS xxiii
1. BEFORE YOU START 1-1
1.1 BACKGROUND: PURPOSE AND DEVELOPMENT OF
THE MODEL 1-1
1.1.1 Description of the Model 1-1
1.1.2 Simulation of Childhood Lead Exposure and
Retention 1-3
1.1.3 Historical Evolution from Slope Factor Models to the
Integrated Exposure Uptake Biokinetic Model 1-5
1.1.4 Using the Integrated Exposure Uptake Biokinetic
Model for Risk Estimation 1-9
1.1.5 Validation of the Integrated Exposure Uptake
Biokinetic Model 1-10
1.1.5.1 The Model Is Biologically and Physically
Plausible 1-11
1.1.5.2 The Model Is Computationally Accurate 1-12
1.1.5.3 Emphirical Comparisons of the Model 1-12
1.2 ORGANIZATION OF THE MANUAL 1-13
1.2.1 Increasing Levels of Guidance and
Technical Assistance 1-13
1.2.2 Additional Documentation 1-14
1.3 GETTING READY TO USE THE MODEL 1-15
1.3.1 Preparing a Site-Specific-Exposure Scenario 1-15
1.3.2 Understanding How the Biokinetic Component of the
Model Works 1-17
1.3.3 Understanding Limitations of the Model 1-18
1.4 RUNNING THE MODEL 1-19
1.4.1 Your Responsibilities 1-19
1.4.2 Exploring Model Options 1-20
1.4.3 Documentation of Input Parameter and Data Files .... 1-21
1.4.4 Documentation of Model Output 1-22
1.4.4.1 Selecting Output Alternatives 1-22
1.4.4.2 Understanding the Output 1-23
1.4.4.3 Interpreting the Output and Communicating
the Results 1-24
1.5 REFINEMENTS AND ENHANCEMENTS 1-28
1.6 GETTING MORE HELP 1-29
vii
-------
TABLE OF CONTENTS (cont'd)
Page
2. A GUIDED TOUR THROUGH THE LEAD MODEL 2-1
2.1 THE LEAD MODEL IS DRIVEN BY MENUS 2-1
2.2 DETAILED DESCRIPTION OF MENUS 2-3
2.2.1 Help Menu 2-3
2.2.1.1 General Help 2-3
2.2.1.2 Information Menu 2-4
2.2.1.3 Other On-Line Help Menus 2-4
2.2.2 Parameter Input Menus 2-4
2.2.2.1 Air Lead 2-4
2.2.2.2 Dietary Lead 2-7
2.2.2.3 Drinking Water Lead 2-8
2.2.2.4 Soil and Dust Lead 2-10
2.2.2.5 Alternate Source 2-14
2.2.2.6 Bioavailability of Lead in Food, Drinking
Water, Soil, and Dust 2-17
2.2.2.7 Maternal-Fetal Lead Exposure 2-17
2.2.2.8 Save and Load Options 2-18
2.2.3 Computation Menu 2-20
2.2.3.1 Run a Single Simulation of the Model 2-20
2.2.3.2 Run Multiple Simulations of the Model for
a Range of Media Lead 2-20
2.2.3.3 Multiple Simulation Runs of a Medium To
Find Concentration of Lead in the Medium
That Produces a Specified Blood Lead 2-21
2.2.3.4 Batch Mode Multiple Simulation Runs
Using Input Data Files 2-22
2.2.3.5 Statistical Analyses of Batch Mode
Data Sets 2-26
2.3 BUILDING AN EXPOSURE SCENARIO 2-27
2.3.1 Air Lead Menu 2-27
2.3.1.1 Default Air Lead Exposure Parameters 2-27
2.3.1.2 Ventilation Rate 2-27
2.3.1.3 Indoor/Outdoor Activity Patterns 2-28
2.3.1.4 Lung Absorption 2-29
2.3.2 Dietary Lead Menu 2-29
2.3.2.1 Total Dietary Lead Exposure 2-29
2.3.2.2 Dietary Lead Exposure by Additional
Pathways 2-31
2.3.3 Drinking Water Lead Exposure Menu 2-33
2.3.3.1 Drinking Water Lead Default Exposure
Parameters 2-33
2.3.3.2 Alternate Drinking Water Exposure by Age . . . 2-36
2.3.4 Soil/Dust Lead Exposure Menu 2-37
Vlll
-------
TABLE OF CONTENTS (cont'd)
Page
2.3.4.1 Soil and Dust Lead Default Exposure
Parameters . 2-38
2.3.4.2 Exposure to Soil and Dust 2-38
2.3.4.3 Sources of Dust Exposure 2-40
2.3.4.4 Fraction of Exposure as Soil or Dust 2-42
2.3.4.5 Bioavailability of Lead in Soil and Dust 2-44
2.3.5 Alternate Source Exposure Menu 2-45
2.4 STARTING AND RUNNING THE MODEL 2-45
2.4.1 Loading and Starting the Model 2-45
2.4.2 Running the Model 2-46
2.4.2.1 Computation Options 2-46
2.4.2.2 Output Options 2-46
3. QUICK REFERENCE FOR THE EXPERIENCED USER 3-1
3.1 FINDING YOUR WAY THROUGH THE MENUS 3-1
3.2 PARAMETER LIST WITH DEFAULT VALUES 3-1
3.3 BATCH MODE INPUT FORMAT 3-2
3.4 OUTPUTS FOR DOCUMENTATION, BRIEFING, AND
PRESENTATION 3-9
3.4.1 Overview of Output Options 3-9
3.4.1.1 Plotting 3-9
3.4.1.2 Uses of Batch Mode Analysis 3-10
3.4.2 Detailed Instructions on Output Options 3-11
3.4.2.1 Save Output from a Single Run 3-11
3.4.2.2 Save Output from Multiple Runs for
Probability Plots 3-11
3.4.2.3 Save Output from Multiple Runs for
Media-Level Plots 3-11
3.4.2.4 Save Output from a Batch Mode Run 3-12
3.4.2.5 Probability Plots for Single Runs 3-12
3.4.2.6 Probability Plots for Multiple Runs 3-13
3.4.2.7 Multi-Level Plots for Blood Lead Versus
Media Lead 3-13
3.4.3 Recommendations on Multi-Level Soil Lead
Exposure Scenarios 3-13
4. MORE ABOUT THE MODEL 4-1
4.1 LEAD BIOAVAILABILITY 4-1
4.1.1 Background 4-1
4.1.2 Definitions 4-1
4.1.3 Literature Sources on Bioavailability 4-2
4.1.4 Lead Absorption-Bioavailability Relationships 4-3
4.1.5 Cellular and Subcellular Mechanisms of Lead
Absorption 4-3
4.1.6 Factors Affecting Lead Absorption 4-5
ix
-------
TABLE OF CONTENTS (cont'd)
'3age
4.1.7 Bioavailability of Lead in Soils and Dusts 4-7
4.1.7.1 Biophysico-Chemical and Environmental
Features of the Exposure Matrix 4-7
4.1.7.2 Is There a Better Way To Classify
Lead-Contaminated Sites? 4-9
4.1.7.3 Methodological Approaches To Quantifying
Bioavailability 4-10
4.1.7.4 Determination of Absolute Bioavailability .... 4-10
4.1.7.5 Absolute Versus Relative Bioavailability 4-12
4.1.7.6 Quantitative Experimental Models
of Human Lead Bioavailability 4-13
4.1.7.7 Summary and Advisory Overview for Lead
in Soils and Dust 4-16
4.1.8 Bioavailability of Lead in the Diet 4-16
4.1.9 Bioavailability of Lead in Water 4-19
4.1.10 Bioavailability of Lead in Air 4-20
4.2 USING THE INTEGRATED EXPOSURE UPTAKE
BIOKINETIC MODEL FOR RISK ESTIMATION 4-21
4.2.1 Why Is Variability Important? 4-21
4.2.1.1 Intent of the Model and the Measure 4-21
4.2.1.2 Individual Geometric Standard Deviation .... 4-21
4.2.2 Variability Between Individuals Is Characterized by
the Geometric Standard Deviation 4-23
4.2.3 Statistical Methods for Estimating the Geometric
Standard Deviation from Blood Lead Studies 4-25
4.2.4 Choosing the Geometric Standard Deviation:
Intra-Neighborhood Variability 4-26
4.2.5 Basis for Neighborhood Scale Risk Estimation 4-27
4.2.6 Relationship Between Geometric Standard Deviation
and Risk Estimation 4-28
4.2.7 Risk Estimation at a Neighborhood or Community
Scale 4-30
4.2.7.1 What Do We Mean by "Neighborhood" or
Community" Risk? 4-30
4.2.7.2 Neighborhood Risk Estimation as the Sum
of Individual Risks 4-31
4.2.7.3 An Example for the "Sum of Individual
Risks" Approach 4-32
4.2.7.4 Assessment of Risk Using Grouped Data for
a Neighborhood 4-34
4.2.7.5 Assessment of Risk with Neighborhood or
Neighborhood-Scale Input 4-36
4.3 ENVIRONMENTAL PATHWAY ANALYSIS 4-37
-------
TABLE OF CONTENTS (cont'd)
?age
4.3.1 Concept of Pathway Analysis 4-37
4.3.2 Pathway Analyses by Linear Regression 4-38
4.3.3 Pathway Analysis Using Structural Equation Models . . . 4-39
4.3.4 Regression Analyses for Multiple Exposure Pathways:
Soil Example 4-41
4.4 USE OF DATA FROM BLOOD LEAD STUDIES 4-42
4.4.1 Overview 4-42
4.4.2 Data Quality 4-45
4.4.3 Age of the Population Tested 4-46
4.4.4 Time of the Year When Testing Was Done 4-46
4.4.5 Concurrent Characterization of Lead Sources 4-47
4.4.6 Demographics and Behavioral Factors That Affect
Lead Exposure 4-48
4.4.7 Effect of Public Awareness or Educational
Intervention 4-48
4.4.8. Comparison of Observed and Predicted Blood
Lead Concentrations 4-49
4.4.8.1 Were Important Sources of Lead Exposure
Overlooked? 4-49
4.4.8.2 Are There Interrupted or Enhanced Exposure
Pathways at the Site? 4-50
4.4.8.3 Are the Assumptions About Site-Specific Intake
Rates and Uptake Parameters Valid? 4-50
4.5 ASSESSING THE RELATIONSHIP BETWEEN SOIL/DUST
LEAD AND BLOOD LEAD 4-51
4.5.1 Assessing Reductions in Blood Lead 4-51
4.5.2 Situations in Which the Use of the Integrated
Exposure Uptake Biokinetic Model Is Uncertain 4-53
4.5.2.1 Assessment of Risk with Community or
Neighborhood-Scale Input 4-53
4.5.2.2 Use of Surrogate Input Data from Models
or Surveys 4-53
4.5.2.3 Use of the Model To Assess Risk of Elevated
Blood Lead at the Regional or State Level . . . 4-53
4.5.2.4 Use of the Model To Assess Trigger Levels
for Soil Abatement at the Community,
Regional, or State Level 4-54
4.5.3 Factors That Constrain or Limit the Use of the
Model 4-54
4.5.3.1 Data and Data Sets Used as Input for
the Integrated Exposure Uptake
Biokinetic Model 4-54
XI
-------
TABLE OF CONTENTS (cont'd)
Taee
4.5.3.2 Biological and Exposure Parameters Used in
the Integrated Exposure Uptake Biokinetic
Model Bioavailability of Soil Lead 4-56
4.6 WHAT YOU NEED TO KNOW ABOUT BIOKINETICS 4-58
4.6.1 Description of the Biokinetic Model 4-58
4.6.2 Consequences of Biokinetic Parameters for
Site-Specific Risk Assessment 4-60
4.7 ISSUES IN USE OF THE MODEL FOR PAINT CHIPS 4-61
4.7.1 Inappropriateness of Use of the Integrated Exposure
Uptake Biokinetic Model for Paint Chip Ingestion .... 4-61
4.7.2 Daily Intake of Paint Chips 4-63
4.7.3 Relationship of X-Ray Fluorescence Lead Paint
Surface Loading to Lead Paint Concentration 4-64
4.7.4 Dissolution of Paint Chips in Acid Environments 4-64
4.7.5 Absorption of Lead Paint In Vivo 4-65
5. APPLICATIONS WITH EXAMPLES 5-1
5.1 APPLICATIONS FOR POPULATION ESTIMATES 5-1
5.2 APPLICATIONS WHERE ENVIRONMENTAL LEAD
CONCENTRATIONS CHANGE OVER TIME 5-1
5.3 APPLICATIONS FOR PROBABILITY AND
RISK ESTIMATION 5-18
5.4 BATCH MODE INPUT AND STATISTICAL
ANALYSES OF OUTPUT 5-21
5.5 SOIL LEAD ABATEMENT EXAMPLES 5-28
6. REFERENCES 6-1
APPENDIX A: How to Calculate the Geometric Standard Deviation from
Blood Lead Data, If You Must A-l
APPENDIX B: Summary of Revisions to Lead Uptake Biokinetic Model
Software Versions B-l
Xll
-------
LIST OF TABLES
Number Page
2-1 Dietary Lead Intake for U.S. Children by Age, for Each Year
from 1978 to Present 2-31
2-2 Estimates of Lead Intake from Consumption of Local Produce
by Children, Ages 2 to 6 Years, in Kellogg, Idaho 2-33
2-3 Estimates of Lead Intake from Consumption of Local Fish by
Children, Ages 2 to 6 Years, in Kellogg, Idaho 2-34
2-4 Average Daily Water Intake in U.S. Children 2-37
2-5 Tap Water Intake by Age Category 2-37
2-6 Daily Intake of Soil and Dust Estimated from Elemental
Abundances 2-39
2-7 Age-Specific Soil and Dust Intake 2-40
2-8 Minimum Percentage Soil Intake as a Function of Age in Dutch
Children in Daycare Centers 2-44
3-1 Default Values for Model Parameters 3-3
3-2 Format for Batch Mode Input Data File 3-7
4-1 Piecewise Linear Regression Models for Blood Lead Versus
Water Lead in Three Studies 4-20
4-2 Example of Neighborhood Risk Estimation with Grouped Data . . . 4-35
4-3 Example of Neighborhood Risk Estimation with Coarsely
Grouped Data 4-35
4-4 Percentage Increase in Blood Lead Levels in Infant Male
Wistar Rats with 48-Hour Oral Exposure to Lead Acetate,
and to Lead Octoate and Lead Chromate Paints of Different
Particle Sizes 4-66
4-5 Percentage Increase in Blood Lead Levels in Infant Male
Baboons with Chronic Exposure to Lead Paint, Lead Acetate,
and Other Lead Compounds 4-66
Xlll
-------
LIST OF TABLES (cont'd)
Number
4-6 Percentage Increase in Blood Lead Levels in Juvenile
Baboons with Chronic Exposure to Lead Paint, Lead Acetate,
and Other Lead Compounds 4-67
5-1 User-Selected Entries for Integrated Exposure Uptake
Biokinetic Model Worksheet for Example 5-2, Child
Born in 1975 5-3
5-2 User-Selected Entries for Integrated Exposure Uptake
Biokinetic Model Worksheet for Example 5-2, Child
Born in 1975 5-3
5-3a Soil Lead Data Entry Worksheet for Child Exposed to
2000 /xg/g Since Age 0 (Birth) 5-6
5-3b Soil Lead Data Entry Worksheet for Child Exposed to
2000 /xg/g Since Age 1 5-6
5-3c Soil Lead Data Entry Worksheet for Child Exposed to
2000 /xg/g S;nce Age 2 5-7
5-3d Worksheet for Yearly Soil Lead Concentration for Hypothetical
Children Moving from a Residence Where Soil Concentration is
100 /xg/g to a Residence Where Soil Concentration is
2000 ^g/g 5-7
5-4 Predicted Annual Average Blood Lead Concentrations for
Hypothetical Children Moving from a Residence Where Soil
Concentration is 100 /*g/g to a Residence Where Soil
Concentration is 2000 /xg/g 5-8
5-5a Soil Lead Data Entry Worksheet for Child with Soil Abated to
100 /xg/g Since Age 0 (Birth) 5-9
5-5b Soil Lead Data Entry Worksheet for Child with Soil Abated to
100 /xg/g Since Age 1 5-10
5-5c Soil Lead Data Entry Worksheet for Child with Soil Abated to
100 /xg/g Since Age 2 5-10
5-5d Worksheet for Hypothetical Children in a Neighborhood
Where Soil Concentration is Reduced from 2000 /xg/g
to 100 /xg/g 5-11
xiv
-------
LIST OF TABLES (cont'd)
Number Page
5-6 Predicted Blood Lead Concentrations for Hypothetical
Children in a Neighborhood Where Soil Concentration
Is Reduced from 2000 jtg/g to 100 jtg/g 5-11
5-7 Neighborhood Identifiers and Distance from Stack
for Kellogg, Idaho, Study 5-14
5-8 Observed and Estimated Air, Soil, and Dust Lead
Concentrations for Use in Historical Exposure
Reconstructions in Silver Valley Communities 5-15
5-9 User-Selected Entries for Integrated Exposure Uptake
Biokinetic Model Worksheet for Example 5-5, Child
Born in Kellogg, Idaho, in 1983 5-16
5-10 User-Selected Entries for Integrated Exposure Uptake
Biokinetic Model Worksheet for Example 5-5, Child
Born in Smelterville, in Kellogg, Idaho, in 1983 5-17
5-11 User-Selected Entries for Integrated Exposure Uptake
Biokinetic Model Worksheet for Example 5-5 5-17
5-12 Effects of Geometric Standard Deviation on the
Probability of Exceeding 10 /xg/dL, Using Only
Default Exposure Parameters, for Children
Ages 24 to 35 months 5-19
5-13 Range Finding Run for Target Soil Lead Concentration 5-30
5-14 Focused Run for Target Soil Lead Concentration 5-30
5-15 Verification Run for Target Soil Lead Concentration 5-31
A-l Cells of Blood Lead Levels in 165 Midvale
Children, by Paint Removal Status, Age, and
Intervals of 250 pg/g in Soil and Dust Lead A-3
A-2 Geometric Mean and Geometric Standard Deviation
of Blood Leads in Cells or Groups, by Paint
Removal Status, Age, and Intervals of 250 /*/g
in Soil and Dust Lead A-8
A-3 Stem and Leaf Plot of Geometric Standard Deviation
for Midvale Children A-13
xv
-------
LIST OF TABLES (cont'd)
Number Page
A-4 Stem and Leaf Plot of Geometric Standard Deviation
for Midvale Children (Weighted by Degrees of Freedom) A-14
B-l Summary of Revisions to Lead Uptake Biokinetic Model
Software from Lead 0.2 to Lead 0.4 B-2
B-2 Summary of Revisions to Lead Uptake Biokinetic Model
Software from Lead 0.4 to Lead 0.5 B-3
B-3 Summary of Revisions to Lead Uptake Biokinetic Model
Software from Lead 0.5 to Lead 0.99d B-4
xvi
-------
LIST OF FIGURES
Number Page
1-1 Conceptual diagram of the movement of environmental lead into
and through the human body 1-4
1-2 Components of the Integrated Exposure Uptake Biokinetic Model,
showing environmental exposure sources and pathways,
absorption compartments, critical body tissue compartments, and
elimination pathways 1-8
1-3 Categories of application of the Integrated Exposure Uptake
Biokinetic Model 1-26
2-1 Schematic diagram of the overall functions of the lead model .... 2-1
2-2 Decision diagram for the air lead menu options 2-6
2-3 Decision diagram for the dietary lead menu options 2-8
2-4 Decision diagram for the drinking water lead menu options 2-10
2-5 Decision diagram for the soil/dust lead menu options 2-12
2-6 Decision diagram for the alternate lead source menu
options 2-16
2-7 Decision diagram for the absorption/bioavailability menu
options 2-18
2-8 Decision diagram for the multiple simulation menu options 2-22
2-9 Decision diagram for the batch mode menu options 2-25
2-10 Historical relationship between lead in gasoline and lead in air
in the United States 2-28
2-11 Integrated Exposure Uptake Biokinetic Model sample worksheet .. 2-47
3-1 Lead model menu tree 3-2
4-1 Schematic drawing of the enterocyte showing possible mechanisms
for lead absorption 4-4
4-2 Dose-dependent relationship between dietary lead (formula
mixed with water) and blood lead in infants 4-6
xvu
-------
LIST OF FIGURES (cont'd)
Number Page
4-3 The time-course of bioavailability of lead in the blood and
in the brain of juvenile rats following a single dose 4-11
4-4 Kinetics of absorption during repeated dosing 4-11
4-5 Under conditions of equilibrium, the amount of lead as the
free ion is limited by mass balance dissolution of the solid
phase 4-15
4-6 Under physiological conditions, free lead ion is removed from
solution by active and passive absorption mechanisms
potentially shifting the equilibrium of the dissolution
process far to the left 4-15
4-7 The impact of the relative positions of the level of concern
and the geometric mean on the proportion of children
"at risk" for two populations with different geometric
standard deviations 4-24
4-8 Probability density of blood lead in houses 1 to 4 4-33
4-9 Exposure pathways of lead in the environment 4-37
4-10 Biokinetic compartments, compartmental lead flows, and
uptake pathways in the Integrated Exposure Uptake
Biokinetic Model 4-60
XVlll
-------
LIST OF SCREENS
Number Page
2-1 The main menu 2-3
2-2 The general help menu 2-4
2-3 The information menu 2-5
2-4 The air lead menu 2-5
2-5 The dietary lead main menu 2-7
2-6 The alternative dietary source menu 2-9
2-7 The drinking water lead main menu 2-9
2-8 The age-specific drinking water consumption menu 2-11
2-9 The soil and dust main menu 2-11
2-10 The multiple dust source menu 2-13
2-11 The alternative indoor dust menu 2-14
2-12 The soil/dust ingestion rate menu 2-15
2-13 The alternate source lead menu 2-15
2-14 The absorption/bioavailability menu 2-19
2-15 The maternal/fetal lead exposure menu 2-19
2-16 Single simulation using the program processing menu 2-20
2-17 Multiple simulation using the program processing menu 2-21
2-18 Selection of media for multiple range run 2-23
2-19 Range selection during multiple processing 2-23
2-20 Using multiple simulation to find acceptable media concentrations
for a predetermined blood lead concentration 2-24
2-21 Running the model in batch mode 2-24
xix
-------
LIST OF SCREENS (cont'd)
Number
2-22 Data entry for air 2-30
2-23 Using dietary lead intake for a child born in 1983 2-34
2-24 Using dietary lead intake from local vegetables and
fish in Kellogg 2-35
5-1 Multiple runs probability density function for soil
lead = 1,000 pg/g, dust lead = 0 to 1,500 j*g/g, by
steps of 250 /ng/g (Runs 1 through 7) in Example 5-6 5-20
5-2 Multiple runs probability of exceedance of blood lead
levels for soil lead = 1,000 /*g/g, dust lead = 0 to
1,500 fig/g, by steps of 250 jwg/g (Runs 1 through 7)
in Example 5-6 5-21
5-3 Relationship of predicted blood lead to dust lead
in Example 5-6 5-22
xx
-------
TECHNICAL REVIEW WORKGROUP FOR LEAD
Harlal Choudhury
U.S. Environmental Protection Agency
Environmental Criteria and Assessment
Office
26 West Martin Luther King Dr.
Cincinnati, OH 45268
Barbara Davis
U.S. Environmental Protection Agency
(5204G)
401 M St. SW
Washington, DC 20460
Rob Elias
U.S. Environmental Protection Agency
(MD-52)
Environmental Criteria Assessment Office
Research Triangle Park, NC 27711
Susan Griffin (Chair)
U.S. Environmental Protection Agency
Region 8 (8 HWM-SM)
999 18th St., Suite 500
Denver, CO 80202
Karen Hogan
U.S. Environmental Protection Agency
(7404)
401 M St. SW
Washington, DC 20460
Mark Maddaloni
U.S. Environmental Protection Agency
Region 2
Emergency and Remedial Response
Division
26 Federal Plaza
New York, NY 10278
Allan Marcus
U.S. Environmental Protection Agency
(MD-52)
Environmental Criteria and Assessment
Office
Research Triangle Park, NC 27711
Roy Smith
U.S. Environmental Protection Agency
Region 3 (3 HW15)
Hazardous Waste Management Division
841 Chestnut St.
Philadelphia, PA 19107
Pat Van Leeuwen
U.S. Environmental Protection Agency
Region 5 (HSRLT-5J)
Waste Management Division
77 West Jackson Blvd.
Chicago, IL 60604
Chris Weis
U.S. Environmental Protection Agency
Region 8 (8 HWM-SM)
999 18th St., Suite 500
Denver, CO 80202
Paul White
U.S. Environmental Protection Agency
(8603)
Office of Health and Environmental
Assessment
401 MSt., SW
Washington, DC 20460
xxi
-------
GLOSSARY OF MODEL TERMS
Absorbed dose - The amount of a substance penetrating an absorption barrier (the exchange
boundaries) of an organism, via either physical or biological processes.
Absorption barrier - Any of the exchange barriers of the body that allow differential
transport of various substances across a boundary. Examples of absorption barriers are the
skin, lung tissue, and gastrointestinal tract wall.
Accuracy - The measure of the correctness of data, as given by the difference between the
measured value and the true or standard value.
Ambient - Surrounding conditions.
Ambient measurement - The measurement (usually of the concentration of a chemical or
pollutant) taken in an ambient medium, normally with the intent of relating the measured
value to the exposure of an organism that contacts that medium.
Ambient medium - One of the basic categories of material surrounding or contacting an
organism (e.g., outdoor air, indoor air, water, or soil) through which chemicals or pollutants
can move and reach the organism. (See biological medium, environmental medium.)
Arithmetic mean - The sum of all the measurements in a data set divided by the number of
measurements in the data set.
Background level (environmental) - The concentration of substance in a defined control area
during a fixed period of time before, during or after a data gathering operation.
Bias - A systematic error inherent in a method or caused by some feature of the
measurement system.
Bioavailability - The fraction of intake at a portal of entry into the body (lung, gut, skin) that
enters the blood. Bioavailability is typically a function of chemical properties, physical state
of the material that an organism ingests or inhales, and the ability of the individual organism
to physiologically absorb the chemical. The absorption rate varies widely by type of
substance and can greatly influence the toxicity of lead over that acute timeframe.
Biokinetics - processes affecting the movement of molecules from one internal body
compartment to another, including elimination from the body.
Biological measurement - A measurement taken in a biological medium. For the purpose of
exposure assessment via reconstruction of dose, the measurement is usually of the
concentration of a chemical/metabolite or the status of a biomarker, normally with the intent
of relating the measured value to the internal dose of a chemical at some time in the past.
xxin
-------
(Biological measurements are also taken for purposes of monitoring health status and
predicting effects of exposure). (See ambient measurement.)
Biological medium - One of the major categories of material within an organism (e.g., blood,
adipose tissue, or breath) through which chemicals can move, be stored, or be biologically,
physically, or chemically transformed. (See ambient medium, environmental medium.)
Body burden - The amount of a particular chemical stored in the body at a particular time,
especially a potentially toxic chemical in the body as a result of exposure. Body burdens can
be the result of long term or short term storage, for example, the amount of a metal in bone,
the amount of a lipophilic substance such as PCB in adipose tissue, or the amount of carbon
monoxide (as carboxyhemoglobin) in the blood.
Comparability - The ability to describe likenesses and differences in the quality and relevance
of two or more data sets.
Compartment - A distinct anatomical organ, tissue, fluid pool, or group of tissues within the
body that are regarded as "kinetically homogeneous."
Dose - The amount of a substance available for interaction with metabolic processes or
biologically significant receptors after crossing the outer boundary of an organism. The
potential dose is the amount ingested, inhaled, or applied to the skin. The applied dose is the
amount of a substance presented to an absorption barrier and available for absorption
(although not necessarily having yet crossed the outer boundary of the organism). The
absorbed dose is the amount crossing a specific absorption barrier (e.g., the exchange
boundaries of skin, lung, and digestive tract) through uptake processes; internal dose is a
more general term denoting the amount absorbed, without respect to specific absorption
barriers or exchange boundaries. The amount of the chemical available for interaction by
any particular organ or cell is termed the delivered dose for that organ or cell.
Environmental medium - One of the major categories of material found in the physical
environment that surrounds or contacts organisms (e.g., surface water, ground water, soil, or
air) and through which chemicals or pollutants can move and reach the organisms. (See
ambient medium, biological medium.)
Exposure - Contact of a chemical, physical, or biological agent with the outer boundary of an
organism. Exposure is quantified as the concentration of the agent in the medium in contact
integrated over the time duration of that contact.
Exposure pathway - The physical course a chemical or pollutant takes from the source to the
organism exposed.
Exposure route - The way a chemical or pollutant enters an organism after contact (e.g., by
ingestion, inhalation, or dermal absorption).
xxiv
-------
Exposure scenario - A set of facts, assumptions, and inferences about how exposure takes
place that aids the exposure assessor in evaluating, estimating, or quantifying exposures.
Geometric mean - The nth root of the product of n values. Also, the exponential function of
the mean or expected value of the natural logarithm of a variable.
Geometric standard deviation (GSD) - The exponential function of the standard deviation of
the natural logarithm of a variable.
Guidelines - Principles and procedures to set basic requirements for general limits of
acceptability for assessments.
Intake - The process by which a substance crosses the outer boundary of an organism without
passing an absorption barrier (e.g., through ingestion or inhalation). (See also "potential
dose").
Internal dose - The amount of a substance penetrating across the absorption barriers (the
exchange boundaries) or an organism, via either physical or biological processes.
Matrix - A specific type of medium (e.g., surface water, drinking water) in which the analyte
of interest may be contained.
Median value - The value in a measurement data set such that half the measured values are
greater and half are less.
Monte Carlo technique - A repeated random sampling from the distribution of values for
each of the parameters in a generic (exposure or dose) equation to derive an estimate of the
distribution of (exposures or doses in) the population.
Pathway - The physical course a chemical or pollutant takes from the source to the organism
exposed.
Pharmacoldnetics - The study of the time course of absorption, distribution, metabolism, and
excretion of a foreign substance (e.g., a drug or pollutant) in an organism's body.
Potential dose - The amount of a chemical contained in material ingested, air breathed, or
bulk material applied to the skin.
Precision - A measure of the reproducibility of a measured value under a given set of
conditions.
Probability samples - Samples selected from a statistical population such that each sample has
a known probability of being selected.
Random samples - Samples selected from a statistical population such that each sample has an
equal probability of being selected.
xxv
-------
Range - The difference between the largest and smallest values in a measurement data set.
Reasonable worst case exposure or risk range - The lower portion of the "high end" of the
exposure, dose or risk distribution. The reasonable worst case conceptually should be
targeted at above the 90th percentile in the distribution, but below about the 98th percentile
("maximum exposure or risk range").
Representativeness - The degree to which a sample is, or samples are, characteristic of the
whole medium, exposure, or dose for which the samples are being used to make inferences.
Risk - The probability of deleterious health or environmental effects.
Route - The way a chemical or pollutant enters an organism after contact (e.g., by ingestion,
inhalation, or dermal absorption).
Sample - A small part of something designed to show the nature or quality of the whole.
Exposure-related measurements are usually samples of environmental or ambient media,
exposures of a small subset of a population for a short time, or biological samples, all for the
purpose of inferring the nature and quality of parameters important to evaluating exposure.
Scenario evaluation - An approach to quantifying exposure by measurement or estimation of
both the amount of a substance contracted, and the frequency/duration of contact, and
subsequently linking these together to estimate exposure or dose.
Structural Equations Model - A statistical model of a process in which several regression
equations are solved simultaneously, and outputs or responses from one equation may be
used as inputs or predictors in another equation. Useful in pathway modeling.
Surrogate data - Substitute data or measurements on one substance used to estimate
analogous or corresponding values of another substance.
Uptake - The process by which a substance crosses an absorption barrier and is absorbed into
the body.
xxvi
-------
1. BEFORE YOU START
1.1 BACKGROUND: PURPOSE AND DEVELOPMENT OF THE
MODEL
The Integrated Exposure Uptake Biokinetic (IEUBK) Model for Lead in Children is a
stand alone, PC compatible software package. It allows the user to estimate, for a
hypothetical child or population of children, a plausible distribution of blood lead
concentrations centered on the geometric mean blood lead concentration predicted by the
model from available information about children's exposure to lead. From this distribution,
the model calculates the probability that children's blood lead concentrations will exceed the
user selected level of concern (default 10 /xg/dL). The user can then explore an array of
possible changes in exposure media that would reduce the probability that blood lead
concentrations would be above this level of concern.
The model should be viewed as a tool for making rapid calculations and recalculations
of an extremely complex set of equations that includes scores of exposure, uptake, and
biokinetic parameters. This Guidance Manual concisely describes key features of the
conceptual underpinnings of the IEUBK model, its evolution and development, its
capabilities, and its limitations. The Manual then goes on to offer guidance on the use of the
model as a risk assessment tool while cautioning against a number of possible misapplications
of the model. A detailed description of the equations and parameters used in the model is
provided in the Technical Support Document: Parameters and Equations Used in the
Integrated Exposure Uptake Biokinetic Model for Lead in Children (a companion document
to this Guidance Manual).
1.1.1 Description of the Model
The IEUBK Model is a simulation model. As a risk assessment tool, it can be a useful
component of remediation strategies for lead in the human environment. The simulation of
childhood lead exposure and retention is only one part of the risk assessment process. It is
important to note that the model alone does not determine the level of cleanup required for a
specific site. Rather, it predicts the likely blood lead distribution for children given the
exposure to lead at that site, and the probability that children exposed to lead in that
environment will have blood lead concentrations exceeding a health-based level of concern.
1-1
-------
Blood lead concentrations are not only indicators of recent exposure, but also are the
most widely used index of internal lead body burdens associated with potential health effects.
Health effects of concern have been determined to be associated with childhood blood lead
concentrations at or below 10 /ig/dL (U.S. Environmental Protection Agency, 1986, 1990;
CDC, 1991). The probability that children will have blood lead levels exceeding this level
of concern is an important consideration for a risk assessor in compiling and evaluating all
information applicable to a site to enable remediation decisions.
The IEUBK model can be applied at several different scales of application, but the
interpretation of the model output and the form of the model or subsequent risk estimates is
different for each application. In most uses of the model, a site is a spatial domain that is
appropriate for remediation decisions, typically a residential yard with a single housing unit,
or an equivalent area for multi-unit buildings or for undeveloped lots. The home and its
surrounding yard is the basic unit for risk analysis because lead exposure for pre-school
children commonly occurs within this domain. In Sections 1.4.4.2 and 4.2 we will describe
an array of applications of the IEUBK model based on aggregating clusters of sites. The
array is:
A: One location
Al: one living unit, one child;
A2: one living unit, more than one child;
A3: more than one living unit, more than one child, homogeneous media
concentrations;
B: Multiple locations, one neighborhood, homogeneous media concentrations
C: Multiple locations, one neighborhood, heterogeneous media
concentrations;
D: Multiple locations, more than one neighborhood, heterogeneous media
concentrations;
In category A, risk is calculated as the probability that, in a single child at a single site
with the specified exposure scenario, the child's blood lead concentration will exceed the
level of concern. The probability distribution describes the likely variability in blood lead
for a child with a given exposure scenario. The best single-number prediction of blood lead
concentration is the geometric mean of the distribution of blood lead concentrations that may
occur for a child with the specified exposure scenario. This single-child assessment is used
to evaluate remediation options on a house-by-house or yard-by-yard basis.
1-2
-------
In categories B, C, and D, a frequency distribution of the individual risk of exceeding a
blood lead level of concern is obtained. The percentage of children in multiple sites that are
likely to have a blood lead concentration exceeding the level of concern can then be
calculated. For category B, where all children of the same age have the same exposure
scenario, this can be done with a single run of the IEUBK model. For categories C and D,
where distinct subgroups have different exposure scenarios, risk must be calculated by
aggregating the results from a number of model runs. Risk estimation for more than one
neighborhood, for category D, has the added complication that a variety of model parameters
may differ between neighborhoods, and within each neighborhood. Therefore, environmental
lead concentrations may differ between neighborhood subgroups.
1.1.2 Simulation of Childhood Lead Exposure and Retention
Lead is a naturally occurring nonnutrient metal that follows environmental pathways
similar to those of nutrient metals such as calcium. In the human environment, these
pathways or routes of exposure transfer lead from sources such as food, drinking water, air,
soil, and dust, to the human body by means of ingestion or inhalation. There are important
analogies to be made between lead and calcium that contribute to our understanding of the
biological behavior of lead. These analogies have aided in the formulation of the lead
model. In particular, the nature of gut absorption of lead and calcium may be similar.
Childhood growth and development of bone and soft tissue which require calcium influence
the uptake of environmental lead from the gut. In addition to similarities in absorption, both
lead and calcium are stored in quantity and subsequently released from bone tissue.
Shown conceptually on Figure 1-1, inhaled or ingested lead is absorbed through the
lungs or gut into the blood stream where it is transferred to body tissues, including bone
tissues. After a period of time, this lead returns to the blood stream where it is transferred
to other tissues or eliminated with urine. Lead may also be eliminated from the body with
sweat, hair or sloughed epidermal tissue, or it may be transferred through the liver and bile
duct back to the gut where it passes out of the body with feces.
In Figure 1-1, the oval shapes show environmental lead media, and some of the
pathways between them. The large rectangle shows the compartment that is central to lead
distribution in the child, the blood plasma pool and associated extra-cellular fluid. Each
lower rectangle shows a compartment in the child's body where lead may be retained. The
excretion of lead from the body is shown by the circles.
1-3
-------
Air
Lung
Gut
*\
Extra
Plasma -- Cellular
Fluid
Feces
Figure 1-1. Conceptual diagram of the movement of environmental lead into and
through the human body. The oval shapes show environmental media and
the pathways of uptake. The large rectangle is the blood plasma
compartment central to the distribution of lead in the body.
The foundation of the present IEUBK model is the construction of a detailed and
thorough exposure scenario for children aged 0 to 84 months that can be adjusted to match
the exposure of any child. The user starts with exposure information specific to these
children and accepts generalized assumptions about any additional information required to
complete the exposure scenario. The site-specific information usually consists of
environmental media concentrations such as soil lead concentrations.
The model inserts default values whenever site-specific information is not used. The
default values (e.g., dietary lead concentrations and consumption values) are typical of a
child's environment hi the sense that they are broad-based estimates of the expected
1-4
-------
environment of a child. These default values are not necessarily appropriate for every site
and should be reviewed by the user for every site-specific application.
This model uses standard age-weighted exposure parameters for consumption of food,
drinking water, soil, and dust, and inhalation of air, matched with site-specific concentrations
of lead in these media, to estimate exposure for the child. The model simulations represent
chronic exposure and do not incorporate the variability in consumption patterns and media
concentrations on a daily or seasonal basis. The model includes continuous growth of the
child and simulates the changing environment of the child on a yearly basis. In theory, the
exposure component of the model would apply to a single child or to any number of children
with the same lead exposure scenario. With the proper substitution for media concentrations,
the exposure component (but not the biokinetic component) would also apply to any other
substance with sources and pathways of exposure similar to lead.
The model simulates lead uptake, distribution within the body, and elimination of lead
from the body. The uptake portion of the model takes into consideration two mechanisms of
absorption of lead in the digestive tract: saturable and non-saturable. Elimination of lead is
modeled through several routes: urine, gastro-intestinal excretion, and sloughing of
epidermal tissue, including hair and nails.
1.1.3 Historical Evolution from Slope Factor Models to the IEUBK Model
An explicit mathematical method for estimating the likely risk of elevated blood lead
concentrations in young children has previously been used by the Environmental Protection
Agency as one of its tools for developing the National Ambient Air Quality Standard for
Lead and the National Primary Drinking Water Regulation for Lead. The method has
historically been based mainly on an estimation of relationships between lead concentrations
in children's blood and lead concentrations in specific individual environmental media such
as air, water, soil and dust, based on empirical observations derived from experimentally
controlled human exposure, animal toxicological studies, and epidemiological analyses. Such
relationships also provide a basis for estimating the probability that elevated blood lead
concentrations exceed a level of concern due to exposure to environmental lead in these
media.
A mathematical approach of this type was used to evaluate potential alternative air lead
standards based on health effects criteria (U.S. Environmental Protection Agency, 1977,
1-5
-------
1978, 1989a). The relationship between blood lead and lead in environmental media was
estimated statistically, both for adults and children (U.S. Environmental Protection Agency,
1986, 1989a). While the relationship was somewhat non-linear at blood lead concentrations
above about 40 /xg/dL in adults and 30 /wg/dL in children, it was nearly linear at lower blood
lead concentrations of interest. The relationship between blood lead and environmental lead
concentrations in different media (air, water, soil, dust, food) was estimated using a model
linear in lead concentrations. The linear regression coefficients between blood lead and lead
in each of the environmental media have since become known as the slope factors for the
media.
As more evidence has become available, it has become clear that these slope factors can
not be regarded as universal constants that are the same everywhere, for all children at all
sites. Some of the problems involved in the use of slope factors have been discussed by the
U.S. Environmental Protection Agency (1989a) and by Brunekreef et al. (1984). In the
development of unproved lead models (U.S. Environmental Protection Agency 1986, 1989a),
the following points were discussed:
(1) Slope factors are a function of many factors: media ingestion rates;
bioavailability and absorption of lead from the medium; and biological
kinetics of lead retention and elimination in the child. Biological and
physical differences between sites and study populations cannot be
incorporated explicitly and quantitatively into regression slope factors
from different studies.
(2) Slope factors for a single medium, such as lead in air or lead in soil, may
provide only a very incomplete picture of total lead exposure from a
particular source, even if the source is identified with the medium.
A single medium such as household dust may contain lead from many
sources, and lead from a single source such as exterior lead-based paint
may contribute to several exposure media pathways to the child.
Therefore, in 1985, the EPA Office of Air Quality Planning and Standards (OAQPS)
initiated a project that would allow the calculation of blood lead concentrations in children
exposed to differing arrays of concentrations of lead in air, soil, and dust. This model,
called the Uptake/Biokinetic (or UBK) model for lead, was a computer simulation model
based on the biokinetic model for lead in children developed by N. Harley and T. Kneip
(1985). The biokinetic parameters for the UBK model were extrapolated from long-term
feeding studies on infant and juvenile baboons (Mallon, 1983), autopsy data on human
children, human infant feeding studies, and other sources. The exposure model that was
1-6
-------
coupled to the biokinetic model was developed by OAQPS. Model calibration and validation
was done using data from the 1983 EPA/CDC/Montana study on children in East Helena,
Montana, who lived in the vicinity of a large primary lead smelter. The modeling approach
was reviewed and approved by EPA's Clean Air Science Advisory Committee (CASAC) in
1990.
The overall framework of both the UBK and IEUBK models is shown in Figure 1-2.
The oval shapes show environmental lead concentrations and the funnel-shaped symbols show
lead intake from the environment at the portals of entry, the lung and the gut. These are the
exposure/intake components of the IEUBK model. The next large rectangle shows the gut
not only as the main portal of entry for lead from most exposure media, but also as the site
for key absorption/uptake components of the IEUBK model for the evaluation of lead from
soil, dust, diet, and drinking water. The very large rectangle shows the child's blood lead,
partitioned into plasma-extracellular fluid and red blood cells. The two boxes to the right of
the blood lead pool sketch the bone and soft tissue pools, and the elimination pathways are
shown as circles. The right-hand box shows the blood lead concentration in the child, and
the subdivisions show the estimated contribution of each medium to the child's blood lead
concentration. In the example in Figure 1-2, we have assumed that all external lead media
have been used in the IEUBK model, as have all internal lead sources. There is no
unattributable component called "background". The attribution of specific fractions of blood
lead to uptake from specific media is not as subject to statistical artifacts, since pathways
from soil lead and air lead to dust lead are also included in the IEUBK model.
In all particulars, the present version of the model, the IEUBK model, may be
considered an enhancement and extension of the UBK model. Theoretically, in situations
where the child has constant long-term or chronic lead exposure, both the slope factor
approach and the UBK model (now the IEUBK model) should produce similar results when
sufficient data exist to correctly characterize lead exposure, absorption, and biokinetics.
The IEUBK Model addresses three emerging paradigms of environmental risk
assessment.
(1) Assessments that recognize the multimedia nature of exposures to
environmental toxicants are a significant improvement in assessing health
risks. Assessments restricted to single pathways of exposure can overlook
situations where integrated multimedia exposures are high enough to
trigger health concerns. The lead model is structured to integrate
exposures occurring through air, water, food, soil, and dust in estimating
the blood lead levels in children in realistic environmental settings.
1-7
-------
Cortical and
Trabecular
Bone
Plasma+
Extra-
Cellular
Fluid
Red
Blood
Cells
From Air
From Dust
From Soil
From Water
From Diet
From Other
Liver,
Kidney,
Other
Soft
Tissues
Environment
Exposure/ Absorption/
Intake Uptake
Blood
Tissues Elimination
Biokinetics
Blood
Lead
Figure 1-2. Components of the IEUBK Model, showing environmental exposure sources
and pathways, absorption compartments, critical body tissue
compartments, and elimination pathways.
(2) Pharmacokinetic information can strengthen the validity of environmental
health assessments in comparison with more traditional methods that
address only external dose or intake of a compound. Internal measures of
dose that are pertinent to the biological effects exerted by a compound
form an improved metric for risk assessment. The IEUBK estimates of
blood lead concentrations as an internal indicator of potential health risk
are based on pharmacokinetic modeling of lead absorption, transport,
redistribution, and elimination.
(3) Environmental assessments need to address the substantial variability in
exposure and risk resulting from these factors. Single point estimates of
exposure or risk are of limited utility. Individuals differ in their
surroundings, behavior, and physiological status. The Lead Model
addresses variability through the estimation of probability distributions of
blood lead levels for children exposed to similar environmental
1-8
-------
concentrations of lead. Through systematic application of the model, data
on the variability of levels of environmental lead contamination can be
translated into estimates of the distribution of blood lead levels within
populations of children.
1.1.4 Using the IEUBK Model for Risk Estimation
The IEUBK Model for lead is designed to facilitate: (a) rapid delineation of the
relationship between environmental lead and blood lead in children; and (b) calculation of the
risk of elevated blood lead (i.e., the probability of a given child or a group of children
having blood lead concentrations exceeding a specified level of concern). As such, the
IEUBK Model provides a tool for site-specific risk assessment for young children exposed to
lead from different media and through different pathways in their environment, with
particular emphasis on lead in air, water, soil, and household dust. Many other applications
are possible. The intended applications of the IEUBK model are to:
(1) Provide a summary of children's long-term, primarily residential,
exposure to lead;
(2) Provide a best estimate of the geometric mean blood lead concentration
for a typical child aged 6 to 84 months, assumed to reside at a given
residence;
(3) Provide a basis for estimating the risk of elevated blood lead (i.e., for
exceeding a designated blood lead concentration of concern) for a
hypothetical child of specified age with given site-specific residential lead
exposure;
(4) Provide a basis for estimating the risk of elevated blood lead
concentrations among early pediatric populations in a given neighborhood
by aggregating the individual residential risk estimates;
(5) Predict likely changes in risk of elevated blood lead concentrations from
exposure to soil, dust, water, or air lead following abatement actions
designed to reduce exposure levels from one or more environmental
media;
(6) Provide assistance in determining appropriate soil or dust lead target
cleanup levels at specific residential sites;
1-9
-------
(7) Provide assistance in estimating blood lead concentrations associated with
soil or dust lead concentrations at undeveloped residential sites that may
be developed in the future.
Each of these applications is discussed in more detail in Chapters 2, 4, and 5. The IEUBK
model has been used for many purposes in addition to those for which it was originally
intended. We are sure that the IEUBK model will continue to be used in many unintended
and unexpected applications, just like any other new tool that has multiple uses. Some of
these new applications are valid, others are demonstrably invalid, and the validity of many
applications is simply unknown.
The risk estimates are calculated for a hypothetical child or a hypothetical population of
children who could be occupying the specific household at the time of the measurements or
at some future time. The IEUBK model can therefore be used to estimate the risk of
elevated blood lead even when there are no children currently living at a house, or if there
exist only environmental lead data for the dwelling unit. The model does not require that a
neighborhood or community blood lead study be carried out. The user should be aware that
a site-specific risk assessment requires site-specific soil and dust concentrations, and some of
the absorption parameters may depend on specific characteristics of the soil and dust at the
site. The IEUBK model accepts user inputs for site-specific differences in bioavailability of
lead in different media, and site-specific differences in environmental lead pathways for
different lead sources.
1.1.5 Validation of the IEUBK Model
What does it mean to say that a computer simulation model is "valid"? In general, we
interpret this to mean that:
• the model is biologically and physically plausible and incorporates the best
available empirical data on parameters;
• the model uses numerically accurate algorithms and the accuracy of the
computer codes for these algorithms has been verified;
• the model provides some satisfactory empirical comparisons of model
output with real-world data.
We believe that the scientific basis and computational correctness of the IEUBK Model is
sound, and that the IEUBK model provides valid prediction of observed blood lead
1-10
-------
concentrations from representive populations of children with typical exposure. The
empirical comparisons in which there are differences between observed and predicted blood
lead concentrations underscore the importance of valid exposure scenarios as input. They
also show the importance of valid blood lead data from truly representative population
sampling methods when interpreting these empirical comparisons.
1.1.5.1 The Model Is Biologically and Physically Plausible
The parameters and equations used in the model are documented in the Technical
Support Document: Parameters and Equations Used in the Integrated Exposure Uptake
Biokinetic Model for Lead in Children. The exposure model component is based on data for
human children in most instances, with lead exposures that are characteristic of children in
the U.S. since about 1980. The ingestion parameters are based on surveys for drinking
water and tap water (Ershow and Cantor, 1989), market basket estimates of dietary intake
(Pennington, 1983; Gartrell, 1986), and on observational studies of soil and dust ingestion
for children in the U.S. (Binder et al., 1986; Calabrese et al., 1989, 1992a,b, 1993; Davis
et al., 1990). While these studies have not resolved all of the uncertainty in childhood lead
exposure, especially from sources such as lead-based paint, they have provided a much more
realistic basis for quantitative modeling. The exposure component of the IEUBK model
extends the UBK model assumptions (U.S. Environmental Protection Agency, 1989a) that
have been reviewed by CAS AC (1990).
An absorption component was developed for the IEUBK model based on evidence
discussed in Section 4.1. This evidence includes in vivo data in infant and juvenile baboons
and human infants whose intake of lead is observed and known (Mallon, 1983; Sherlock and
Quinn, 1986). The model has two modes for absorption, saturable and non-saturable. In the
non-saturable mode, absorption of lead is a constant fraction of the total lead ingested for a
specific medium. The saturable mode follows the Michaelis-Menten kinetics for saturable
absorption as proposed by Aungst and Fung (1981). Development of the algorithm is also
based on data from lead balance and feeding studies in human infants and children
(Alexander, 1974a,b; Ryu et al., 1983, 1985; Ziegler et al., 1978).
The compartmental structure of the earlier biokinetic model is based on compartmental
models for lead in adults as discussed in detail in the Air Quality Criteria Document for Lead
(U.S. Environmental Protection Agency, 1986). The model was verified and extended based
on studies in infant and juvenile baboons (Mallon, 1983) whose age (5 to 26 months) and
size (2.5 to 6 kg) are only slightly smaller than those of human children. The biokinetic
distribution and elimination parameters use ratios of lead concentrations in tissues and blood
1-11
-------
following chronic exposure. The ratios of lead concentrations in tissues of human children
from autopsy data (Barry, 1975, 1981) were used to adjust the baboon's biokinetic
distribution parameters to human infants and children (Harley and Kneip, 1985). The
biokinetic parameters for baboons were re-estimated using the compartmental structure of the
current IEUBK model (Marcus, 1992). The tissue-to-blood concentration ratios from the
human child autopsy data were incorporated in the IEUBK model, assuring complete
consistency with the best available data.
1.1.5.2 The Model Is Computationally Accurate
The IEUBK model uses a fast and accurate one-step numerical integration method
known as 'backward Euler', with user-adjustable time steps to verify numerical accuracy of
the solution. Coding of the model equations was verified by a separate receding of the
model in another programming language. Independent code verification will be described in
forthcoming Technical Memoranda (see Section 1.2.2).
1.1.5.3 Empirical Comparisons of the Model
Comparison of the IEUBK model output with empirical human blood lead data has two
requirements. The first requirement is that the child's total lead exposure is completely and
accurately characterized by the empirical data, including site-specific data on environmental
lead concentration, media ingestion, and bioavailability. The second requirement is that the
blood lead data from the field study are accurate and typical for that exposure scenario.
A typical child may not have the exposure described by the measured and default parameters
of the model, or a child may also respond atypically to the measured and default parameters.
The solution is to find the correct set of parameters (measured or site-specific alternatives to
default) that describes the child's site-specific exposure or response to exposure.
Environmental lead concentrations and blood lead measurements are subject to
measurement errors such as repeat sampling variability and analytical error. Without careful
attention to quality assurance/quality control (QA/QC) procedures, there may be systematic
biases in blood lead measurements. The results of the blood lead field study may also differ
from the model predictions for typical children if the blood lead sample is not representative
of the population being sampled.
Validation by empirical comparisons with paired data sets of good quality is an ongoing
process. In earlier versions of the model, empirical comparisons indicated satisfactory
agreement between observed and predicted blood lead concentrations. Several data sets have
been identificed that are of adequate data quality for evaluating the validity of the IEUBK
1-12
-------
Model, and more data sets are expected to become available in the future. The Field Study
Data Set Comparisons document referred to in Section 1.2.2 will discuss the results of these
analyses. Comparisons of empirical data with the IEUBK model require appropriate site-
specific exposure scenarios, valid assumptions about bioavailability, and demonstrated
representativeness of the sample of children recruited into the study in relation to the target
population from which they were drawn.
Our preliminary analyses of several data sets so far indicate that the model satisfactorily
predicts blood lead concentrations for the overall sample populations in specific
neighborhoods. Further analyses will be needed to determine if empirical comparisons are as
strong for subpopulations defined by factors such as differences in age, differences in contact
or behavior that affected the amount of soil ingested, suspected or possible differences in
bioavailability, differences in contribution of soil to household dust, and identifiable biases in
recruitment of children. More extensive evaluation of these data sets will be described in the
Field Study Data Set Comparisons document described in Section 1.2.2.
Careful determinations should be made by users with regard to how well default values
specified by this manual for key exposure and demographic parameters apply to the particular
sample of children (or subpopulations) being evaluated. Appropriate adjustments made in
pertinent default values may notably improve the fit of the model to empirical data. We
caution the user not to arbitrarily select alternate values for the default parameters, but rather
to obtain site specific or population specific data on important parameters.
1.2 ORGANIZATION OF THE MANUAL
1.2.1 Increasing Levels of Guidance and Technical Assistance
This manual is designed to provide you with the information you need at several levels
of detail. The further you read into manual the more specific guidance you will find for
using the model. By the time you have finished reading Chapter 1, you should have a
general understanding of how the model works and what it can do. You may want to install
the model and then work your way through Chapter 2 as you become more familiar with
each feature of the model. Instructions for installing the model are found in Section 2.4.
As you explore the various features of the model, you will become familiar with the
menus and their options. An overview of the menu system is in Section 2.1, and a detailed
1-13
-------
description of these menus can be found in Section 2.2. This is the section that the novice
user will want to follow closely. In a guided tour through the menu system, you will find
that each menu option becomes a part of the process of constructing a model "run," and that
these runs may be as simple as determining the blood lead concentration using only default
exposure conditions, or as complicated as neighborhood risk estimation calculated as the sum
of individual risks. Many of these options were suggested by comments received during the
extensive review of drafts of this Guidance Manual.
As you begin to apply the model to a specific risk assessment situation, you will find
that Section 2.3 contains detailed recommendations for building an exposure scenario. This
section also contains a helpful worksheet for planning model runs. Follow this section
closely, as it contains many helpful suggestions on the appropriate use of the model, as well
as warnings of improper applications. In Chapter 4, you will find a detailed discussion on
assessing the relationship between soil/dust lead and blood lead. This chapter also describes
the biokinetics of the model and specific issues in the use of the model for the ingestion of
paint chips. If you need more help, turn to Chapter 5, where several specific examples are
available to guide you through some of the more complicated procedures. As you become
more experienced, you will find Chapter 3 a quick and ready reference to the various menu
options. This chapter also contains a comprehensive review of default parameters.
1.2.2 Additional Documentation
Additional technical documents are or soon will be available to supplement the IEUBK
Model and this Guidance Manual. These are:
• Technical Support Document: Parameters and Equations Used in the
Integrated Exposure Uptake Biokinetic Model for Lead in Children—a
description and documentation of all equations and parameters in the model;
• Field Study Data Set Comparisons—a description of several validation
exercises that have been or will shortly be carried out;
• Sampling Manual—approaches and protocols for environmental and
biological sampling for collection of data compatible with the IEUBK
Model;
• Technical Memoranda—occasional technical updates that will be released to
explain some features in greater detail or to alert the user to possible
misapplications of the model.
1-14
-------
1.3 GETTING READY TO USE THE MODEL
1.3.1 Preparing a Site-Specific Exposure Scenario
The use of the IEUBK model requires input data that are appropriate to the site(s) and
subject(s). The most convenient way to do this is to construct a multi-media, site-specific
exposure scenario using the exposure scenario worksheet (Figure 2-11; see Section 1.4.3).
For most assessments of lead-contaminated soils, the minimal site-specific data are the
soil lead and indoor dust lead concentrations for the residential exposure unit. Additionally,
it would be helpful to include estimates of specific exposures from diet, drinking water, air,
maternal exposure, or other sources that could replace the default exposure parameters
believed to be of concern at the particular site.
There may be potentially important differences among sites, and predictions of blood
lead values are expected to become more accurate as more site-specific data are added.
Children at highest risk are those with the highest exposures to some lead-containing
medium. Data should be collected at a site so as to identify locations in the residence or
community where young children may be exposed to elevated levels of lead in soil, dust,
water, or air. Household-level data are useful because proposed soil, dust, and paint
abatements are usually based on the house and yard as the most likely sources of lead
exposure in preschool children. High exposures from lead in the household water
distribution system are also possible, and this source has been identified in some childhood
lead poisoning cases (Cosgrove et al., 1989). The preferred level of environmental input
data for the model can be derived from a comprehensive multimedia household
environmental lead study.
The households studied should be representative of housing or sites where young
children currently reside, as well as the places where young children may live in the future.
In many applications, you will also need to include existing homes not occupied by children.
These can usually be addressed in the same manner as housing currently occupied by
children, using specific measurements of various environmental media lead concentrations.
Risk assessments addressing as yet unbuilt housing should use existing residential site soil
concentration data.
Predictions of blood lead concentrations may improve with better information on lead
concentrations where the child spends time during the day, or on child-specific behavior.
1-15
-------
Activity pattern analysis, based on data taken from questionnaires and family interviews, can
be useful in identifying children currently at risk, and in determining site-specific differences
in behavior or access to lead sources that may differ in bioavailability. Public education and
parental awareness of lead hazards may reduce the amount of lead in soil and dust ingested
by the child, and quantitative studies of the effects of such actions are currently in progress.
Exposure of children in day care centers, playgrounds or open areas may substantially
affect total exposure to lead when potential lead exposures in such areas are high. These
cases need to be considered in site risk assessment. The IEUBK Model allows dust and
drinking water ingestion components to be separated into household and non-household
sources by allocating a percentage of dust and water intake outside the home to sources with
other concentrations. Time-weighted average air lead exposures are believed to be adequate
indices of lead intake by inhalation in home and non-home settings under most circumstances
and are used in the model. However, there is presently little information on the use of time-
weighted averages for ingestion of soil, dust, or water away from the home. Soil and dust
ingestion depends on children's activities, on hand-to-mouth behavior, and on intensity of soil
contact related to sources and pathways away from home.
In addition to exposure, the IEUBK Model also allows site-specific information on the
bioavailability of lead from various sources to be taken into account. Bioavailability
describes the relationship between the potentially available lead intake from environmental
media and the amount of lead entering the body through the lungs or the gut and then into
systemic circulation.
You should be alert to the possibility that there may be site-specific differences in
bioavailability of lead at different sites, particularly with respect to soil and paint. Some
factors that may affect bioavailability include chemical speciation of lead in soil or paint, size
of particles, mineral matrix of the particles, and whether the particles are: likely to be
ingested by the child along with meals or on an empty stomach. These are discussed in
Section 4.1. Many of the issues are subtle and should be referred to the EPA Technical
Review Workgroup for Lead.
In some cases, relatively non-available environmental lead in soil or paint can be
converted into readily available lead particles in household dust by physical and chemical
processes in the environment. A housing unit with lead in paint or soil will continue to
generate household dust lead exposure as long as paint deteriorates or is disturbed by
1-16
-------
remodeling, and as long as outside soil and surface dust are moved into the house by pets
and by human activities like gardening and remodeling.
The model default value for the Geometric Standard Deviation (GSD) (reflecting
variability among individuals who have contact with a fixed lead concentration) is based on
analyses of data from neighborhoods having paired sets of environmental concentration and
blood lead data. The recommended default GSD of 1.60 is believed to be very widely
applicable. Only when reliable site-specific paired data from a sufficiently large study are
available, should the substitution of a site-specific GSD be made using guidance given in
Section 4.2.
1.3.2 Understanding How the Biokinetic Component of the Model Works
The general term "biokinetic" is used to describe the movement of lead through various
parts of the human body as a kinetic process. Current blood lead concentrations depend on
prior exposure history as well as present exposure. With constant lead exposure, a near
steady-state blood lead concentration level is achieved because there is a dynamic near-
equilibrium between lead moving out (from blood plasma to peripheral tissues and through
excretory routes), and lead moving in (to plasma from gastrointestinal uptake and
remobilization into plasma from peripheral tissues and long-term bone storage).
The IEUBK Model assumes that skeletal lead turnover occurs relatively more rapidly in
children than in adults. The lead in a child's blood is thus a mixture of lead taken up from
recent environmental exposure and lead released from skeletal stores that reflect historical
exposures. However, the faster turnover time assumed for children compared to adults
implies that the lead burden in the skeleton is a smaller fraction of total body burden in
children than in adults. The skeletal contribution to blood lead thus increases as the skeletal
fraction of total body burden of lead increases.
The blood lead concentrations in children achieve nearly a steady state relationship with
exposure within a period of months after changes in exposure. The situation in children is
more complicated than in adults because the kinetic parameters also change with the child's
growth and with changes in behavior that affect lead intake, absorption, distribution, and
elimination. The model is adequate to estimate childhood blood lead concentrations in near-
equilibrium or in slowly changing exposure settings, as may be attained some time (months)
after abatement occurs. The gradual phase down of lead in gasoline would be an example of
1-17
-------
changes that occurred slowly enough in most urban areas to permit accurate modeling of
blood lead concentration changes accompanying the air lead concentration changes.
1.3.3 Understanding Limitations of the Model
The IEUBK Model is designed to evaluate relatively stable exposure situations, rather
than rapidly varying exposures. The model does not report each iterative calculation; rather,
it reports one-year average blood lead concentrations. Because the IEUBK Model allows
changes in exposure to environmental lead concentrations only at one year intervals, and
provides output at only one year age intervals, changes in exposure are smoothed over one
year. The model cannot be used to predict the effects of short term exposure episodes, such
as exposure over a few days or weeks to lead dust and airborne particles that may be
generated during lead paint abatement. The IEUBK Model should provide reasonable
accuracy for blood lead concentration prediction as long as the changes in these
environmental lead concentrations can be approximated by annual average values.
The model is intended to describe a single residential-level exposure setting. The
dwelling unit could be a detached single-family home, a separate home in a multiple-unit
building such as a row house or duplex, or an apartment in a multiple-unit building. There
is an implicit assumption that the input parameters characterize long-term residential
exposure scenarios in such settings. While exposure changes daily in response to changes in
the child's diet and activity, there is presumably a true mean exposure level that can, in
principle, be estimated from real-life samples. For this reason, the IEUBK model allows
changes in air, food, dust, and soil lead exposure input parameters only at 1-year intervals.
Although water lead exposure could, in principle, be handled in similar detail, the IEUBK
model does not allow annual changes in drinking water lead during the model run. The
IEUBK model includes some capabilities for dealing with lead exposures outside the home,
such as by use of separate dust ingestion parameters and concentrations at day care centers,
schools, and secondary residences.
We recommend using a simple average or arithmetic mean of soil lead concentrations
from a representative area in the child's yard, and an average of dust lead concentrations
from representative areas frequented by children inside the house. This rationale is
appropriate for areas that are sufficiently small so that any part of the area may be accessible
to a typical child living at a random residence located within the area.
1-18
-------
The IEUBK model calculates blood lead and tissue lead burdens for all ages from 0 to
84 months. However, the blood lead concentrations in children less than 6 months of age
will still be affected by pre-natal lead exposure and are likely to show little influence from
exposure to soil, dust, and paint, which are the media currently of greatest interest. The
results of the model simulation are therefore not reported for children younger than
6 months.
There are many reasons why individual blood lead concentrations may differ from the
predicted geometric mean blood even though the predicted mean accurately describes the
population. Some of the components of individual differences are discussed in Section 4.2.
The GSD is the only parameter in the model that characterizes the combined variability in
blood lead attributable to inter-individual differences and "random" temporal variability in
absorption and biokinetics, "random" behavioral changes and inter-individual differences
affecting ingestion rate, and measurement errors in environmental lead concentration. The
strength of this approach is that GSD estimates are based on empirical data on the variability
of blood lead levels in children exposed to similar concentrations of lead. Other approaches
to evaluating the effects of variability, such as Monte Carlo simulation, were deferred for the
present version of the IEUBK Model, because they demanded excessive computation and
require much greater amounts of model input data. Monte Carlo methods, however, are still
being evaluated as a possible enhancement of the IEUBK model, as discussed in Section 1.5.
1.4 RUNNING THE MODEL
1.4.1 Your Responsibilities
The IEUBK model provides a great deal of flexibility in describing site-specific or age-
dependent exposure scenarios. The price for this level of flexibility is that no exposure
scenario is appropriate for every application of the IEUBK model, and this is particularly
true of the "default" parameters. The responsible use of the IEUBK model requires input
data that are appropriate to the site(s) and subject(s). The most convenient way to do this is
to use the exposure scenario worksheet (Figure 2-11; see Section 1.4.3).
The most sensitive parameters for most applications involving soil lead exposure are the
soil-to-indoor dust transfer coefficient, the soil and dust ingestion parameters, the soil lead
absorption fraction, and the Geometric Standard Deviation. You should always review these
parameters.
1-19
-------
Factors affecting transport of soil lead into household dust should be noted when
appropriate. For example, houses with very small grass-covered yards are likely to have a
smaller contribution of the yard's soil lead concentration to household dust lead concentration
than houses with large yards, no grass cover, and fine uncompacted surface soils that are
easily blown or carried into the house by humans and outdoor pets. While the concentration
of lead in exterior dust derived from the soil may be a useful measure of exposure, these
data are not usually available because exterior surface dust samples are not usually collected.
You are always responsible for the decision to use default values in ^ace of either measured
dust lead concentrations or dust lead concentrations estimated from soil lead concentrations.
The proportion of intake in the form of soil vs. dust should be considered carefully, as
there may be differences in the bioavailability of lead in soil vs. lead in house dust even
when much of the dust is derived from soil. In spite of considerable efforts to determine the
ingestion intake of soil and dust by children, these values are still subject to uncertainty.
Site-specific data on soil ingestion by children are rarely available, but would be valuable in
modeling site-specific exposure to lead. Only limited information is available about the
effects of the child's micro-environment on soil and dust ingestion, with evidence suggesting
much larger intakes of soil for children in intrinsically dirty environments such as
campgrounds, and lower soil intake for children who spend much of their time in cleaner
environments such as day care centers.
You are responsible for the choice of non-default bioavailability parameters.
Bioavailability parameters may differ among sites. Non-default bioavailability parameters
may be justified by experimental studies with the actual site materials, assessments of other
sites with similar materials, or site specific information on properties of particles that may
affect bioavailability.
The Geometric Standard Deviation is not considered a highly site-specific parameter,
and should normally be kept at its default value of 1.60. If you use some other value, you
should document the reasons for this modification, since risk estimates are typically very
sensitive to the GSD value used.
1.4.2 Exploring Model Options
The IEUBK model has a large number of options. You are encouraged to explore these
options before doing any substantive analyses, because there are often several alternative
methods that can be used to obtain model outputs. These options are identified in Chapter 2.
1-20
-------
They include alternative source menus for soil and dust lead, dietary lead, and lead in
drinking water. The soil/dust lead menu includes options for air-to-dust and soil-to-dust
transfer coefficients, as well as for non-household sources.
There are options beyond single runs of the model. These include multiple runs for
overlay plotting of probability curves, for plotting blood lead vs. environmental media lead
concentration, and for multiple runs (batch mode input) for each of a group of individual
children of different ages using child-specific data.
The multi-media bioavailability menu includes options for changing the passive vs.
facilitated absorption of lead from all media. The half-saturation uptake, a parameter that
determines the extent of non-linear or saturable absorption, may also be changed from the
normal default value of 100 /xg Pb/day.
Run options include the choice of an iteration time step. With low exposure and no
year-to-year change in concentration, as used in the "Default" option, there should be no
differences in output using other iteration time steps. Differences in blood lead of a few
percent may occur with higher and rapidly changing exposures. For a single run, almost any
PC (XT or later) will produce a solution within 60 seconds, even without a math
coprocessor, with the default iteration time of 4 hours. However, with a batch mode input
file of several hundred records, the simulation run may take many hours. In this case, you
may select a longer iteration time and speed up the run for a preliminary analysis. If you
use a longer time step, you should verify accuracy using records with high exposure or large
changes in exposure.
1.4.3 Documentation of Input Parameter and Data Files
By reviewing every adjustable parameter in the model and noting which ones have been
modified in a particular run, you have a permanent record of the input. An electronic copy
of the exposure input parameters can be made using the parameter SAVE option. Distinctive
names for parameter files ([name].SV3), input data files ([name].DAT), simulation run files
(RESULTS.TXT), batch mode output files ([name].TXT and [name].ASC), probability plot
overlay files ([name].LAY) and blood lead vs. media concentration files ([name].MED) may
be used to document input specifications as well as output.
The worksheet provides a convenient format for noting reasons for use of non-default
parameters, or justification for use of default parameters. For example, soil lead
1-21
-------
concentrations and dust lead concentrations could be measured values at each house.
Repeated values of household data would be used to weight the statistical results from batch
mode files. Missing value imputation methods should be identified, for example, "KID ID
= 17,22,35, missing dust lead concentration estimated by PbD = 180 + 0.28 * PbS." This
is critical information in allowing other users to reproduce your results (including yourself,
since it is unlikely that most users will be able to recall over one hundred model parameters
after the passage of some months or years).
1.4.4 Documentation of Model Output
1.4.4.1 Selecting Output Alternatives
Results of IEUBK model simulations may be saved in several forms. You should select
in advance the most useful of these forms, since the results of some interactive simulations
cannot be recovered once you have bypassed the opportunity to save the results. Choices
are:
(1) A sequence of single simulation runs. Sequential runs can be interactively
appended to the file named RESULTS.TXT. The average of the
geometric mean blood lead concentrations for children in sequential
one-year age intervals, the input concentrations for several media, and the
media-specific daily lead uptake for each year are saved. You must use
the "Save" option at the end of each run to be saved, but this allows you
to drop results from non-informative runs rather than save them.
(2) A sequence of graphics overlay simulation runs. The multiple plot option
saves input data for blood lead probability plots for a range of evenly
spaced media lead concentrations. For example, you may generate plot
data for soil lead concentrations of 250, 500, 750, and 1000 ug/g, for
children of ages 12 to 24 months. The data in the [name].LAY overlay
file includes the geometric mean blood lead for children in the age range,
the lead concentration in soil and in other media. The actual plots of
probability density or cumulative distribution functions depend on the
GSD value selected, and these plots include the probability of exceeding
the user-specified LOG for use in risk estimation. Probability plots may
be printed on standard laser printers.
(3) A sequence of blood lead vs. media lead simulation runs. The media
range option saves input data for blood lead vs. media lead plots for a
range of evenly spaced media lead concentrations. For example, you may
generate plots of blood lead vs. soil lead concentrations smoothly
interpolated from calculated values at 250, 500, 750, and 1000 ug/g, for
children of ages 12 to 24 months. The data in the [name].MED overlay
1-22
-------
file includes the geometric mean blood lead for children in the age range
at the selected media lead concentrations, the lead concentration in soil
and in other media. Plots may be printed on standard laser printers.
(4) Batch mode simulation runs. The batch mode option requires an input
data file, as described in Chapter 2. Output consists of user-named files
[name].ASC and [name].TXT that contain predicted blood lead
concentrations for each case or record (child) in the input data file. The
output files also document the missing value imputations when some of the
input data on residential lead concentrations in air, water, soil, or dust are
missing. The files may be used as input for the statistical analysis
programs in the companion PBSTAT program, which produce statistical
and graphical comparisons of the observed and predicted blood lead
concentrations.
1.4.4.2 Understanding the Output
You should carefully review the output options described in Section 1.4.4.1. Each
option allows you to examine a different aspect of the IEUBK simulation. The numerical
simulation component of the IEUBK model produces an estimate of a geometric mean blood
lead concentration for children of a given yearly age. This is the average of the estimates for
children during that one-year interval. The IEUBK model arrives at these estimates by
calculating at each time step an updated estimate of all compartment lead masses, or
equivalent tissue lead concentrations. The update algorithm combines uptake of lead from
the environment with all of the movements of lead into each compartment from another
compartment, or out of each compartment, either into another compartment or by elimination
from the child's body. In this version of the IEUBK model, the output consists of the daily
uptake rate (intake rate times fraction absorbed) for each medium, and the blood lead
concentration, as annual averages.
The output from a single simulation run may be displayed in several forms. Most users
wish to see the variability associated with a predicted blood lead concentration. This range
can be demonstrated graphically by selecting the intrinsic variability GSD and then plotting a
cumulative probability distribution. The range of plausible blood lead values may be
determined graphically as defined by upper and lower percentiles of the distribution. For
example, the 5th and 95th percentiles of the distribution will include 90 percent of the
children with the given site-specific or household-specific exposure scenario. Since
"plausible range" requires a subjective choice of percentiles, you are free to choose any
appropriate values. Since the predicted geometric mean blood lead concentration is based on
1-23
-------
an a priori mathematical simulation and not on a data-driven statistical estimate, this plausible
range should never be considered as equivalent to a confidence interval.
The other output characteristic that many users wish to see is the estimated probability
of exceeding the specified blood lead level of concern, corresponding to the given exposure
scenario or scenarios (for multiple runs in a given medium). This also requires a GSD
value. This probability may be interpreted as the percentage of children with the given
household-specific exposure scenario who are expected to exceed the level of concern.
If applied to a single site or residence, it may also be interpreted as the probability of
exceeding the level of concern for any single child who may reside at that site in the future.
1.4.4.3 Interpreting the Output and Communicating the Results
The model calculates the probability that a blood lead concentration derived from the
model's specified parameters will exceed a level of concern specified by the user. There are
two valid interpretations for the output:
(1) The output of the model may be considered to be the best estimate of a
plausible range of blood lead concentrations for a hypothetical child with
a specific lead exposure scenario. The range of values is centered on the
geometric mean blood lead concentration expected for a typical child with
this exposure scenario. The upper tail of the probability distribution
provides an estimate of the risk of exceeding some blood lead level of
concern for a typical child of that age residing in the same household and
with the same exposure history.
(2) The output of the model may also be considered to be the predicted
geometric mean blood lead of a population of children with the same lead
exposure scenario, and the upper tail of the probability distribution to be
the fraction of children exceeding the chosen blood lead level of concern
when all of these children have the same exposure history.
The array of applications for which the IEUBK model can be validly used is:
A: One location
Al: one living unit, one child;
A2: one living unit, more than one child;
A3: more than one living unit, more than one child, homogeneous media
concentrations;
B: Multiple locations, one neighborhood, homogeneous media concentrations
1-24
-------
C: Multiple locations, one neighborhood, heterogeneous media
concentrations;
D: Multiple locations, more than one neighborhood, heterogeneous media
concentrations;
A single run of the IEUBK model is sufficient for categories A and B. A classification
or disaggregation of the neighborhood into distinct exposure subgroups is required in
categories C and D, with the possibility of different ingestion or absorption parameters for
different neighborhoods in category D. Neighborhood-scale and community-scale risk
estimation requires aggregating the risk estimates for individuals or subgroups.
The differences between these levels is sketched in Figure 1-3. Category A requires
calculating only a single blood distribution. Category B requires calculating a blood lead
distribution for each child, but since each child of the same age has the same exposure
scenario in category B, a single run of the model is sufficient to characterize risk for this
subgroup. In category C, there are different exposure scenarios for each subgroup. Risk
estimates must be calculated for each such subgroup, then added up across sites and children.
The model output in category A: Single child, single site of exposure, includes a blood
lead concentration, a distribution of blood lead concentrations, and a probability of exceeding
the blood lead level of concern. Since children in environments with the same lead exposure
may have a range of blood lead concentrations, we describe the likely variability in blood
lead for a child with a given exposure scenario by a probability distribution. The predicted
blood lead concentration is the geometric mean of the distribution of blood lead
concentrations that may occur for a typical child with the specified exposure scenario. Risk
is calculated from this distribution as the probability that a hypothetical child living at this
site, with the specified exposure scenario, will have a blood lead concentration exceeding the
blood lead level of concern. This single-child assessment is necessary in order to use the
model to evaluate remediation options on a house-by-house or yard-by-yard basis. The
single-child assessment also provides a criterion for model testing and validation using
epidemiology data.
The model output in category B: Multiple children, single site or equivalent sites of
exposure, is the predicted blood lead concentration for each child as the geometric mean of
the distribution of blood lead concentrations that may occur for each child with the specified
exposure scenario. Risk is calculated by aggregating the calculated risk for each child as the
percentage of hypothetical children living at this site or at these sites, with the specified
exposure scenario, that will have a blood lead concentration exceeding the blood lead level of
1-25
-------
A. Single child
Single site of exposure
Probability
Blood Lead
Multiple children or population
Single site or comparable sites
Child 1 Child 2 Child 3
C. Multiple children or population in neighborhood
Multiple sites with different exposures
Sitel Site 2 Site 2 Site 3
Child 1 Child 2 Child 3 Child 4
D. Multiple children or population in heterogeneous community
or region, different neighbohood exposure scenarios
Figure 1-3. Categories of application of the IEUBK Model.
concern. The calculation is exactly the same as the single-child assessment, but there is an
important shift in inteipretation of the output.
There are situations in which a single site really can have multiple children of the same
age with the same exposure scenario. A single housing unit may be occupied by several
households with pre-school children of the same age. Rental properties may be occupied in
succeeding years by different families, each of which may have a pre-school child of the
same age with virtually the same exposure as occupants in other years. In general, the
multiple-child or population exposure scenarios would be applied to a hypothetical population
of occupants.
Neighborhood-scale risk estimation is discussed in Section 4.2, with examples. The
model output in C: Multiple children, multiple sites with different exposure, cannot be
1-26
-------
obtained by a single run of the IEUBK model. It is necessary to construct an exposure
scenario for each distinct exposure subgroup in the population. For each child or exposure
subgroup, risk is calculated in a single run of the IEUBK model with the specified exposure
scenario. The risks for each exposure subgroup are aggregated across all subgroups,
weighted by the number of children with that exposure scenario or by the percentage or
likelihood of the exposure scenario.
There is no one-step method by which neighborhood-scale risk estimation can be done
using this version of the IEUBK model. The problem of risk estimation for children in a
large community or a region is even more difficult when different subgroups of children may
have very different exposure scenarios, including differences in behavior that affect
ingestion, and differences in lead absorption due to behavioral or nutritional differences.
A common misinterpretation of the IEUBK Model is that it predicts community
geometric mean blood lead and the fraction of children at risk when the input is the mean or
geometric mean of household-specific environmental lead concentrations. That mis-step can
be misleading, particularly when the environmental variables have a wide distribution among
the neighborhoods of the community. This misinterpretation is especially dangerous for post-
abatement settings intended to eliminate the higher exposures when there are multiple
exposure media. A correct approach requires applying the model to each individual home or
site using the lead concentrations seen at that site and combining these results as an aggregate
of sites in several neighborhoods to form an estimate of community risk. A second useful
approach is based on subdividing a community into neighborhoods and clusters of residence
units with similar media lead concentrations. Specific information on building appropriate
neighborhood exposure scenarios is given in Section 2.3, Building an Exposure Scenario.
Examples are provided in Section 4.2.
We should emphasize that the IEUBK model is intended to provide a best estimate of
geometric mean blood lead. The IEUBK model is not intended to be used in a worst-case
scenario, as the model does not apply any uncertainty factors or modifying factors in making
risk estimates. If, as usual, there is some uncertainty about model parameters, these can be
evaluated using sensitivity analyses. Remember that you are responsible for documenting
plausible non-default values.
Uncertainty about parameters is not the same as the intrinsic variability in
environmental data and blood lead responses. The components of variability are discussed in
1-27
-------
Section 4.2 on the blood lead Geometric Standard Deviation (GSD), which plays a critical
role in risk estimates.
1.5 REFINEMENTS AND ENHANCEMENTS
The biokinetic component of the IEUBK model is based on an age-dependent
compartmental model with identifiable physiological compartments: red blood cells, plasma
and extracellular fluids, kidney, liver, other soft tissues, trabecular and cortical bone
(Figure 1-1). There are many compartmental models in the literature; some with fewer
compartments (Rabinowitz et al., 1976), others with many more compartments (Leggett,
1993). The Technical Review Workgroup for Lead was aware of important research in the
development of physiologically-based pharmacokinetic (PB-PK) models for lead in humans,
primates and rats that took into account the slow diffusion of lead through the bone matrix
(O'Flaherty, 1992a,b,c, 1993a,b). However, the Workgroup chose to develop a
compartmental model that uses transfer times or transfer rates between compartments instead
of physiologically based compartmental coefficients. The transfer rates can be estimated
from data in non-human primates, especially the studies on infant and juvenile baboons that
were done at New York University (Mallon et al., 1983; Harley and Kneip, 1985).
The IEUBK biokinetic model was based on:
(1) empirical kinetic data on blood lead in baboons of similar weight and
developmental stage to human infants and young children;
(2) kidney, liver, tibia and femur lead concentrations in baboons after the end
of the lead exposure study;
(3) autopsy data for lead levels in young children who died from causes not
related to lead exposure;
(4) extrapolations from studies in human adults;
(5) lead feeding and lead balance studies in human infants.
There is, in principle, a degree of similarity between these approaches, since the
compartments in the IEUBK model are defined by real anatomical and physiological
properties. The transfer times from the PB-PK model can be calculated from blood flow
rates to organs and tissue groups, volumes of these organs, partition coefficients across
1-28
-------
membranes, and solid state diffusion coefficients for the bone matrix. The principal
difference between the biokinetic components of the IEUBK and PB-PK models is that, in the
absence of suitable physiological data, empirical data were used in estimating transfer times
in the IEUBK model. Future development of the IEUBK Model is expected to continue in
the direction of physiologically based biokinetic components similar to PB-PK models.
Many users have expressed interest in tools that allow a more detailed investigation of
the effects of non-environmental variability on the distribution of blood lead concentration.
The Monte Carlo approach would allow every parameter in the model to be assigned a
random variation at every iteration of the computation. For example, each parameter could
be multiplied by a random factor (mean value 1) at every iteration. This would require that
adequate data would be available to support the input distributions. An extremely large
amount of computing would be necessary. A substantial amount of additional study is
needed before Monte Carlo methods can be added to the IEUBK model.
The IEUBK model currently evaluates children from birth to age 84 months. Many
users have requested extension of the model to other populations, including older children
and adults, with emphasis on populations at special risk. Both the physiological and
biokinetic parameters of adults are at least as well known as those of children, with the
possible exception of lead distribution within the human maternal-fetal unit. Transfer of lead
from the mother to the neonate during lactation would also be of interest.
1.6 GETTING MORE HELP
As scientific knowledge advances, this Guidance Manual will be updated and revised.
If you have questions regarding the site-specific application of the IEUBK Model, you may
direct your inquiries to the appropriate EPA Regional Toxics Integration Coordinator.
Comments on the technical content of the manual or suggestions for its improvement may be
brought to the attention of members of the EPA Technical Review Workgroup for Lead listed
in the front of this document.
1-29
-------
2. A GUIDED TOUR THROUGH THE LEAD MODEL
2.1 THE LEAD MODEL IS DRIVEN BY MENUS
Environmental Protection Agency's Integrated Exposure, Uptake, and Biokinetic Model
for Lead in Children (IEUBK Model) is a microcomputer program that performs many
different functions related to estimating blood lead levels in children. The overall model
functions are sketched in Figure 2-1.
V
Nn -L*
/^attiways\D9fe
Pathways^
Select
Pathways/Input
Menus
V
/ Enter Model /
/ Parameters /
/(Options 1-6,L,S)/
L
r
Run Model
Simulations
(Options R.MAB)
T T
ult Default
->• Values
V
/Risk Estimates/ / Risk Estimates / / Estimate / /View Output/
/ for Single Runs/ / for Multiple Runs/ / Lead Levels/ / Data File /
/ (Option P) / / (Option P) / / (Option B)/ / (Option V)/
V
12-17 1
V M
2-181 1 2-201
V V
V
Statistical
Comparison of
Model Simulations
(Option T)
v
Figure 2-1. Schematic diagram of the overall functions of the lead model. Numbers in
pentagons indicate sections in this document containing more detailed
information.
2-1
-------
The oval shapes are terminal steps (i.e., the beginning or end of a function or option).
Rectangles show internal processes and rhomboids show user data entry operations or
functions. Diamond-shaped figures are decision points where the user must choose one of
the model options on a list. The "NO" branch usually follows the model's baseline or
"default" parameters and functions. Horizontal and vertical arrows refer the user to another
figure or page.
The IEUBK model is menu-driven, with on-line help available in almost any menu.
The main menu, where any use of the IEUBK model begins, is shown in Screen 2-1. There
are five numbered options:
1. Parameter Input Menu
1: Air lead menu
2: Dietary lead menu
3: Drinking water lead menu
4: Soil/Dust lead menu
5: Alternative lead source menu
6: Maternal lead menu
L: Load pre-saved parameter input menu
R: Return to Main Menu
2. Computation Menu
1: Run a single model simulation
2: Multiple simulation runs with a range of values
3: Blood lead versus media with a range of values
4: Multiple simulation runs with batch input (input data file for each child
or household)
R: Return to Main Menu
3. Output Processing Menu
1: Save program parameters to file
2: Plot graphs of blood lead distributions
R: Return to Main Menu
4. Help Menu
1: General information
2: Information about menus plus help in other menus
R: Return to Main Menu
5. Quit
Q: Return to DOS prompt.
2-2
-------
2 - COMPUTflTION Menu.
3 - OUTPUT Menu.
4 - HELP Menu.
D - QUIT (EXIT to DOS).
To select, press lumber lor letter) of selection OR use
arrow keys to highlight selection and press return.
Screen 2-1. The main menu.
We will briefly discuss the options in each of the input menus. Scientific justifications
for the options and guidance values are provided in Section 2.3.
2.2 DETAILED DESCRIPTION OF MENUS
2.2.1 Help Menu (4)
2.2.1.1 General Help (1)
The General Help menu provides on-line information on the data or parameter entry
menus, menu selections for running single or multiple model simulations, and use of
keyboard keys. This information is shown in Screen 2-2.
2-3
-------
To RUN the Model: The menu that allows single rune of the
Uptake/Biokinetic Model is accessed by two methods:[1] by direct
U PgUp/On Home/End
Screen 2-2. The general help menu.
2.2.1.2 Information Menu (2)
The Information Menu provides on-line information on the parameter save and load file
options, on multiple-run and output processing menus. The information is presented here in
Screen 2-3.
2.2.1.3 Other On-Line Help Menus
Most menu screens contain additional information on the lower part of the screen.
Additional information screens are available on specific menu options.
2.2.2 Parameter Input Menus
2.2.2.1 Air Lead (1)
The Air Lead input parameter menu is shown in Screen 2-4 and schematically in
Figure 2-2. The air lead concentration is set initially to a typical 1993 urban value of
0.1 fig/m (U.S. Environmental Protection Agency, 1991c). It is assumed that the indoor air
2-4
-------
L - LOBD.
Use LOAD to load a file which uas saued using the SflUE
function This can saue time when utilizing conplex data seta.
Exit:ESC Scroll: H PgUp/On Home/End
Screen 2-3. The information menu.
Press either
N for NO
V for VES
Wary Sir Cone, by Vear '??
Outdoor Air Lead Concentration (ug/m3)
Indoor flir Pb Cone. (Percentage of Outdoor)
Uieu/Change Time Spent Outdoors
Uieu/Change Uentilation Rates
i y.
NO
0.1QO
30.0
Default
Default
HELP WINDOW
If you uant to vary air concentration for each year or wieu current
variable data > Press V for VES
NO is displayed in this field if all yearly concentrations are equal.
VES 18 displayed in this field if all yearly concentrations are NOT equal.
NO MUST be selected to enter a single cone for Outdoor Air Lead Cone.
Esc: MfllN Menu F1• General HELP
PgUp/PgDn: Media Switch
Screen 2-4. The air lead menu.
2-5
-------
' Enter Air Pb
Concentration/
By Year /
Enter
Indoor/Outdoor/
Concentration/
Ratio
View/Change
Lung Absorption?
Relevant
Menus
Figure 2-2. Decision diagram for the air lead menu options.
lead concentration is 30% of the outdoor concentration (i.e., 0.03 jttg/m ) initially. The time
spent outdoors and ventilation rate are assumed to depend on the child's age. These
parameters allow a time-weighted air lead intake to be calculated; 32 % of that intake is
absorbed through the lungs into the child's blood. All parameters except the indoor/outdoor
air lead concentration ratio may be changed by entering YES in the first line. Some are
age-specific values.
2-6
-------
2.2.2.2 Dietary Lead (2)
The Dietary Lead input parameter menu is shown in Screen 2-5 and schematically in
Figure 2-3. The daily dietary lead intake values for each age apply to a typical U.S. child in
a typical setting in the United States after 1990. These dietary lead values may be altered by
entering YES to the query "View/Change Dietary Pb Intake?" During the period 1982-1989
there was a distinct reduction in food lead generally attributed to the replacement of lead-
soldered cans and the removal of lead from gasoline. Since 1990, food lead in U.S.
supermarket food has remained relatively constant. Dietary lead ingestion for years prior to
1990 are given in Section 2.3.2.
Data Entru for DIET
Uieu/Change Dietary Pb Intake "m - Default
Use alternate Diet Ualues -m NO
GI Ualuee/Bioflvailabilitu/" :
HELP WINDOW
Exposure to lead in food is based on the typical American diet and
quarterly surveys of Pb concentrations in this diet. If you want to
change or view the current dietary intake levels of lead for each yearly
period (based upon these surveys) • Press > V for YES
The program default values (in ug Pb/day) are 5.53, 5.78, 6.19, 6 2t,
6.81. 6.3t, 7.00 for the respective yearly periods.
F1: General HELP
PgUp/PgDn' Media Suitch
Screen 2-5. The dietary lead main menu.
If the dietary lead sources are non-standard, usually because of suspected contamination
of fruits, vegetables, fish and meat raised locally or otherwise lead-contaminated, the user
can enter specific values by responding YES to the query, "Use Alternate Diet values?" This
invokes the alternative menu shown in Screen 2-6.
2-7
-------
Dietary Lead
Input Menu
(Option 2)
/ E
/ Age-D
/ Dieta
/ E
/ Cono
/ /P<
/ forS
inter / ™s
epenueni /^ •-
ry Intake/
Inter /
entration / Yes
ircent / ^ ^
>ources /
/view/Change
x. Dietary Lead
Xjntake?/
> No
^ I1W
i
XUse^
/ Alternate
\ Dietary Lead
^W Mn
Multimedia
Absorption
Menu
Figure 2-3. Decision diagram for the dietary lead menu options.
2.2.2.3 Drinking Water Lead (3)
The Water Lead input parameter menu is shown in Screen 2-7 and schematically in
Figure 2-4. The water lead concentration is set initially to a typical 1990 urban value of
4 /ig/L (Marcus and Holtzman, 1990). The age-specific ingestion of tap water is described
in Section 2.3.3.2. Consumption may be modified by responding YES to "View/Change
Drinking Water Intake?" and entering new values, as shown on Screen 2-8.
Alternative information may be available in the form of measured lead concentration
and percentage of tap water intake from water fountains or other outside sources, and water
2-8
-------
DataEntrufo^JIET
Alternate Diet Entr
Home Grown Fruits
Home Grown Uegetables
Fish from Fishing
Game Animals from Hunting
Ethnic Preferences
of all fruits)
of a11 veg.)
of all meat)
of all meat
National consumption surveys of homegrown vegetables show a consumption
rate from ^ to 75 percent fl typical value for communities that might be
influenced by this factor is 25%, and a worst case value is judged to be
40%. Comparable values for fruits are 20% and 39% For these two food
types, enter the percent consumption within these ranges and the average
lead concentration
ESC to EXIT (return to previous level)
Screen 2-6. The alternative dietary source menu.
Enter Lead Cone in Drinking Water (ug/L)
lew/Change Drinking Water Intake*?
Us* alternate Water Ualues w»
HELP UINDOUI
Enter the concentration of lead in the drinking water (ug Pb/L)
This concentration is held constant over the entire seven year period.
The program default concentration is —> H.00 ug Pb/L
Drinking water is defined as the tap water consumed in a glass or from a
drinking fountain. Tap water used for mixing beverages or cooking is part
of the dietary exposure and changes here are reflected in tap water used
for cooking and mixing THIS FIELD NOT ACCESSIBLE IF ALTERNATE WATER IS VESi
Fl: General HELP
Screen 2-7. The drinking water lead main menu*
2-9
-------
Water Lead
Input Menu
(Option 5)
No
Yes / Enter Alternate
->•/ Source
Lead Levels
.*•
No 1
/ Enter /
/ Constant /
/ Lead Level /
i
Enter Age-
>/ Dependent
Tap Water Intake/
Figure 2-4. Decision diagram for the drinking water lead menu options.
consumed at home in first-draw or flushed modes. This may be entered by responding YES
to "Use Alternate Water Values?"
2.2.2.4 Soil and Dust Lead (4)
The Soil and Dust Lead input parameter menu is shown in Screen 2-9 and schematically
in Figure 2-5. The soil and dust lead concentrations are set initially to a value of 200 |Wg/g.
The age-specific ingestion intake of soil and dust combined was estimated from the
EPA/OAQPS staff paper on Exposure Assessment Methodology and Validation for the first
2-10
-------
Data Entru for DRINKING UflTEP,
Water Consum fl_/d):§0 20 |p 50 ^0 52 lp.53 ^D 55
Press ESC to EXIT
Press F8 to Recall Defaults
Use the arrow keys or Tab key to change years
Enter the daily consumption volume of uater (Liters/day) for each age
Pb Intake from water - consumption volume X concentration
NOTE ALL Changes are Immediate and Stay in Effect Unless Re-Changed »'
Press ESC to EXIT freturn to previous level!
Screen 2-8. The age-specific drinking water consumption menu.
Data Entry for SOIL/DUST
| Soil Pb Levels (ug/g) 206.6
Indoor Dust Pb Levels (ug/g)1 20B.0
Press either
1 - Enter a constant value
2 • Enter variable values
gSoiI/Dust Ingest ion Weighting ;
Factor (percent soil]1 45.0
Uiew/Change flmt of Soil/Dust
Tnne»«ted Daily •>•>•> Default
GI Ualues/Bioflvailabilit
The Pb concentrations of outdoor soil that are used by the Uptake Model
are selected by •--> Pressing the selection number from the menu above.
The 'constant value' selection allows a single cone to be entered, which is
then used for all ages The 'variable value* selection allows a different
soil cone to be entered for each age
The program default constant cone is > 206 ug Pb/g soil.
Esc: MftIN Menu Fl- General HELP
Screen 2-9. The soil and dust main menu.
2-11
-------
From Air Lead
/EnterConstant / Constant
Soil Lead
Conoontrstion
EnterConstant /Constant
Dust Lead
Concentration
/ Enter Dust Pb
/ Concentration,
/ Percent for home/
' daycan, etc.
/ Enter
/ Soil/Dust
Ingeston Rate
Go To
I \ Multimedia
| /Absorption
— Menu
Figure 2-5. Decision diagram for the soil/dust lead menu options.
version of the UBK model (U.S. Environmental Protection Agency, 1989a). Both
concentration and intake may be modified by the user.
As shown in Screen 2-9, both soil lead and dust lead concentrations may be changed on
a yearly basis by user selection "2", allowing the user to construct reasonable site-specific
scenarios.
The multiple-source option ("3") on the dust entry line allows the user to use
information about the contribution of soil lead, air lead, and other sources to household dust
2-12
-------
lead. The Data Entry Screen for the Multiple Source Analysis (Screen 2-10) has three data
entry lines. The first line is the fraction of the soil lead concentration that contributes to the
concentration of lead in household dust. If there were no other sources, this would be the
ratio of the dust lead concentration to the soil lead concentration. The current default value
of 0.70 is appropriate to neighborhoods or residences in which loose particles of surface soil
are readily transported into the house. The second data entry line is the contribution to
household dust from the deposition of airborne lead, over and above the soil lead
contribution. The current default value is an additive increment of 100 /xg/g lead in house
dust for each /xg Pb/m air.
Data Entry for SOIL/DUST
Multile Source final
Contribution of Soil Pb to Indoor
y-tousehold Dust Pb [conversion factor]: 0 85
^ Contribution of Outdoor ftirborne
^ Pb to Indoor Household Dust Pb
[conversion factor]; 100 0
SSSSSM^ V" "V
Consider Alternate Indoor Dust
Pb Sources *rr> : NO
Press ESC to EXIT
Lead in Household dust is usually comprised of Pb from soil sources and
airborne fallout sources Other sources, such as Pb-based paints in the
home, can also contribute to Pb in nousehold oust
Enter the fraction of household oust that comes from soil.
The program default is > 0 28 (ug Pb/g dust per ug Pb/g soil)
Press ESC to EXIT [return to previous level]
Screen 2-10. The multiple dust source menu.
The third line asks whether the user wants to add other sources. If "Yes", then the
Multiple Source Analysis Screen is replaced by the Alternative Indoor Dust Entry screen
(Screen 2-11). The user may assign both the concentration and percentage of dust intake to
baseline household dust, secondary occupational dust, dust at school, daycare, or second
home, and the exposure to lead in dust from household paint measured as a percentage of
total dust ingestion and its concentration. The default dust lead concentration in the
2-13
-------
Data Entry for SOIL/DUST
Multiple Source Analysis
fllternate Indoor Dust Entr
Household Dust Caug)'||10 0
Socondary Occupational Duet-f|l200.0
« Dust at School-||200 0
Dust at Oaucare-l§200
Second Home Dust
Lead-based Paint in Home-§^1200.0
Prose ESC to EXIT
to Recall Defai
HELP UIINDOLJ
The Household Dust Cone, and Percent can NOT be accessed' They are shown
for informational purposes. The Household Dust Percent uill change uhen
the other percent fields are changed Its value is 100 minus the total of
the other percent fields
Percent refers to Total Pb Exposure from Indoor Dust Example- if flLL
dust exposure occurs in a home with Pb paint, the ualue in the paint field
should be 100 percent
Press ESC to EXIT (return to previous leuell
Screen 2-11. The alternative indoor dust menu.
Alternative Indoor Dust Entry screen is 100% of household dust at 150 /wg/g. If the
Alternative Source Analysis is not used, then the default dust contribution consists of 70% of
the soil concentration plus 100 times the air lead concentration. For default conditions, the
total dust lead concentration equals 150
If non-residential exposures to soil/dust are important, the user may access the multiple
non-residential source menu.
The combined soil/dust ingestion rate (grams total soil + dust per day) can be changed
from the current default values in Screen 2-12.
2.2.2.5 Alternate Source (5)
The alternate exposure source menu is shown in Screen 2-13 and schematically in
Figure 2-6. The default daily lead intake value for each age is set to 0 /*g Pb/day. The user
has the option to input any source not otherwise covered by other menus. Examples might
be the direct ingestion of lead-based paint, cosmetics or home remedies. In this case, the
amount of lead per day needs to be calculated from the information available. If the
2-14
-------
Press ESC to EXIT
Press F8 to Recall Defaults
Screen 2-12. The soil/dust ingestion rate menu.
Data Entry for OTHER (fllternate Source
View/Change Other Lead Intake '"" : Default
Change GI Values/Biofluailabilit/" :
Screen 2-13. The alternate lead source menu.
2-15
-------
Alternate Lead
Input Menu
(Options) )
View/Change
Alternate Lead
Intake?
Enter Lead
Intake Values
Go To
Multimedia
Absorption
Menu
Figure 2-6. Decision diagram for the alternate lead source menu options.
alternative source is lead-based paint (LBP), this exposure would be in addition to exposure
to lead-based paint in house dust, which is Option 4 in the multiple source menu of soil and
dust. See Section 4.7.1 for a discussion on issues in the use of the model for paint chips.
Building an exposure scenario using this option should be done with care. The model
assumes all entries represent chronic exposure. In the example above, the child would
require immediate medical attention. Remember that the model output represents only those
children defined by the exposure scenario.
2-16
-------
2.2.2.6 Bioavailability of Lead in Food, Drinking Water, Soil, and Dust
Bioavailability or absorption of intake from the gut or lung into the blood is a key
element in relating external exposure to body burden. Lead intake from media with low
bioavailability poses much less of a hazard than does the same intake from media with high
bioavailability. The bioavailability of lead from normal infant diet is known to be very high
(Alexander et al., 1974a,b; Ziegler et al., 1978; Ryu et al., 1983), with at least 40 to 50%
of the dietary lead intake passing into the child's blood. See Section 4.1 for a discussion of
bioavailability.
The main functions of the bioavailability menu are shown in Figure 2-7. The model
calculates lead absorption from the gut as a function of two components. The passive
component does not depend on lead concentration in the gut and is not saturable. The
facilitated or active component may become saturated when the total concentration of lead in
the gut from total gut intake by all media is sufficiently large, which is a kinetically
non-linear absorption mechanism. The data entry Screen 2-14 allows the user to specify the
parameters for intake from soil, dust, drinking water, diet, and alternate sources. The total
absorption percentage is the sum of the passive and facilitated absorption components. The
default value of absorption for alternate sources is 0%, which requires that the user must
enter the bioavailibilty of any specific alternative source, such as lead-based paint.
The total absorption from any medium is then divided into two components, and the
user specifies a small fraction of the total absorption percent for the passive or non-saturable
(i.e., high-dose) component. The default is 20% of the total available for absorption. The
percentage absorption in the larger saturable component is the remainder of the total
available for absorption. For example, with a dietary lead intake of 50%, the absorption
fraction for the passive component is 20% of 50%, or 10% of dietary lead intake, and the
saturable component is 80% of 50%, or 40% of dietary lead intake.
2.2.2.7 Maternal-Fetal Lead Exposure (6)
The maternal lead exposure input parameter menu is shown in Screen 2-15. The lead is
transferred from the mother to the fetus in utero. The lead that is stored in the tissues of the
newborn child is calculated by entering the maternal blood lead value at birth (default =
2.5 jug/dL). The IEUBK model assumes that the infant's blood lead at birth is a fraction of
the maternal blood lead level, and the amounts of lead in the blood and other tissues in the
newborn infant are calculated so as to be consistent with concentration ratios observed in
autopsies of newborn infants (Barry, 1981).
2-17
-------
(Bioavailability\
Input Menu
Accessed Through
Options 1-6) J
/Enter Absorption
at Low Intake
for Each Medium /
Enter Passive
' Absorption Fraction/
for Each Medium /
Figure 2-7. Decision diagram for the absorption/bioavailability menu options.
2.2.2.8 Save and Load Options
If the user wishes to use a certain set of model parameters as the starting point for
another analysis, the parameter set created from use of Options 1 through 6 should be saved
using the "S" option on the output menu accessed from the main menu, or the F6 option on
any of the parameter input menus. The user may create an 8-character or shorter name for
the file, which will be stored in the form [NAME].SV3. If a saved parameter set is needed
later, it may be loaded from the "L" option on the Parameter Input Menu.
2-18
-------
TOTflL
fWHILflBLE FOR
flBSORPTION
(percent)
RflTIO OF PBSSIVE TO TOIHL HBSORP110N 0 20
(ratio of lead absorbed by
passive pathway relative to total
lead available for absorption)
HRLF-SHTURflTION LEVEL (ug/day]
Press ESC to EXIT
gTE__flLL Changes are Mediate and Stay in Effect Unless Re-Chanped lit
Screen 2-14. The absorption/bioavailability
menu.
Enter the Mother", blood Pb leue
at time of birth
mine default Pb
Screen 2-15. The maternal/fetal lead exposure
menu.
2-19
-------
2.2.3 Computation Menu
2.2.3.1 Run a Single Simulation of the Model (1)
The menus for the Run command are shown in Screen 2-16. This option uses only the
currently loaded parameter set. The user may view or change the time step for the
numerical iteration. The default is four hours. We recommend setting the iteration time to
the lowest convenient selection, and verifying all "important" solutions by rerunning the
model with the shortest possible time step (currently 15 min). An output option (Option 2)
allows plotting of results and calculation of probability of elevated blood lead.
RUN the MODEL has been Chosen
Select from menu-
1 - RUN the Model with current parameters.
C - Change Biokinetic default parameters.
G - General Information about the Model.
S - Select Nuntier of Time Steps for iteration.
X - for internal use only
R - RETURN to Main Menu or Data Entry screen
To select, press number (or letter) of selection OR use
arrow keys to highlight selection and press return
Screen 2-16. Single simulation using the program processing menu.
2.2.3.2 Run Multiple Simulations of the Model for a Range of Media Lead (2)
The menus for the Multiple Run command are shown in Screen 2-17 and schematically
in Figure 2-8. More detailed menus for range selection and output are shown in
Screens 2-18 and 2-19. This option uses only the currently loaded parameter set, except that
it repeats the run for each new value of a medium concentration (e.g., soil lead
concentration) or intake (dietary lead as jug Pb per day). The user may view or change the
2-20
-------
RANGE SELECTION MENU
1 - Select MEDIfl for Range Ualues
2 - Enter RP.NGE UQLUES for Selected Media
3 - RUN Model for Selected Media & Ualues.
- Change OUTPUT CHOICES for the Model Run
H - HELP Information
R - RETURN (Exit).
Current Selected Media- Soil
Model Run Output Choices:
Current Range Ualues
for Selected Media-
Send To Overlay File ? . NO
Display Sumnary Outputs'3: VES
Nuntoer of Runs for Range 7
Start Ualue: Q BOH ug/g
End Ualue : 1 GOO.000 ug/g
Screen 2-17. Multiple simulation using the program processing menu.
time step for the numerical iteration during the run step. We recommend verifying all of the
"important" solutions by rerunning the model with the shortest possible time step (currently
15 min). Since only one medium can be changed in each use of the "2" option, the user who
wants to look at a range of soil lead values should use the Multiple Source Dust option "3"
and a user-specified dust lead to soil lead concentration ratio. Output data for plotting, with
overlays of results at each concentration in the range, may be saved when the user creates
RANGE#.LAY files.
2.2.3.3 Multiple Simulation Runs of a Medium To Find Concentration of Lead in the
Medium That Produces a Specified Blood Lead (3)
This option is similar to Option 2. The menus for the Multiple Run command are
shown in Screen 2-20 and schematically in Figure 2-8. This option uses only the currently
loaded parameter set, except that it repeats the run for each new value of a medium
concentration (e.g., soil lead concentration) or intake (dietary lead as jug Pb per day) until
the specified age-dependent geometric mean blood lead level is achieved exactly by that
concentration.
2-21
-------
/Multiple Simulation Runs
With a Range of Media
Values (Option B)
Find Media
Concentration for
Specified B
Enter Range
of Values
for Medium
' Select Number
of Output Runs
)
1
/ Select
/ Medium /
/
)
^
/ Select
/ Age Range /
>
f
Rnd Lead
Concentration
in Medium
Age Range
)
1
Run Multiple
Simulations
Figure 2-8. Decision diagram for the multiple simulation menu options.
Since only one medium can be changed in each use of the Multiple Simulation Run "3"
option, the user who wants to look at a range of soil lead values should use the Multiple
Source Dust option "3" and a user-specified dust lead to soil lead concentration ratio. Output
data for plotting may be saved when the user creates *.PBM files.
2.2.3.4 Batch Mode Multiple Simulation Runs Using Input Data Files (4)
This option is similar to Option 2. The menus for the Batch Mode Run command are
shown in Screen 2-21 and schematically in Figure 2-9. This option uses the currently loaded
default parameter set, but repeats the run for using the new values for the five exposure
2-22
-------
1 - Select MEDIB for
2 - Enter RBN6E
3 - RUN Model f
14 - Change OUT
H - HELP In-for
R - RETURN (Exi
Screen 2-18. Selection of media for multiple range run.
1 - Select MEDIA for Range Ualues
2 - Enter RANGE UflLUES for Selected Media
3 - RUN Model for Selected Media & Values
Change Range Ualues:
Screen 2-19. Range selection during multiple processing.
2-23
-------
BLOOD Pb us MEDIA CQNC SELECTION MENU
1 - Select MEDIA.
2 - Enter RANGE UALUES for Selected Media.
3 - RUN Model SERIES for Selected Values.
- Change OUTPUT CHOICES for Model Runs.
A - Change AGE RANGE.
F - FIND Media CONC Associated with a Blood Pb
H - HELP Information.
R - RETURN (Exit).
Current Selected Media: Soil
Current Range for Selected Medi
Model Run Output Choices
Output to Files'- NO
Number of Rune: 8
Start Ualue: 3.008 ug/g
End Ualue : 1006.668 ug/g
AGE Range: 21 to 36 Months
Screen 2-20. Using multiple simulation to find acceptable media concentrations for a
predetermined blood lead concentration.
Select BATCH MODE Data Input File:
COPPERSM.DAT
MINESITE.DAT
RCRAS2.DAT
SMELTER3.DAT
URBANH1.DAT
URBANLO DAT
URBANMED DAT
| To Select:
Highlight file and
press Return key.
use H, Home, End,
PgUp,PgDn to shift
highlight bar...
Press ESC to Abort
NOTE: ALL Paint Chip
ualues not included
in input file MUST be
previously setlll
TE' a properly formatted
data file uith file
extension '.DAT' is
REQUIRED!!
Screen 2-21. Running the model in batch mode.
2-24
-------
Multiple Simulation
Runs Using Site-
Specific Batch Data
File (Option A) J
Create '.DAT Input
Data Rle with Multiple
Environmental Variables,
'Set Non-specific
Parameters with
Input Menus 1-6, L/
No
i
f
Run Multiple
Simulations
for Each Set
of Input Data
/ Rename
/ *.7XT,*.ASC
/ Output Files
Figure 2-9. Decision diagram for the batch mode menu options.
parameters (soil concentration, dust concentration, drinking water concentration, air
concentration and alternate source consumption) for each child in the data set. The input
data are entered one line at a time from a data set with a specified list of input variables.
These must be created by the user in a special *.DAT file in the Lead Model directory.
Each line of data may include:
The child code or case;
2-25
-------
The "family" identifier for individuals at the same living unit;
An area, block, or neighborhood identifier;
Each line of data must include:
The child's age, in months, as of the end of the data collection period;
The soil lead concentration, in ^g/g;
The house dust lead concentration, in /ig/g;
The drinking water lead concentration, in /ig/L;
The air lead concentration, hi jug/m ;
The alternate source intake rate, hi /tg/day;
The child's blood lead level at specified age, in /xg/dL.
The child's age must be entered. Either a soil lead or a dust lead value is needed for the
simulation, along with a stand-in value (imputation rule) if one of them is missing (for
example, if the user does not fill in missing dust lead values, the current default is to replace
a missing dust lead concentration by the soil lead concentration). The user may prefer to
create an input data file with missing dust lead concentrations replaced by some fraction of
the soil lead concentration. Missing values of air, water, and alternate lead are replaced by
default values. If there is no actual child blood lead data, then Option 1 produces output data
sets with *. ASC and *.TXT extensions that contain all of the input data, including imputed
values, and predicted blood lead levels for each line of data.
The batch mode option can be used to perform statistical analyses of simulated
community blood lead distributions, even without observed blood lead levels (for example, if
an investigator has carried out a multimedia environmental lead study at a site, without blood
lead data being collected). However, this option will be even more useful if blood lead data
from a well-conducted study are available for model comparisons using statistical tests in
Option 5. Output data files may be reviewed using Option 2, as demonstrated in later
sections.
2.2.3.5 Statistical Analyses of Batch Mode Data Sets (4)
A set of statistical procedures for analyzing batch mode data sets exists as a separate
module in the IEUBK Lead Model. Although the Option 1 data sets can be edited and used
2-26
-------
in any other statistical programs the user may have available, we have included in Option 5
some of the most commonly used statistics, statistical hypothesis tests, and graphical data
displays for comparing observed and modelled blood lead levels. We recommend using a
variety of graphical and statistical techniques in evaluating the output of Batch Mode model
runs. This will also be demonstrated in Section 5.
2.3 BUILDING AN EXPOSURE SCENARIO
2.3.1 Air Lead Menu
2.3.1.1 Default Air Lead Exposure Parameters
The default air lead concentration is 0.1 /xg/m3, which is approximately the average
1990 urban air lead concentration (U.S. Environmental Protection Agency, 1991b). During
the period 1970-90, ambient air lead concentrations dropped drastically in the United States
due to the phasedown of lead in gasoline (Figure 2-10). When adequate monitoring data
exist to define concentrations higher or lower than the default outdoor lead concentrations,
these should be used. Current air lead levels are low in most places in the United States,
and do not require year-to-year specification. Elevated air lead levels have been reported
around some point sources in the United States and Europe (Davis and Jamall, 1991) and
lead modeling for changes in these sources requires year-by-year input data.
A constant air lead value larger than 0.1 jug/m may be appropriate for assessment at
locations in the vicinity of active point sources of lead emissions such as lead smelters or
battery plants. In such cases, an appropriate estimate of annual average air lead
concentration must be available.
An example of a striking increase over time was the air lead levels in Kellogg/Silver
Valley, Idaho, following a September 1973 baghouse fire. These levels remained elevated
for a sufficiently long time such that the use of these values in predicting blood lead
concentrations for 1974-1975 from the Lead Model was justified (Agency for Toxic
Substances and Disease Registry, 1988).
2.3.1.2 Ventilation Rate
The intake of air increases from infancy to adulthood. The range of values for child
ventilation rates was established by EPA (U.S. Environmental Protection Agency, 1989a) as
2-27
-------
Gasoline
Ambient Air
1976 1978 1980 1982
Year
1984
1986
1988
Figure 2-10. Historical relationship between lead in gasoline and lead in air in the
United States.
Source: U.S. Environmental Protection Agency (1986), with updating.
3 33
2 to 3 m /day at age 0 to 1 years, 3 to 5 m /day at age 1, 4 to 5 m /day at ages 2 and 3,
-l ^
5 to 7 m /day at ages 4 and 5, and 6 to 8 m /day at age 6. The Lead Model uses midrange
2
values of 2, 3, 5, and 7 m /day at ages 0+, 1, 2 to 4, and 5 to 7 respectively. Children
who exercise more than average will have a correspondingly greater intake, and those who
are very inactive will have a lower ventilation rate. The higher intakes may be useful in
modeling children who spend time at playgrounds or outdoor play areas near an air lead
source. Changes in activity pattern can change ventilation rate in a child or in a
neighborhood.
2.3.1.3 Indoor/Outdoor Activity Patterns
The range of values for outdoor time was established by EPA (U.S. Environmental
Protection Agency, 1989a) as 1 to 2 h/day in the first year of life, 1 to 3 h at age 1, 2 to 4 h
at age 2, and 2 to 5 h/day from ages 3 to 7. The default values in the Lead Model are 1, 2,
3, and 4 h/day at ages 0+, 1, 2, and 3 + , respectively, roughly at the middle of these
ranges. The outdoor air lead concentration provides a large part of the total air lead
exposure, because the indoor air lead concentration is typically only about 30% of the
outdoor concentration (U.S. Environmental Protection Agency, 1986). Site-to-site
2-28
-------
differences may exist due to natural ventilation, climate, season, family activity, and
community access to outdoor play activities.
2.3.1.4 Lung Absorption
The range of values for child lung absorption was established by EPA (U.S.
Environmntal Protection Agency, 1989a) as 25 to 45% for young children living in non-point
source areas, and 42% for those living near point sources. The default value used in the
Lead Model is 32 %. Changes in the source of airborne particulates may also affect lung
absorption. No quantitative recommendations can be made.
EXAMPLE 2-1: Characterizing Effects of Air Lead Phasedown on Inhalation Intake
If the Lead Model were to be used to estimate blood levels of children living in an
urban area in previous decades, when the predominant sources of lead exposure for many
U.S. children were air lead from combustion of leaded gasoline and dietary lead from lead-
soldered food cans, it would necessary to use community air lead levels during that period of
time. Representative values of air lead concentrations were presented in the EPA Air Quality
Criteria Document for Lead (U.S. Environmental Protection Agency, 1986, Chapter 7,
Table 7-2) for urban center or suburban locations in nine metropolitan areas for 1970 through
1984. The reductions in air lead from 1977 through 1988 attributable to the phasedown of
leaded gasoline are quite evident in both urban centers and suburban areas. For example, for
a retrospective estimate of blood lead levels in children in 1981, one would need to start with
1975 air lead levels to include prenatal exposure of children up to age 7 in 1981.
Figure 2-10 shows that air lead exposure in 1981 would be at 0.48 pg/m , and so on. For a
5-year old child in 1981, air lead exposure, at age 0+ in 1975 is 1.2 /*g/m , and so on.
This adjustment in air lead concentration does not estimate the indirect effects of air lead
changes on blood lead through gradual changes in soil and dust lead. This example is
generic, not site-specific, however. The air lead data entry screen for children born in 1975
is shown in Screen 2-22.
2.3.2 Dietary Lead Menu
2.3.2.1 Total Dietary Lead Exposure
Data assembled from a variety of sources, including Market Basket Surveys
(Pennington, 1983) and representing changes in consumer behavior over time, were used to
construct dietary lead intake estimates as described in Chapter 7 of the EPA Air Quality
Criteria Document for Lead (U.S. Environmental Protection Agency, 1986). The method is
2-29
-------
fllr Cone (ug/m3)'|8 IBB
Outdoors Chr/day) Sl.B
Uent.Rate (m3/day) :|2.0
Luna Absoru. fXl :S32.0
Press ESC to EXIT
Press F8 to Recall Defaults for CURRENT ROil
Use the arrow keys or Tab key to change years.
Enter the outdoor air cone, (ug Pb/m3) for each age If all ualues are the
same, the model assumes a constant [not variable) air concentration.
NOTE: ftLL Changes are Immediate and Stay in Effect Unless Re-Changed •!•
Press ESC to EXIT (return to preuious level)
L
Screen 2-22. Data entry for air. The user may input data from historical records of air
lead concentrations on this screen.
based on U.S. FDA Market Basket samples in 231 food categories and has been updated to
1988 (U.S. Environmental Protection Agency, 1989a). Because two major sources of lead in
food (lead-soldered cans and air deposition on food crops) have been greatly reduced or
eliminated, dietary lead is believed to be relatively constant since 1990, especially for
children under seven years.
Table 2-1 shows how estimated mean dietary lead intake depends on the child's age,
and that this intake has changed very drastically with the near-elimination of lead solder from
food cans and other food packaging in the United States since the 1970s. Where site-specific
dietary levels are not available, it is recommended that the values from Table 2-1 be used for
the appropriate years and ages, and that the most recent values (1988) be assumed for all
future years. Seasonal effects are omitted here since the Lead Model uses annual values for
dietary exposure parameters. For alternate exposure scenarios with seasonal intakes, the user
may need to calculate time-weighted annual averages from seasonal data.
If the Lead Model is used in connection with historical exposures, for such purposes as
model validation or retrospective dose reconstruction, the dietary intake data should be
2-30
-------
TABLE 2-1. DIETARY LEAD INTAKE (/ig/day) FOR U.S. CHILDREN BY
AGE, FOR EACH YEAR FROM 1978 TO PRESENT
19781
19791
19801
19811
19822
19832
19842
19852
*
1986
1987-Present3
6-11 Mo
NE
NE
NE
NE
19.2
14.4
19.0
10.2
7.9
5.5
1 Year
45.8
41.2
31.4
28.3
25.0
18.3
22.7
10.6
8.2
5.8
2 Years
52.9
48.0
36.9
33.8
27.5
21.9
26.4
12.3
9.4
6.5
Age
3 Years
52.7
47.8
36.9
33.7
27.4
21.4
26.0
11.9
9.1
6.2
4 Years
52.7
47.8
36.9
36.8
27.2
21.1
25.7
11.8
8.9
6.0
5 Years
55.6
50.3
38.7
35.8
28.6
22.3
27.1
12.4
9.4
6.3
6 Years
NE
NE
NE
NE
31.6
24.8
29.9
13.6
10.3
7.0
NOTES: NE = Not estimated.
1 = Estimated by J. Cohen and D. Sledge, Table A-2 (U.S. Environmental Protection Agency,
1989a).
2 = U.S. Environmental Protection Agency (1986), updated with data from the FDA Market
Basket Survey.
3 = Average of 1986 Q4 through 1988 Q3. Further decreases in food lead concentrations since
1987 are believed to be negligible.
* = Linear extrapolation between 1985 and 1987.
adjusted to the year when the data were collected. For prediction in future years, the most
recent default value for each age may be used.
2.3.2.2 Dietary Lead Exposure by Additional Pathways
For some children, there are important dietary lead sources that are not characterized
by the FDA Market Basket Survey data summarized in Table 2-1. Child-specific or site
specific data will be needed to verify these alternative dietary lead sources. Local sources of
fruit and vegetables are used in many small towns and in rural areas. Some individuals
obtain much of their produce from their own gardens. If the local or home-grown produce is
grown in soils with high concentrations of lead, or if the edible leafy portions are
contaminated by airborne lead particles, then some fraction of the environmental lead may be
added to the child's diet. The additional intake of lead in diet may become important if the
2-31
-------
environmental lead concentrations are sufficiently high. This was important in evaluating the
Bunker Hill Superfund site in Kellogg ID and was included in the Risk Assessment Data
Evaluation Report (the "RADER") prepared by EPA (U.S. Environmental Protection
Agency, 1990c). Additional pathways of dietary lead exposure are discussed in
Example 2-2.
Dietary lead exposure is the product of the amount of food consumed in each category
and the concentration of lead in the food item. Normal intakes are reported by Pennington
(1983). To adjust for home gardens, a fraction of this intake may be allocated by the
Alternate Diet Entry Menu to local produce, and the rest to Market Basket produce that is
not grown locally.
Local game animals feeding on plants contaminated by lead in soil may also have
elevated lead concentrations. Lead contamination of rivers and lakes by deposition and by
erosion of leaded soils may also increase lead concentrations in local fish. Some rural
families may use hunting and fishing as a significant supplement or even as their primary
source of animal protein. See Baes et al. (1984) for a comprehensive approach to estimating
pathways of trace elements in the food chain. A fraction of the meat and fish intake may be
allocated by the Alternate Diet Entry Menu to local game and fish.
Other consumer products may have nontrivial potential for dietary lead exposure.
These include lead-glazed or soldered cooking and food preparation utensils, ethnic or
regional preferences for food products with high lead content, and the use of oral ethnic
medicines such as "empacho" or "azarcon" that have high concentrations of lead and are
known to have caused cases of acute lead poisoning in children (Trotter, 1990; Sawyer et al.,
1985). No general recommendations about parameter values for these sources of lead can be
made at this time. Approximate intake for oral medicines may be estimated from
recommended or customary doses for young children.
EXAMPLE 2-2: Characterizing Indirect Dietary Lead Intakes for an Old Lead Smelter
Community
Some data from the Human Health Risk Assessment (Jacobs Engineering, 1989) and the
RADER for Kellogg ID may be useful (U.S. Environmental Protection Agency, 1990c).
Table 2-2 shows that a large percentage of the population uses local produce, that the use of
local produce increases toward the more rural Pinehurst area but the lead concentration
decreases, and that the lead levels in local produce hi 1983 were enormously higher than in
2-32
-------
TABLE 2-2. ESTIMATES OF LEAD INTAKE FROM CONSUMPTION OF LOCAL
PRODUCE BY CHILDREN, AGES 2 TO 6 YEARS, IN KELLOGG, IDAHO
Area
Smelterville
Kellogg
Pinehurst
National
Market
Basket
Percent Using
Local Produce
16%
36%
46%
~
Number of
Gardens
2
17
20
—
Consumption
(g/day)
* *
Leafy Root
25
25
25
25
15
15
15
15
Concentration (/tg/g
Leafy
6
3
0
wet
wt.)
Root
.1
.5
.017
4.
4.
2.
0.
5
5 +
2
041
Intake (/ig/day)
In Summer
220
220 +
121
1
**
NOTES: Leafy vegetables are lettuce and spinach. Root vegetables are carrots and beets.
Average of Kellogg and Smelterville.
Annual average is 1/4 of this.
Source: RADER Tables 5-8 and 5-4 (U.S. Environmental Protection Agency, 1990).
Jacobs Engineering (1988) Table 7-16.
the National Market Basket samples for the same period (1982 to 1984). The calculated
increment of daily dietary lead intake during the summer months was 220 /xg/day in the
report. However, for the purposes of this example, we will assume that this total
consumption occurs over the course of the year and includes fresh as well as frozen or
canned produce to give an annual average increment of 55 jtig/day.
Table 2-3 shows that the lead concentration in fish in nearby Lake Coeur d'Alene in
1985 was much higher than in the Columbia River and higher than fish at the average
National Pesticide Monitoring Station lead concentration for 1976/1977. A moderate rate of
consumption is two 2-oz fish portions per week, or 114 g/week = 16 g/day on average. The
incremental intake from local fish is equal to the concentration difference, 0.80 to
0.34 = 0.46 jig/g times 16 g/day =7.5 /*g/day.
Screen 2-23 shows dietary lead intakes for a typical child born in 1983, and
Screen 2-24 shows the extra exposure for intake from contaminated fish.
2.3.3 Drinking Water Lead Exposure Menu
2.3.3.1 Drinking Water Lead Default Exposure Parameters
Water sampling methods may be as first draw standing samples, partially flushed
samples, or fully flushed samples. The highest lead concentrations at the tap are usually
obtained for lead pipes, lead-alloy solder on copper pipes, or lead-alloy brass faucets and
2-33
-------
TABLE 2-3. ESTIMATES OF LEAD INTAKE FROM CONSUMPTION OF LOCAL
FISH BY CHILDREN, AGES 2 TO 6 YEARS, IN KELLOGG, IDAHO
Source
Lake Coeur d'Alene
(1985)
Columbia River
(1986)
National Pesticide
Monitoring Stations
(1976-1977 August)
Concentration
Og/g wet wt.)
0.80
0.34
0.34
Fish Consumption
(g/day)
16
16
16
Lead Intake
(/xg/day)
13.0
5.5
5.5
NOTES: ^Two-ounce portions, twice a week.
For annual average, multiply by fraction of year when local fish are consumed.
Source: RADER Tables 5-8 and 5-4 (U.S. Environmental Protection Agency, 1990).
Jacobs Engineering (1989) Table 7-16.
Data Entry for DIET
ftGE- 8-1 1-2 2-3 3-4 t-5 5-6
Press ESC to EXIT
Press FS to Recall Defaults
Use the arrou kegs or Tab key to change years.
Enter the Dietary Pb Intake (uq Pb/day) for each age group.
NOTE: flLL Changes are Immediate and Stay in Effect Unless Re-Changed !•!
Press ESC to EXIT (return to previous level
\
Screen 23. Using dietary lead intake for a child born in 1983 (see Example 2-2).
2-34
-------
Alternate Diet Ent
CONC fuo Pb/ol^ Percent (of food cl
of allTnjits]
of all veg.l
of all meat]
of all
Home Grown Fruits:
Home Grown Vegetables
Fish from Fishing'
Game Animals from Hunting'
Ethnic Preferences:
lonal Prefences:
Press ESC to EXIT
National surveys have determined that sports fishermen may consume an
average of 30 g/bay, with a worst case of IfQ g/day. This program
the substitution of fish for other meat dishes In estimating exposure due
to recreational fishing, select 10/C for the average fraction of the meat
diet, or as much as 5QX for the worst case Rlso enter the average Pb
concentration of the fish which are consumed
return to previous level
L
Screen 24. Using dietary intake from local vegetables and fish in Kellogg (see
Example 2-2).
fittings in contact with corrosive water for several hours. The new EPA National Primary
Drinking Water Regulation for Lead (NPDWR) requires public water systems to collect first
draw samples, standing a minimum of 6 h, from a sample of homes targeted as potentially at
risk. Water lead concentrations can be significantly different for different sampling
protocols, depending on the sources of lead in water drawn through the tap. First draw
samples generally have higher lead concentrations than flushed samples. The typical effects
of different water sampling procedures are discussed in the Sampling Manual that is to
accompany this model.
Drinking water lead concentrations in the Lead Model are held constant during the
entire seven years of the child's exposure. In the Case Studies below, household-specific
water lead concentrations are used. If no household-specific or relevant community water
lead data are available, we recommend using the default value of 4
If a substantial fraction of the child's activity is spent outside the home, it may be
useful to separate drinking water exposure into primary residence and secondary residence or
daycare. A large number of U.S. children spend time during the weekday at daycare centers
2-35
-------
or secondary residences. If adults and older children in the household are either at work or
at school during the day, there may be two stagnation periods for drinking water during the
day—overnight, and midday. In this case, a larger fraction of the child's water lead
exposure can occur at the higher "first-draw" concentrations. Some exposure scenarios are
discussed by Marcus (1991) in evaluating the risk from lead leached out of newly installed
brass faucets. The default scenario is defined by setting 50% of the child's water intake to
household first-draw consumption. The remaining intake consists of partially flushed intake
inside the home (35%) and water consumed outside the home (15%). The total intake of
lead in drinking water would then be:
PbW = 0.5 x PbW (first draw) + 0.35 x PbW (flushed) + 0.15 x PbW (fountain).
There is no general rule for estimating the amount of water ingested from water coolers
in day care centers or other non-home locations. Since the child's activities outside the home
are likely to be different than inside the home, it is unlikely that the ratio of non-home to
home water intake is proportional to the amount of time spent away from the home versus at
home. Two drinks per day, each about 60 mL (2 oz) or 120 mL, is a reasonable upper limit
for day care intake. The default is 15% of the daily tap water intake, v/hich ranges from
75 mL at age 1 year to 90 mL at age 6 years.
2.3.3.2 Alternate Drinking Water Exposure by Age
The default values in the IEUBK model (Table 2-4) are taken from the U.S. EPA
Exposure Factors Handbook (U.S. Environmental Protection Agency, 1989c). A survey of
drinking water consumption in U.S. children was reported by Ershow and Cantor (1989) in a
study for the National Cancer Institute. These values have been smoothed and disaggregated
into yearly values shown in Table 2-4. The range of values from the Ershow-Cantor data in
Table 2-5 show that the default values for the IEUBK model are similar to but somewhat
lower than the median values, but also contain information about the percentiles of the
distribution of tap water intake, about gender differences in intake and other factors that you
may find useful. A plausible scenario for elevated exposure to lead in drinking water would
be to use larger tap water intakes, such as the 90th percentile values in Table 2-5. Note that
for children receiving formula reconstituted with tap water, consumption of tap water would
be much higher, perhaps closer to one liter per day. In an assessment addressing risks from
lead in drinking water, the exposure to infants consuming reconstituted formula requires
specific attention.
2-36
-------
TABLE 2-4. AVERAGE DAILY WATER INTAKE IN U.S. CHILDREN
*
Ershow-Cantor Study
Total (L/day)
Age (Months)
0-5
6-11
12-47
48-84
M
0.992
1.277
1.409
1.551
F
1.035
1.238
1.300
1.488
Tap (L/day)
M
0.250
0.322
0.683
0.773
F
0.293
0.333
0.606
0.709
IEUBK Model
Age
(Mo)
0-5
6-11
12-23
24-35
36-47
48-59
60-71
72-84
Tap
Water Intake
(L/day)
0.20
0.20
0.50
0.52
0.53
0.55
0.58
0.59
^Ershow and Cantor (1989).
U.S. Environmental Protection Agency Exposure Factors Handbook (1989c).
TABLE 2-5. TAP WATER INTAKE (L/day) BY AGE CATEGORY
Age Category
(Months)
0-5
6- 11
12-47
48-84
Mean
0.27
0.33
0.65
0.74
10
0
0
0.24
0.30
Percentiles
50 (Median)
0.24
0.27
0.57
0.66
90
0.64
0.69
1.16
1.30
Source: Table 2-5, Ershow and Cantor (1989).
2.3.4 Soil/Dust Lead Exposure Menu
One of the most important uses of the IEUBK model is to compare risks among
alternative soil lead and dust lead exposure scenarios. Many of these scenarios arise in
assessing exposure reduction strategies. For example, in evaluating soil lead abatement at a
particular residential yard, we might be interested in the following sequence of comparisons:
(1) Calculate the risk of an elevated blood lead level for the present soil and
dust lead levels;
2-37
-------
(2) Calculate the risk of an elevated blood lead level for the replacement of
soil lead with soil having a lower lead concentration, along with cleaning
up household dust;
The first scenario describes risk to occupants with present exposure levels. The second
scenario describes risk to occupants in the distant future after lower new lead levels have
been achieved by abatement. The IEUBK model can accept input data describing both of
these exposure scenarios.
2.3.4.1 Soil and Dust Lead Default Exposure Parameters
The natural concentration of lead in soil, from weathering of crustal materials, is
estimated as about 10 to 25 /Jg/g. A plausible urban background is 75 to 200 /ig/g (U.S.
Environmental Protection Agency, 1989a; HUD, 1990).
It is expected that lead concentrations in undisturbed soils may persist for many
thousands of years. However, urban areas are hardly undisturbed environments and available
data (von Lindern, 1991; Jacobs Engineering, 1990) suggest that near-surface soil lead
concentrations may decrease by a few percent over a decade or so. It is usually adequate to
assume a constant soil lead concentration unless soil abatement is included in the exposure
scenario.
It is also possible that the soil becomes recontaminated over time, for example if
surface soil is abated and then is recontaminated by ongoing atmospheric lead deposition
from non-abated sites near by or by contamination from deteriorating exterior lead-based
paint. Changes in soil concentration can be incorporated on an annual basis in developing
the exposure scenario. This is done with the Option "2" on the Soil/Dust Data Entry Menu.
2.3.4.2 Exposure to Soil and Dust
The default value for total intake of soil and dust depends on age, and ranges from
85 to 135 mg/day. These values are within the ranges identified in the OAQPS staff paper
that supported the first UBK model and have been reviewed by the EPA Clean Air Science
Advisory Committee. Recent investigations by Binder et al. (1986), Clausing et al. (1987),
Calabrese et al. (1989, 1991b), van Wijnen et al. (1990), and Davis et al. (1990) apply the
trace element approach to quantify ingestion rate. These investigations currently constitute
the most appropriate basis for estimating the quantity of soil ingested. The results are
summarized in Table 2-6. The van Wijnen et al. data are discussed in Section 2.3.4.4. It is
likely that the intake rate depends on the child's age, activity pattern, and the total
2-38
-------
TABLE 2-6. DAILY INTAKE OF SOIL AND DUST ESTIMATED FROM
ELEMENTAL ABUNDANCES
Soil/Dust Intake, mg/day
Study
Davis et al. (1991)
Ages 2-7 years
Calabrese et al. (1989)
Ages 1-4 years
Binder et al. (1986)
Ages 1-3 years
Clausing et al. (1987)
Ages 2-4 years
Element
Al
Si
Ti
Al
Ti
Y
Zr
Al
Si
Ti
Al
Ti
AIR
Median
25
59
81
30
30
11
11
121
136
618
92
269
106
Mean
39
82
246
154
170
65
23
181
184
1,834
232
1,431
124
Maximum
904
535
6,182
4,929
3,597
5,269
838
1,324
799
17,076
979
11,620
302
AIR = Acid Insoluble Residue.
accessible dust and soil in the environment. It is recommended that soil and dust intake be
defined by an age-dependent scenario shown in Table 2-7, as reviewed by the Clean Air
Science Advisory Committee (U.S. Environmental Protection Agency, 1990b).
Two of the studies, Davis et al. (1991) and Calabrese (1989), measured the dietary
(including medication) intake of the trace elements and subtracted this quantity in estimating
soil ingestion. These studies therefore provide the most complete quantitation of ingestion.
Because Binder et al. (1986) did not measure dietary intake, the results for this study are
likely to provide an upper bound on ingestion among those subjects. Van Wijnen et al.
(1990) did not measure dietary intake but attempted to compensate for this approach by using
the lowest observed tracer result for each child and subtracted out a value obtained for
hospitalized children who were assumed not to ingest soil or dust. The combination of these
two techniques may lead to a downward bias in ingestion estimates.
2-39
-------
TABLE 2-7. AGE-SPECIFIC SOIL AND DUST INTAKE
Age (Years)
0-
1 -
2 -
3 -
4-
5 -
6-
<1
<2
<3
<4
<5
<6
<7
Intake (g/day)
0 -
0.080 -
0.080 -
0.080 -
0.070 -
0.060 -
0.055 -
0.085
0.135
0.135
0.135
0.100
0.090
0.085
Adopted for
Guidance Manual
0.085
0.135
0.135
0.135
0.100
0.090
0.085
Source: U.S. EPA (1989a).
The reader should also note that there are statistical problems in interpreting an
observed median value from these studies. For example, in a population of children who all
ingested very small amounts of soil on most days but occasionally ingested larger quantities,
the median from a short term measurement study will be below the average daily quantity
ingested by any of the children. The mean value is not subject to this bias, and therefore is
judged to be a more meaningful measure of ingestion.
It should be noted that the 200 mg/d ingestion value presented in Superfund guidance
can be supported as, roughly, an upper bound on mean ingestion considering the values seen
in different ingestion studies. The values recommended for use in the model (85 to
135 mg/d) represent a more central value within the range of values seen in different studies.
The smaller study of Clausing et al. (1987) used methods similar to the later study of
van Wijnen et al. The values shown for soil ingestion hi Table 2-7 are unconnected for
dietary intake. The paper presents additional estimates using acid insoluble residue and
tracer excretion by hospitalized children.
2.3.4.3 Sources of Dust Exposure
Contribution from Atmospheric Deposition and Soil
We recommend collecting household dust data. If that has not been done, then
Option 3 may be used to estimate dust lead concentrations. The OAQPS Exposure Analysis
and Methodology Validation (U.S. Environmental Protection Agency, 1989a), used for the
earlier version of the model on which the current IEUBK Model is based, calculates the
2-40
-------
contribution of atmospheric deposition and soil to house dust by linear regression models
between dust lead, soil lead, and air lead. There is a relationship between dust lead
concentration in jiig/g (denoted as PbD), soil lead concentration in /xg/g (denoted as PbS),
and air lead concentration in /ig/m (denoted as PbA). In a number of studies, statistically
significant relationship of the form:
PbD = 00 + 0S PbS + 0A PbA
This equation suggests that house dust lead concentration consists of three components:
a soil component, which is the fraction j3s of the soil concentration, an air component,
consisting of a coefficient 0A relating /*g/g lead in dust to /ig/m of lead in air, and a third
component of /3O coming from unidentified sources.
As a default value in the model, we used /3A = 100 /ig/g per ^g/m based on several
analyses. We recommend a default soil-to-dust coefficient of 0.70, which represents some
real sites where soil is a major contribution to household dust. The reader should be aware
that other values have been identified for other site-specific exposure scenarios.
Dust Lead Increment from School Dust
Dust ingestion while at school may be significant, depending on the amount of exposure
on the floor or playground. While the IEUBK model deals primarily with preschool
children, some children may be in school and subject to a more structured regimen of
hygiene and reduced dust exposure. The amount of dust ingested and its implicit fraction of
total dust ingestion is not necessarily proportional to length of time at the facility. Hygiene
and dust loading are additional predictive factors. Playground geometric mean dust lead
levels of 170 - 3,700 /-ig/g were reported by Duggan et al. (1985) in a sample of 11 British
schools.
Dust Lead Increment from Day Care
Dust ingestion while at daycare (including nursery school and kindergarten) may be
significant, depending on the amount of exposure on the floor or outside play area. Dutch
children who spent a considerable amount of time at a daycare center were known to ingest a
large quantity of dust and soil, although apparently much less in rainy weather than in good
weather (van Wijnen et al., 1990).
2-41
-------
Dust Lead Increment from Second Home
Children often spend several hours per day in the home of a relative or in an informal
daycare setting. Dust exposure information can often be collected and used in the same
manner as for the primary home.
Dust Lead Increment Remaining from Primary Residence
When the Multiple Source Analysis option is selected on the Soil/Dust Data Entry
Menu, the IEUBK model offers the opportunity to change the soil and air parameters of the
regression equation set at 0.70 and 100, respectively as default values. The selection of a
default value for the soil-to-dust coefficient was based on empirical data. In sites where soil-
to-dust coefficients have been measured and where paint does not contribute greatly to dust,
the range was from 0.09 to 0.85. Among the sites where soil-to-dust coefficients have been
measured are the following: East Helena, 0.85 (0.81 and 0.89); Midvale, 0.70 (0.68, 0.72);
Butte, 0.26; and Kellogg, 0.09. Recent data suggest the coefficient decreases over time at
some sites where major sources of soil lead deposition are no longer active. The user is
cautioned, however, that the contribution of soil to dust concentration varies greatly from site
to site, and site-specific soil and dust data should be collected for use in the model. The user
may choose to enter values for alternate sources of dust, including both an estimate of
concentration (jwg/g) and relative contribution (%) for each source. Of the five alternate
sources, two (secondary occupational dust and lead-based paint in home) represent
contributions to house dust lead within the home, and three (dust at school, dust at daycare,
and second home dust) represent exposure outside the primary home. If no selection is made
from any of these five, the house dust concentration remains as calculated from the linear
equation. If any of the five options are selected, this percentage is subtracted from the house
dust component, the contribution from all sources is calculated, and the average is shown on
the Multiple Source Average line of the Soil/Dust Data Entry Menu. This line appears only
if the Multiple Source Option is selected.
2.3.4.4 Fraction of Exposure as Soil or Dust
We recommend using the default assumption that 45 % of the total dust intake is derived
from soil. The ratio of soil intake to dust intake is not simply proportional to the ratio of the
number of waking hours that the child spends outdoors versus indoors. Children spend only
15 to 30% of their waking hours playing outside but are more likely to be in contact with
bare soil areas, in locations with large amounts of accessible loose particles, and are likely to
wash their hands less often than when they are indoors. The default 45/55 ratio in the model
represents our best judgement of a properly weighted ratio for this parameter.
2-42
-------
The issue of intake of soil and dust has not been properly resolved hi the scientific
literature. The distinction is important because there is some indication that even if soil lead
is the principal source of dust lead, there may be chemical or physical differences between
soil and dust that may affect bioavailability. Calabrese (1992) has found that most of the soil
and dust intake in a soil pica child was in the soil component, but this is hardly
representative of a larger population that may have large differences in relative exposure to
soil and dust.
Section 2.3.4.4 discusses the option to select the amount of dust and soil consumed by
the child each day. The default values are age weighted from 85 to 135 mg/day, and this
dust is ingested either during kitchen preparation of food or by hand-to-mouth activity during
indoor and outdoor play activity. This section discusses the option to allocate a portion of
the ingested dust to dust derived from soil that is ingested during outdoor play activity. This
distinction is important when there are differences between the bioavailability of dust derived
from soil and dust in the home, and when there are large differences in the concentration of
dust from the two sources. When house dust is thought to be mostly of soil origin and each
are expected to have similar bioavailability, the designation of this fraction is a moot point.
It is in cases where house dust differs significantly from soil derived dust that the soil/dust
ratio becomes important. One example might be the presence of interior lead-based paint.
In this case the parameter can be effective in separating soil derived dust and paint derived
dust into two components where both the amount ingested and percent absorbed can be
correctly input into the model.
There is some evidence that the soil intake is very responsive to exogenous factors,
such as weather and location. Data reported by van Wijnen et al. (1990), summarized hi
Table 2-8, show the lowest soil and dust intakes at daycare centers occurred in rainy
weather, when the children had the least amount of outdoor activity.
There is an implicit assumption that the exterior dust that a child ingests during outside
play activity is derived completely from soil, and we use soil as a surrogate for exterior dust
exposure. These intakes were measured during a 3 to 5 day sampling period, when soil and
dust intake estimates ranged from 33 to 88 mg/day for children aged 1 to 2 years and from
12 to 62 mg/day for children older than 3 years. The intake of soil and dust is describe in
detail in Section 2.3.4.2.
In the absence of any better data, we have reanalyzed and reinterpreted the van Wijnen
et al. data based on the assumption that the rainy-weather intake is only interior dust, and
2-43
-------
TABLE 2-8. MINIMUM PERCENTAGE SOIL INTAKE AS A FUNCTION OF AGE
IN DUTCH CHILDREN IN DAYCARE CENTERS3
Estimated Geometric Mean LTM, mg/day
Age (years) Good Rainy Difference (mg/d)
< 1 102 (4) 94 (3) 8
1 - <2 229(42) 103 (18) 126
2- <3 166(65) 109(33) 57
4- <5 132 (10) 124 (5) 8
aMinimum daily ingestion of acid insoluble residue or other tracers, denoted LTM (Limiting Tracer Method)
from Table 4 in Van Wijnen et al. (1990). Number of children shown in parenthesis.
that the good-weather intake is both interior and exterior dust although probably with a
smaller amount of interior dust than in rainy weather. The authors also made the distinction
between soil and dust in their discussion of the study. For our reanalysis, we took the rainy-
weather intake by age as dust and the good-weather intake as soil plus dust, to estimate an
age-related difference of 8 to 126 mg/day soil (Table 2-8). The difference between LTM
during good weather and LTM during mostly rainy weather is believed to be a lower bound
on the soil intake. The combined intakes of soil and dust estimated by other authors are of a
similar order of magnitude, such as the median soil and dust intake of 25 to 81 mg/day found
by Davis et al. (1991) for children of ages 2 to 7 years in Richland-Pasco-Kennewick,
Washington. We therefore assume that a substantial fraction of the combined soil and dust
intake in U.S. children is in the form of soil, as suggested by the large difference in
Table 2-8 between good and rainy weather intakes, and a substantial fraction is in dust, as
suggested by the large intake during rainy weather, in Table 2-8. The minimum intake,
denoted LTM for Limiting Tracer Method, has not been corrected for food intake.
However, it is likely that the differences between LTM intakes do not depend on food intake.
2.3.4.5 Bioavailability of Lead in Soil and Dust
The current assumption in the Lead Model is that 30% of dust and soil lead intake is
absorbed into the blood. This is assumed to be partitioned into a nonsaturable component of
6% and a saturable component of 24%. Some investigators (Steele et al., 1990) argue that
the bioavailability of lead in soil from some old lead mining sites is much less than that of
dissolved lead salts for several reasons: (1) large lead particles may not be completely
dissolved in the GI tract; (2) the solubility of chemical species commonly found in mine
wastes, particularly lead sulfide, is much lower than that of other lead salts. These
2-44
-------
hypotheses are based on studies in small laboratory animals such as rats (Barltrop and Khoo,
1975; Barltrop and Meek, 1979), and while the results may be qualitatively relevant to
humans, it is not clear how they should be extrapolated to humans or to other large animals
with similar physiological properties such as baboons or swine.
2.3.5 Alternate Source Exposure Menu
One possible use of the Alternative Source Exposure Menu is the direct ingestion of
chips of lead-based paint (LBP). The user might assume that a child with pica for paint
ingests one paint chip per day. If this chip weighs 0.3 grams and contains lead at 10%
(100,000 jug/g), then the calculated ingestion is 100,000 /*g/g X 0.3 g/day, or 30,000 jug/day
each day for a year. Note that this exposure would be in addition to exposure to lead-based
paint in housedust, which is Option 4 in the Multiple Source Menu of Soil and Dust. The
limited information available on the bioavailability of lead in paint chips suggests that at
doses this high, absorption mechanisms may be largely saturated (Mallon, 1983), which
would indicate appropriate adjustments in bioavailability. The user is referred to
Section 4.7, and is encouraged to review the literature on this topic prior to making a risk
assessment decision. Similar calculations can be made for the ingestion of soil or other non-
food items.
2.4 STARTING AND RUNNING THE MODEL
2.4.1 Loading and Starting the Model
The IEUBK Model is a stand alone software package that requires only an IBM
compatible PC with DOS. The diskette accompanying this manual contains the following
files:
LEAD99d.EXE (the main program file)
PBHELP99.HLP (a help file)
PBSTAT.EXE (the statistical package)
Several *.BGI files (for graphic output)
One or more EX AMPLE*. DAT (sample data sets)
2-45
-------
Copy all files into a directory of your choice, then type LEAD99d at the DOS prompt
to start the program. The initial screen gives the model name and version number. Several
information screens with recent developments and other news items then follow. The Main
Menu gives the user access to all of the menus described in this chapter.
While the LEAD99d.EXE file occupies only about 160 KB on the hard drive, it will
expand to a much larger size when loaded into RAM. Normally, a PC with 640 KB has
enough RAM to run the program, but there may be some conflicts with TSR (Terminate and
Stay Resident) programs. It may be necessary for the user to remove some TSR programs in
order to run the IEUBK Model.
The Model does not require a math co-processor, but calculations may take up to
20 times longer without a co-processor.
2.4.2 Running the Model
The user should fill in the worksheet in Figure 2-11, which defines the exposure
scenario, before proceeding with the parameter entries and the computations.
2.4.2.1 Computation Options
The computation menus present the user with a set of computation choices. One choice
is the iteration time step, Selection "2". These choices range from 15 minutes to 30 days.
The default of 4 hours is adequate for most purposes. Setting this option on the RUN Menu
also sets the iteration time for other computation modes, including the Batch Mode.
2.4.2.2 Output Options
At any time during the session, the program may be saved to a designated file. This
gives the user the option of retrieving a set of parameters at some future session without
reentering the parameters individually. After each model run, the user can select one of
several plot options, which can be viewed on the screen, printed to file or sent to a printer.
Most plots generated by the model can be printed by using the F10 key on the keyboard.
The program presently interfaces with nine standard printer types or orientations. The
Graphics Menu selection "7" allows user-specified scaling of the X-axis variable. Future
versions of the model may have additional output options.
2-46
-------
IEUBK MODEL WORKSHEET
SITE OR PROJECT:
Model Run Control Number:
PARAMETER
Model
Version:
Date:
Site Description:
DEFAULT
VALUE
USER SELECTED
OPTION
UNITS
AIR (constant)
Outdoor air lead concentration
Ratio of indoor to outdoor air lead concentration
0.10
30
Mg/m
%
AIR (by year)
Air concentration
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
Time outdoors
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-7 years (36-83 mo)
Ventilation rate
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
Lung absorption
.10
.10
.10
.10
.10
.10
.10
1
2
3
4
2
3
5
5
5
7
7
32
/ig/m
h/day
m /day
%
DATA ENTRY FOR DIET (by year)
Dietary lead intake
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
5.53
5.78
6.49
6.24
6.01
6.34
7.00
Hg Pb /day
Figure 2-11. Integrated exposure uptake biokinetic model sample worksheet.
2-47
-------
DATA ENTRY FOR ALTERNATE DIET SOURCES (by food class)
Concentration:
home-grown fruits
home-grown vegetables
fish from fishing
game animals from hunting
Percent of food class:
home-grown fruits
home-grown vegetables
fish from fishing
game animals from hunting
0
0
0
0
0
0
0
0
Mg Pb/g
%
DATA ENTRY FOR DRINKING WATER
Lead concentration in drinking water
Ingestion rate
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
4
0.20
0.50
0.52
0.53
0.55
0.58
0.59
Mg/L
liters/day
DATA ENTRY FOR ALTERNATE DRINKING WATER SOURCES
Concentration
first-draw water
flushed water
fountain water
Percentage of total intake
first-draw water
flushed water (not a user entry; calculated
based on entries for first-draw and fountain
percentages)
fountain water
4
1
10
50
100 minus
first draw and
fountain
15
Mg/L
%
DATA ENTRY FOR SOIL/DUST (constant)
Concentration
Soil
Dust
Soil ingestion as percent of total soil and dust
ingestion
200
200
45
Mg/g
%
Figure 2-11 (cont'd). Integrated exposure uptake biokinetic model sample worksheet.
2-48
-------
DATA ENTRY FOR SOIL/DUST INGESTION (by year)
Soil/dust ingestion
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0.085
0.135
0.135
0.135
0.100
0.090
0.085
g/day
DATA ENTRY FOR SOIL (by year)
Soil lead concentration
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0
0
0
0
0
0
0
Mg/g
DATA ENTRY FOR DUST (by year)
Dust lead concentration
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0
0
0
0
0
0
0
Mg/g
DATA ENTRY FOR SOIL/DUST MULTIPLE SOURCE ANALYSIS (constant)
Ratio of dust lead concentration to soil lead
concentration
Ratio of dust lead concentration to outdoor air
lead concentration
0.70
100
unitless
/ig Pb/g dust
per ug
Pb/m air
DATA ENTRY FOR SOIL/DUST MULTIPLE SOURCE ANALYSIS WITH ALTERNATIVE
HOUSEHOLD DUST LEAD SOURCES (constant)
Concentration
household dust (calculated)
secondary occupational dust
school dust
daycare center dust
second home
interior lead-based paint
150
1,200
200
200
200
1,200
Mg/g
Figure 2-11 (cont'd). Integrated exposure uptake biokinetic model sample worksheet.
2-49
-------
Percentage
household dust (calculated)
secondary occupational dust
school dust
daycare center dust
second home
ulterior lead-based paint
100 minus all
other
0
0
0
0
0
%
BIOAVAILABIIITY DATA ENTRY FOR ALL GUT ABSORPTION PATHWAYS
Total lead absorption (at low intake)
diet
drinking water
soil
dust
alternate source
Fraction of lead absorbed passively at high intake
diet
drinking water
soil
dust
alternate source
50
50
30
30
0
0.2
0.2
0.2
0.2
0.2
%
unitless
DATA ENTRY FOR ALTERNATE SOURCES (by year)
Total lead intake
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0
0
0
0
0
0
0
jtg/day
DATA ENTRY MENU FOR MATERNAL-TO-NEWBORN LEAD EXPOSURE
Mother's blood lead level at time of birth
2.5
Mg/dL
DATA ENTRY MENU FOR PLOTTING AND RISK ESTIMATION
Geometric standard deviation for blood lead,
GSD
Blood lead level of concern, or cutoff
1.6
10
unitless
Mg/dL
COMPUTATION OPTIONS
Iteration time step for numerical integration
4
h
Figure 2-11 (cont'd). Integrated exposure uptake biokinetic model sample worksheet.
2-50
-------
3. QUICK REFERENCE FOR THE
EXPERIENCED USER
3.1 FINDING YOUR WAY THROUGH THE MENUS
The Lead Model is a menu driven program. There is no need to remember special
commands, just a cursory understanding of the menu structure. Often you can find what you
want just by exploring the various menu options, and familiarity with the model is the best
way to shorten this journey. For your reference, a complete menu tree is given in
Figure 3-1. Use caution with unfamiliar menu options. Detailed explanations are given in
Chapter 2, and more complete documentation may be found in Chapter 4 for most menu
options. These should be reviewed before final decisions are made on critical model runs.
3.2 PARAMETER LIST WITH DEFAULT VALUES
The values in Table 3-1 have been assumed for the parameters of the model. These are
our best estimates for urban residents with no unusual lead exposures. The estimated blood
lead levels with the default parameters represent our best estimate of the blood lead
"background" levels that cannot be avoided. The adjustable parameters are listed by screen
in the order in which they appear in the model.
Default values are provided for the convenience of the user, but these values may not
be appropriate for specific applications. The user has the ultimate responsibility for
justification of values used in the applications of the model. We recommend that the user
review each of the parameter values in Table 3-1. Most of the parameters will not need to
be modified, but the user should be aware of them. Sensitivity analyses on parameters will
be useful in documenting results. Many default parameters in the model have only a minor
effect on the results (i.e., a 10% change in the air lead concentration parameter will change
blood lead levels by less than 1%), but some parameters may be more influential.
3-1
-------
^E^ll^iiSlJsSiaAil
Ddtrlxitlon H Flto H Rwfl"
Probability I '
Figure 3-1. Lead model menu tree.
3.3 BATCH MODE INPUT FORMAT
You may find a number of circumstances ,in which it is convenient to enter input data
for many similar exposure scenarios in a single run of the model. Option 4 of the
Computation Menu allows you to use a different age or a different value of the lead
concentrations in soil, dust, water, air, and alternate lead intake sources for each exposure
scenario. The media intake and absorption parameters are the same for every exposure
scenario in the run and must be specified before using this option, unless default values are
used. These situations may include, but are not limited to:
3-2
-------
TABLE 3-1. DEFAULT VALUES FOR MODEL PARAMETERS
PARAMETER
DEFAULT VALUE
UNITS
AIR (constant)
Outdoor air lead concentration
Ratio of indoor to outdoor air lead concentration
0.10
30
3
%
AIR (by year)
Air concentration
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
Time outdoors
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-7 years (36-83 mo)
Ventilation rate
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
Lung absorption
.10
.10
.10
.10
.10
.10
.10
1
2
3
4
2
3
5
5
5
7
7
32
3
h/day
m /day
%
DATA ENTRY FOR DIET (by year)
Dietary lead intake
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
5-6 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
DATA ENTRY FOR ALTERNATE DIET SOURCES
Concentration:
home-grown fruits
home-grown vegetables
fish from fishing
game animals from hunting
5.53
5.78
6.49
6.24
6.01
6.34
7.00
(by food class)
0
0
0
0
Hg Pb /day
jig Pb/g
3-3
-------
TABLE 3-1 (cont'd). DEFAULT VALUES FOR MODEL PARAMETERS
PARAMETER DEFAULT VALUE UNITS
Percent of food class:
home-grown fruits 0 %
home-grown vegetables 0
fish from fishing 0
game animals from hunting 0
DATA ENTRY FOR DRINKING WATER
Lead concentration in drinking water
Ingestion rate
Age = 0-1 year (0-11 mo), 0.20 liters/day
1-2 years (12-23 mo) 0.50
2-3 years (24-35 mo) 0.52
3-4 years (36-47 mo) 0.53
4-5 years (48-59 mo) 0.55
5-6 years (60-71 mo) 0.58
6-7 years (72-84 mo) 0.59
DATA ENTRY FOR ALTERNATE DRINKING WATER SOURCES
Concentration /•"•S/L
first-draw water 4
flushed water 1
fountain water 10
Percentage of total intake
first-draw water 50 %
flushed water 100 minus first draw
and fountain
fountain water 15
DATA ENTRY FOR SOIL/DUST (constant)
Concentration
soil 200 /xg/g
dust 200
Soil ingestion as percent of total soil and dust ingestion 45 %
_ DATA ENTRY FOR SOIL/DUST INGESTION (by year)
Soil/dust ingestion
Age = 0-1 year (0-11 mo), 0.085 g/day
1-2 years (12-23 mo) 0.135
2-3 years (24-35 mo) 0. 135
3-4 years (36-47 mo) 0.135
4-5 years (48-59 mo) 0. 100
5-6 years (60-71 mo) 0.090
6-7 years (72-84 mo) 0.085
3-4
-------
TABLE 3-1 (cont'd). DEFAULT VALUES FOR MODEL PARAMETERS
PARAMETER DEFAULT VALUE UNITS
DATA ENTRY FOR DUST (by year)
Dust lead concentration
1-2 years (12-23 mo) 0 /xg/g
2-3 years (24-35 mo) 0
3-4 years (36-47 mo) 0
4-5 years (48-59 mo) 0
5-6 years (60-71 mo) 0
6-7 years (72-84 mo) 0
DATA ENTRY FOR SOIL/DUST MULTIPLE SOURCE ANALYSIS (constant)
Ratio of dust lead concentration to soil lead concentration 0.70 unitless
Ratio of dust lead concentration to outdoor air lead concentration 100 jtg Pb/g dust per
fig Pb/m air
DATA ENTRY FOR SOIL/DUST MULTIPLE SOURCE ANALYSIS WITH ALTERNATIVE
HOUSEHOLD DUST LEAD SOURCES (constant)
Concentration
household dust 150 /*g/g
secondary occupational dust 1,200
school dust 200
daycare center dust 200
second home 200
interior lead-based paint 1,200
Percentage
household dust 100 minus all other %
secondary occupational dust 0
school dust 0
daycare center dust 0
second home 0
interior lead-based paint 0
BIOAVAILABILITY DATA ENTRY FOR ALL GUT ABSORPTION PATHWAYS
Total lead absorption (at low intake)
diet 50 %
drinking water 50
soil 30
dust 30
alternate source 0
Fraction of lead absorbed passively at high intake
diet 0.2 unitless
drinking Water 0.2
soil 0.2
dust 0.2
alternate source 0.2
3-5
-------
TABLE 3-1 (cont'd). DEFAULT VALUES FOR MODEL PARAMETERS
PARAMETER DEFAULT VALUE UNITS
DATA ENTRY FOR ALTERNATE SOURCES (by year)
Total lead intake
Age = 0-1 year (0-11 mo), 0 /xg/day
1-2 years (12-23 mo) 0
2-3 years (24-35 mo) 0
3-4 years (36-47 mo) 0
4-5 years (48-59 mo) 0
5-6 years (60-71 mo) 0
6-7 years (72-84 mo) 0
DATA ENTRY MENU FOR MATERNAL-TO-NEWBORN LEAD EXPOSURE
Mother's blood lead level at time of birth 2.5 /ug/dL
DATA ENTRY MENU FOR PLOTTING AND RISK ESTIMATION
Geometric standard deviation for blood lead, GSD 1.6 unitless
Blood lead level of concern, or cutoff 10 /ig/dL
COMPUTATION OPTIONS
Iteration time step for numerical integration 4 h
(1) comparison of predicted values from the Integrated Exposure/Uptake
Biokinetic (IEUBK) model with actual blood lead levels observed in a
blood lead and environmental lead field study such as illustrated in
Table 3-2.;
(2) risk estimation using predicted values from the IEUBK model with actual
environmental lead levels observed in an environmental lead field study;
(3) estimation of the effect of variability hi environmental lead levels on the
distribution of blood lead levels in the population;
(4) sensitivity analyses on the impact of environmental lead exposure.
The input data file for a batch run must be created outside the IEUBK model using whatever
text editor the user prefers. The following conventions MUST be observed in creating the
batch file:
3-6
-------
TABLE 3-2. FORMAT FOR BATCH MODE INPUT DATA FILE
(SIMULATED DATA)
Header line 1 > INPUT DATA FILE FOR A MONTE CARLO SIMULATION
Header line 2 > FIRST THREE COLUMNS RANDOM NOS. FOR SOIL, DUST, BLOOD
Header line 3 > ID FAM NBHD AGE PBS PBD PBW PBA ALT PBB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
FP-1
FP-2
FP-3
FP-4
FP-5
FP-6
FP-7
FP-8
FP-9
FP-10
FR-1
FR-2
FR-3
FR-4
FR-5
FR-6
FR-7
FR-8
FR-9
FR-10
FR-11
FR-12
FR-13
FR-14
FB-1
FB-2
FB-3
FB-4
FB-5
FB-6
FB-7
FB-8
FB-9
FB-10
FB-11
FB-12
FB-1 3
FB-14
FB-15
FB-15
PIONEER_HILL
PIONEER_HILL
PIONEER_HILL
PIONEER_HILL
PIONEER_HILL
PIONEER_HILL
PIONEER HILL
PIONEERJHLL
PIONEER HILL
PIONEER_HILL
RIVERSIDE
RIVERSIDE
RIVERSIDE
RIVERSIDE
RIVERSIDE
RIVERSIDE
RIVERSIDE
RIVERSIDE
RIVERSIDE
RIVERSIDE
RIVERSIDE
RIVERSIDE
RIVERSIDE
RIVERSIDE
BRIDGE ST
BRIDGE_ST
BRIDGE ST
BRIDGEJST
BRIDGE_ST
BRIDGE_ST
BRIDGE ST
BRIDGE_ST
BRIDGE_ST
BRIDGE_ST
BRIDGE_ST
BRIDGE_ST
BRIDGE_ST
BRIDGE_ST
BRIDGE_ST
BRIDGE ST
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
510.0
200.0
373.1
1042.5
519.2
3123.6
1938.2
287.5
1227.0
321.7
631.9
55.7
336.3
666.9
2005.6
394.8
1183.3
450.8
210.2
650.6
238.4
256.6
636.2
100.3
678.6
249.5
291.5
750.9
89.3
943.2
1668.6
1399.1
1226.8
1905.1
366.3
62.5
75.8
393.5
461.6
461.6
457.8
159.8
475.3
1361.6
332.7
1117.6
1295.8
649.9
1997.5
405.5
58.4
23.9
658.6
221.5
1532.1
92.7
692.8
83.6
94.7
682.0
314.2
48.0
219.6
31.5
605.4
136.3
127.8
2514.8
135.4
282.4
369.2
503.0
503.0
978.3
2727.8
464.7
49.3
88.6
883.7
883.7
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4,0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.6
2.3
3.4
8.7
2.9
6.3
4.9
11.8
23.3
4.4
5.6
2.0
4.1
3.2
26.2
2.3
2.7
2.4
1.5
4.5
2.0
2.0
4.4
3.2
10.0
2.0
4.1
4.5
3.3
6.7
11.9
6.0
5.9
4.6
13.4
3.9
5.4
1.2
1.7
1.9
3-7
-------
• The input data file must be an ASCII file with no special characters.
• The data set must have a DAT extension (i.e., [name].DAT).
• The first three lines of the input data file can be any identifiers that the user
requires; we usually use the first line for the run name, the second line for
modelling options used in the run, and the third line as headers for
variables in the data set;
• The data fields are entered format-free, although the use of regular spacings
and alignment of decimal points are recommended to improve readability;
• Maximum width 80 columns;
• Variable values should be separated by spaces;
• Missing values must be shown by an isolated decimal point as in some
examples in Chapter 5;
• Each line in the input data file must contain the following 10 variables:
1. Child identifier or code
2. Family or residence unit identifier or code
3. Area or neighborhood identifier code
4. Child's age in months
5. Soil lead concentration in jicg Pb/g
6. Dust lead concentration in /ig/g
7. Drinking water lead concentration in jwg/L
8. Air lead concentration in /tg/m3
9. Daily intake of lead from other sources, fig Pb/day
10. Observed child blood lead level.
In Chapter 5 we will demonstrate an approach to using the IEUBK model in the batch
mode option. Soil, dust, and blood lead "data" were simulated using realistic parameters for
sites with active air lead point sources. The attached Table 3-2 shows a batch mode input
data file for 40 children. All were assumed to be 18 months old and had default air, water,
and alternate source lead values. The first three columns in Table 3-2 are the child
identifier, the family identifier (note simulated twins as ID 39 and 40), and the
"neighborhood". These fields could be any alpha-numeric identifiers defined by the user.
3-8
-------
3.4 OUTPUTS FOR DOCUMENTATION, BRIEFING, AND
PRESENTATION
3.4.1 Overview of Output Options
The IEUBK model includes some output options that facilitate presentation of the
results and testing of the results for parameter sensitivity. These options may be a useful
part of the documentation for decisions in which the IEUBK model plays a role. We will
first describe the output options, and then show some of their applications. Some sensitivity
analyses can be facilitated by the use of these options.
3.4.1.1 Plotting (Option 2)
Single-Plot Options (Selections 2 and 3)
Some plot options can be used with single runs of the IEUBK model, and others require
multiple runs. Single-run options include:
• Plotting the log-normal probability density function of blood lead levels for
a single exposure scenario predicting geometric mean blood lead
(Selection 3);
• Plotting the cumulative probability distribution for exceeding any user
specified blood lead level of concern for a single exposure scenario
(Selection 2). This is sometimes called the exceedance probability
distribution.
The probability density function gives most users a better idea of the spread of blood
lead levels for children exposed to a single set of environmental lead concentrations. The
exceedance probability distribution may be used to visually estimate the fraction of children
above a blood lead level of concern for the single-exposure case (e.g., what fraction of
children are above 10 jug/dL), or to visually estimate the blood lead level corresponding to a
specified fraction of children (e.g., what blood lead concentration encompasses 95% of the
children). The user may route the probability plots to a printer after viewing the display.
Multiple-Plot Options (Selections 1, 4, and 5)
There are additional features that allow the user to combine output from several runs
onto single plots. These multiple-run options include:
3-9
-------
• Overlaid plots of the log-normal probability density functions of blood lead
levels for multiple exposure scenarios, where each run increases the lead
concentration in a specified medium by a user-defined amount (Selection 5);
• Overlaid plots of the cumulative exceedance probability distributions of
blood lead levels for multiple exposure scenarios, where each run increases
the lead concentration in a specified medium by a user-defined amount
(Selection 4);
• Plots of geometric mean blood lead levels versus environmental lead levels
for the medium whose values are varied (Selection 1).
Selection 1 cannot be used unless the user has previously created an output file from the
Computation Menu, designated *.PBM. Selections 4 or 5 cannot be used unless the user has
previously created an output file from the Computation Menu, designated *.LAY.
The overlaid probability density functions give most users a better idea of how the
probability of exceeding a blood lead level of concern increases with each increment in
environmental lead. The exceedance probability distributions may be used to estimate the
increases in the fraction of elevated blood lead levels or to visually estimate the
environmental lead levels corresponding to a specified fraction of non-protected children
above the level of concern.
3.4.1.2 Uses of Batch Mode Analysis (Option 4)
Results from multiple exposure scenarios can be accumulated using the batch mode
options. For sensitivity analyses involving constant concentrations in air, water, dust, soil,
or an alternate medium whose intake is constant, it is possible to create a batch mode input
file in which each line represents a different case for the analysis. However, with Option 4
it is only possible to carry out sensitivity analyses in which the cases differ on the basis of
concentration. Other types of sensitivity analyses require the accumulation of single runs in
an overlay file.
Another application in which batch mode methods are useful is a Monte Carlo analysis
in which all the modelled variability is assigned to differences in the environmental
concentrations. The results can be stored in a batch mode output file which may then be
used for statistical analyses.
3-10
-------
3.4.2 Detailed Instructions on Output Options
3.4.2.1 Save Output from a Single Run
1. Develop an exposure scenario.
2. Run the model (Selection 2 on Computation Menu, or F5 from any Data
Entry Menu).
3. Save results (Selection 2 after running the model). Results may be
appended to the file RESULTS.TXT, or added to an overlay plot file
defined by the user with name [name].LAY.
4. Results may be sent to a printer.
3.4.2.2 Save Output from Multiple Runs for Probability Plots (Option 3 on Range
Selection Menu)
1. Develop an exposure scenario
2. Use the Multiple Runs option (Option 2) on the Computation Menu.
1- Select the medium (Soil, Dust, Air, Water, Diet)
2- Select the lower and upper values for the medium
4- OUTPUT CHOICES
- Select number of steps from small to large
- Send results to overlay file RANGE(#+1).LAY, where the output
file is automatically named by increasing the index of the largest
numbered (#) current RANGE#.LAY file.
3. Results may be sent to a printer.
3.4.2.3 Save Output from Multiple Runs for Media-Level Plots (Option 3 on the
Computation Menu)
1. Develop an exposure scenario
2. Use the Media Run option (Option 3) on the Computation Menu.
1- Select the medium (Soil, Dust, Air, Water, Diet)
2- Select the lower and upper values for the medium
4- OUTPUT CHOICES
- Select number of steps from small to large
- Change the age range for calculating mean blood lead.
- Send results to overlay file MEDBLD(#+1).PBM, where the output file is
automatically named by increasing the index of the largest current
MEDBLD#.PBM file.
3-11
-------
3. Results may be sent to a printer.
3.4.2.4 Save Output from a Batch Mode Run (Option 4 on the Computation Menu)
1. Create a batch mode input data file, [name]. DAT.
2. Use the Batch Mode Run option (Option 4) on the Computation Menu.
3. Load the [name].DAT file you have created.
4. Rename the output data file [newname].* if required in the Run step.
5. Run the model.
The output data sets are named [newname].ASC and [newname].TXT, or
[name].ASC and [name].TXT if not renamed.
3.4.2.5 Probability Plots for Single Runs
For Current Runs Using Option 2 on Output Menu:
1. Run the model with user-defined exposure scenario in Option 2 on Output Menu.
2. Then choose Selection 2 or 3 in the Graphics Selection Menu.
3. Choose the Age Range.
4. Print graph, without exiting, by using the F10 key, then selecting printer type.
If the user has exited from the current Run, but has not done any further runs, then
Steps 3 through 5 can be executed by returning to the Graphics Menu and executing
Option 2 or 3.
For Current Runs Using Option 2 on Output Menu:
1. Run the model with user-defined exposure scenario in Option 2. If the user has
aborted the runs in Option 2, but has not done any runs since, then the last
complete run may be plotted as above. The procedure is:
2. Select Option 3 in the Main Menu, then option 2 on the Output Menu.
3. Then choose Selection 2 or 3 in the Graphics Selection Menu.
4. Choose the Age Range.
5. Print graph, without exiting, by using the F10 key, then selecting printer type.
3-12
-------
3.4.2.6 Probability Plots for Multiple Runs
1. Run the model with user-defined exposure scenario in Option 2 on the
Computation Menu. Do not abort the runs in Option 2.
2. Select Option 3 in the Main Menu, then Option 2 on the Output Menu.
3. Then choose Selection 4 or 5 in the Graphics Selection Menu.
4. Identify the *.LAY data set to plot.
5. Choose the Age Range.
6. Print graph, without exiting, by using the F10 key, then selecting printer type.
3.4.2.7 Multi-Level Plots for Blood Lead Versus Media Lead
1. Run the model with user-defined exposure scenario in Option 2 on the
Computation Menu.
2. Select Option 3 in the Main Menu, then Option 2 on the Output Menu.
3. Then choose Selection 1 in the Graphics Selection Menu.
4. Identify the MEDBLD#.PBM data set to plot.
5. Print graph, without exiting, by using the F10 key, then selecting printer type.
Batch mode files can be used for display or documentation through the statistics
module on the Batch Mode Menu, Option 4.
3.4.3 Recommendations on Multi-Level Soil Lead Exposure Scenarios
The IEUBK model carries out multi-level analyses by increasing the concentration of
the user-specified medium by equal steps at each run, and holding everything else constant.
When evaluating different soil lead exposure scenarios, it may be preferable to keep a
constant soil-to-dust coefficient so that dust lead exposure increases with increasing soil lead
exposure. This can be done by first invoking the Multiple Source Analysis for dust and
defining the dust lead to soil lead coefficient. This is particularly important if some
component of the soil lead abatement is expected to permanently alter the soil to dust
pathway.
3-13
-------
4. MORE ABOUT THE MODEL1
4.1 LEAD BIOAVAILABILITY
4.1.1 Background
The concept of bioavailability is important for site-specific risk assessments for lead.
The concept springs from the fact that lead potentially available to produce harm and found
in exposure pathways or in body receiving compartments (lung, skin, gut) must reach the
biological sites of action in order for an adverse health effect to occur in exposed humans or
ecological biota.
This section focuses primarily on the bioavailability of inorganic lead from soils and
dusts. Lead bioavailability from air and drinking water is also important and is discussed in
limited detail below. In order to provide coherent and useful guidance to the reader and user
of this chapter, it is subdivided into (1) introductory material that includes definitions of
bioavailability and resource material in the technical literature; (2) the close lead absorption-
bioavailability relationships, including the physiological and biochemical mechanisms of lead
absorption and the many, complex factors that influence such uptake; (3) the main focus of
the chapter, bioavailability as it relates to human and experimental toxicology, including the
various biophysico-chemical and environmental aspects of the lead exposure matrix,
methodological approaches in toxicology for quantifying bioavailability, the increasingly
important question of relevant experimental animal models for quantifying lead bioavailability
in humans; and, finally, (4) a summary and critical overview, which attempts to spell out the
appropriate uses of bioavailability information and limits to use this information in site-
specific risk assessment.
4.1.2 Definitions
A clear agreement on a definition of bioavailability should be established before one
presents a detailed discussion of this topic. The difficulty here is that there are various
This chapter is intended to provide guidance on some technically advanced applications of the model. We have
attempted to provide the best scientific documentation available but recognize that new information may become
available in these rapidly advancing fields. The user is referred to Section 1.6 for information on how to get
additional and more up-to-date assistance with specific applications of the model.
4-1
-------
definitions of bioavailability depending on the scientific discipline using the term and the
technical context of use.
Typically, the pharmacologist or toxicologist or others in biomedical disciplines are
concerned with measuring bioavailability as that fraction of the total amount of material in
contact with a body portal of entry (lung, gut, skin) that then enters the blood. For the
purpose of describing the Integrated Exposure Uptake Biokinetic (IEUBK) Model, this is the
definition to be used in this manual. However, an aquatic biologist may define
bioavailability as that fraction of material solubilrzed in the water column under certain
conditions of hardness and pH. An aquatic toxicologist might consider contaminants which
are soluble under specific stream conditions to be bioavailable to fish or benthic organisms.
A biochemist or biochemical toxicologist would consider bioavailability with reference to that
fraction of a toxicant which is available at the organ or cellular site of toxicity.
The above definitions can be viewed as dosimetrically descriptive. There are
quantitative methodological definitions that figure as well. As described later, bioavailability
can be defined as being absolute or relative (comparative). Absolute bioavailability, for
example, is the amount of substance entering the blood via a particular biological pathway
relative to the absolute amount that has been ingested. Relative bioavailability of lead is
indexed by comparing the bioavailability of one chemical species or form of lead with that of
another form of lead. A second methodological description for bioavai]ability that is used by
toxicologists is the ratio of areas under the dose-response curve for either of two forms of
lead, or two methods of administration. Typically, the latter involves comparing injected
with orally administered doses.
4.1.3 Literature Sources on Bioavailability
More detailed reviews and discussions of the topic of lead bioavailability in humans and
experimental animals have been presented by Mushak (1991) and Chaney et al. (1988).
As is evident from these reviews, our present understanding of lead bioavailability has
developed from both human and animal studies. For further in-depth discussion of the
various components of bioavailability, for example, lead absorption, the: reader is also
referred to the following documents: (1) the Air Quality Criteria Document for Lead (U.S.
Environmental Protection Agency, 1986), and (2) the Proceedings of the Symposium on the
Bioavailability and Dietary Exposure of Lead (1991).
4-2
-------
Citations of key specific studies are provided in the relevant sections and subsections of
this chapter rather than here, so as to be less disruptive to the reader.
4.1.4 Lead Absorption-Bioavailability Relationships
By definition, the absorption (uptake) of lead into the circulation is the critical kinetic
component of the overall process called bioavailability. Not only the amount, but also the
rate of uptake of that given amount is important, particularly under acute or subacute
exposure conditions, and when dealing with lead-containing media in the gastrointestinal (GI)
tract. Such material is itself moving through the GI tract within a relatively short time
period. Consequently, the biological and physiological characteristics of absorption, the
subcellular mechanisms of absorption, and the factors influencing its occurrence must be
understood in order to understand the resulting phenomenon. The focus of this chapter is
soil and dust lead ingested (swallowed) by populations at risk, requiring that lead uptake
phenomena in the gastrointestinal tract be given most of the attention.
Species-specific anatomical and physiological determinants of GI absorption are the
macroscopic factors that provide the basic means by which lead absorption occurs. As noted
in more detail in Section 4.1.5, there are major structural differences in the anatomy of the
GI tract of various mammalian species that would affect lead absorption. Similarly, it is the
physiology of the mucosal lining (epithelium) of the mammalian GI tract that is the first
dynamic determinant of lead movement from the GI tract to the bloodstream.
4.1.5 Cellular and Subcellular Mechanisms of Lead Absorption
Lead absorption is believed to proceed by several cellular mechanisms involving the
enterocytes, cells lining the intestinal wall (Figure 4-1) (e.g., Mushak, 1991). Absorption
also entails complex interactions with the uptake of essential nutrients such as calcium, iron
and phosphate (Barton et al., 1978, 1981; Mahaffey-Six and Goyer, 1972).
The first uptake mechanism may be diffusion through the gut lumen driven by a
concentration gradient from the luminal surface lining the intestine to the basolateral surface
(vascular side). This mechanism is likely to depend to some extent on the concentration of
2 +
ionic or unbound lead ion (Pb ), and consequently would depend on the solubility
characteristics of lead species of interest. This may be a passive diffusion process requiring
no energy input. It involves either intracellular or paracellular movement of lead across the
4-3
-------
Pb++
Tight Junctioi
Intermediate
Junction
Microvilli
Tight Junction
Terminal Web
Enterocyte
Desmosome
Intercellular
Space
Basement
Membrane
Figure 4-1. Schematic drawing of the enterocyte showing possible mechanisms for lead
absorption. Possible mechanisms include: (1) an active or facilitated
component; (2) a transcellular component perhaps involving pinocytotic
mechanisms; and (3) a diffusion-driven paracellular route across tight
junctions.
Source: Mushak, 1991, adapted from Morton et al. (1985).
wall. Paracellular transport would entail movement across the area between cells called
"tight junctions."
In the second possibility, lead may enter the gut tissue (but not necessarily the
bloodstream) by pinocytosis or other vesicular mechanisms. In pinocytosis, lead-bearing
media in a liquid micro region of the gut are engulfed by the (enterocyte) cell membrane.
Such encapsulating may involve lead in either a truly soluble or an emulsified/suspended
form that is then carried to blood or to sites of toxic action. This process is biochemically
analogous to handling of solid particles in phagocytosis.
Perhaps the quantitatively most important transport mechanism in environmental
exposures typical for most individuals is energy-driven active transport, exploiting
homeostatic transport mechanisms in place for calcium and iron transport (e.g., calcium
binding protein [CaBP] or calbindin D), and under control of an enzyme—calcium,
2 -f 2 +
magnesium-dependent ATPase (Ca , Mg -ATPase)—involved in the absorption and
regulation of blood calcium levels and located in the basolateral membrane of mucosal
epithelial cells. This active component of lead absorption displays a strong age dependence,
4-4
-------
being more important at younger ages. It is interesting that some of the transport systems
that bring calcium into the body seem to have an even higher affinity for lead than for
calcium (e.g., Fullmer et al., 1985).
While the results of experimental studies can be described quantitatively, the precise
nature of biological and biochemical mechanisms in lead bioavailability is not yet completely
understood. There is, however, a useful characterization of lead absorption mechanisms as
either saturable (facilitated) or nonsaturable (passive). These various and complex
biochemical/cellular mechanisms obviously have important implications for experimental
models of human lead bioavailability, particularly with reference to comparison of in vivo to
in vitro simple chemical simulation models.
4.1.6 Factors Affecting Lead Absorption
Lead uptake, especially from the GI tract, does not occur in a physiological vacuum but
is the outcome of a complex set of interactions with other inorganic and organic substances,
particularly such nutrients as calcium, iron, phosphate, vitamin D, fats, etc., as they occur in
meals or with intermittent eating. In addition, uptake is a function of developmental stage
(age), administered dose, the chemical species and the particle size of the lead-containing
media.
It is well known that lead uptake is markedly lower with consumption of meals than
under fasting conditions in adults (e.g., James et al., 1985; Rabinowitz et al., 1980) and
presumably in children as well. Human data, in the aggregate, indicate that calcium, iron
and other cations interact strongly as competitors to lead uptake so that lead uptake generally
increases as dietary levels of these nutrients decrease (Mushak, 1991; U.S. Environmental
Protection Agency, 1986). In rats, Garber and Wei (1974) showed that fasting increased the
amount of lead taken up by the gut. Children are likely to be exposed to lead under a
variety of fed or fasted (between meal) conditions. Therefore, any interpretations of lead
bioavailability studies of site-specific characteristics should include the effect on uptake of
food and time since eating.
There is a developmental or age dependency for the extent of lead absorption in both
humans and experimental animals (Mushak, 1991; U.S. Environmental Protection Agency,
1986). Prepubertal children absorb more lead than do adults (Alexander et al., 1973; Ziegler
et al., 1978). Experimental animal studies support the human data. Studies using rats
showed that pre-weanling animals absorb 40 to 50 times more of a given dose of lead than
4-5
-------
do adult animals (Kostial et al., 1971, 1978; Forbes and Reina, 1972), while infant monkeys
will absorb 16 to 21 times more lead than adult monkeys (Munro et al., 1975). Possible
mechanisms for this age dependence have been discussed (Weis and LaVelle, L991; Mushak,
1991). The design or interpretation of bioavailability studies, aimed at assessing lead
absorption for children, must consider age dependence of uptake of lead in any adjustments
of the bioavailability parameter in the UBK model.
Human data indicate a dose dependence to the absorption of lead (Sherlock and Quinn,
1986). In duplicate diet studies of bottle-fed infants (5 to 7 kg) exposed to lead in water and
in formula mixed with contaminated water, Sherlock and Quinn were able to quantify the
dose dependence of lead absorption. Over the exposure range investigated in the study (40 to
3,000 /*g/week), these investigators determined that the relationship between blood lead
concentration and lead intake was curvilinear (Figure 4-2). This opportunistic human data
describing the dose-dependence of lead absorption was considered by the Agency when
establishing the kinetic approach to lead absorption used in the IEUBK Model.
35
e-
I 30
IS
^ 25
I
-------
dose-dependent inhibition of uptake is consistent with an active transport mechanism that
requires lead-inhibited enzyme(s) for its operation and which also becomes saturated at
higher lead dosings (Aungst and Fung, 1981; Mykkanen and Wasserman, 1981). Design and
interpretation of studies to assess bioavailability of lead should also consider dose dependency
in site-specific assessments.
Finally, the metal species and particle size may influence the solubility, and because of
that, the bioavailability of lead. Experimental studies using relatively simple lead species
showed that lead as the sulfide, chromate, naphthenate or octoate was less bioavailable (44 to
67%) relative to the more soluble carbonate (Barltrop and Meek, 1975). Barltrop and Meek
(1979) also demonstrated an inverse relationship between lead uptake from leaded paint and
particle size.
On the other hand, other investigators have documented that lead species that are
relatively insoluble under simple in vitro conditions are as bioavailable as soluble salts under
conditions of fasting (LaVelle et al., 1991; Rabinowitz et al., 1980).
4.1.7 Bioavailability of Lead in Soils and Dusts
Quantitative approaches to estimating bioavailability for purposes of the IEUBK model
require consideration of three issues. The first, of course, is the physicochemical nature of
the site-specific environmental media containing lead and what this suggests for behavior of
lead-containing media in the GI tract (i.e., biophysico-chemical behavior). As noted earlier,
particle size and chemical species are important. Equally important is the environmental
matrix within which some particular chemical species of lead is to be found. The
physicochemical complexity of these environmental matrices (e.g., dusts and soils, mining
and process waste) considerably exceeds that of simple, laboratory forms. The second aspect
is methodological: how one can quantify bioavailability in experimental or observational
studies? Finally, it is critical that users of this manual and model understand the merits and
the limits of the various types (classes) of bioavailability studies that can be done on a site-
specific basis.
4.1.7.1 Biophysico-Chemical and Environmental Features of the Exposure Matrix
Types of Soil Lead Contamination
Environmental lead is found in a variety of chemical and physical forms. Lead-
contaminated areas could be categorized according to the type of industry or lead-generating
4-7
-------
processes associated with the site. Since we are concerned principally with lead in dusts and
soils, these are the media of most site-specific concern.
Urban area sites are typically contaminated with those chemical forms arising from
either the combustion of leaded gasoline (alkyl lead species such as tetraethyl lead used as
anti-knock agents) at high levels in past years or from flakes, chips and dusts from exterior
and interior lead-based paint.
Dust or soil lead originating from auto exhaust typically begins as lead-mixed halides
(chloride, bromide) but undergoes transformation quickly to the oxide or sulfate (U.S.
Environmental Protection Agency, 1986), two relatively bioavailable forms. Auto emission
paniculate is typically of small diameter (one micron or less), especially on residential
surfaces farther away from roadways, where distant atmospheric transport is more favored
than for the heavier particles that are deposited closer to the traffic sources. Such particles
are also readily breathed into the lungs and readily stick to the hands of children, to family
pets, etc. (U.S. Environmental Protection Agency, 1986; Mushak, 1991).
Paint lead is typically found as carbonate, chromate or octoate, and the element may
represent up to 70% of the weight of the dried paint product. While lead-paint surfaces are
intact, leaded paint would only become available as young children chew on accessible
surfaces like painted furniture. In older structures, with surface aging, window and door
frame abrasion, and deterioration of leaded paint surfaces, paint will flake, chip, chalk
(interior) or weather (exterior) and become an important source of lead exposure for
children. The nature of this material, especially as small adherent flecks and fine dusts, and
its significant solubility are factors likely to favor significant bioavailability. The greatest
numbers of lead-painted residential units are found in urban areas, but any unit anywhere
built before 1978 may have lead-based paint.
Battery recycling plants, typically containing secondary lead smelting capacity, are
often found as localized sources of environmental lead. Waste byproducts of this kind of
lead processing include lead sulfate (sulfuric acid) on casings, and battery sulfuric acid itself,
mobilizing lead into and through soils of limited buffering capacity. Lead from this material,
either as feedstock or from secondary smelter stack emissions, is apt to be of small particle
size as well. These factors warrant estimating bioavailability at the upper end of the range.
In nonferrous mining areas, lead is commonly found in a variety of material produced
by hard rock mining, milling, and smelting processes. It is beyond the scope of this chapter
4-8
-------
to present a detailed discussion of lead contamination with nonferrous mining, milling and
smelting. The reader is referred to the review by Mushak (1991) for further details.
Mining waste can be broadly characterized as: (1) waste rock; (2) mill tailings; and
(3) smelting waste. Waste rock is that material removed from the mine but having
insufficient mineral economic value to warrant processing. This material is typically
discarded at openings to the mine, consists of larger particles, and may or may not be
enriched in heavy metals.
Mill tailing is material that has been processed by a variety of physical grinding,
separating and enrichment processes. This material typically has smaller particle size than
the less processed wastes and the material is enriched in toxic elements, including lead.
Mineral content depends on the characteristics of the ore body and the milling process, and
g
may range from soluble carbonates (K^ ~ 10" ) to extremely insoluble phosphates
-80
(Ksp » 10" ) of lead. Furthermore, lead that is associated with mining waste may either be
freely exposed at the particle surface or entirely encapsulated, so that the lead is not available
to be dissolved in simple solvents like water.
Smelting waste may exist in many forms. Air- and water-quenched slags are strikingly
different in their physical nature. Water-quenched material is typically of fine particle size,
while air quenching results in large chunks of oxidized slag. Chemically, these slags consist
of various metal oxides and include lead and silicon oxides. Bag house dust consists of the
fine paniculate matter trapped in the emissions stream by a simple bag filter prior to leaving
the stack. This material is very high in toxic metal content, including lead, and occurs in
very small particle size. These small particles include lead sulfate and oxide species. Dross
is the foam or lighter fraction of the liquid product of the floatation process. When cool, it
may be discarded, resulting in a potentially important exposure source.
4.1.7.2 Is There a Better Way To Classify Lead-Contaminated Sites?
It is often convenient to discuss lead-contaminated sites by classifying them as mining,
smelting, urban or battery sites. As our understanding of the complexities of lead-
contaminated sites improves, it becomes less and less useful to use these simplified
descriptions. For example, mining areas typically are associated with present or historical
milling and smelting. Significant smelter-related contamination may remain at closed and
operating mines that can contribute to typical mine waste exposure concerns.
4-9
-------
Mine wastes may consist of lead in a multitude of physical and chemical forms as
discussed above, making generalizations about exposure or potential exposure (and
bioavailability) inappropriate without additional applied research data. Mining and smelting
areas may share exposure sources often associated with such urban areas. Adequate
characterization of lead-contaminated media, for the purpose of estimating bioavailability,
should include assessment of physical and chemical parameters (e.g., particle size and
appropriate media solubility) as well as biophysico-chemical characterisljcs. Generalizations
regarding the source of lead contamination which do not address risk-specific details of the
physicochemical and biochemical nature of the waste are not as useful for predicting health
risks from exposures.
4.1.7.3 Methodological Approaches to Quantifying Bioavailability
While lead can have severe toxic effects following a single very high exposure, we are
primarily concerned in this chapter with relatively low levels of average exposure and
average blood lead concentration (see Figures 4-3 and 4-4 for single versus multiple
exposures and target organ concentrations).
The average near steady state (pseudoequilibrium) of an accumulating toxicant such as
lead in blood following chronic (repetitive) exposure is proportional to the amount absorbed
during each exposure. At low ingestion rates, where absorption and bickinetic processes are
nearly linear, the following relationship applies between changes in blood lead and changes
in chronic exposure:
A PhR = ^ Pb-abs./day * mean residence time in blood pool
volume of distribution in blood pool
Methods used to describe the fraction absorbed from exposure are well established and
will be the primary focus of the following discussion.
4.1.7.4 Determination of Absolute Bioavailability
The methodology for quantifying absolute bioavailability in toxicology commonly
compares (a) the area under the time-versus-blood-concentration curve (AUC) following
intravenous (IV) injection with (b) an equivalent dose and a similar AUC measurement
following ingestion of the substance being investigated. The ratio of AUCora to AUC is
then taken as a measure of percent absorption in the gut. From this, absolute bioavailability
over a short time frame may be defined as:
4-10
-------
1.0
50
100
Time (h)
150
200
Figure 43, The time-course of bioavailability of lead in the blood (•) and in the brain
(•) of juvenile rats following a single dose. Note that accumulation of lead
in the target tissue (brain) continues as blood lead decreases. The
significance of brain levels indicated is unknown.
Source. Adapicd iu>iu Moni'iilovic and Kostial (1974).
Mean Blood Lead
Time (arbitrary units)
4 4 Khirtics of absorption during repeated dosing. At steady state, the area
under the curve described by one dosing interval is equivalent to the area
under the curve following a single, bolus dose.
4-11
-------
Absolute bioavailability = r- x 100%.
(AUC)IV (DOSE)oral
While careful attention must be given to presystemic elimination (the amount of
chemical excreted via the GI tract prior to entry into the systemic circulation), this simple
approach can provide, with appropriate sampling and analytical quality control, an effective
estimate of the percent absorption into the blood following oral exposure. The longer term
kinetics (concentration versus time) of chronic lead absorption are likely to be influenced by
the accumulation of lead in peripheral compartments such as bone. Thus, bioavailability
estimates conducted with longer-term exposures are preferable in developing quantitative
estimates of lead bioavailability. The reader is referred to Gibaldi (1982) for a more detailed
discussion of the kinetics of absorption and distribution of toxicants.
4.1.7.5 Absolute Versus Relative Bioavailability
It is usually the case that bioavailability is quantified in absolute teims: it is presumed
to be equal to the absorbed fraction for a specific substance. For example, if CdCl2 were
6% absorbed from some medium and CdS were 3% absorbed at equimolar concentrations,
the absolute bioavailability for these compounds would be 6 and 3%, respectively.
There are occasions, however, where bioavailability may be specified not in absolute
but in relative terms, relative to the bioavailability of some reference compound. Using the
earlier examples, if CdCl2 were the reference compound, then the relative bioavailability of
the sulfide would be 50% (3%/6% X 100). This approach has much practical value,
because one may not have direct bioavailability data for oiher than one or two forms when
estimating risks.
This approach would therefore have value for comparative exposure risk when adjusting
risk calculations at Superfund sites. Here, risks are usually calculated from Reference Doses
(RfDs) and cancer slope factors that are nearly all based on administered, rather than
absorbed, doses. If site-specific exposures involve different chemical/physical forms, it may
be necessary to adjust intake dose to uptake dose values in order to account for differing
bioavailability in estimating toxicity levels. In such cases, absolute bioavailability
measurements may be useful for site-specific forms but are not required for relative risk
determinations. While the lead model uses absolute bioavailability as the input parameter,
knowledge of the relative bioavailability of ingested materials may be applied. If the relative
bioavailability of the material of interest is known relative to a second material whose
4-12
-------
absolute bioavailability can be assessed, then the absolute bioavailability of the first can also
hr estimated.
j,i a,-,niton to establishing the distinction between absolute and relative bioavailability, it
!» ru'-ccwuy to distinguish between biovailability and solubility. Solubility is a metabolically
passive ''impiifiod, in vitro characteristic of a substance that constitutes but one element in
unavailability. This distinction is explored in the following section.
4.1.7.6 Quantitative Experimental Models of Human Lead Bioavailability
Site-specific bioavailability studies of lead in soil have been conducted for several
hazardous, waste sites in the western United States (LaVelle et al., 1991; Freeman et al.,
1991; Weis et al., 1994). In cases where (1) current exposure is significant, (2) soil
i IK n. icrisi.c^ preclude simple extrapolation from existing studies, and (3) estimated cleanup
ci •'sis are sufficiently high, such studies may improve the accuracy and the reliability of the
risk assessment process. Site-specific bioavailability studies can be expensive, can require
if-i••„ Fur oinpleuon, and do require considerable technical expertise for the design and
i.ud'.iuU of ih«-, studies. This means that the remedial project manager (RPM) or risk
n sic is •> mi-ill manager needs to obtain advice from individuals with training and experience in
'!H>, area. If .xperirnental studies are needed, the toxicology expert may recommend studies
ui .iic of me following levels, in order of increasing cost and complexity.
Lias* I Study
Mudit-.s in this class consist of simplified, in vitro approaches in which one determines
d-.iu.-'H'S solubility of lead from various solid species. This approach has little utility for
quj mitral i ve human bioavailability assessments. First, solubility itself is but one factor, and a
;; :ruue
-------
speaking, induces a shift in intraintestinal equilibria among lead forms in the direction of
greater dissolution (to compensate for the lead removed by active transport). Such active
uptake produces a complex process that yields more bioavailability than predicted in simple
in vitro approaches. This shift in equilibrium is compelled by a simple, widely-known
principle of chemical processes, Le Chatelier's Principle, that states (CRC, 1978):
If some stress is brought to bear upon a system in equilibrium, a change occurs,
such that the equilibrium is displaced in a direction that tends to undo the effect of
the stress.
In the present case, the stress is active intestinal uptake and the displacement to undo the
effect is to dissolve more lead during its passage through the gut. Such a shift, relative to a
simple bench-top system, is depicted in Figures 4-5 and 4-6.
Class II Study
Class n and Class ffl studies involve in vivo animal models of human bioavailability of
lead. They differ in their experimental specifics. Class n investigations are intermediate
in vivo studies (i.e., carried out over a relatively short time). Such studies examine the
bioavailability of lead within a time frame in which the dosing ends before pseudoequilibriurn
in the central (blood) compartment is reached. Since lead accumulates in critically important
peripheral compartments such as bone and this accumulation will influence longer term
uptake and distribution values, longer term studies are desirable for assessing target tissue
bioavailability of lead in mammals.
Class n studies are useful in terms of providing a relative index of lead bioavailability,
that is, comparison of several lead forms. Class n studies should, of course, consider all the
factors already noted that influence any in vivo lead study, including the target population
and pathway specifics for the site, age, concentration dependence of lead uptake in the dosing
regimen, nutrition, physiology and anatomic structural characteristics.
In terms of model biology, physiology, and behavior, an appropriate selection for
human simulation would take account of eating/feeding habits, human versus animal
gastrointestinal tract differences, comparative biochemistry, etc.
Class III Study
Bioavailability investigations that have as their purpose the site-specific adjustment of
the default bioavailability parameters in the IEUBK model may require: a more complex
approach. Such advanced studies should only be conducted after consultation with qualified,
4-14
-------
PbX
(s)
PbX
(aq)
Figure 4-5. Under conditions of equilibrium, the amount of lead as the free ion
is limited by mass balance dissolution of the solid phase (PbX).
B.
Blood
PbX
PbX
(aq)
(s)
Intestinal mucosa
Figure 4-6. Under physiological conditions, free lead ion (Pb +) is removed from
solution by active and passive absorption mechanisms potentially shifting
the equilibrium of the dissolution process far to the left.
4-15
-------
experienced individuals and should be subject to the most rigid quality assurance/quality
control (QA/QC) protocols for study management. This especially applies to preserving the
original physicochemical form of the lead-containing test materials from a particular site.
The design and duration of Class HI studies should be such that they assure achievement of
near steady state (pseudoequilibrium) for the blood concentration versus time curves.
As with Class n studies, Class IQ investigations need to take account of the site-specific
target population and exposure pathways, age of subjects, nutritional and physiological state
of the animal, etc.
4.1.7.7 Summary and Advisory Overview for Lead in Soils and Dust
Unavailability studies are intended to provide valid information about the associations
of site-specific physical and chemical properties of exposure media with bioavailability at a
target tissue site. Properly designed studies can elucidate differences traceable to such
factors as the physicochemical properties of the site's lead-containing media, lead chemical
form, matrix species, particle size, mixture effects from other metals or other chemical
species from matrix, diet, and such, and study animal or human population variables such as
age and levels of exposure. These studies need to meet two fundamental qualifications:
(1) Doses used need to be low enough to be comparable to human exposure
situations that are to be assessed. Basing calculations on high doses of
lead may greatly weaken the utility of an experimental study.
(2) Animal models need to be carefully examined for their appropriateness to
represent human gut processing and absorption of lead. The
demonstration that absolute bioavailability is low in an animal model is of
limited significance unless that model can be supported as being
quantitatively relevant to humans.
Bioavailability factors can be validly adjusted to account for site-specific lead exposure
characteristics in the IEUBK model. However, selection of a site-specific bioavailability
parameter other than the model default value of 30% for soils and dusts requires considerable
caution and warrants review by qualified technical experts.
4.1.8 Bioavailability of Lead in the Diet
The absorption of lead from food and liquid diet by infants up to six months old is
known to be very high (Ryu et al., 1983; Marcus, 1989a), and much lower in adults
4-16
-------
(Chamberlain et al., 1978; Blake and Mann, 1983; Rabinowitz et al., 1980; James et al.,
1985). Less is known about changes in lead absorption from diet for older infants, toddlers,
and children. A value of 50% was selected as an intermediate level in children and infants
(U.S. Environmental Protection Agency, 1990b).
The exact form of the dietary lead absorption coefficient in humans is not known.
There is evidence that the absorption of lead in food by infants is quite high, at least 40 to
50%. The range cited by the U.S. Environmental Protection Agency (1989a) is 42 to 53%.
While this probably decreases after infancy, we have no direct evidence on how to
interpolate this range for children of ages 2 to 6. A smoothing of the absorption data from
infant to juvenile baboon in the studies by Harley and Kneip (1985) has been proposed as a
basis for extrapolation by the U.S. Environmental Protection Agency (1989a). In view of the
uncertainty about this, we have chosen to keep the same default value of 50% for ages
1 to 6. This value will, at worst, slightly overestimate dietary lead uptake in older children.
Lead absorption from diet depends on the lead concentration in the stomach, and on a
host of other dietary cofactors such as zinc, iron, vitamins, and phytate. When dietary lead
intake during meals is sufficiently high, absorption of lead through the gut lumen decreases,
probably due to competition for the limited anionic lead-binding sites on the gut wall.
The absorption of lead has some similarities with the absorption of other metals
(Mushak, 1991), especially alkaline earths such as calcium and strontium. Calcium
researchers have hypothesized three possible mechanisms of gut absorption. The first is a
type of saturable active transport. This may be a secondary process because the enzyme
requiring energy input is on the basolateral membrane and not on the membrane of the gut
lumen. It would be more accurate to describe this as a facilitated diffusion process.
A second saturable facilitated process involving pinocytic mechanisms has also been
hypothesized by calcium researchers, but is not well understood. These saturable diffusion
processes are the dominant modes of transport at low concentrations. Processes requiring
carriers are often called facilitated diffusion processes. For convenience, we may call either
of these saturable processes facilitated diffusion processes. The third process, the dominant
mode of transport at high concentrations, is probably a simple diffusion through tight
junctions on the lumenal side and is not saturable. Binding and transport of calcium across
the gut lumen involves a protein called calbindin. We have described this as a passive
diffusion process. The last two processes have no specific inhibitors and are difficult to
study. The extent to which lead absorption shares these calcium processes, or is
quantitatively different, is not known. The study by Aungst and Fung (1981) on transport of
4-17
-------
dissolved lead across the gut lumen in vitro in everted rat intestines shows that lead
absorption is likely to consist of two distinct processes. The first process depends on a
passive diffusion mechanism that is independent of gut concentration. The second process
depends on a facilitated diffusion mechanism that is saturable, with a half-saturation
concentration of about 120 /xg/L (0.59 ^mol). The quantitative extrapolation of this value to
human children in vivo is uncertain.
The Glasgow duplicate diet study reported results on infant blood lead and dietary lead
intake at a single time point, age 3 months. There appeared to be a very large non-dietary
background source contributing about 12 /*g/dL blood lead to these infants. This is attributed
in part to the inhalation of leaded gasoline, which was still widely used in the United
Kingdom, and in part to residual exposure pre-natally. The dietary lead intake in these
infants is believed to constitute almost all of the ingested lead, since children at this young
age are believed to have minimal contact with soil, house dust, or paint. Some small
contribution of inhaled lead particles may be transferred to the ingestion route by mucociliary
transport.
A non-linear regression model was fitted to the Sherlock and Quinn data in a form that
is directly comparable to the Michaelis-Menten formula used to describe in vitro studies
(Aungst and Fung 1981ab). The model that was fitted to all data was:
log(Blood lead) = log(B + L * Pblntake + K * Pblntake / (1 + Pblntake / M)).
The parameters have the following interpretation:
B = background lead concentration from pre-natal and inhalation exposure;
L = linear (passive) uptake coefficient between blood lead and dietary lead intake;
K = non-linear (facilitated) uptake coefficient between blood lead and dietary lead
intake;
M = Michaelis-Menten type (non-linearity) parameter, the daily dietary lead intake
rate at which the facilitated component of lead uptake is half saturated.
Three methods were used to estimate the parameters. The first two methods are based on
weights for the grouped data shown in Figure 4-2 of Sherlock and Quinn (1986), shown in
this document as Figure 4-1. The first set of weights was based on the estimated sample size
within each bar on the graph. The second method was based on the normalized coefficients
of variation from the standard error bars for each group. The third method was based on
4-18
-------
using within-cell geometric mean blood lead and dietary lead intake values from Table 4-2 in
their paper, with cell counts used as weights. The first method appears to be the most
accurate, both absolutely and relatively. The fitted blood lead model is:
Blood lead = 10.85 + 0.0090 Pblntake + 0.2981 Pblntake / (1 + Pblntake / 90.33)
The total blood lead to lead intake regression coefficient at low intake levels (much less than
the Michaelis-Menten coefficient M = 90 /ig/day) is K + L = 0.0090 + 0.2981 =
0.307 /ug/dl per /zg/day. The goodness of fit of this non-linear lead uptake model of
Michaelis-Menten form to 3-month old human children, combined with the similar piecewise
linear model that could be fitted to the water lead studies, and the goodness of fit of the
Michaelis-Menten model found for the data on blood lead and lead intake data in infant and
juvenile baboons presented by Mallon (1983) support the use of this model for lead
absorption in older children as well. The suggestion by Chamberlain (1984) that absorption
in adults is greatly reduced at intake rates above 300 /ig/d is also consistent with the infant
estimate of 90 jug Pb/day.
4.1.9 Bioavailability of Lead in Water
The bioavailability of dissolved lead salts in drinldng water is very high when
consumed by adults between meals (James et al., 1985), and very low when consumed with
meals. The maximum retention of lead in children probably exceeds that of adults, which is
about 60% on an empty stomach, and absorption is likely to be only somewhat smaller than
retention. Thus the value of 50% is recommended as plausible. A range of values for water
lead absorption from the U.S. EPA/OAQPS Staff Paper (1989a), shown in Table 4-1, should
be used as a basis for age-variable absorption coefficients.
The volume of water in a typical United States faucet is about 90 to 125 milliliters, and
at least two or three faucet volumes must be drawn before the tap water lead concentration
decreases to the level of the source water and water distribution line lead concentrations
(Schock and Neff, 1988; Gardels and Sorg, 1989; Marcus, 1991a). The sample volume of
first-draw water specified in U.S. EPA's drinking water regulation is 1 L (U.S.
Environmental Protection Agency, 1991c). Water lead concentrations in most U.S. water
supply systems are low (<5 /-ig/L), but geometric means may exceed 10 to 20 jug/L in
first-draw samples from systems with highly corrosive water and a great deal of lead
plumbing, which is not uncommon in older urban areas in the northeastern United States.
4-19
-------
TABLE 4-1. PffiCEWISE LINEAR REGRESSION MODELS FOR
BLOOD LEAD VERSUS WATER LEAD IN THREE STUDIES
Parameter
Intercept
Glasgow
Infants
N = 91
12.82
Edinburgh
School Children
N = 495
6.841
!!;•,,•
Children and Adult
N - 180
*1,2
CONC. for
SLOPE CHANGE
16.4
Intercept depends on other covariates.
Not given.
Source: Marcus (1989b) and Maes et al. (1991).
15.0
15.0
SLOPE (< CHANGE)
pg/dL per fj.g/L
SLOPE (> CHANGE)
/xg/dL per /ig/L
0.254
0.0426
0.161
0.0318
0.130
0.0242
Even if the community mean is low, lead in drinking water in some households may be
sufficiently high to cause overt lead poisoning (Cosgrove et al., 1989).
4.1.10 Bioavailability of Lead in Air
Lead on aerosol particles must be inhaled and deposited before pulmonary absorption
can occur. Particles inhaled but not deposited may be exhaled or trapped by the mucociliary
lift mechanism and ingested. The number of inhaled particles of a given size range varies
with the ambient concentration and size distribution and the breathing rat;. The breathing
rate varies with age and physical activity. Inorganic lead in ambient air consists primarily of
paniculate aerosols with a size distribution determined largely by the nature of the source
and proximity to it. In rural and urban environments, This size distribution is usually from
0.05 to 1 micron. Near point sources, particles greater than 10 microns prevail.
Deposition in the respiratory tract can be by inertial impaction in the nasopharyngeal
regions, where the airstream velocity is high, or by sedimentation and inierception in the
4-20
-------
tracheobronchial and alveolar regions, where the airstream velocities are lower. In the
alveolar region, diffusion and electrostatic precipitation also become important.
Particles greater than 2.5 microns are deposited in the ciliated regions of the
nasopharyngeal and tracheobronchial airways, where they are passed to the gastrointestinal
tract by the mucociliary lift mechanism. Particles small enough to penetrate the alveolar
region can be dissolved and absorbed into systemic circulation or ingested by macrophagic
cells. Evidence that lead does not accumulated in the lungs suggests that lead entering the
alveolar region is completely absorbed (Barry, 1975; Gross et al., 1975). Rabinowitz et al.
(1977) found about 90% of the deposited lead was absorbed daily. In the IEUBK model the
default assumption is that 35 % of the inhaled lead is bioaccessible (reaches the absorbing
surface), and 100% of this is absorbed.
4.2 USING THE INTEGRATED EXPOSURE UPTAKE BIOKINETIC
MODEL FOR RISK ESTIMATION
4.2.1 Why Is Variability Important?
4.2.1.1 Intent of the Model and the Measure
The Geometric Standard Deviation (GSD) as used in this manual is a measure of the
relative variability in blood lead of a child of a specified age, or children from a hypothetical
population, whose lead exposures in a specified dwelling are known. The GSD is intended to
reflect the jive types of individual blood lead variability identified below, not variability in
blood lead concentrations where different individuals are exposed to substantially different
media concentrations of lead.
The IEUBK Lead Model is intended to be used for individual children who live at a
residence, or for a hypothetical population of children who may live there in the future, or
for hypothetical children who may some day live in a house built on a plot of now vacant
land of appropriate size for future construction of a single residential dwelling unit.
4.2.1.2 Individual Geometric Standard Deviation
Why do different children have different blood lead levels? The answer to this question
has two parts. The first part of the answer is that children are exposed to different levels of
lead in their community environment. The second part of the answer is that individual
4-21
-------
children, exposed to exactly the same measured levels of lead, will still have different h'^nd
lead levels for the following reasons:
— Different Environmental Context. Carpeting, other furnishings, end accessibility of
yard soil affect contact with environmental lead in ways that are not easily
measured.
— Behavioral Differences. Interaction \vith caretakers, with ••" >;\: artd playmates.
and other factors that affect mouthing behavior and play activity vill nmdify lead
intake from dust and soil.
— Different Exposures. The children will have different exposures due to differences
in contact with soil, dust, water, and other environmental media that vary at
different locations and different times, so that no single sample of environmental
lead in any medium can be said to completely characterize the child's actual
activity-weighted exposure to lead in that medium.
— Measurement Variability. The environmental lead measurements, are not perfectly
reproducible due to sampling location variability, repeat samplinjs variability, and
analytical method error, so that equality of measured sample lead concentrations
does not imply equality of the true exposure concentrations.
— Biological Diversity. Children are biologically diverse so that even children of the
same age, weight, and height are expected to have differences in the biokinetic
distribution and elimination of lead.
— Food Consumption Differences. A number of factors, including nutritional status
and time of ingestion of lead relative to meal times, affect the uptake or absorption
of lead ingested from a medium.
While sociodemographic factors underly many of these differences, it is not appropriate
to assume any specific effect for future residents. Risk estimates should bs applicable to any
hypothetical resident, and this requirement adds to the variability associated with the
estimate.
4-22
-------
4.2.2 Variability Between Individuals Is Characterized by the Geometric
Standard Deviation
Inter-individual variability is the starting point for risk analysis using the IEUBK
model. Even if we knew the correct value for all of the environmental exposure variables,
we could at best predict only the typical blood lead level expected for a child of a certain age
who had that exposure. We will therefore assume that individual child blood lead levels can
be divided into two parts, a predicted blood lead and a random deviation from the predicted
blood lead level. A statistical model that has proven to be very useful and fits all of the
blood lead studies we have analyzed is based on the following three assumptions:
(Assumption 1) Observed blood lead = (Predicted blood lead) * (Random deviation);
(Assumption 2) The random deviation is log-normally distributed with geometric mean or
median = 1, and a geometric standard deviation (GSD) defined by
GSD = exp (standard deviation of In (blood lead). Here, exp(.) denotes
the exponential function and ln(.) denotes the natural logarithm;
(Assumption 3) The GSD is the same for all values of the predicted blood lead (i.e., for
all values of environmental exposure).
Risk is the probability of exceeding the blood lead level of concern. The IEUBK model
calculates risk from these three assumptions. The user provides an exposure scenario from
which the IEUBK model calculates a predicted blood lead. Then the user provides a blood
lead level of concern, whose default value is now defined as 10 /ig/dL based on health
effects criteria, but can be modified by the user. This risk is calculated as the probability
that a standardized, normally distributed random variable exceeds the level Z, where
Z = In (blood lead level of concern/predicted blood lead) / In (GSD).
If Z = 1.645, the risk is 5%. If Z = 1.96, the risk is 2.5%. If the GSD is increased, then
Z is decreased, and the risk of a blood lead level exceeding the level of concern is increased
(provided that the blood lead level of concern is larger than the predicted blood lead, which
is usually true). This is illustrated in Figure 4-7. The default value of Z is
Z = In (10/predicted blood lead) / In (1.6)
= (2.3026 - In (predicted blood lead)) / 0.47.
4-23
-------
LOC
LOC
Figure 4-7. The impact of the relative positions of the level of concern (LOC) and the
geometric mean (GM) on the proportion of children "at risk" for two
populations with different GSDs. If LOC > GM, then the area for
children at risk (shaded plus solid) for GSD = 1.7 is greater than the area
(solid) for GSD = 1.4. If LOC < GSD, then the area for children not at
risk (shaded plus solid in lower tail) for GSD = 1.7 is larger than the area
for children not at risk (solid in lower tail) for GSD = 1.4.
The GSD has been estimated in a number of ways. The statistical model has the same
form as the model used as the basis for estimation of slope factors reported in the Air
Quality Criteria Document for Lead (U.S. Environmental Protection Agency, 1986). The
GSD values were estimated by exp(s), where s denotes the residual standard deviation of the
fitted In (geometric mean blood lead) as a function of the environmental lead concentrations
and of demographic cofactors. The residual standard deviation estimate for In (blood lead) in
a system of structural equations for lead was also used to estimate in some more recent
studies.
Estimates of GSD for lead mining and smelter sites have increased towards larger GSD
values as the geometric mean blood lead levels at those sites have decreased. This probably
reflects the fact that at low to moderate levels of exposure, lead levels are likely to be
influenced by several media with similar media-specific uptake rates, rather than by a single
dominant medium. This condition tends to magnify individual differences in intake behavior
or in biokinetics, and increases the GSD. The GSD estimates for severa.1 mining and smelter
sites ranged from 1.30 to 1.79 (Marcus, 1992). We chose a value smaller than the
4-24
-------
maximum that is consistent with the remaining variability after differences in the usual site
specific soil lead and dust lead measurements have been removed. The remaining sources of
variability include not only biological and behavioral variability in the children, but also
repeat sampling variability, sample location variability, and analytical error. For empirical
support in selecting a site specific GSD see Appendix A.
The default value is:
GSD = 1.6.
This default value is based on calculations of GSDs from specific sites. The median GSDs
weighted for sample size within cells were estimated as 1.69 for Midvale, 1.53 for the
Baltimore data of the Urban Soil Lead Abatement Demonstration Project, and 1.60 for the
Butte study. This type of adjusted GSD calculation was chosen because of its treatment of
outliers. Other types of adjusted GSDs, such as those derived from structural analyses are
described below.
We must discourage the user from changing the GSD value by use of empirical site-
specific data from a blood lead study. As discussed in Section 4.5 below, blood lead studies
may be subject to subtle sampling biases and changes in child behavior in response to the
study. The GSD value reflects child behavior and biokinetic variability. Unless there are
great differences in child behavior and lead biokinetics among different sites, the GSD values
should be similar at all sites, and site-specific GSD values should not be needed.
The user may wish to demonstrate that the variability in a specific well-conducted
blood-lead study is consistent with the default assumption. In the next section, we will
describe how to estimate a site-specific, inter-individual GSD when necessary. These
analyses should be done only when necessary, and with thorough documentation of the
reasons why the site may have more or less variation among child behavioral and biological
parameters than at most other sites. We must remind the user that it is not necessary to have
site-specific blood lead data in order to appropriately use the model with the default GSD.
4.2.3 Statistical Methods for Estimating the Geometric Standard
Deviation from Blood Lead Studies
We have used several statistical methods to estimate GSD values recommended here.
Two methods are described in detail in Appendix A. The first method is a direct method in
4-25
-------
which environmental lead levels are fixed in ranges or intervals, and blood lead variability
for children exposed to these concentrations is calculated directly. The second method is a
statistical regression method appropriate to the generally skewed distribution of blood lead
values and estimates the variability in blood lead concentrations after an empirical estimate of
blood lead concentrations expected at each environmental lead concentration. The two
methods give reasonably consistent results. The regression method uses child-specific age
and lead concentration. The regression method crudely mimics the IEUBK model.
4.2.4 Choosing the Geometric Standard Deviation: Intra-Neighborhood
Variability
There have been some cases in which the IEUBK Lead Model or a preceeding model
was used to estimate the distribution of blood lead in a community when only community-
level input was available, such as geometric mean soil, dust, and air lead. Further
experience with the IEUBK model suggests that this application may be appropriate under
some conditions in which certain mathematical assumptions are approximately correct.
It also suggests that there are some other situations in which this approach is incorrect
because the necessary mathematical assumptions are not satisfied. At this time, we
recommend using the IEUBK Lead Model for neighborhood and individual blood lead
assessment, but not for communities or for larger scale blood lead assessments without
carefully evaluating the input assumptions. The neighborhood scale assessment requires
stratifying the neighborhood by intervals of soil and dust lead.
A neighborhood is a spatially contiguous area that often has identifiable physical or
geographical boundaries. For the purposes of this manual, a neighborhood is characterized
according to the following guidelines:
• Boundaries such as a highway, railroad right-of-way, river, or by non-residential
land uses such as commercial, industrial, agricultural, or park;
• Approximately 400 households with about 100 children;
• Church, school, and retail establishments within walking distances;
• Diameter about 1.5 kilometers (1 mile).
4-26
-------
The neighborhood concept is used here to classify small areas of relatively similar
childhood lead exposure, and will rarely be the same as a census tract, political locale w 'i
as a precinct or ward, or community association membership area.
Input parameters for the model at a neighborhood scale should be some measure that
characterizes typical exposure concentration in a medium, such as the arithmetic tman or
geometric mean, or the median. When activity pattern or behavior weighted exposme
information is unavailable, we recommend use of the arithmetic mean to characterise soil
lead concentrations in areas that are sufficiently small that any part of the area may be
accessible to a typical child living at a random residence located within the area. This will
certainly be applicable to the yard and adjacent play areas of a single residence.
Our recommended approach for risk estimation involves more calculations than the
single-input soil and dust lead, but much less calculation than the use of each individual yard
or housing unit. Our approach requires the division of the neighborhood into units th^t are
larger than single yards or other sites, but smaller than the whole neighborhood, and dearly
must depend on the scale of a risk assessment. Risk within a neighborhood can he assessed
in a single model run only if media concentrations of lead are relatively homogeneous
between different residential sites.
There is no definition of a "community" for model use. It is expected that older
children will be able to play anywhere within a neighborhood, but are limited to their own
neighborhood within the community. An alternative approach is to define "neighborhoods"
by isopleths or contours of soil lead concentrations, but this is more likely to be useful in the
vicinity of active or inactive smelter or battery recycling plants, where soil lead deposition
has a definite point source pattern. No specific approach based on Geographic Information
Systems (GIS) data bases has yet been adopted. The definition of neighborhood scale
suggested here is roughly equivalent to an area of 4 to 10 city blocks in many urban areas
(160 to 240 meters square). A neighborhood should not be larger than a one kilometer
square.
4.2.5 Basis for Neighborhood Scale Risk Estimation
The basis of the neighborhood approach is that a few important environmental
parameters largely determine the predicted geometric mean blood lead. Since the
environmental lead concentrations are known to have some measurement error, there should
be little loss of accuracy in grouping the environmental lead concentrations by small
4-27
-------
intervals. For example, the interval ranges for soil lead concentration could be
0 to 249 /tg/g, 250 to 499 fig/g, 500 to 749 jug/g, etc. Soil lead levels in an interval, for
example from 250 to 499 /*g/g, would be described by a single number in that range, such as
the midpoint of the interval at 375
One of the most important determinants of blood lead concentration in children is lead
in household dust. It is necessary to use small intervals of dust lead concentration along with
small intervals of soil lead concentration. There are many other sources of lead in household
dust in addition to soil lead, including dust lead from air lead deposition, from interior lead-
based paint, and from workplace dust carried home by adults residing in the house. The
actual range of dust lead concentrations corresponding to a soil lead interval is therefore
generally much wider than the range of soil lead concentrations.
There may be circumstances in which other lead exposures in a neighborhood are
known, and vary over a wide range. For example, there may be information on water lead
concentrations in different houses. Some of the houses may have sufficiently high water lead
concentrations that lead in water becomes another significant source of lead exposure.
Additional stratification or classification of sites by this variable may also be useful.
Neighborhoods defined by small geographic areas are also much more likely to be
homogeneous with respect to sociodemographic factors that affect blood lead variability.
There should be some similarity in child activity patterns, household environmental contexts,
behavioral patterns, and nutritional patterns within a neighborhood. Therefore, the individual
GSD may be applied plausibly to the relatively homogeneous subpopulation within a
neighborhood. If the neighborhood defined initially is very heterogeneous, then a larger
GSD may be needed. It would be better to subdivide the neighborhood defined initially into
more homogeneous subareas. This requires knowledge about the neighborhood residents, or
an assumption about future residents.
4.2.6 Relationship Between Geometric Standard Deviation and Risk
Estimation
The GSD is a very sensitive parameter for risk estimation. In this model, we use
"risk" in the following specific ways:
4-28
-------
• Individual risk is the probability that a hypothetical child living in a particular
house or dwelling unit characterized by its environmental lead levels will have a
blood lead concentration that exceeds a user-specified level of concern;
• Neighborhood or community risk is the fraction of children in a neighborhood or
community characterized by a specified distribution of environmental lead
concentrations that are expected to have blood lead concentrations exceeding a
user-specified level of concern.
The assessment of potential health risk from environmental exposure to toxicants is one
of EPA's most significant activities. We are using only part of this process. An elevated
blood lead concentration (however one defines "elevated") is an index of internal exposure or
body burden of lead. It is a useful index precisely because it changes in response to changes
in exposure, with characteristic time scales of a few days or so in plasma and red blood
cells, reflecting deeper changes of a few months in soft tissues, and years in hard bone.
An elevated blood lead concentration is not precisely an adverse health effect by itself, but
has been a very useful predictor of an increased likelihood of neurobehavioral deficits in
children. The "risk" involved here is the risk of an increase in an easily measured index of
lead exposure that is a predictor of adverse health effects.
The most general form of the model is multiplicative:
Blood lead = controllable factors * random factors.
For a single child, with defined sources of exposure, the IEUBK model estimates the
geometric mean blood lead, or typical blood lead (i.e., the median when variability is log-
normal, as it usually is). The model then is given by:
Blood lead = GM * exp( Z * In(GSD))
where GM is the model-predicted geometric mean blood lead, exp(.) is the exponential
function, ln(.) is the natural logarithm function, and Z is a normally distributed random
variable. Therefore risk, defined as a probability for a single child, is calculated by the
equation
Risk = Probability{Blood lead > level of concern for given exposure)
4-29
-------
= Probability{Z > (ln(level of concern) - ln(GM)) / In(GSD)}.
When the level of concern is greater than the expected or typical blood lead at that exposure,
then risk increases when GSD increases. Figure 4-7 illustrates the difference in "at risk"
children for two populations, one with a GSD of 1.4 and another with 1.7. When the level
of concern is above the geometric mean, the population with the higher GSD has a greater
proportion of the children at risk. When the level of concern is less than geometric mean,
the population with the lower GSD has a greater proportion of children at risk.
4.2.7 Risk Estimation at a Neighborhood or Community Scale
4.2.7.1 What Do We Mean by "Neighborhood" or "Community" Risk?
Representative questions of interest in assessing the risk of elevated child blood lead in
a neighborhood are:
• What is the frequency distribution of risk of exceeding a blood lead concentration of
concern, such as 10 /ig/dL, within the neighborhood?
• What fraction of a hypothetical or actual population of children would be expected to
exceed some specified blood lead concentration of concern if they resided in the
representative sample of houses in this neighborhood for which we have soil and
dust lead data?
• How much could we reduce high individual risk or the fraction of children with
elevated blood lead concentrations by cleaning up soil to some specified level?
• What is the distribution of risks for a hypothetical population of children if housing
units were constructed on soil at this vacant site?
The implicit definition of risk in these questions is the fraction of children living in a
dwelling unit anywhere in the neighborhood who have elevated blood lead levels. We see
that the neighborhood or community risk level has two distinct components of variability:
(1) Inter-individual differences, as in Section 4.2.4; and
(2) Inter-dwelling unit differences in lead exposure.
4-30
-------
In some circumstances, these two can be combined and the same approach used to
estimate the fraction of children at risk in a neighborhood. But, if there is a broad
distribution of inter-dwelling unit differences, as is commonly observed, then a simplistic
application of the IEUBK model may substantially under-estimate the real risk from the most
contaminated parts of the neighborhood. Whatever the distribution of inter-dwelling unit or
intra-neighborhood exposure levels, the "sum of risks" approach can always be applied.
Note that there is a subtle difference between inter-dwelling exposure and intra-neighborhood
exposure. Inter-dwelling exposure distribution would be the distribution of exposures
measured in each home and would assume that the individual exposure is within the property
boundaries of the dwelling unit. Intra-neighborhood exposure would include additional
exposure from nonproperty sources, such as parks, schools and playgrounds.
4.2.7.2 Neighborhood Risk Estimation as the Sum of Individual Risks
Neighborhood risk is based on the expected number of children in the neighborhood
who have elevated blood lead levels, here taken as greater than 10 /ug/dL. Using the
computer model, some of these questions can be addressed by the following procedure:
1. Set up a batch mode file in which each line represents the age and environmental
lead exposure of each child in the real or hypothetical population.
2. Use the IEUBK Lead Model to estimate the geometric mean blood lead for each
child in the batch mode file.
3. Apply an individual GSD to estimate the probability of exceeding the blood lead
level of concern for each child or each household in the batch mode file.
4. Calculate the expected number of blood lead values exceeding the level of concern
by adding up the probability of exceeding the blood lead level of concern across all
children in the batch mode file.
Note that even houses with low lead concentrations have a small positive risk for
resident children. In houses with high lead concentrations, the risk of elevated blood
lead is much larger, but some children (even in those high lead houses) will not have
elevated blood lead concentrations. The total of all such risks characterizes
neighborhood exposure.
4-31
-------
5. Neighborhood risk is the ratio of the calculated expected number of blood lead
values exceeding the level of concern to the total number of children in the batch
mode file. This last point is illustrated in the following narrative.
4.2.7.3 An Example for the "Sum of Individual Risks" Approach
Suppose that there are data on four households with children in a neighborhood.
Residents of each household are exposed to lead-contaminated soil. The first house has
250 jug/g lead in soil, the second has 250 pg/g, the third has 1000 /Ag/g, and the fourth house
has 1000 /ig/g. We have assumed dust lead concentrations as 70 percent of the soil lead
concentration in houses 2 and 4, and as 15 percent of the soil lead concentration in houses
1 and 3. We have added 10 ^g/g to dust lead as an estimate of the air lead contribution to
dust lead at 0.1 /ig Pb per cubic meter of air. The respective dust lead concentrations are
thus 47.5 /ig/g, 185 pg/g, 160 /xg/g, and 710 /*g/g.
The neighborhood is usually not just 4 houses. We may have samples at only these
4 houses, or there may be 100 houses at each of these 4 soil and dust lead concentrations.
The assumption is that the samples are representative of the exposure distribution in the
neighborhood. We are showing calculations for four houses only for the purposes of
illustration. The risk estimates are intended to be unbiased estimates of potential risk for
other years in which different children, not in the current sample, may occupy the same or
other houses in the neighborhood. Obviously, a reliable estimate of neighborhood risk will
require many more than 4 houses.
All other parameters are set to default values. We used a soil and dust absorption
model with 30% absorption of lead from both dust and soil. (Smaller values of soil lead
absorption may be needed for some sites—see Section 4.1). We assumed GSD = 1.6; larger
values of GSD may be needed at some sites. The probability density of blood lead for four
houses is shown in Figure 4-8.
For the house with soil lead at 250 /ig/g and dust lead at 47.5 /xg/g, we expect 0.55%
of children to exceed 10 ju.g/dL. For the house with 250 /ig/g soil lead and 185 /ug/g dust
lead, we expect 1.99% to exceed 10 /ig/dL. For the house with soil lead at 1,000 /xg/g and
dust lead at 160 /ig/g, we expect 21.06% of children to exceed 10 /-eg/dL. For the house
with 1000 j«g/g soil lead and 710 j«g/g dust lead, we expect 42.68% to exceed 10 /zg/dL.
The sum of the risks for these four houses is 0.55% + 1.99% + 21.06% +42.68% =
66.28% children = 0.6628 children expected to exceed 10 /xg/dL, or an average risk for the
neighborhood of 66.28% 74 = 16.57%, which is greater than the 5% neighborhood risk
4-32
-------
0.4
0.3
0)
Q
o.2
0.1
0.0
House 4
House 3
House 2
House 1
5 10 15 20
Blood Lead, ug/dL
Figure 4-8. Probability density of blood lead in houses 1 to 4.
25
target used in this example. However, the major part of the risk falls in the one house with
high soil and dust lead concentrations.
The use of aggregate neighborhood input data requires that we compare the probability
density function (PDF) and elevated blood lead (EBL) risk calculated from aggregate
parameters with the correct PDF and EBL risk functions, which are the mathematical
composites of the individual PDF and risk functions. Expressed mathematically:
true neighborhood PDF = (PDF(site 1) + PDF(site 2) +...)/N
true neighborhood risk = (risk(site 1) + risk(site 2) +...)/N
(Equation 4-1)
The approach we have outlined here does not require any mathematical assumptions about
the distribution of soil and dust lead concentrations, nor of any other parameters or variables
except for blood lead. We have assumed that the conditional distribution of individual blood
lead is log-normal with a constant GSD (given specified values of lead exposure variables
that determine the geometric mean blood lead for individuals with that exact environment).
The method suggested here is the most convenient and flexible framework we have found for
neighborhood assessment of the effect of soil lead abatement.
4-33
-------
4.2.7.4 Assessment of Risk Using Grouped Data for a Neighborhood
The example in the preceding section had a "neighborhood" with only 4 houses, so that
the amount of work required was not very burdensome. In the real world, the site manager
or risk assessor may be dealing with relatively homogeneous neighborhoods or small
communities with several hundred households. These calculations can be simplified by
grouping soil and dust lead levels into small cells with fixed ranges of values. The grouped
data within each cell are all assigned the same value, such as the midpoint of the interval.
Each cell is then assigned a statistical weight. The statistical weights could be:
(1) The number of housing units with soil and dust lead concentrations in the
interval;
(2) The number of children observed or expected to live in housing units with soil
and dust lead concentrations in the interval;
(3) The fraction of housing in a neighborhood that is expected to have soil and dust
lead concentrations in the interval;
(4) The fraction of area in as-yet-undeveloped neighborhoods with soil and dust lead
concentrations in the interval.
The probability density function (PDF) and risk of EBL children in then the weighted sum of
the cell PDF or cell risks. If the respective weights are denoted weight (cell 1), weight
(cell 2), etc., and the PDFs are denoted PDF (cell 1), PDF (cell 2), etc., and the risks are
denoted risk (cell 1), risk (cell 2), etc., then:
neighborhood PDF = [weight (cell 1) * PDF (cell 1) +• weight (cell 2) *
PDF (cell 2) + etc.] / [weight (cell 1) + weight (ceil 2) + etc.]
neighborhood risk = [weight (cell 1) * risk (cell 1) + weight call (cell 2) *
risk (cell 2) + ...] / [weight (cell 1) + weight (cell 2) + ....]
The following hypothetical example may illustrate these points. Suppose that a random
sample of 250 houses and apartments has been obtained in a neighborhood. The number of
houses in each interval of 250 /ig/g soil and 250 ^g/g dust lead is shown in Table 4-2. This
4-34
-------
TABLE 4-2. EXAMPLE OF NEIGHBORHOOD RISK ESTIMATION WITH
GROUPED DATA
Hypothetical example of grouped data for a neighborhood with dust and soil samples of
250 sites of house yards. Intervals are 250 jig/g hi soil and in dust lead.
Soil
Interval
0-249
0-249
0-249
250-499
250-499
250-499
250-499
500-749
500-749
500-749
500-749
750-999
750-999
1000-1249
TOTAL
Soil
Midpoint
125
125
125
375
375
375
375
625
625
625
625
875
875
1125
Dust
Interval
0-249
250-499
500-749
0-249
250-499
500-749
750-999
250-499
500-749
750-999
1000-1249
1000-1249
1750-1999
1250-1499
Dust
Midpoint
125
375
625
125
375
625
875
375
625
875
1125
1125
1875
1375
Statistical
Weight
30
50
20
10
40
30
20
10
20
10
3
4
1
2
250
Blood Lead1
(Mg/dL)
2.9
4.3
5.7
4.1
5.4
6.7
7.9
6.5
7.7
8.8
9.9
10.8
13.6
12.5
Risk2
Percent
0.39
3.45
10.61
2.70
9.36
18.62
28.52
16.45
26.86
38.16
47.56
52.78
72.73
66.93
14.28
Calculated from IEUBK model with default parameters.
Assuming GSD = 1.6.
TABLE 4-3. EXAMPLE OF NEIGHBORHOOD RISK ESTIMATION
WITH COARSELY GROUPED DATA
Hypothetical example of grouped data for the same neighborhood as in Table 4-1,
with intervals of 500 /ig/g in soil and dust lead.
Soil
Interval
0-499
0-499
500-999
500-999
500-999
500-999
1000-1499
TOTAL
Soil
Midpoint
250
250
750
750
750
750
1250
Dust
Interval
0-499
500-999
0-499
500-999
1000-1499
1500-1999
1000-1499
Dust
Midpoint
250
750
250
750
1250
1750
1250
Statistical
Weight
130
70
10
30
7
1
2.
250
Blood Lead1
fcig/dL)
4.2
6.8
6.4
8.7
10.9
12.8
12.4
Risk2
Percent
3.05
19.81
15.45
36.05
55.50
66.92
64.01
14.41
Calculated from IEUBK model with default parameters, ages 6 to 84 mo.
Assuming GSD = 1.6.
4-35
-------
same example is shown in Table 4-3 in intervals of 500 /ig/g in soil and 500 /*g/g in dust.
There is no requirement that there be equal interval lengths in either soil or dust.
The user may then calculate neighborhood risk in three ways:
• Sum of risks for 250 housing units;
• Sum of risks for 14 cells or groups of width 250 jug/g soil and dust;
• Sum of risks for 7 cells or groups of width 500 /xg/g in soil and dust.
The results of calculations are shown in the Tables 4-2 and 4-3. The total risk in Table 4-3
is calculated as:
(130 * 3.05% + 70 * 19.81% + 10 * 15.45% + 30 * 36.05% +
7 * 55.50% + 1 * 66.92% + 2 * 64.01 %)/250 = 14.41%
The risk calculation in Table 4-2 is similar. If there are not too many cells, the amount of
calculation can be strikingly reduced. However, as the intervals are made longer, there is a
corresponding loss of accuracy in the neighborhood risk estimate. The extra effort in
calculating risks with 250 /xg/g intervals (14 cells) is probably compensated by the increased
precision, with an estimate of 14.28% instead of 14.41%. The actual risk for the ungrouped
sample with 250 simulated houses in 14.13%.
4.2.7.5 Assessment of Risk with Neighborhood or Neighborhood-Scale Input
There are situations in which it is either inconvenient or impossible to apply the IBUBlv
model at the intended household residence scale. For example, if only neighborhood mean
values or geometric mean values of input parameters such as soil and dust lead are available,
the model estimate may be far less reliable than if individual residential measurements were
made. Another possibility is that there are a substantial number of soil and dust lead
measurements at a site, but not at houses or locations within the site where blood lead and
EBL risk estimates are needed, for example, to compare with blood leads observed at
residences where there are no environmental data. There are some circumstances in which
this is clearly not a valid application of the model. As we do not clearly understand the
range of conditions under which the IEUBK model may be used with large-scale input data at
this time, we must discourage use of the IEUBK model except with single-residence or
residential lot-sized input data, or with data grouped into cells as in Section 4.2.7.4.
4-36
-------
4.3 ENVIRONMENTAL PATHWAY ANALYSIS
4.3.1 Concept of Pathway Analysis
Environmental pathways for lead have been a subject of interest for EPA for a long
time. Methods for analyzing with multi-media exposure pathways from air lead were used in
developing slope factors for blood lead versus air lead, dust lead, and soil lead in EPA's Air
Quality Criteria document (U.S. Environmental Protection Agency, 1986). Even though the
focus was on exposure to air lead as a primary source, it was clearly recognized that
whatever the source of lead in air, paint, or soil, the primary exposure vector for young
children was through fine particles of surface soil and household dust that adhered to the
children's fingers and were ingested in the course of normal hand-to-mouth contact at ages
one to five years. Thus the total impact of air lead exposure had to be evaluated as the sum
of exposure over several pathways (Figure 4-9).
Auto
Emissions
Crustal
Weathering
Paint,
Industrial
Dusts
Figure 4-9. Exposure pathways of lead in the environment.
4-37
-------
The IEUBK model assists the user in defining critical pathways for each exposure
scenario. For example, there are two places on the Multiple Source Analysis Menu for
household dust where pathway information may be inserted:
(1) the soil-to-to-dust pathway coefficient may be entered in the first line of the
Multiple Source Analysis Data Entry Menu, replacing the default value of
0.70 fig/g dust per /xg/g soil lead;
(2) the direct air-to-dust pathway coefficient is on the second line of the Multiple
Source Analysis .Data Hntry Menu, replacing the default value of 100 /xg/g dust
2
lead per /*g/m air Pb.
The following paragraphs provide the basis for some of the default parameters used in
(lit iliUBiv model, and suggest some methods for estimating alternate coefficients from site-
specific data, provided the user has some knowledge of statistical regression. While physical
measurement methods such as a comparison of stable lead isotope composition ratios have
been used tor source apportionment studies (Yaffee et al. 1983; Rabinowitz 1987), most
users will probably have to infer site-specific pathway parameters by statistical analyses of
4,3.2 Pathway Analyses by Linear Regression
liic slope factor approach, described in Section 1.5, determines the linear relationship
n r,vo pathway components. This methods was used for the EPA Exposure Analysis
illation (U.S. Environmental Protection Agency, 1989a) to show that there
• c- i )u «jsliip bt-i..vt m lead in air (PbA), lead in soil (PbS), and lead in dust (PbD).
::-[ :i.uu,0imp may be approximately linear, depending on properties of soil and dust lead
paiticles; if not linear, then it is at least a positive cause and effect relationship. The
relationship was established using data from air lead point sources such as primary and
secondary lead smelters, other non-ferrous metal smelters, and lead battery plants. The
analysis, using mean values, found the relationship:
PbD = bDO + bDA PbA + bDS PbS (Equation 4.3-2)
M'hrirc: bDS = 0.364 for all point source communities, but = 0.894 for the East Helena
primary lead smelter community. This suggests that there may be substantial differences
among communities in terms of soil-to-dust transfer.
4-38
-------
The direct air-to-soil relationship was also estimated:
PbS = bso + bSA PbA (Equation 4.3-3)
When estimating slope factors by a sequence of regression equations the user should be
aware that the "measurement errors" in pathway equations will almost certainly attenuate the
size of the regression coefficients, and could even reverse the sign of the coefficients
(Kupper 1984). Structural equation modeling techniques attempt to resolve this problem by
the simultaneous estimation of coefficients in pathway models in the face of measurement
errors.
4.3.3 Pathway Analysis Using Structural Equation Models
Systems of linear equations in which the output of one equation (such as PbD predicted
from PbS and XRFI) is used as the input or predictor in another equation (such as PbB from
PbD) can be reliably estimated using a method known as structural equation models (Bollen
1990). This method was introduced in the analysis of blood lead data by Bornschein et al.
(1985) and Clark et al. (1985) in their analyses of data from the Cincinnati Prospective
Childhood Lead Study. Several authors have extensively explored applications of the method
to environmental lead data (Marcus 1991; Burgoon and Menten 1991). Two different
approaches were compared, and found to produce very similar results.
The first approach uses linear equations without logarithmic transformation, but with a
robust method of estimation that is not sensitive to skewness or to instability of measurement
error variances. (The software implementation in the EQS program (Bentler 1989) was
particularly convenient.) For a lead mining community, or an urban area in which air lead
levels are so low as to be negligible predictors of blood lead, A typical small system of
equations might be:
PbB = bgo + bBS PbS + bgj) PbD + b^ XRFI (Equation 4.3-6)
PbD = bDO + bDXI XRFI + bDS PbS (Equation 4.3-7)
PbS = bso + bSXE XRFE + bSage House-age (Equation 4.3-8)
where: PbB = blood lead concentration (/xg/dL)
4-39
-------
PbS = soil lead concentration (jwg/g)
PbD = dust lead concentration
XRFI = interior measurement of lead-based paint by XRF (mg/cm2)
XRFE = exterior measurement of lead-based paint by XRF (mg/cm )
bnm = raw regression coefficient where the subscript refers
to: response variable n on predictor m,
BO = intercept
BD = blood to dust
BXI = blood to XRFI
DO = intercept
DXI = dust to XRF (interior)
DS = dust to soil
SO = intercept
BS = blood to soil
SXE = soil to XRF (exterior)
Sage = soil to house age
This model assumes that there is direct ingestion of interior lead paint, which also contributes
to household dust, but no direct ingestion of exterior lead-based paint. The exterior
lead-based paint contribution is subsumed in the paint to soil to dust pathway. Because of
the linear equation formulation, partial effects of lead source terms are preserved:
PbB = CBO + CBS PbS + CBXI XRFI (Equation 4.3-9)
PbB = dBO + dBXE XRFE + dBXI XRFI (Equation 4.3-10)
CBS = bBS + bfiD bps (Equation 4.3-11)
dBXl = bBXl + bBD boxi (Equation 4.3-12)
dBXE = (bss + bfiD bos) bsxE (Equation 4.3-13)
where: cBm = composite regression coefficient for blood on predictor m
d^ = composite regression coefficient for indirect pathways from predictor m
to response n
4-40
-------
The second approach uses logarithmic transformation of the equations in the system
Equations 4.3-6 through 4.3-8:
log(PbB) = logCbeo + bES PbS + bBD PbD + bBXi XRFIXEquation 4.3-14)
log(PbD) = log(bDO + bDXI XRFI + bDS PbS) (Equation 4.3-15)
log(PbS) = log(bso + bSXE XRFE + bSage House-age) (Equation 4.3-16)
This system can be estimated using SAS PROC MODEL or similar programs for non-linear
systems modelling. All of the coefficients were constrained to be non-negative, since
negative coefficients are non-interpretable. However, the appearance of a negative estimate
for an intrinsically positive coefficient should be taken as a diagnosis of some statistical
problem, such as multi-colinearity or the omission of important predictive variables.
4.3.4 Regression Analyses for Multiple Exposure Pathways: Soil Example
The variables for regression analyses were described briefly in Section 1.5. The use of
a regression coefficient in risk assessment is a complicated matter, because one can use either
aggregate regression coefficients, which combine information on all exposure pathways, or
disaggregate regression coefficients in which each exposure pathway has its own slope
coefficient. The exposure of young children to air lead includes soil and dust pathways, as
well as direct inhalation. This is discussed in detail in the OAQPS staff papers (U.S.
Environmental Protection Agency, 1989ab) based on earlier work by Brunekreef (1984).
The aggregate blood lead regression coefficient for air lead, including soil and dust exposure
pathways, is CBA = 5 to 6 ^.g Pb/dL blood per /ig Pb/m air, whereas the direct inhalation
coefficient b^ is only about 2 ^g Pb/dL blood per jug Pb/m" air. For a simple soil lead
pathway model,
soil -*• dust -* hands -» child
soil -*• hands -» child
whose equations are given by
PbB = bgQ + bgs PbS + b^ PbD (Equation 4.3-17)
PbD = bDO + bDS PbS (Equation 4.3-18)
4-41
-------
the aggregate blood lead vs. soil lead regression coefficient should be
CBS = bss + bBD bos (Equation 4.3-19)
An empirical regression coefficient approach would use only the three coefficients bgs,
and bDS. In the absence of data from a well-conducted child blood lead study at the same
site or at some similar site, including both soil and dust lead data matched to each child's
total lead exposure, there is no basis for calculating the aggregate soil lead coefficient.
However, the use of a model like the IEUBK model allows estimation of the parameters DQS
and bgQ in Equation 4.3-8 from a synthesis of many diverse studies and! does not require
blood lead data at the site. Any additional information about site-specific exposure and soil
or dust lead characteristics would progressively refine the model predictions, even without a
child blood lead study. Site-specific soil and dust lead data are needed in either approach.
The IEUBK model has a parameter in the Multiple Source Analysis for Dust in which the
soil-to-dust coefficient bDS can be inserted.
4.4 USE OF DATA FROM BLOOD LEAD STUDIES
4.4.1 Overview
In general, data from well-conducted blood lead studies of children at a site can provide
useful information to the risk assessor and site decision maker. The purposes of this chapter
are to explain what type of information a well-conducted blood lead study can provide, how
blood lead study data can be used when assessing exposure to lead, and how to interpret
model predictions when blood lead data for a site are also available.
Proper design and conduct of a blood lead study are critical if the results of the study
are to be considered by the risk assessor. Blood lead data alone, without environmental lead
exposure data and without elements of study design that control rapid changes in exposure
prior to sampling, or without adequate control for sampling and analysis, should not be used
to assess risk from lead exposure or to develop soil lead cleanup levels. However, a
well-designed and conducted blood lead study can be useful in conjunction with site-specific
environmental data in evaluating risk to children.
Blood lead concentrations are widely held to be the most convenient, if imperfect, index
of both lead exposure and relative risk for various adverse health effects (U.S.
4-42
-------
Environmental Protection Agency, 1986). In terms of exposure, however, it is generally
accepted that blood lead concentrations yield an index of relatively recent exposure because
of the rather rapid clearance of absorbed lead from the blood. Such a measure, then is of
limited usefulness in cases where exposure is variable or intermittent over time, as is often
the case with pediatric lead exposure.
According to the EPA Science Advisory Board in its 1992 report, "Review of the
Uptake Biokinetic Model for Lead" (U.S. Environmental Protection Agency, 1992a), blood
lead concentrations are responsive to abrupt or unanticipated changes in recent lead exposure
for children. Since internal exposure is a function of lead intake (concentration multiplied by
intake rates) and uptake, these changes can be environmental, behavioral, and physiological.
For example, leaving a child in a house where lead-based paint has just been sanded would
likely result in a significant elevation in that child's blood lead concentration. Reduction in a
child's blood lead concentration may result from altered beha.vior that reduces exposure to
lead (i.e., more frequent house cleaning, more attention to child's cleanliness, etc.). Cross-
sectional blood lead studies (all done within a short time interval) are most useful when there
is no reason to believe that child lead exposure has changed significantly within the last few
months due to changes in environment or behavior.
A blood lead value may say little about any excessive lead intake at an early age, even
though early childhood exposure may have resulted in significant irreversible toxicity.
On the other hand, analyses that are retrospective in nature such as whole tooth or dentine
analyses can only be done after the particularly vulnerable age in children—under 4-5 years-
has passed. Such a measure, then, provides little basis upon which to implement regulatory
policy for environmental or clinical intervention.
Furthermore, over a relatively broad range of lead exposure through some medium, the
relationship of lead in the external medium to lead in blood is curvilinear, such that relative
change in blood lead per unit change in exposure level generally becomes increasingly less as
exposure increases. This behavior may reflect changes in tissue lead kinetics, reduced lead
absorption, or increased excretion. In any event, modest changes in blood lead
concentrations with exposure at the higher end of this range are in no way to be taken as
reflecting correspondingly modest changes in body or tissue uptake of lead, (U.S.
Environmental Protection Agency, 1986).
Data from good quality blood lead studies can be useful in examining the predictiveness
of the model. The IEUBK Model predicts blood lead concentrations in children younger
4-43
-------
than 84 months based on environmental inputs for soil, house dust, water, air, and dietary
lead intake. It would be logical to assume that the distribution of blood lead concentrations
predicted by the model using site-specific data would be generally similar to those measured
in the population, provided that the actual blood lead study was well designed and conducted.
The IEUBK Model may not be able to account for all sources of exposure. If the predicted
blood lead concentrations are not similar to those observed, an attempt should be made to
identify the reasons for those differences.
It is important to recognize that most implementations of the lead model now rely on
the assumption that exposure to lead in soil and dust is primarily residential in nature.
However, in an actual population of children, there will be substantial opportunity for non-
residential lead exposures. Periods of time spent away from the home may also have the
effect of reducing the residential exposures that would otherwise occur. The fact that the
model applications cannot now track all aspects of nonresidential of lead sources that a child
may encounter implies that a precise match between calculated and predicted blood lead
distributions cannot be expected. Nevertheless, due to the importance of residential
exposures to lead in children, a reasonable overall agreement should be anticipated in such
comparisons. These considerations argue that reliance on P-values from statistical tests is not
an appropriate basis for judging the comparability between observed and predicted blood lead
concentrations. It should be noted that calculations of blood lead concentrations on the
assumption of residential exposure is a useful endpoint in site risk evaluation, as many
children will indeed experience primarily residential exposures to lead.
It is important to understand that the model should not necessarily be expected to
reproduce the observed blood lead concentrations exactly. The model predicts the geometric
mean blood lead level corresponding to a given set of exposure inputs. Probability
distribution estimates produced by the model for a given GSD can be used to define a
prediction interval for blood lead concentrations. As long as the interval includes the
observed blood lead corresponding to the same exposure inputs, the model has performed
adequately. Even when a predicted blood lead interval for a set of exposure inputs does not
overlap an observed blood lead level, there may be plausible explanations owing to the
complex nature of multi-media exposures and the difficulty in characterizing all the relevant
determinants of these exposures, and the degree of inter-individual variation in blood lead
concentrations that is known to exist even when exposure is very well characterized.
4-44
-------
4.4.2 Data Quality
The quality of blood lead data can be specified by Data Quality Objectives (DQOs)
which are established prior to the data collection effort. This DQO effort, as outlined in the
Guidance for Data Useability in Risk Assessment, Part A (U.S. Environmental Protection
Agency, 1992b), should result in a sampling and analysis plan which details the chosen
sampling and analysis options, and provides goals for confidence intervals. The data quality
indicators of completeness, comparability, representativeness, precision, and accuracy can
provide quantitative measures of data quality of both sampling and analysis for blood leads
concentrations.
The data quality indicators for sample collection and analysis are presented in detail in
the Centers for Disease Control and Prevention protocols for blood collection and analysis.
Those protocols also cover the elements of QA/QC for specimen collection, specimen
preservation and shipping, analytical method performance, bench and blind quality control
material, and data integrity. The following guidance is given by CDC on selecting a
proficient laboratory and interpreting the results from that laboratory (Centers for Disease
Control and Prevention, 1991):
"Laboratories where blood is tested for lead levels should be successful
participants in a blood lead proficiency testing program, such as the program
conducted jointly by CDC, the Health Resources and Services Administration,
and the University of Wisconsin. In interpreting laboratory results, it should
be recognized that a proficient laboratory should measure blood lead levels to
within several /xg/dL of the true value (for example, within 4 or 6 /xg/dL of a
target value). The blood lead level reported by a laboratory, therefore may be
several /xg/dL higher or lower than the actual blood lead level."
In terms of evaluating the design of a sampling plan for blood lead, perhaps the most
important data quality indicator is that of representativeness. Representativeness is the extent
to which the data defines the true risk to human health for the population living at that site.
For consideration in the risk assessment process, the sampling must adequately represent
each exposure area and exposure scenario. Sampling that is nonrepresentative increases the
potential for false negative or false positive results. A statistically based sampling plan is
needed in order to achieve representativeness. Most studies have tried to include all children
less than 84 months of age, or a random subsample of that age group. A substantial non-
response rate, or attrition rate in a longitudinal study, will undermine the reliability of the
4-45
-------
study findings. Opportunistic or selective samplings may occur with a medical referral
program, a daycare center recruitment, or a community-wide request for volunteer
participation, and are likely to be non-representative for the whole population. Studies in
which respondent families are identified by telephone may miss families, without phones, that
may include transient populations and poorer populations who are possibly at greater risk.
4.4.3 Age of the Population Tested
The IEUBK Model contains uptake parameters and pharmacokinetic algorithms for
children younger than 84 months of age, and predicts blood lead levels only for those ages.
Infants and children younger than 84 months of age, that is, 6 months to 7 years old, have
been identified as the subpopulation most susceptible to the adverse effects of exposure to
low levels of lead (U.S. Environmental Protection Agency, 1986). For this reason, the
blood lead study data that are to be evaluated in conjunction with the results of the IEUBK
Model should consist only of those children younger than 84 months of age. If age groups
older than 84 months are included in the study, it will be necessary to remove the data for
these children from the data set, and to remove their contribution to the statistical results.
4.4.4 Time of the Year When Testing Was Done
Blood lead concentrations show seasonal fluctuations due to factors such as the
relatively short half-life of lead in blood, reduced outdoor exposures in the wintertime, and
perhaps to physiological (hormonal) changes. Cold weather, attending school, and snow
cover tend to reduce the amount of time a child spends outdoors, and the child's direct
contact with contaminated soil. The amount of this fluctuation is variable depending on
physiological and behavioral factors as well as climatic ones. Seasonal fluctuations in blood
lead concentrations as great as 4 to 6 jig/dL have been observed in some studies (Stark et al.,
1982; Rabinowitz et al., 1984; Menton et al., 1994).
Hence, a blood lead study conducted in August would not be comparable to one
conducted in March. In the 1979 to 1982 Boston lead study (Rabinowitz et al., 1984;
Menton et al., 1994), blood lead concentrations associated with fluctuations in air lead and
dust lead (probably from combustion of leaded gasoline) were at their maximum during the
May to August period. Depending on the climatic conditions at a site, the peak summer
months are an optimum time to conduct blood lead testing when soil lead is the primary
source. The children are more likely to have been playing outdoors for 2 to 3 months and
have had the greatest opportunity to be exposed to outdoor sources of lead.
4-46
-------
The amount of time a child has been exposed to a specified environment is also a
concern when evaluating the testing period. Because of the relatively long amount of time
required for a child to come to nearly complete equilibrium with his or her environment, it is
recommended that children who have lived at their current residences for less than
three months or who spend more than 80 % of their time away from their residences be
excluded from the statistical analyses if only environmental lead data from their current
residence are available. Blood lead results for these children may not be representative of
the true health risk at the current residential site.
Because there are few data to quantify the impact of seasonal fluctuations on childhood
blood lead, the model was calibrated using data collected during the peak summer months.
Blood lead studies conducted at other times of the year should be adjusted to compensate for
this seasonal difference.
4.4.5 Concurrent Characterization of Lead Sources
If a blood lead study is to be evaluated in the risk assessment process, it is important
that all of the sources of lead exposure at the site be characterized and quantified. The most
useful data bases contain "paired" data sets (i.e., each child's blood lead would be paired
with the environmental data that represents the child's integrated exposure to lead). This
pairing of environmental data with blood lead data allows the risk assessor to examine the
relationship between a child's blood lead and his or her sources of exposure. At a minimum,
the environmental data would include the lead concentration in soil and in house dust at the
child's residence.
When the blood lead concentrations predicted by the model vary significantly from
those observed in the population, this pairing of environmental and biological data provides
the risk assessor with a tool by which to examine those differences. For example, were all
of the children's predicted blood lead values systematically higher or lower than those
observed? If so, perhaps an important source of lead exposure in the community was
overlooked, perhaps assumptions about intake rates or uptake may be invalid, or perhaps
unidentified behavioral variables affecting the source lead-blood lead relationship are
operating. If a few individual children show particularly striking deviations of observed
blood lead from predicted blood lead, then the contaminant concentrations or
demographic/behavioral data for those children should be re-examined.
4-47
-------
4.4.6 Demographics and Behavioral Factors That Affect Lead Exposure
Prior to sample collection, a well-designed blood lead study will have obtained
information on the demographics and behavioral factors that affect lead exposure in a
community. Such a community survey asks families about occupations, hobbies, social and
economic status, house cleanliness, interior/exterior paint condition, children's mouthing
behavior, etc. All of these questions are designed to identify factors that can modify the
extent to which a child is exposed to the concentrations of lead in his or her environment
(i.e., in media around the child). The Demographics Workplan for the California Gulch
Study Area is an example of one such survey (Woodward-Clyde, 1991).
The results of the community survey can be used to evaluate differences between blood
lead concentrations predicted by the model and those observed. Affirmative answers to
"Have you sanded the paint in your home recently?" or "Does your child eat paint chips
frequently?" may highlight why some predicted and observed levels differ. With the
information from these surveys, a risk assessor can evaluate differences between observed
and predicted blood concentrations due to behavioral or demographic factors.
4.4.7 Effect of Public Awareness or Educational Intervention
Whether or not a community's awareness of the hazards of lead exposure can cause its
members to act to alter blood lead levels is an unresolved question. It is possible that an
enhanced awareness of lead exposure in a community could prompt that community to alter
behaviors to reduce lead exposure, and subsequently, reduce blood lead concentrations in that
community. However, the empirical data on this phenomenon are very limited. Anecdotal
evidence suggests that one-on-one counseling and educational intervention targeted
specifically toward high risk children is effective in reducing individual blood lead
concentrations (personal communications: R. Bornschein, 1992; I. Von Lindern, 1992).
We are not aware of any study that has been designed specifically to test the effectiveness of
educational intervention. A good study design is needed to avoid both statistical and
sampling biases.
Whether or not a general type of awareness in a community may elicit a similar
response has yet to be determined. The differential effectiveness of public awareness
campaigns about soil and dust lead hazards in different subpopulations has also not been
investigated. A study in Raleigh, NC, found that the greatest response to the city's offer to
4-48
-------
test tap water for lead (at no cost to the water customer) was from the higher income
neighborhoods of the city (Simmons, 1989).
Therefore, when a risk assessor is evaluating a blood lead study, he or she should keep
in mind the potential effect of public awareness on blood lead concentrations. If active
educational intervention and counseling programs are being conducted at a site prior to blood
lead collections, or if there is a high level of citizen concern about contaminated sites, the
results of that blood lead study may be different than it would have been otherwise.
4.4.8 Comparison of Observed and Predicted Blood Lead Concentrations
4.4.8.1 Were Important Sources of Lead Exposure Overlooked?
Unless site-specific data are provided by the user to the IEUBK Model for soil, house
dust, air, drinking water, and diet, the model will assume a standard default value for intake
from each medium. For example, at a site where the soil lead concentrations are elevated
and homegrown fruits and vegetables are a large part of the diet, the diet pathway may be
contributing more significantly than the model assumes to total lead exposure. The standard
diet default value in the model is based on recent FDA market-basket survey information and
pertains to lead concentrations in store-bought food. It doesn't consider the contribution of
lead from homegrown fruits and vegetables, which may vary from site to site depending on
the soil lead concentration, soil conditions, type of produce, climate, etc. Communities that
have large ethnic minority populations may also have unique sources of childhood lead
exposure in folk medicines or cosmetics that use lead compounds, or in foods imported in
lead-soldered cans.
Ingestion of paint chips is another source of exposure that may be overlooked.
Exposure to lead occurs from deteriorating house paint via ingestion of paint chips, and via
ingestion of fine particles of paint in household dust. Exposure to fine particles of lead-based
paint hi dust and soil is handled through the soil/dust menu. For ingestion of paint chips,
however, the IEUBK Model assumes a standard default of 0 jug/day for lead from paint chips
and other alternate sources.
In addition to examining the possibility of overlooking an important source of lead
exposure, the risk assessor should examine the representativeness and accuracy of the
environmental data that were collected. For example, is the model input value for lead
concentration in drinking water based on first draw tap samples, groundwater samples, or
estimates from public water company records? A weighted combination of first draw and
flushed tap water samples (plus water from school or day care fountains, if applicable)
4-49
-------
provides the most appropriate representation of the average lead values in a child's drinking
water. The farther away you move from these sources of information, the less accurate and
more uncertain your input to the model will be.
Is the soil lead input based on the average of soil lead concentrations over the entire
yard or is it based on composite samples from a child's yard? If there is substantial
variability in soil concentrations at different locations about the yard, as is often true, an
average of the entire yard may not be an accurate estimate of risk. An integrated assessment
using the perimeter, play areas, and bare areas from each child's residence would provide an
alternative basis for estimation.
Ideally, the inputs to the model should represent the integrated daily exposures each
child might be expected to have. The absence of data specifically collected to estimate the
integrated exposure will limit the accuracy of an analysis. Refining the accuracy and
representativeness of the environmental data values provided to the model may be useful in
resolving differences noted between estimated and observed blood lead concentrations.
4.4.8.2 Are There Interrupted or Enhanced Exposure Pathways at the Site?
A mistake that is often made is equating contaminant concentration with exposure or
risk, where the risk assessor assumes potential exposure is actual exposure. Briefly, if there
is no exposure, there is no risk. If an exposure pathway is diminished or enhanced, then
regardless of contaminant concentration, the resulting exposure or risk is also diminished or
enhanced. For example, at the same concentration of lead in soil, exposure to bare soil may
be greater than if the soil has a good vegetation cover.
4.4.8.3 Are the Assumptions About Site-Specific Intake Rates and Uptake Parameters
Valid?
Internal (systemic) exposure for humans is a function of contaminant concentration,
intake rate and uptake. Environmental sampling can be designed and conducted to obtain a
reasonably accurate representation of the lead exposures a child might experience at a site,
thereby reducing some of the uncertainty in the exposure estimate. However, it is more
difficult to reduce the uncertainty about the site-specific intake rates (i.e., soil ingestion rate,
water ingestion rate) and uptake parameters.
At this time, the empirical evidence on these assumptions is limited and variable.
In other words, there is a degree of imprecision and uncertainty in the intake rates and
uptake parameters. For example, bioavailability of lead from soil is one uptake parameter to
which the model is very sensitive. The model assumes a standard default of 30% for
4-50
-------
absorption of lead from soil in the gastrointestinal tract, yet existing bioavailability studies in
animals show values ranging from 5 to 40 %. Concerns exist about the design of, and animal
models used in, these studies (Section 4.1). Site-specific adjustments in the uptake
parameters require strong justification.
A risk assessor should first explore all of the other rationales for differences between
observed and predicted blood lead concentrations (i.e., sources of lead that were overlooked,
incorrect assumptions about pathways, inaccurate estimates of environmental intake, and
inadequate information about important or relevant demographic/behavioral factors). The
risk assessor should then have strong site-specific justification before exploring non-default
assumptions about uptake parameters.
4.5 ASSESSING THE RELATIONSHIP BETWEEN SOIL/DUST AND
BLOOD LEAD
4.5.1 Assessing Reductions in Blood Lead
The IEUBK model can be used to estimate the change in geometric mean blood lead
from reducing lead exposure, provided the exposure has remained stable for at least three
months and there is a sufficiently detailed characterization of post-reduction lead exposure.
This means that it is necessary to calculate the post-reduction levels for the controlled
medium, the recontamination of the controlled medium by sources of lead exposure that are
left after reduction, changes in the other exposure media from different pathways, and
changes in physical or chemical properties of all media that may affect access, intake, and
bioavailability to children.
There are not many data on post-abatement environmental lead concentrations for
nonurban sites, such as smelters or lead mining sites. As an example, suppose that a
primary lead smelter has been closed down. This immediately reduces or eliminates
air-borne leaded particulates. Over the next few months, fine surface particles in household
dust not otherwise trapped by carpets, upholstered furniture or inaccessible nooks and
crannies, will be gradually swept, washed, or blown out of the house. If replaced by new
surface soil particles, these will be much lower in lead than before the smelter was shut
down, so that the household dust lead concentration may be expected to decrease within
characteristic time scales of a few months to a new quasi-equilibrium value. The surface soil
that had high concentrations of lead before the smelter was shut down may gradually be worn
4-51
-------
away by wind or water erosion, but over a period of many years. This pattern is an informal
description of what has actually been observed at the Bunker Hill site in northern Idaho.
The IEUBK model may be used with long-term post-abatement values to predict blood
lead concentrations in children occupying these residences long after abatement has been
carried out, without worrying about the dynamics of soil and dust lead changes over time.
However, the post-abatement soil and dust at a specific site may not be the same as pre-
abatement soil and dust at the same site. If highly aggregated soil is replaced by loosely
consolidated fine particles in clean fill soil, and is not adequately covered by grass or sod,
then the post-abatement soil may be both more easily transported into the house and more
bioavailable than before abatement. Conversely, if the grass or sod cover is maintained well
after abatement, then the post-abatement soil-to-dust lead coefficient in the IEUBK model
may be different than the pre-abatement value. The validity of the IEUBK model predictions
for post-abatement risks is limited by the validity of the input parameter assumptions for
post-abatement exposures.
At present the definition of elevated blood lead (EBL) is the level of concern of
10 jwg/dL defined by USEPA (1990b) as the lower limit of the range of known possible
adverse neurobehavioral effects in young children. The protection level most often used in
practice is a maximum 5 percent risk of elevated blood lead (EBL) for children in a given
household.
The user has the responsibility for using model input parameters that are appropriate to
the site. Collecting an adequate number of representative soil and dust samples, and
determining their lead concentrations and physical or chemical properties that affect transport
and bioavailability, are generally the minimum site-specific data collection and analyses that
are needed. The ideal input data includes (1) a multimedia household environmental lead
study that includes soil, dust, paint, water and air; (2) information on lead exposures outside
the child's home; and (3) information on family demographics and child behavior patterns in
the community that may affect access to lead sources; (4) characterization of physical and
chemical properties that affect bioaccessibility and bioavailability.
Interest has been growing in the potential uses of the IEUBK model for sites at which
there is presently no residential housing, or at sites at which children may be exposed
without residential dwelling units being physically on the site. Since the IEUBK model
calculates expected geometric mean blood lead concentrations and EBL risks for hypothetical
populations of children, the model can be used for these applications. This can be done only
4-52
-------
if there is sufficient information on child exposure to estimate time-weighted or activity-
weighted soil lead and dust lead concentrations, combining both residential and on-site
exposures.
4.5.2 Situations in Which the Use of the Integrated Exposure Uptake
Biokinetic Model Is Uncertain
4.5.2.1 Assessment of Risk with Community or Neighborhood-Scale Input
There are situations in which it is either inconvenient or impossible to apply the IEUBK
model at the intended household residence scale. For example, only mean values or
geometric mean values of input parameters such as soil and dust lead may be available for a
group of households. Another possibility is that there are a substantial number of soil and
dust lead measurements at a site, but not at houses or locations within the site where blood
lead and EBL risk estimates are needed. We have little information on applications of the
IEUBK model with larger-scale input data, and we must caution the user against using the
IEUBK model for this purpose, because little is known about blood lead variability in such
situations.
4.5.2.2 Use of Surrogate Input Data from Models or Surveys
When modeled or survey data is to be used as input in the Lead Model, the user should
consider the collection time and scale of the data in order to obtain maximum predictability
in the output. Applicability to the individual home, neighborhood area or community should
also be demonstrated. For example, housing age can provide a useful screening variable for
field measurements of lead in tap water and lead-based paint, but it is not likely to be an
adequate substitute for the lead concentration data unless a quantitative predictive relationship
can be established by other studies in the same home, neighborhood or community. Such
screening variables may be useful in screening for areas of concern for lead exposure
sources. At the same time, the output values should not be construed as accurate
representations of the actual child blood lead levels in these areas.
4.5.2.3 Use of the Model To Assess Risk of Elevated Blood Lead at the Regional or
State Level
There is no empirical basis whatever for using the present version of the IEUBK model
at this scale. We have serious concerns that large-scale input data may be totally inadequate
characterizations of the spatially confined exposure for any individual child.
4-53
-------
4.5.2.4 Use of the Model To Assess Trigger Levels for Soil Abatement at the
Community, Regional, or State Level
Use of the present version of the IEUBK model at this scale is discouraged, because
risks cannot be estimated adequately.
4.5.3 Factors That Constrain or Limit the Use of the Model
4.5.3.1 Data and Data Sets Used as Input for the Integrated Ey jsure Uptake
Biokinetic Model
Residential Versus Commercial/Industrial Sites
The IEUBK Model uses site-specific data on the lead concentrations in air, water, soil
and household dust, and average daily intake of lead from diet and from direct ingestion of
paint chips, to estimate the geometric mean blood lead in children exposed to environmental
sources of lead. The data input requirements assume a residential exposure, and thus the
output of the IEUBK Model with default assumptions is probably not predictive for industrial
or commercial sites at which exposures for small children are restricted, except perhaps in
assessment of future use scenarios, or as additive components to a residential exposure
scenario. Development of model estimates in such situations would require adequate
specification of soil and dust ingestion derived from the contaminated site.
Age Group for Which Data Is Available
The IEUBK Model contains data and algorithms to determine intake;, absorption,
excretion and movement of lead between body pools for children from 6 months to 7 years
of age. The IEUBK Model is only predictive for children in this specified age range or any
subinterval within this range. Future versions of the IEUBK Model may be expanded to
include data on metabolic processes in older children and adults, and thus allow
characterization of blood lead levels in these populations.
At present, the IEUBK Model cannot be used to characterize blood lead levels in
children older than seven years or in adult populations.
Other Critical Subpopulations
The IEUBK Model does not predict the blood lead levels of pregnant women, given
either default or site-specific exposures. A parameter input for the maternal blood lead level
has been provided in the IEUBK Model to capture the effect of prenatal exposure in unusual
circumstances of exposure, i.e., in occupational settings. In general, maternal lead exposure
during pregnancy is not well characterized for changes that occur from pre-pregnancy
baseline. The adverse effect of prenatal lead exposure on neurobehavioral and physical
4-54
-------
development is highly significant, and future versions of the IEUBK Model may include a
prenatal exposure component based on the transfer of lead from the mother's blood to the
fetus at the time of birth.
The IEUBK Model contains no specific data to differentiate the adverse effects of lead
on different racial or ethnic groups, nor is there sufficient published data to develop this
component. However, exposure scenarios for a specific subpopulation may be provided by
the user if data are available.
Residency Requirements
The IEUBK Model does not allow entering rapid time-varying changes in exposures to
lead sources. The IEUBK Model has been developed using blood lead data from children
who have had at least a three-month exposure to their residential sources prior to blood
sampling for lead analysis, that is, a minimum three-month residential requirement for
inclusion in blood lead studies. The three month residency requirement guarantees that
predicted blood lead attributable to the current residential exposure will be nearly at a steady
state level. If residency requirements have not been met or if lead exposures are changing
rapidly, the IEUBK Model can be expected to give less than accurate predictions, because
exposures at prior residences may still be a major determinant of blood lead.
Timing of Data Collections
Because of the variability in child blood lead levels with seasonal exposure and the
corresponding variability in environmental lead levels (i.e., changes in household dust lead
levels with seasonal and activity changes) strict attention should be paid to the timing of data
collections if the data is to be used as input in the IEUBK Model to make predictions about
individual or community blood lead levels in children. This is especially important if the
predicted blood lead levels are to be compared with the results from a community blood lead
study, to assure that the two studies measure the same population at the same period in time,
same season of the year. The parameters for the IEUBK model were developed from diverse
animal and human studies. Collectively, these studies reported ranges of values for these
parameters. The first stage in model validation was a calibration stage, using paired data-
measurements of lead in environmental media and in blood collected from children under the
age of six, taken within a short period. Comparison of observed and predicted blood levels
suggested modifications of the parameters, within the range of plausible values suggested by
the literature or by our analyses of research data. After there adjustments, the model
obviously could not appropriately be tested again using the same set of data. Therefore,
4-55
-------
validation tests were performed using the sets of community blood lead data paired with
environmental exposure data for the same child.
4.5.3.2 Biological and Exposure Parameters Used in the Integrated Exposure Uptake
Biokinetic Model Bioavailability of Soil Lead
The bioavailability of lead from different sources may be variable due to differences in
lead concentration, lead speciation, particle size and mineral matrix (Barltrop and Meek,
1975; Barltrop and Meek, 1979; Heard and Chamberlain, 1982; Rabinowitz et al., 1980;
Cotter-Howells and Thornton, 1991; Aungst and Fung, 1981). Additionally, bioavailability
may vary as a function of physiological parameters such as age, nutritional status, gastric
pH, and transit time. The IEUBK Model uses a default of 30 percent lead absorption from
soil, which is constant across all concentrations and soil sources. Site-specific data on the
soil and dust bioavailability may improve the accuracy of the blood lead level predictions.
Other Lead Exposure Inputs
Child default values for dietary lead intake are provided by year and by age of the child
in the IEUBK Model. The use of default values is appropriate unless the dietary lead intake
is very high, due perhaps to a high intake of home-grown fruits and vegetables or the intake
of lead-contaminated ethnic food or drugs.
Exterior lead-based paint can make a significant contribution to soil lead, and is usually
considered as part of this exposure source. The contribution of lead-based paint to indoor
household dust is harder to estimate because the condition of the paint varies from house-to-
house and the rate of incorporation into house dust is variable. If the household lead-based
paint contribution is highly variable in a community, care should be taken to avoid
combining all homes in a single run of the IEUBK Model, as the output results may not be
applicable to the population.
Children can eat chips or strips of deteriorating lead-based paint directly from painted
surfaces, even when the total area of lead-painted surfaces is so small that the total
contribution of lead-based paint to interior household dust or exterior soil is too small to
identify. Paint chip intake reflects child-specific behavior, including observed ingestion of
paint chips, observed contact of the child's mouth with painted surfaces and the frequency of
mouthing of non-food objects.
Blood Lead Variability
The variability of individual blood lead levels with respect to the geometric mean blood
lead level predicted by the IEUBK Model is characterized by a single number: the geometric
4-56
-------
standard deviation. The GSD is used as a single number to characterize the relative
variability of the log-normal distribution representing the aggregate uncertainty hi all sources
of population variability: biological, uptake, exposure, sampling and analytical.
A common misconception is that the IEUBK Model predicts the community geometric
mean blood lead and the fraction of children at risk when the input is the arithmetic mean or
geometric mean across households of household-specific lead concentrations. This use of the
IEUBK Model may cause seriously misleading interpretations of the output of the model,
when the true extent of variability is not known. A correct approach to neighborhood risk
estimation is given in Section 4.2.5.
Prior Body Burden
Child blood lead level predictions obtained using the IEUBK Model reflect the
contributions from lead sources entered into the model; they do not take into account any
existing body burden which may be the result of prior exposures not known to the user.
Current blood lead levels depend on prior exposure history as well as present exposure.
If past exposure levels have been greatly elevated, the results obtained from the IEUBK
Model may not be accurate. Where children have had high prior exposures, that prior
exposure affects blood lead levels for at least three months after the exposure ends, a
"washout" period. Future estimates are based on present conditions. If those conditions
change (e.g., deteriorating paint that might change house dust lead concentrations), the
exposure and consequent risk will be different.
Alternate Exposure Locations
Child blood lead levels obtained using the IEUBK Model refect input lead sources at
the household level or neighborhood level. They do not necessarily take into account
increased or reduced lead exposures which may have taken place at parks, preschool, homes
of babysitters, neighbors or relatives, or other locations frequented by the child, unless these
exposures are measured and explicitly entered into the model as inputs. Thus, the results
obtained from the IEUBK Model may not be accurate unless the child's activity patterns have
been well documented.
Socioeconomic Status
The blood lead levels of two children with identical lead exposure scenarios, but living
in different family behavior patterns might vary greatly. The difference in socioeconomic
status might be reflected in differences in household repair and cleaning, washing of
children's hands and toys, food preparation methods, concern for balanced meals and
4-57
-------
improved nutritional status, more regular eating patterns, etc., all of which may impact blood
lead levels. Use of the IEUBK Model should be proceeded by adequate characterization of
information on behavioral and other socioeconomic differences, and advice from regional
offices on appropriate adjustments, if warranted.
Intervention/Public Education Programs
Intervention and public education programs can inform the community of the adverse
health effects of lead exposures and how to reduce them. These activities may result in
reductions in blood lead levels in portions of a community that may be temporary, depending
on how well the information is conveyed and received. These temporary changes in blood
lead concentrations might occur during a one-tune blood lead survey and cannot be predicted
using the IEUBK Model. Some of the examples in Chapter 5 describe the correct application
of the model in this situation.
4.6 WHAT YOU NEED TO KNOW ABOUT BIOKINETICS
4.6.1 Description of the Biokinetic Model
The IEUBK model has a very detailed biokinetic modelling component. This
component of the model is not accessible to the user because, in our judgement, most users
will neither wish to change the biokinetic parameters nor have the need to change any of the
biokinetic parameters. The biokinetic parameters are used to define intrinsic biological
variables that do not change from one exposure scenario to another, once a child's age is
specified. The basis for the biokinetic parameters are described in the Technical Support
Document: Parameters and Equations Used in the IEUBK Model for Lead in Children (see
Section 1.2.2).
The biokinetic model is a compartmental model, in that it assumes that all of the lead in
the child's body can be attributed to one of seven kinetically homogeneous compartments and
that transfer between these compartments occurs through normal physiological processes.
The compartments in this model are:
Plasma and extra-vascular or extra-cellular fluids (denoted ECF);
Red blood cells
Kidney
Liver
Other soft tissues
4-58
-------
Trabecular (spongy) bone
Cortical (compact) bone
The distribution of lead in the body is only approximated by a compartmental structure,
even in so-called physiologically-based pharmacokinetic models, because no tissue
compartment is, in reality, completely homogeneous. However, the compartmental method
is so useful and accurate that it has been almost universally adopted.
Realistic growth equations are used for each organ or tissue pool (biokinetic
compartment) from newborn status to age 7 years. The transfer times (equivalently,
fractional transfer rates) among compartments are scaled according to organ volume or
weight, or body volume or weight, using allometric scaling consistent with organ or body
surface-area scaling. The basis for the compartmental transfer times are the reanalyses we
have done for data from studies in infant and juvenile baboons, using data in (Mallon 1983)
and (Harley and Kneip 1985). A wide variety of studies in human children and adults, in
other species, and in other metals was used to estimate biokinetic parameters not estimatable
from the baboon studies. Growth equations were derived from Altman and Dittmer (1962),
Spector (1956), and Harley and Kneip (1984). The literature review revealed 17 adult and
3 pediatric studies for evaluating the transfer time from blood to urine. An allometric
scaling factor, based on the correlation between body surface area and glomerular filtration
rate (West, 1948), was applied to the transfer time composited from the 17 adult studies to
provide an estimate of the blood to urine transfer in children. An estimate of transfer from
blood to feces and blood to urine for adults was taken from Chamberlain et al. (1978) and
Rabinowitz et al. (1976), and for transfer from blood to soft tissues from Rabinowitz (1976),
and equations for compartment to blood lead concentration ratios from Barry (1981).
The flow of lead from external media into the body and the distribution and elimination
of lead is shown graphically in Figure 4-10. Transfer of lead to and from plasma and
extravascular fluids is governed by first-order kinetics, in that the rate of change of the lead
content in each compartment is a function of the current state of the system as defined by the
lead content of each of the compartments. If the dependence of the rate of change of lead
content is a linear function of the contents of all of the compartments, then the biokinetic
model is described as a first-order linear kinetic model. The IEUBK has almost linear
kinetics, except that we assume that the lead-binding capacity of the red blood cells can be
saturated when lead uptake into the body is very high. Uptake of lead can occur through the
lungs into the plasma-ECF pool, or through the gut into the plasma-ECF pool. While the
4-59
-------
Extra
Plasma -- Cellular
Fluid
Trabecular
Bone
Cortical
Bone
Other
Soft Tissue
i
Liver
r
Sweat
Skin
Hair,
Figure 4-10. Biokinetic compartments, compartmental lead flows, and uptake pathways
in the integrated exposure uptake biokinetic model.
plasma-ECF pool may be viewed as the central pool or compartment in this system, the usual
observable is the blood lead concentration, which combines both the lead in plasma-ECF
pool and the lead in the red blood cells.
4.6.2 Consequences of Biokinetic Parameters for Site-Specific Risk
Assessment
The exposure scenarios that can be used in the IEUBK model change only once a year.
Since most of the transfer times for children are on the order of 1 hour to 1 month, the
IEUBK calculations of lead levels in blood and soft tissues may be assumed to be in
a quasi-steady state condition with respect to exposure. The quasi-steady-state condition may
allow the use of simple linear approximations to blood lead vs. media concentration or media
intake of lead at different ages. But the time scale for release of lead from bone is much
4-60
-------
longer, on the order of 1 to 3 years, so that the bone lead level quickly builds up and the
skeleton contains 60 to 70 percent of the total body burden of lead by age 2 years. There is
not be a true steady-state blood lead level during the 7-year interval used in the IEUBK
model for young children. As bone lead burdens increase, so will there be a growing
component of lead in blood that comes from release of lead from the skeleton or from
resorption of bone in growing children.
Because of this release of bone lead, there may be a large component of blood lead
levels in children that will respond only slowly to any changes in environmental lead
exposure. This is particularly noticeable in evaluating soil lead abatement studies and
strategies, where a child may have accumulated a large body burden of lead before the
abatement. In the first year or two after the abatement, the internal or endogenous source
oflead stored in the skeleton may cause a moderately elevated blood lead level to persist in
the child. Children who were never exposed to the elevated environmental lead, or who did
not accumulate a large body burden of lead before the abatement even though the
environmental exposure was high, will not have this residual elevation of blood lead from
resorbed bone lead.
4.7 ISSUES IN USE OF THE MODEL FOR PAINT CHIPS
4.7.1 Inappropriateness of Use of IEUBK Model for Paint Chip Ingestion
The IEUBK model, Version 1.0, does not contain an explicit component for lead-based
paint ingestion outside of the Alternate Source Option in the Soil/Dust Menu. The correct
use of the IEUBK model is to estimate geometric mean blood lead levels and distributions of
blood lead levels in young children who have long-term chronic exposures to lead. It has
long been known that the ingestion of even tiny quantities of paint chips on a single occasion
can cause serious lead intoxication. Chisolm and Harrison (1956) show photographs of small
paint chips weighing several grams that can easily be removed and eaten by a child. Since
old lead-based paints can contain hi excess of 50 percent lead, the child may ingest several
million micrograms of lead in a single episode. The IEUBK model is not intended to address
this situation. The IEUBK model is intended to address the situation where the child ingests
typical quantities of household dust that have been contaminated by leaded soils and by
deterioration of old lead-based paint from interior surfaces. The inclusion of lead-based paint
in the dust menu implicitly assumes that paint has fallen off the painted surface as fine
particles, or has fallen off as discrete flakes or chips of paint and has been reduced to small
4-61
-------
particles in situ on the floor, carpet, furniture or other surfaces. Interior lead-based paint
may not wear as rapidly as exterior paint due to the near-absence of sunlight on most
household surfaces, but common observation finds many deteriorated lead-painted interior
surfaces in older housing, especially in wet rooms such as kitchens, bathrooms, or laundry
rooms (HUD, 1991).
The following data are presented to assist the user who wishes to develop an exposure
scenario in which there is long-term ingestion of chips of lead-based paint, in addition to the
interior household dust lead contribution that is already included in the IEUBK model.
An exposure scenario with paint chip ingestion can be entered in the Other Source Menu of
the model. The data for construction of an alternative lead-based paint chip menu were
reviewed by the EPA Technical Review Work Group, who concluded that these data were
not adequate to be recommended as default values. There are greater uncertainties about
paint chip exposure and uptake than about other exposure media. These uncertainties
include:
(1) The quantity of paint chips ingested on a long-term or chronic basis is
unknown; however, even small quantities of ingested paint chips can
produce a lead intake of millions of micrograms per day, overwhelming
all other sources.
(2) Lead levels in housing are most typically measured as surface loadings
using portable XRF analyzers. While there are several proposed
relationships between lead paint surface loading and daily lead intake,
these require making assumptions about other uncertain relationships, such
as the "area" of surface ingested the child, or the thickness of the paint
chip and the relationship between lead concentration and lead loading.
We will describe these relationships, but we believe that the;/ do not yet
have an adequate empirical basis.
(3) Paint chips are, by definition, discrete units. Even if paints chips are at
least one millimeter in diameter, or even larger, they may not be
completely dissolved in the stomach or completely absorbed in the
intestines. Observations of child fecal samples sometimes find discrete
paint chips. Radio-opaque samples in stool may be lead or some mixture
of lead with other heavy metals such as barium or chromium commonly
found in leaded paint pigments.
4-62
-------
(4) Lead paint absorption by rats has been found to depend significantly on
particle size and chemical speciation of paint particles. Many chemical
species are found in lead-based paint, most often including lead octoate
(as a dryer), lead carbonate, and lead chromate. The sequence of
absorption or bioavailability is probably
carbonate _>_ octoate > chromate
based on rodent studies (Barltrop and Meek, 1979). While the ranking is
probably similar in human children and other primates , direct evidence is
limited to baboons (Cohen, 1975; Mallon, 1983). Studies are currently in
progress using miniature swine as closely analogous models of human
gastro-intestinal absorption of nutrients and contaminants, but results on
absorption of lead from actual lead paints have not yet been reported.
It is clear in any case that estimates of lead bioavailability in paints may
require a much more complete site-specific characterization by particle
size and chemical speciation than does soil.
4.7.2 Daily Intake of Paint Chips
The American Academy of Pediatrics (1972) has used a provisional estimate of one
2
square inch (6.25 cm ) of paint surface ingested per day. This appears to be a nominal value
for purposes of risk estimation, and no empirical basis for this value has been provided.
2
They cite evidence that 1 cm of one layer of ulterior paint may weigh 5.0 to 8.2 mg
(average 6.5 mg), and that six layers of paint weighed 37.0 to 40.6 mg (average 38.8).
Thus, using data that may represent Providence PJ in 1972, where six layers of paint were
9
typical, ingestion of 6.25 cm of painted surface through a single painted layer would
correspond to 40.6 mg/day intake, and a thick chip containing six layers would average
233 mg/d paint chip intake. Even if the ingested paint chips were square-inch monolayers
with one percent lead, the daily lead intake would be 400 ug Pb/d. We cannot provide any
realistic estimate of the uncertainty of this estimate. It is likely that there is some correlation
between the size, thickness, and lead content of ingested paint chips, since additional lead is
reported to add a sweet taste to the chips that may appeal to a child with pica for lead paint
chips.
These estimates were also cited in a report by the National Academy of Sciences
(NAS, 1973) to the Consumer Product Safety Commission (CPSC). They concluded that the
4-63
-------
quantitative evidence was inadequate "to promulgate a standard based on knowledge of the
essential quantitative relations that link the lead content of paint to symptoms of intoxication.
However, this is not unusual in public-health practice. Many useful standards have been
established by informed people who make judgments based on whatever facts are available"
(NAS, 1973, pp. 25-26).
In view of the lower quality of information on paint chip intake than on intake of soil
and dust, diet, and drinking water, and the usefulness of providing baseline risk assessments
in the absence of lead-based paint, we have used a default value of 0 jwg/dL in the model.
4.7.3 Relationship of XRF Lead Paint Surface Loading to Lead Paint
Concentration
The estimate of lead intake from paint chip ingestion depends on a lead concentration
for the ingested chips. However, this is not available in field samples without removing a
piece of paint from the wall or trim. Therefore, the use of non-destructive field sampling
methods such as portable XRF analyzers has become the common method for determining
paint hazard. We can calculate
2
lead concentration (/ig/g) = 0.001 (/xg/mg) * lead loading (mg/cm ) /
thickness of paint (cm) * paint density (g/cm ).
Calculations from the EPA Lead Reference Materials Workshop (EPA 1991) assuming a
seven-layer thickness of paint (40 mil = 1 mm) and a density of 2 g/cm calculates
5,000 /ig/g equivalent to 1 mg/cm2. This is reasonably concordant with some analyses of
measurements of paint loading and concentration that we had calculated from data in the
Boston Brigham and Women's Hospital Longitudinal Lead Study. However, this relationship
is likely to vary so greatly from house to house that we cannot recommend its use without
site-specific verification.
4.7.4 Dissolution of Paint Chips in Acid Environments
Not all of the lead in a large lead paint chip may be available for absorption. Roberts
et al. (1974) report that "20 to 60 percent of the lead in surface soil was extractable in
0.1N HC1 compared with less than 10 percent extractable from paint samples." Particle
dissolution is a component of lead bioavailability.
4-64
-------
4.7.5 Absorption of Lead Paint In Vivo
The absorption of lead-based paint particles by rats is described in (Barltrop and Meek
1979). They conclude that "The physical form of particles derived from paint film would
seem to modify the availability of Pb compounds contained in them for absorption. Little is
known of the physical or chemical changes which paint flakes undergo after ingestion,
although it is known that some paint flakes remain relatively intact when swallowed by a
child and may be observed radiographically in the gut lumen, or on inspection of feces.
In spite of this, sufficient absorption of Pb resulting in childhood poisoning is known to
occur, and in many cases the ingested flakes become too finely divided to be visible
macroscopically. Thus the composition of the paint and the chemical nature of the added Pb
compounds may determine its stability in the gut and hence the availability of Pb for
absorption. Long-term feeding of paint flakes identical to those used in this work, but of
larger size (500 to 1,000 microns) have been reported to result in minimal absorption by the
rat (Barltrop and Meek, 1975)."
Table 4-4 summarizes their results. Lead absorption can be characterized by the
difference in blood lead levels between exposed and control rats. The increase in blood lead
for rats fed lead octoate in particles between 500 and 1,000 microns diameter is about
60 percent of the absorption of lead octoate particles < 50 microns, and absorption of lead
chromate paint in particles of 500 to 1,000 microns is about 45 percent of the absorption of
lead chromate paint in particles of 500 to 1,000 microns. For particles <50 microns, the
increase in blood lead for lead octoate particles is about 60 percent of the increase from lead
acetate. It is not clear how these results can be used quantitatively for humans to determine
absolute or relative bioavailability of LBP.
Juvenile and infant baboons were exposed to oral intakes of lead salts and prepared lead
paint samples from New York City (Mallon, 1983). The lead salts and paint samples were
fed in gelatin capsules to sedated baboons. The relative bioavailability could be estimated
from differences in the steady-state blood lead levels achieved after 5 or 6 months of chronic
exposure. These are shown in Tables 4-5 and 4-6. The increase in blood lead in infant
baboons (age 6 months at the start of the study) was 23 /ig/dL (no s.e.) for 2 baboons
exposed to lead acetate and 6.125 /xg/dL for 8 baboons exposed to New York city paint at a
controlled dose of 100 /xg/kg/day (roughly 250 to 350 /xg/day in baboons who grew from
2.5 to 3.5 kg body weight). At higher doses, the increases in blood lead were clearly
nonlinear with respect to dose rate. In juvenile baboons (ages 20-24 months at the beginning
of the study) the increase in blood lead was 11.7 /xg/dL (no s.e.) for 2 baboons exposed to
4-65
-------
TABLE 4-4. PERCENTAGE INCREASE IN BLOOD LEAD LEVELS IN INFANT
MALE WISTAR RATS WITH 48-HOUR ORAL EXPOSURE TO LEAD ACETATE,
AND TO LEAD OCTOATE AND LEAD CHROMATE PAINTS OF DD7FERENT
PARTICLE SIZES
Paint Chip
Size (mm)
-
-
0.5-1
<0.05
0.5-1
<0.05
Lead
CONTROL
ACETATE
OCTOATE PAINT
OCTOATE PAINT
CHROMATE PAINT
CHROMATE PAINT
Dose Rate
/tg/kg/d
0
330001
330001
330001
330001
330001
Blood Lead
(S.E.) /itg/dL
8.1 (1.9)
38.3 (4.0)
19.3 (3.7)
27.2 (4.0)
14.5 (3.2)
22.8 (2.2)
Blood Lead -
Control /*g/dL
30.2 (4.4)
11.2(4.2)
19.1 (4.4)
6.4 (3.7)
14.7 (2.9)
Percent of
PbAc
-
-
37.1
63.2
21.1
48.7
Calculated as 0.02% lead in diet, per 31 to 33 g diet in 48 h, per 96 g body weight (range 90 to 103 g).
Source: Adapted from Barltrop and Meek (1979).
TABLE 4-5. PERCENTAGE INCREASE IN BLOOD LEAD LEVELS IN INFANT
BABOONS WITH CHRONIC EXPOSURE TO LEAD PAINT, LEAD ACETATE, AND
OTHER LEAD COMPOUNDS
Age Lead
5-6 mo CONTROL
ACETATE
ACETATE
ACETATE
CARBONATE
OCTOATE
PAINT
Dose Rate
jtg/kg/d
0
100
200
1000
1000
100
100
Blood Lead (N)
pg/dL
9 (1)
32 (2)
42 (2)
72 (1)
69 (1)
90 (1)
15.12(8)
Blood Lead -
Ctrl. /tg/dL
-
23
33
53
60
81
6.12
Percent of
PbAc
-
-
-
-
95.2
352
26.6
Source: Adapted from Mallon (1983).
lead acetate and 33.7 jig/dL for 1 baboon exposed to lead octoate at 100 /wg/kg/d, but only
3.7 jug/dL (no s.e.) in 2 baboons exposed to New York city paint at a controlled dose of
200 jug/kg/day. The increase in blood lead was 31.7 /*g/dL (no s.e.) for 2 baboons exposed
to lead acetate and 93.7 /xg/dL for 1 baboon exposed to lead octoate at 500 /xg/kg/d, but only
12.7 jiig/dL in 2 baboons exposed to New York city paint at a controlled dose of
500 /ig/kg/day. Therefore, the bioavailability of lead in actual paint samples was at most
4-66
-------
TABLE 4-6. PERCENTAGE INCREASE IN BLOOD LEAD LEVELS IN JUVENILE
BABOONS WITH CHRONIC EXPOSURE TO LEAD PAINT, LEAD ACETATE, AND
OTHER LEAD COMPOUNDS
Age Lead
20-24 mo CONTROL
ACETATE
ACETATE
OCTOATE
OCTOATE
PAINT
PAINT
Dose Rate
/tg/kg/d
0
100
500
100
500
200
500
Blood Lead (N)
Jtg/dL
12.33 (3)
24 (2)
44 (2)
46 (1)
106 (1)
16 (2)
25 (1)
Blood Lead -
Ctrl. jtg/dL
-
11.67
31.67
33.67
93.67
3.67
12.67
Percent of
PbAc
-
-
-
288.6
295.8
31.41
40.0
Calculated relative to 100 jig/kg/d lead acetate.
Source: Adapted from Mallon (1983).
25 to 40 percent of the bioavailability of lead acetate administered during chronic exposure
studies at dose rates roughly comparable to those assumed in the American Academy of
Pediatrics report. The much higher relative bioavailability of the pure lead octoate
compound remains to be explained. The absolute bioavailability of lead acetate in diet
estimated by Mallon was estimated by Mallon was 24 percent at a dose rate of 12 /*g/kg/d,
8 percent at 100 /ig/kg/d, and 6 percent at 200 pig/kg/d in infant baboons; 12 percent at
12 /ig/kg/d, 3 percent at 100 /ig/kg/d, and 1 percent at 1,000 /xg/kg/d in juvenile baboons.
The estimates of absolute bioavailability of oral lead acetate developed by Marcus (1992)
using a saturable absorption mechanism to account for the bioavailability were higher, about
28 percent and 20 percent at dose rates that were much less than 200 /xg/kg/day. The
bioavailability of these lead-based paints must then be taken as less than 7 percent and
5 percent respectively. A detailed characterization of the chemical composition and size
distribution of the prepared paint samples would have been useful, but was not presented.
4-67
-------
5. APPLICATIONS WITH EXAMPLES
5.1 APPLICATIONS FOR POPULATION ESTIMATES
The purpose of this chapter is to provide concrete examples complete with explanations
that can guide the user through specific applications of the model. These examples are taken
in part from past applications of the model, but they have been modified for the purposes of
illustration and do not represent any specific site or risk management decision. While the
user should find some guidance in these examples, they are not meant to be comprehensive
of all possible model applications, nor should they be generalized to any particular site.
EXAMPLE 5-1. Default Values
As stated earlier in this manual, the model can predict geometric mean blood lead
levels in a population of children with residential and neighborhood exposures, provided that
the distribution of environmental lead parameters is not widely dispersed. The following is
an example of a simple simulation using only default values.
From the main menu shown in Screen 2-1, enter "2" (Computation), then on the
Computation Menu enter "1" (Run). The results shown on the monitor display the average
of monthly geometric mean blood lead concentrations in one-year intervals, along with the
daily lead uptakes from each medium in /xg Pb/day. These results are the geometric mean
blood lead concentrations and lead uptakes within each one-year age interval, assuming
constant environmental lead concentrations from birth through each age interval. They can
be interpreted as representing the results for a "typical" child in contact with these or similar
lead concentrations. See Example 5-4 for an extension of this example to risk estimation.
5.2 APPLICATIONS WHERE ENVIRONMENTAL LEAD
CONCENTRATIONS CHANGE OVER TIME
EXAMPLE 5-2. Reductions in Air and Dietary Lead Levels from 1975 to 1981 Decrease
Baseline Blood Lead Concentrations
This example illustrates the estimation of historical exposures and baseline U.S. blood
lead concentrations from 1975 to 1981.
5-1
-------
•Air: The user should first enter the 1975 air lead levels from Figure 2-10.
•Diet: Then the user should enter the dietary lead values for the same time
period, as in Table 2-1. However, no dietary lead intake values for
children are shown for 1975 to 1977. We estimated the 0-11 month
value for 1975 as 80 percent of the 1-year value for the 1978 value,
that is 80 percent of 45.80 /xg/d = 36.64 ng/d, since the 6-11 month
dietary lead intake values for 1982-1984 are about 80 percent of the
respective 1-year-old values. We then assumed that for a child born
in 1975, the 1975 value was 36.64 jug/d, the 1976 value (age 1 year)
was the same as the 1978 1-year-old value of 45.80 jiig/d, the
1977 value (age 2 years) was the same as the 1978 2-year-old value
of 52.90 /ig/d, the 1978 value (age 3 years) was 52.70 j*g/d as in
Table 2-1, the 1979 value (age 4 years) was 47.30 jug/d as in
Table 2-1, the 1980 value (age 5 years) was 38.70 pg/d as in
Table 2-1. We assumed that the 1981 value (age 6 years) was
110 percent of the 1981 value at age 5 years or 110 percent of
35.80 ng/d = 39.38 /xg/d, since the 1982-1984 6-year-old values are
about 10 percent larger than the respective 5-year-old intake values.
The input values for dietary lead intake are shown in Table 5-1.
•Water: Water lead concentrations were kept at the default values.
•Soil: Adjustments should be made for lead in soil and household dust.
We assumed that soil lead levels, even in areas not heavily impacted
by automobile traffic, would have been somewhat larger in 1975 than
in 1981. In the absence of better information, we assumed that soil
lead concentrations consist of two components, a genuine baseline of
about 200 fj,g/g which is the current default, and a small increment
from air lead deposition that adds about 100 /xg/g soil lead per
/xg/m3 air lead. This assumption implies a relatively small
contribution of 10 /xg/g to soil lead from current air lead levels of
0.1 /xg/m3. Thus the 1975 soil lead level is about 324 jug/g, the
1976 level about 322 /xg/g, and so on, as shown in Table 5-2.
5-2
-------
TABLE 5-1. USER-SELECTED ENTRIES FOR IEUBK MODEL WORKSHEET
FOR EXAMPLE 5-2, CHILD BORN IN 1975
PARAMETER
DEFAULT
VALUE
USER
SELECTED
OPTION
UNITS 1
DATA ENTRY FOR DIET (by year)
Dietary lead intake
Age =0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
5.53
5.78
6.49
6.24
6.01
6.34
7.00
36.64
45.80
52.90
52.70
47.30
38.70
39.38
jwg Pb /day
TABLE 5-2. USER-SELECTED ENTRIES FOR ffiUBK MODEL WORKSHEET
FOR EXAMPLE 5-2, CHILD BORN IN 1975
PARAMETER
DEFAULT
VALUE
USER
SELECTED
OPTION
UNITS
DATA ENTRY FOR SOIL (by year)
Soil lead concentration
Age =0-1 year (0-11 mo) (1975)
1-2 years (12-23 mo) (1976)
2-3 years (24-35 mo) (1977)
3-4 years (36-47 mo) (1978)
4-5 years (48-59 mo) (1979)
5-6 years (60-71 mo) (1980)
6-7 years (72-84 mo) (1981)
0
0
0
0
0
0
0
324
322
320
310
290
256
247
Mg/g
5-3
-------
•Dust: The Multiple Source Analysis method for household dust should be
used, since soil lead and air lead levels are changing over time.
Since particles from leaded gasoline emission are believed to
contribute significantly to surface soil transported into the house
during these years, we have assumed that the soil-to-dust coefficient
is 0.85 appropriate for this historical example, although the current
default is 0.70, and the air-to-dust coefficient is 100. This was
shown in Screen 2-10. These changes are reported to the user in the
main Data Entry Screen for Soil/Dust.
The model can be run by returning to the Computation Menu and using Option 1, or by
pressing the F5 key from any of the main media data entry screens. The results are shown
on the display. The results are reasonably consistent with the decrease in child blood lead
concentrations in the U.S. from about 15 yug/dL to 10 /ug/dL found in the 1976-1980
NHANES n survey (U.S. Environmental Protection Agency, 1986). However, this exposure
scenario follows a single child born in 1975 for six years through 1981. Direct comparison
with NHANES n would require representative blood lead estimates for 1-year-olds in 1976,
2-year-olds in 1977, 3 year-olds in 1978 etc.
The importance of a worksheet in developing and documenting the exposure scenario
should be clear to the reader. The worksheets for this example are shown in Tables 5-1 and
5-2. Since the exposure scenario here is for a typical urban child and is not specific to a site
or neighborhood, the user should not try to extend these results for risk estimation purposes
without incorporating interindividual variability and site-specific information concerning
exposure variability.
The IEUBK model is a biokinetic model, and therefore has the ability to estimate
changes in blood lead in response to yearly changes in enviromnental lead exposure for
children of different ages. The following examples are presented to encourage the user to
explore some of the IEUBK model's capabilities for evaluating age-dependent changes in lead
exposure when this exposure changes over time.
EXAMPLE 5-3. Example for Children Moving From a Lower to a Higher Soil Lead
Concentration
This example demonstrates the effects of change from a constant environmental lead
concentration to a higher constant environmental lead concentration. Assume that a child
5-4
-------
moved into a housing unit with a soil lead concentration of about 2000 /*g/g, from a previous
housing unit with a soil lead concentration of about 100 /xg/g. Assume also that soil is a
significant source of dust in household dust, and that the soil lead contribution to household
dust lead is 70 percent of the soil lead concentration. The user can assess the maximum
effect of new exposure to elevated soil lead (e.g., moving into a new residence). This
assessment is for children of different ages, in an ordered sequence of runs. This sequence
studies the effects of new exposure at later ages.
The work sheet for this example is similar to the segment shown in Table 5-2. In fact,
a sequence of work sheets is needed to study the effects of moving at different ages. There
are two variables to be considered here. The first variable is the age of the child, which is
used in the IEUBK Model calculations, and is entered as the left-hand column of the work
sheets. The second variable is the age at which the child moves into the new exposure
environment. Thus, in Table 5-3(a), if the child moves at age 0 years, the child is exposed
to 2000 /xg/g lead in soil and 1400 /tg/g lead in dust derived from soil from birth through age
6 years. However, if the child moves at age one year, the correct work sheet is shown in
Table 5-3(b). In the work sheet in Table 5-3(b), the child is exposed to 100 /xg/g lead in soil
and 70 /xg/g lead in household dust at age zero years, but to 2000 /xg/g lead in soil and
1400 /xg/g lead in dust from soil at ages 1 through 6 years. Similarly, if the child moves at
age two years, the correct work sheet is shown in Table 5-3(c). In the work sheet in
Table 5-3(c), the child is exposed to 100 /xg/g lead in soil and 70 /xg/g lead in household dust
at ages 0 and 1 years, but to 2000 /xg/g lead in soil and 1400 /xg/g lead in dust from soil at
ages 2 through 6 years.
The worksheets for Tables 5-3 (a-c) are combined and shown as columns 2 to 4 in
Table 5-3(d). The last 4 columns in Table 5-3(d) summarize the soil lead work sheet entries
if the hypothetical child moves at ages 3, 4, 5, or 6 years respectively. For example, in the
extreme right-hand column, if the child moves at age 6 years, he or she is exposed to
100 jixg/g lead in soil from birth through age 5 years, then to 2000 /xg/g at age 6 years.
The IEUBK Model simulation for this example is run 7 times, each run corresponding
to a column in Table 5-3(d) or to a work sheet 5-3(a-c) or analogous work sheets for older
children. The results, as annual averages of predicted geometric mean blood lead
concentration, are shown in Table 5-4 in exactly the same order as in Table 5-3(d).
5-5
-------
TABLE 5-3a. SOIL LEAD DATA ENTRY WORKSHEET
FOR CHILD EXPOSED TO 2000 jig/g SINCE AGE 0 (BIRTH)
PARAMETER
DEFAULT
VALUE
USER
SELECTED
OPTION
UNITS
DATA ENTRY FOR SOIL (by year)
Soil lead concentration
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0
0
0
0
0
0
0
2000
2000
2000
2000
2000
2000
2000
Mg/g
TABLE 5-3b. SOIL LEAD DATA ENTRY WORKSHEET
FOR CHILD EXPOSED TO 2000 jig/g SINCE AGE 1
PARAMETER
DEFAULT
VALUE
USER
SELECTED
OPTION
I
UNITS
DATA ENTRY FOR SOIL (by year)
Soil lead concentration
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0
0
0
0
0
0
0
100
2000
2000
2000
2000
2000
2000
Mg/g
5-6
-------
TABLE 5-3c. SOIL LEAD DATA ENTRY WORKSHEET
FOR CHILD EXPOSED TO 2000 jig/g SINCE AGE 2
PARAMETER
DEFAULT
VALUE
USER
SELECTED
OPTION
UNITS
DATA ENTRY FOR SOIL (by year)
Soil lead concentration
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0
0
0
0
0
0
0
100
100
2000
2000
2000
2000
2000
/*g/g
TABLE 5-3d. WORKSHEET FOR YEARLY SOIL LEAD CONCENTRATION
FOR HYPOTHETICAL CHILDREN MOVING FROM A RESIDENCE
WHERE SOH, CONCENTRATION IS 100 jig/g TO A RESIDENCE
WHERE SOH, CONCENTRATION IS 2000 jtg/g
AGE OF
CHILD
(YEARS)
0
1
2
3
4
5
6
AGE AT TIME OF NEW EXPOSURE (YEARS)
0
2000
2000
2000
2000
2000
2000
2000
1
100
2000
2000
2000
2000
2000
2000
2
100
100
2000
2000
2000
2000
2000
3
100
100
100
2000
2000
2000
2000
4
100
100
100
100
2000
2000
2000
5
100
100
100
100
100
2000
2000
6
100
100
100
100
100
100
2000
5-7
-------
TABLE 5-4. PREDICTED ANNUAL AVERAGE BLOOD LEAD
CONCENTRATIONS (jig/dL) FOR HYPOTHETICAL CHILDREN
MOVING FROM A RESIDENCE WHERE SOIL CONCENTRATION IS
100 fig/g TO A RESIDENCE WHERE SOIL CONCENTRATION IS 2000
AGE OF
CHILD
(YEARS)
0
1
2
3
4
5
6
AGE AT TIME OF NEW EXPOSURE (YEARS)
0
16.2
18.6
17.7
17.3
14.7
12.6
11.3
1
2.8
16.3
17.7
17.3
14.7
12.6
11.3
2
2.8
3.0
14.5
17.2
14.7
12.6
11.3
3
2.8
3.0
2.8
13.5
14.5
12.6
11.3
4
2.8
3.0
2.8
2.6
10.2
12.2
11.2
5
2.8
3.0
2.8
2.6
2.3
8.6
10.8
6
2.8
3.0
2.8
2.6
2.3
2.1
7.5
The changes in exposure scenario are made by first using the parameter selection menu
(Option "1" on the Main Menu), Option "4" on the parameter selection menu, and then
entering selection "2" in the soil concentration box of the Soil/Dust menu. This allows the
entry of separate values for soil lead exposure concentration at each age, The default value
of 200 fjLg/g for each age may be replaced by 100 or by 2000, as indicated by the scenario on
the work sheet. When finished, the user must return to the Soil/Dust menu. In order to
change the dust lead exposure from the default, a constant 200 /*g/g, the; user must move the
cursor down to the dust lead entry box in the Soil/Dust Menu and enter selection "3", the
multiple source menu. The default soil-to-dust coefficient of 0.70 is activated by entering the
Multiple Source Menu, and may be changed as needed. We will not modify either the soil-
to-dust coefficient of 0.7, nor the air-to-dust coefficient of 100 yug/g per wg/m3. The
complete input file may be saved by returning to the Soil/Dust Menu and using the F6 key.
The model may then be run by using the F5 key.
The results of the seven runs are shown in Table 5-4, which is analogous to
Table 5-3(d). The second column shows blood lead concentrations for a typical child
exposed to 2000 /*g/g lead in soil since birth. The peak blood lead concentration of
18.6 jug/dL is reached at age one year. If the initial exposure to 2000 jug/g occurs later, the
peak blood lead concentration is lower.
5-8
-------
Most of the blood lead response to a change in exposure or a change in environmental
lead concentration occurs in the first 3 months after the change. The change in blood lead
during the first three months after changing exposure is at least 50 to 60 percent of the total
difference in quasi-state-state blood lead concentration before and after the change. The
remaining difference will slowly decrease during the next 2 years. We thus suggest that
cross-sectional blood lead studies or baseline blood lead concentrations measured in
longitudinal studies require that all children shall have lived at their present address for at
least 3 to 6 months prior to the blood lead sample.
EXAMPLE 5-4. Example for Children in a Residence Where the Soil Has Been Abated
This sequence of runs considers soil lead exposure decreased from 2000 to 100
and the soil contribution to dust decreased from 1400 to 70 jug/g, at ages 0 (i.e. constant
exposure without soil and dust lead after birth), at age 1, age 2, and so on. This assessment
studies the effects of abatement on children at different ages. The entries for this example
are similar to those of Example 5-3. The summary of seven data entry worksheets is shown
in Table 5-5(d), and the results are shown in Table 5-6.
TABLE 5-5a. SOIL LEAD DATA ENTRY WORKSHEET
FOR CHILD WITH SOIL ABATED TO 100 pg/g SINCE AGE 0 (BIRTH)
PARAMETER
DEFAULT
VALUE
USER
SELECTED
OPTION
UNITS
DATA ENTRY FOR SOIL (by year)
Soil lead concentration
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0
0
0
0
0
0
0
100
100
100
100
100
100
100
Mg/g
5-9
-------
TABLE 5-5b. SOIL LEAD DATA ENTRY WORKSHEET
FOR CHILD WITH SOIL ABATED TO 100 jtg/g SINCE AGE 1
PARAMETER
DEFAULT
VALUE
USER
SELECTED
OPTION
UNITS
DATA ENTRY FOR SOIL (by year)
Soil lead concentration
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0
0
0
0
0
0
0
2000
100
100
100
100
100
100
Mg/g
TABLE 5-5c. SOIL LEAD DATA ENTRY WORKSHEET
FOR CHILD WITH SOIL ABATED TO 100 /tg/g SINCE AGE 2
PARAMETER
DEFAULT
VALUE
USER
SELECTED
OPTION
UNITS
DATA ENTRY FOR SOIL (by year)
Soil lead concentration
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
0
0
0
0
0
0
0
2000
2000
100
100
100
100
100
Mg/g
5-10
-------
TABLE 5-5d. WORKSHEET FOR HYPOTHETICAL CHILDREN IN A
NEIGHBORHOOD WHERE SOIL CONCENTRATION IS REDUCED FROM
2000 Mg/g TO 100 jig/g
AGE OF
CHILD
(YEARS)
0
1
2
3
4
5
6
AGE AT TIME OF ABATEMENT (YEARS)
0
100
100
100
100
100
100
100
1
2000
100
100
100
100
100
100
2
2000
2000
100
100
100
100
100
3
2000
2000
2000
100
100
100
100
4
2000
2000
2000
2000
100
100
100
5
2000
2000
2000
2000
2000
100
100
6
2000
2000
2000
2000
2000
2000
100
TABLE 5-6. PREDICTED BLOOD LEAD CONCENTRATIONS (/ig/dL) FOR
HYPOTHETICAL CHILDREN IN A NEIGHBORHOOD WHERE SOIL
CONCENTRATION IS REDUCED FROM 2000 /tg/g TO 100 |tg/g
AGE OF
CHILD
(YEARS)
0
1
2
3
4
5
6
AGE AT TIME OF ABATEMENT (YEARS)
0
2.8
3.0
2.8
2.6
2.3
2.1
1.9
1
16.2
5.4
2.8
2.6
2.3
2.1
1.9
2
16.2
18.6
6.1
2.7
2.3
2.1
1.9
3
16.2
18.6
17.7
6.6
2.5
2.1
1.9
4
16.2
18.6
17.7
17.3
6.9
2.55
2.0
5
16.2
18.6
17.7
17.3
14.7
6.2
2.4
6
16.2
18.6
17.7
17.3
14.7
12.6
5.8
A sequence of work sheets is needed to study the effects of abatement at different ages.
The two variables to be considered here are the child's age, which is a variable used in the
IEUBK Model simulation, and the age of the child when the abatement was carried out,
which is different for each run in the sequence of 7 runs. In Table 5-5(a), if the soil is
abated at age 0 years, the child is exposed to 100 pg/g lead in soil and 70 jug/g lead in dust
5-11
-------
derived from soil from birth through age 6 years. However, if the soil is abated at age one
year, the correct work sheet is shown in Table 5-5(b). In the work sheet in Table 5-5(b), the
child is exposed to 2000 /xg/g lead in soil and 1400 jug/g lead in household dust at age zero
years, but to 2000 /xg/g lead in soil and 1400 jug/g lead in dust from soil at ages 1 through
6 years. Similarly, if the soil is abated at age two years, the correct work sheet is shown in
Table 5-5(c). In the work sheet in Table 5-5(c), the child is exposed to 2000 /wg/g lead in
soil and 1400 /*g/g lead in household dust at ages 0 and 1 years, but to 100 /ig/g lead in soil
and 70 jxg/g lead in dust from soil at ages 2 through 6 years.
The worksheets for Tables 5-5(a-c) are combined and shown as columns 2 to 4 in
Table 5-5(d). The last 4 columns in Table 5-5(d) summarize the soil lead work sheet entries
for a hypothetical child if the soil is abated at ages 3, 4, 5, or 6 years respectively. For
example, in the extreme right-hand column, if the soil is abated at age 6 years, he or she is
exposed to 2000 jiig/g lead in soil from birth through age 5 years, then to 100 /*g/g at age
6 years.
The IEUBK Model simulation for this example is run 7 times, each run corresponding
to a column in Table 5-5(d) or to a work sheet 5-5(a-c) or analogous work sheets for older
children. The results, as annual averages of predicted geometric mean blood lead
concentration, are shown in Table 5-6 in exactly the same order as in Table 5-5(d).
Abatement at age 1 reduces blood lead from 16.2 to 5.4 /xg/dL in the first year after
abatement, a reduction of 10.8 /xg/dL or 66.7 percent. The effect at age 2 is a reduction
from 18.6 to 6.1 j*g/dL, that is 12.5 Mg/dL or 67.7 percent. Abatement at age 5 has a
reduction of 8.5 /*g/dL or 57.8 percent in the first year. It should be noted that blood lead
concentrations do not reach the post-abatement quasi-steady state level until two years after
the abatement, so that the apparent reduction in blood lead concentration in the first year
after abatement will underestimate the effectiveness of abatement.
EXAMPLE 5-5. Historical Exposure Reconstruction for Soil and Dust Lead
Concentration and Dietary Lead Intake Around an Unused Lead
Smelter
One of the issues that arose in developing validation case studies for the IEUBK model
is that many of the earlier data sets were collected at sites where background lead exposure
differed greatly from current default values, and where both background exposure and
soil/dust exposure were changing substantially during the lifetime of the children in the blood
5-12
-------
lead study. It was therefore necessary to construct an historical exposure scenario for the
children in the blood lead study. The exposure reconstruction for the 1983 East Helena
blood lead study was discussed in the initial validation of the UBK model (U.S.
Environmental Protection Agency 1989). In this example, we will discuss the more
complicated exposure situation for the 1983 companion study in the Silver Valley of Idaho.
We rely heavily on the initial report on Kellogg Revisited (Panhandle District Health
Department 1986), the Human Health Risk Assessment (Jacobs Engineering, 1989, for US
EPA Region X), the Risk Assessment Data Evaluation Report (US EPA 1990), the House
Dust Remediation Report (CH2M Hill 1991 for the Idaho Dept. of Health), the Record of
Decision for the Bunker Hill site (U.S. Environmental Protection Agency 1991), and
personal communications with Dr. Ian Von Lindern of Terragraphics Inc. (1992-1993).
The narrow east-west Silver Valley was divided initially into three residential areas,
Area 1 (Smelterville) about 1.2 to 1.5 km northwest of the smelter complex, Area 2
(Kellogg) about 2.6 to 3.3 km east of the smelter complex, and Area 3 (Pinehurst) about
6 km west of the smelter complex. In subsequent studies this area was extended and
subdivided into 5 to 11 areas or zones. A list of zones and distances is attached as
Table 5-7. The main distinction is that the Page neighborhood which is only 3 km west of
the smelter complex has been distinguished from Pinehurst, and that the Wardner
neighborhood about 3 km southeast of the smelter complex has been separated from the
Kellogg community. The five areas currently defined are closer in size and population to the
"neighborhoods" recommended in Chapter 4.
Silver Valley has a complicated history of lead exposure, including significantly
elevated air and dust lead exposures in 1974 and 1975, and a cessation of lead smelting
activities after December 1991. Therefore, the exposure history reconstruction in
Table 5-8 is a mixture of observed values and interpolated values. The observed values were
sometimes recorded as geometric means and sometimes as arithmetic means, and as estimates
or interpolations enclosed in brackets. The basis for the dust lead interpolation was not
described in more detail in (Jacobs Engineering 1989). The soil lead concentrations were
held at the last measured value until a new observed value had been achieved.
Soil and dust lead concentrations were only observed in 1974, 1975, 1983, and 1986-
1988. Dust lead concentrations have also been observed in these communities since 1990.
There are alternatives to estimating neighborhood soil and dust lead concentrations between
actual observations, such as by linear interpolation, that may provide somewhat different
estimates than the interpolations used in the human risk assessment study.
5-13
-------
TABLE 5-7. NEIGHBORHOOD IDENTIFIERS AND DISTANCE FROM STACK
FOR KELLOGG, ID, STUDY
ZONE
A
B
C
D
E
F
G
H
I
J
K
APPROXIMATE DISTANCE
FROM ZONE CENTER TO
Pb SMELTER STACK
(Km)
1.50
1.15
2.75
3.25
2.60
3.30
3.00
5.70
3.30
DESCRIPTION
Smelterville, south of old Highway 1 10 and west of C
street
Smelterville, east of C street
Kellogg, north of 1-90 and west of Hill street
Kellogg, north of 1-90 and east of Hill street
Kellogg, south of 1-90 and west of Division street
Kellogg, south of 1-90 and east of Division street
Wardner
Pinehurst
Page
Smelterville, (1974-75 only)
Kellogg/Page, (1974-75 (only)
An alternative assumption is that soil and dust lead concentrations decreased linearly
between 1983 and 1986-1988. Thus, in Smelterville the decline in soil lead was 3047 -
2685 = 362 /xg/g in 4 years, or 90 /xg/g per year, whereas in Kellogg it was 2584 -
1988 = 596 /xg/g in 4 years, or about 150 /xg/g per year. The dust lead concentrations in
Smelterville decreased by 3715 - 1203 = 2512 /xg/g in 5 years, or about 250 /xg/g per year.
whereas the dust lead concentration in Kellogg decreased by 2366 - 1450 = 916 /xg/g in
5 years or about 230 /xg/g per year. However, the dust lead concentrations in 1990-1992
were still elevated above the Pinehurst concentration. It would be prudent to assume that the
dust lead concentration was relatively constant for much of the period around and after 1988.
By implication, since soil lead and air lead are sources for dust lead, one might assume that
the soil lead and air lead concentrations for 1988-1992 are relatively constant at the 1988
values.
The soil and dust lead values for a Kellogg child born in 1983, assuming a linear
decrease of 150 /xg/g in soil lead from 2584 /xg/g and a linear decrease of 230 /xg/g in dust
lead from 2366 /xg/g, is shown on the worksheet in Table 5-9. In this example we have
5-14
-------
TABLE 5-8. OBSERVED AND ESTIMATED AIR, SOIL, AND DUST LEAD
CONCENTRATIONS FOR USE IN HISTORICAL EXPOSURE RECONSTRUCTIONS
IN SILVER VALLEY COMMUNITIES
YEAR
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
SMELTER VILLE
PbA1-2
5.7
11.2
16.5
14.3
8.9
9.8
9.1
5.4
6.6
6.2
4.6
0.88
0.20
0.12
0.19
0.30
0.36
0.36
PbS1'2
[6141]
[6141]
[6141]
6141
3991
[3991]
[3991]
[3991]
[3991]
[3991]
[3991]
[3991]
3047
[3047]
[3047]
[3047]
2685
[2685]
PbD1'2
[3530]
[6620]
[12500]
10583
3533
[6030]
[5670]
[3530]
[4020]
[3780]
[2830]
[3715]
[3715
[3715]
[3715]
[3715]
12034
18583
14963
9783
KELLOGG
PbA1-2
8.2
9.6
15.0
14.0
7.4
7.5
6.8
5.4
5.9
5.9
4.1
0.28
0.19
0.12
0.13
0.19
0.17
0.11
PbSu
2514
2586
2584
1988
PbD1'2
6581
4573
2366
14504
19203
15023
12273
PINEHURST
PbA
[6.1]
[6.1]
[6.1]
6.1
3.1
3.4
3.6
2.7
3.1
2.2
1.2
0.16
0.14
0.09
0.10
0.10
0.08
0.08
PbS
765
508
472
PbD
2006
1749
1155
10223
10683
9443
Data Sources:
1. Jacobs Engineering (1989) for data before 1989, Tables 4-5, 4-7, 4-13. PbA values are arithmetic
means of lead in air (fig/m ), PbS and PbD values not in brackets are geometric means of lead in soil and
dust (/ig/g).
2. Jacobs Engineering (1989) for data before 1989. PbS and PbD values in brackets are estimates from
Figure 4-16.
3. I. Von Lindern, personal communication. Arithmetic means of dust lead concentrations.
4. Record of Decision, 1991. Tables 5-1, 5-8.
5-15
-------
TABLE 5-9. USER-SELECTED ENTRIES FOR IEUBK MODEL WORKSHEET
FOR EXAMPLE 5-5, CHILD BORN IN KELLOGG, IDAHO, IN 1983
PARAMETER
YEAR
DEFAULT
VALUE
USER SELECTED
OPTION
UNITS
DATA ENTRY FOR SOIL (by year)
Soil lead concentration
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
1983
1984
1985
1986
1987
1988
1989
0
0
0
0
0
0
0
2,584
2,434
2,284
2,134
1,984
1,834
1,834
Mg/g
DATA ENTRY FOR DUST (by year)
Dust lead concentration
Age = 0-1 year (0-11 mo)
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
1983
1984
1985
1986
1987
1988
1989
0
0
0
0
0
0
0
2,366
2,136
1,906
1,676
1,446
1,446
1,446
/"g/g
treated soil and dust lead concentrations as typical values for the community. Model results
for the distribution of blood lead concentrations using these inputs would be expected to be
more narrow than seen in the actual community due to variability of exposure concentrations
within the community.
The dietary lead intake depends on the age of the child and on the year of interest. For
a child born in 1983, the dietary lead intake data entry worksheet is shown in Table 5-10,
using data from Table 2-1. The two additional dietary exposure scenarios are for children
who consume only home-grown vegetables, or only locally-caught fish. From Table 2-3 we
calculate a weighted average lead concentration of 5.5 /xg/g for leafy and root vegetables
grown in Smelterville. The worksheet is shown in Table 5-11. From Table 2-4 we find a
5-16
-------
TABLE 5-10. USER-SELECTED ENTRIES FOR IEUBK MODEL WORKSHEET
FOR EXAMPLE 5-5, CHILD BORN IN SMELTERVHXE,
IN KELLOGG, IDAHO, IN 1983
PARAMETER
DEFAULT
VALUE
USER SELECTED
OPTION
UNITS
DATA ENTRY FOR DIET (by year)
Dietary lead intake
Age = 0-1 year (0-11 mo),
1-2 years (12-23 mo)
2-3 years (24-35 mo)
3-4 years (36-47 mo)
4-5 years (48-59 mo)
5-6 years (60-71 mo)
6-7 years (72-84 mo)
5.53
5.78
6.49
6.24
6.01
6.34
7.00
14.42
22.67
12.34
9.08
6.01
6.34
7.00
fjig Pb /day
TABLE 5-11. USER-SELECTED ENTRIES FOR IEUBK MODEL WORKSHEET
FOR EXAMPLE 5-5
PARAMETER
DEFAULT
VALUE
USER SELECTED
OPTION
UNITS
DATA ENTRY FOR ALTERNATE DIET SOURCES (by food class)
Concentration:
home-grown fruits
home-grown vegetables
fish from fishing
game animals from hunting
Percent of food class:
home-grown fruits
home-grown vegetables
fish from fishing
game animals from hunting
0
0
0
0
0
0
0
0
5.5
0.80
36
50
j"g Pb/g
%
lead concentration in locally caught fish of 0.80 jwg/g, over twice the national average level
at that time. The data entry for fish is shown in Table 5-11. The assumed percentages for
local vegetables and fish consumption are 36 and 5 percent, respectively.
5-17
-------
The results for elevated soil and dust lead plus baseline dietary lead intake show that if
locally-grown vegetables and fish are consumed in large amounts, there is a modest increase
in blood lead concentration at each age.
We will discuss blood lead estimation for this example in the validation studies that will
reported separately from this manual. We have included this example in the Guidance
Manual to give the reader some "real world" exposure scenarios and to confront the reader
with some of the choices that may need to be made in developing hi .orical exposure
scenarios for blood lead studies.
5.3 APPLICATIONS FOR PROBABILITY AND RISK ESTIMATION
EXAMPLE 5-6. Default Parameters
For the default parameters in Example 5-1, the estimated geometric mean blood lead
for children of ages 24 to 35 months is 4.2 /ig/dL. The user may choose any other age
range. If the user next goes into Option 1 from the bottom menu, then "3" from Graphics
Selection Menu and selects age range 24-36 months (K), the log-normal probability density
should appear on screen. This plot can be printed on a user-specified printer. The user can
save the graphics file for additional review using the Multiple Runs Option M with just a
single run. No default parameters were changed, except for the GSD, which was changed to
1.42. With GSD = 1.42, there is an estimated 0.68 percent risk that a child with the default
exposure scenario will have a blood lead exceeding 10 /jg/dL.
A useful alternative display is shown by selecting the Distribution Probability Percent
"2" among the plot options. This shows the risk of a blood lead exceedance for any possible
blood lead concentration from 0 to 16 /ig/dL, not just the level of concern of 10 /xg/dL, but
the line is too close to zero to be visually distinctive above 12
EXAMPLE 5-7. Sensitivity of Risk Estimates to User-Selected Geometric Standard
Deviation
One way to carry out sensitivity analyses is to carry out each simulation run
individually, but to collect the results for different parameters in cumulative output data sets.
5-18
-------
The IEUBK model does not currently offer options to do this for any parameters except
media concentrations that do not change with age during single simulation run. We will thus
fix all of the model parameters at default values, except for the GSD, which in this example
will take values from 1.42 to 1.90. After running the model as in the preceding example,
we will select "6" in the Graphics Selection Menu. This allows the user to change both the
GSD and the blood lead level of concern, while leaving the geometric mean blood lead level
at the same value, here 4.2 pg/dL. The results for different GSD values are shown in
Table 5-12, for children of ages 24-35 mos.
TABLE 5-12. EFFECTS OF GSD ON THE PROBABILITY OF EXCEEDING
10 /ig/dL, USING ONLY DEFAULT EXPOSURE PARAMETERS,
FOR CHILDREN AGES 24 TO 35 MONTHS
GSD
1.42
1.50
1.60
1.70
1.80
1.90
Probability of Blood Lead >
10 /xg/dL
0.0068
0.0157
0.0324
0.0513
0.0696
0.0870
EXAMPLE 5-8. Effects of Dust Lead Concentration on Risk Estimates for Fixed Soil
Lead Concentration
In this example, we can use Option "2" on the Computation Menu to assess the effects
of different dust lead levels for a fixed soil lead concentration. We will here assume a soil
lead concentration of 1,000 /-tg/g, and dust lead concentrations incremented in the Multiple
Runs Analysis. The soil lead concentration is not a default and must be reset to 1,000 jug/g
in the Soil/Dust Data Entry Menu (4). We will use 7 levels of dust lead, from 0 to
1,500 /xg/g by steps of 250 /^g/g. These should be changed in the Multiple Runs Analysis,
by entering sub-menus 1 (medium = dust), 2 (range set to 0-1500), and 4 (7 levels of dust,
results sent to graphics and results save files). All of the other parameters are set to default
values except for a GSD of 1.70 to illustrate the effect of a larger GSD. Selection 3 runs the
models.
5-19
-------
Return to the Output Menu (3), select Plot (2), select Plot Overlay (Density), highlight
overlay file, select 24-36 months (K), and the plot will appear on the display. The results
are shown in Screen 5-1, which shows the probability density plots for a GSD of 1.70.
We are assuming maximum bioavailability (30%). With no lead in dust, the probability that
a 2-year-old will exceed 10 /wg/dL is estimated as 25 percent. (Run 1), whereas with dust
lead concentration of 1,500 jug/g (1.5 times as large as the soil lead concentration) this
probability increases to 73 percent. We see that there is substantial sensitivity to the soil-to-
dust coefficient and to additional non-soil sources of dust lead in this example.
CutoH: 10.0 ug/dL
Bun1:2S.4«S
Hun2:M.«9%
RunS:4«J2S
Run 4:17.64%
RunI:«UO%
Runt: 70.15%
Hun7:7JJ»%
Screen 5-1. Multiple runs probability density function for soil lead == 1,000 pg/g, dust
lead = 0 to 1,500 /tg/g, by steps of 250 fig/g (Runs 1 through 7) in
Example 5-6.
The cumulative exceedance probability plots (selection 4 in the Graphics Selection
Menu) are shown in Screen 5-2. These plots show a clear increase of risk with increasing
dust lead level at all blood lead levels of concern, and offer the user a visual display that
may help to separate the risk estimates for different dust lead levels.
In order to assess the relationship between geometric mean blood lead and dust lead
concentration, the user must set soil lead to 1000 in Option 4 of the Parameter Input Menu
5-20
-------
Screen 5-2. Multiple runs probability of exceedance of blood lead levels for soil lead
1,000 ftg/g, dust lead = 0 to 1,500 jtg/g, by steps of 250 /tg/g
(Runs 1 through 7) in Example 5-6.
and then go to Option "2" of the Computation Menu. In Option B, enter sub-menus 1
(medium = dust), 2 (range set to 0-1500), and 4 (7 levels of dust, results sent to graphics
and results save files). All of the other parameters are set to default values.
Selection 3 runs the models. The results may be plotted immediately, as shown in
Screen 5-3, or saved in a *.PBM file for later plotting. Note the slight nonlinearity as dust
lead levels exceed 1,000 jug/g, due to saturable absorption effects.
5.4 BATCH MODE INPUT AND STATISTICAL ANALYSES OF
OUTPUT
This section demonstrates the use of the batch mode analysis method with input data
that are typical of the data available to the user in most environmental lead field studies.
Assessment of goodness of fit of predicted and observed blood lead levels (when available)
requires a statistical analysis of the data using a variety of mathematical and graphical
techniques. Output data from the batch mode runs are in ASCII files that can be loaded into
almost any statistical analysis package or spreadsheet program that the user may want to use.
5-21
-------
Dust Pb CONCENTRATION (ug/g)
Screen 5-3. Relationship of predicted blood lead to dust lead in Example 5-6.
The IEUBK batch mode output files will require little or no editing before being imported
into other programs, unless the missing value code (—) is incompatible with the user's
package. We have provided a small special-purpose program called PBSTAT that can be
used after the batch mode output file is created, by exiting from the IEUBK model and
executing PBSTAT, or by Option "5" in the Batch Mode Menu. PBSTAT is provided as a
convenience for the user who may not have or wish to use other programs with the IEUBK
output file. The statistical and graphical methods in PBSTAT are demonstrated in the
following examples. Additional statistical analyses of the batch mode output data files are
not possible using PBSTAT, and must be done with other programs.
EXAMPLE 5-9. Complete Data Set for an Old Mining Community
The input data format for a batch mode input file was described in Section 3.3. The
data input file for this example is shown in Table 3-2. This data set was produced by a
computer simulation and was edited into the format shown in Table 3-2. These are complete
data, i.e., there are no missing values for any of the variables.
5-22
-------
Let us suppose that these data represent the data for a sample cohort of children, all of
whom were 18 months old at the time of blood lead sampling in late October. Let us assume
that the data were collected in the community of "Mountain Pass", an old historic town that
has been the site of active lead mining, ore processing, and smelting operations for over
100 years. These operations stopped about 25 years ago, and after a period of declining
population the town is once more growing as the center of newly developed tourist and
outdoor recreation activities. There is now considerable concern about the potential risk of
elevated lead concentrations in soil and in the interior dust of the older houses in Mountain
Pass. These children were recruited in the first phase of a long-term prospective study on
changes in blood lead concentration in Mountain Pass children during a proposed soil lead
abatement project.
The data set contains blood lead concentration in children, soil and dust lead
concentration in their houses, in four neighborhoods in Mountain Pass. Air lead
concentration were measured by a Total Suspended Particulate (TSP) sampler about ten years
ago and were found to be less than 0.2 /xg/m3, so have not been measured since then. First-
draw and partially flushed water lead samples were collected at each house, but have not yet
been analyzed. Lead-based paint was measured by a portable X-Ray Fluorescence
Spectrophotometer (XRF), but there have been some concerns about the instrument
calibration during the unseasonably cold weather in which the measurements were made and
the site manager has decided not to use the XRF data until the XRF measurements can be
replicated next summer. (Even though this is only a hypothetical example, the reasons why
some data may not be available are real, and are all too likely to occur in any real field
study). The first model run done by the site manager used this data set "as is", with all of
the parameters set to their default values in Table 3-1.
The batch mode run is made from Option 4 in the Computation Menu. The user must
identify the input data set, known here as EXAMPLE1.DAT. The user also has the option
of renaming the data set before running the batch mode analysis. If the user does not rename
the data set, then [name].DAT input file results will be saved in data sets [name].TXT and
[name].ASC—in this case, EXAMPLE1.TXT and EXAMPLE1.ASC. The output data file
EXAMPLE1.TXT may be viewed from Option "2" of the Batch Mode Menu after the batch
run is completed.
Option "5" of the Batch Mode Menu, can be used to examine the differences between
observed and predicted blood lead levels using a variety of graphical and statistical
techniques. The user must leave the main IEUBK model in order to enter the statistical and
5-23
-------
graphical program PBSTAT. Selection 1 in the PBSTAT menu allows the user to load the
ASCII file denoted [name].ASC. Selection 2 displays a screen full of statistical information.
The information on this screen should be useful for many reports. The table includes the
geometric and arithmetic mean blood lead concentrations, as well as the 25th, 50th (median),
75th, and 90th percentiles of observed and predicted blood lead levels. This screen reports
paired-sample T-tests for the equality of geometric mean observed and predicted blood lead
levels in the neighborhood, which is a test of the equality of the mean logarithms (left side of
screen). Tests of the equality of the arithmetic mean blood lead concentration are shown on
the right-hand side. You should not expect that the statistical tests will report agreement
between observed and predicted values (see Section 1.1.5.3). These tests are used to help
diagnose problems.
The two-sample Kolmogorov-Smirnov (denoted K-S) test of the equality of the two
distribution functions is also reported. This is based on a very simple statistic, the largest
absolute difference between the cumulative distribution of the observed blood lead levels and
the cumulative distribution function of the predicted values. We have knowingly violated the
assumption that these values are independent, thus the null hypothesis distribution will not
give valid significance levels. However, we have found that the K-S statistic, together with
the percentiles, provides valuable information about the kinds of discrepancies between the
neighborhood-scale blood lead distribution and the distribution of predicted blood lead
concentration.
Graphical comparisons of observed and predicted blood lead concentrations are very
helpful. If the user exits from the statistics screen and then uses Selection 3 in the PBSTAT
selection menu, for graphics and plots, there are a number of choices. Option 1 in the
PBSTAT graphics selection menu allows plots of cumulative distribution functions, either
singly or combined. Either regular or log-transformed blood lead concentrations may be
plotted. The empirical cumulative distribution functions (CDF's) differ substantially.
Another useful graphical comparison is in Selection 4 of the Graphing Selection menu, "box
and whisker" plots. The boxes show the quartiles of the distribution(s), and the whiskers
show the range of non-outlier blood lead concentrations. Outliers, by internal criteria, are
shown as isolated data points. Observed and predicted values are highly correlated in the
example, as shown by Graphing Selection choice 2. Many other plots may be generated by
use of Selection 3.
In this example the model has somewhat over-estimated the observed blood lead
concentrations. Any one of several factors could explain the difference between observed
5-24
-------
and predicted blood lead concentrations in these children. Are there adequate quality
assurance data for both the blood lead and the environmental lead measurements and do they
show satisfactory performance during the study? Because the narrative for this scenario
stated that blood lead concentrations were collected in late October, which was described as
"unseasonably cold", could the children have been spending much less tune playing in soils
outside? If so, the blood lead data may reflect lower-than-average intake of soil recently, so
that the ingestion rates in the model, which are annual averages, are not representative of the
atypical conditions under which these blood lead data were collected. Were most of these
children placed in some sort of day-care facility? If so, then the children in the day-care
facilities could be analyzed as a separate group with appropriate lead concentration data for
the facilities. Other possibilities, such as lower bioavailability of soil lead at some houses or
in some neighborhoods, should be investigated. In any event, the answers to these questions
are going to be found in site-specific data about child behavior, exposure to soil and dust,
and on the chemical and physical properties of the soil and dust at the site, and not in further
manipulations of model parameters. An analysis of these data, with additional exposure data,
is presented as Example 5-11.
EXAMPLE 5-10. Batch Input Data File with Missing Environmental Lead Data
Some environmental data in a data set may be missing because the samples were not
collected, were lost or damaged during transportation, storage, and sample preparation for
analysis, or were improperly coded and thus not recorded. In any case, the values for
missing data in an IEUBK model batch mode input file may be coded by an isolated decimal
point where the variable value would otherwise be placed. Examples are given in the data
sets EXAMPLE2.DAT and EXAMPLE3.DAT provided on the program disk. Missing
values for water lead, air lead, and paint lead are automatically replaced by default values:
4 /ig/L for water, 0.1 /xg/m3 for air, and 0 /*g/day for alternative sources. The imputation
method for soil and dust lead is different. If soil lead is missing, and dust lead is not
missing, then the missing value of soil lead is set to the dust lead value. If dust lead is
missing, and soil lead is not missing, then the missing value of dust lead is set to the soil
lead value. These cases may be used to estimate or predict blood lead levels. If both soil
and dust lead concentrations are missing, then no data are imputed and the blood lead
concentration is not calculated for this child. The missing values imputed by the model are
earmarked by an asterisk in the [name].TXT output file. The user is responsible for defining
an appropriate data imputation process for any site-specific data set that has missing values.
The file along with any imputed data should be created before it is submitted to the Batch
Mode Option.
5-25
-------
One convenient method for imputation of missing dust lead levels is to invoke the
Multiple Source menu alternative for dust. The default values in this option (soil-to-dust
coefficient of 0.70, air lead contribution of 10 /xg/g to house dust) produce a somewhat
different set of dust lead estimates and correspondingly different predicted blood lead
concentrations.
Note that missing values of blood lead do not affect the prediction of blood lead from
environmental lead data, provided that either a soil lead or a dust lead concentration is
present, or that the user has imputed values for soil and dust lead calculated by some other
method and inserted in place of the missing value.
EXAMPLE 5-11. Lead Exposure in an Old Mining Community Using Site-Specific
Information About Ingestion of Soil and Dust
Suppose that the site manager in Example 5-9 has obtained additional information about
the children hi this sample, and finds that almost all of them have been enrolled in a day care
program in this community. Upon visiting the day care facility, the site; manager observes
that the facility is modern, with easily cleanable floors, entrance surfaces and window sills.
She or he observes that the facility appears to be cleaned often, and that the day care facility
operators are aware of the hazard of childhood exposure to lead in dust and are making
deliberate efforts to reduce the exposure. She or he also learns that most of the children's
parents are employed full-time, and that most of these children spend 8 to 10 hours per day
at the facility.
Is there now enough information to change the parameters of the IEUBK model so as to
possibly provide a closer description of the data? We would not recommend rerunning the
IEUBK Model without additional site-specific data. If predicted blood lead concentrations
tend to be somewhat larger than those observed, any one or more of the following
possibilities could explain the discrepancy:
(i) The soil lead and dust lead concentrations at the day care center may be
much lower than the residential lead concentrations, so that a significant
part of the child's daily ingestion of soil and dust includes much less lead
than if the same quantity were ingested at home;
(ii) The quantity of soil and dust ingested may be smaller than expected
because the child spends a great deal of time away from the: home in a
relatively clean environment, and frequently interacts with adult
5-26
-------
caretakers and with other children, thereby reducing both environmental
and behavioral magnifiers of soil and dust ingestion;
(iii) The bioavailability of lead in soil and dust at home or elsewhere may be
lower than the default values used in the IEUBK Model;
(iv) The children in the sample may represent a non-typical sub-population
with respect to ingestion or absorption;
(v) There may be measurement errors in soil lead, dust lead, or blood lead,
possibly causing a systematic downward bias in lead measurements.
Any manipulation of the IEUBK Model that reduces lead uptake from a medium would
reduce the predicted blood lead concentration and improve the overall fit of the predicted
values to the observed values. This does not prove that the manipulation is valid. Lead
uptake is the product of ingestion rate and absorption from the medium, so that achieving
goodness of fit to the observed values can never prove the correctness of the manipulation of
parameters.
We would recommend that some additional site studies be carried out to evaluate these
possible causes. These studies include, in the same sequence (i-v):
(i) The soil lead, dust lead, and drinking water lead concentrations at the day
care center should be measured;
(ii) The amount of dust in both the residence units and the day care center
should be determined by measuring floor dust loadings;
(iii) Methods for child recruitment should be evaluated for possible sampling
biases. Socio-demographic factors that may affect soil and dust ingestion
should be investigated, including the role of parental awareness and
public information programs. Nutritional differences that may affect lead
bioavailibility, such as deficiency or repleteness of calcium intake, should
be determined where feasible;
(iv) Seasonal biases, biases in sampling locations and in timing of soil and
dust sampling studies should be considered as possible measurement
errors. QA/QC data for analytical procedures for soil lead, dust lead,
blood lead and other media should be reviewed for possible errors,
instrument drift or other systematic biases.
For risk assessment applications, it may be preferable to use the default exposure
scenario for children who do not spend most of their waking day in a clean environment
outside the home. There is no guarantee that other children in this community will not be at
5-27
-------
higher risk than the children in the sample. We are not suggesting the use of conservative
assumptions about ingestion, but rather, the use of realistic assumptions about a plausible
alternative exposure scenario (for example, if the day care facility closes down and is not
replaced by a similar facility).
5.5 SOIL LEAD ABATEMENT EXAMPLES
Example 5-12. Use of the Multiple Runs Selection to Estimate Soil Lead Abatement
Target Levels when Household Dust is Also Allowed to Vary
One of the more frequent applications of the IEUBK model has been to help determine
soil lead concentrations for which abatement may be needed in order to reduce the likelihood
of exceeding a blood lead level of concern (LOG) to some user-defined risk of exceedance
(ROE) of the LOG at the site. These soil lead target concentrations are site-specific variables
and reflect to a greater or lesser degree all of the other parameters that determine childhood
blood lead levels after abatement. Effective soil lead abatement will often include household
dust abatement, both to remove historical reservoirs of contaminated household dust and to
help maintain lower household dust lead concentrations after soil abatement. In this
situation, the post-abatement environment must be characterized by a site-specific soil-to-dust
coefficient so that the soil lead target concentration is connected to a post-abatement dust lead
concentration using the Multiple Source Analysis in the Soil/Dust Data Entry Menu. In this
example, we will assume that all of the parameters in the model have been set to default
values, but even if the default selections in the Multiple Source Analysis for household dust
are invoked, they will not be activated without selecting the Multiple Source option. The
following steps are used to illustrate soil target levels for a soil-to-dust coefficient of
0.70 and an air-to-dust coefficient of 100 jug Pb/g dust per /xg Pb/m air.
1. From the Main Menu, use Option 1: Parameter Menu, then Option 4: Soil/Dust Data
Entry Menu, then tab down to Line 2 (Indoor Dust Pb) and use Option 3: Multiple
Source Analysis.
2. The user may select the soil-to-dust coefficient other than 0.70 and the air-to-dust
coefficient other than 100, but even if the default values are used the user must enter this
menu and then Escape back to the Soil/Dust Entry Menu.
5-28
-------
3. Escape (exit) from the Soil/Dust Data Entry Menu to the Parameter Menu, then to the
Main Menu. Choose Option 2: Computation Menu, then Option 2: Multiple Runs. This
will put the user into the RANGE SELECTION MENU.
4. Set up a range-finding run by using Options 1,2, and 4 in the Range Selection Menu.
In Option 1 (Media), choose Soil and return to the Range Selection Menu. In Option 2
(Range), choose Start = 0 (0 jug/g soil lead) and End = 1500 (1500 /xg/g soil lead) and
return to the Range Selection Menu. In Option 4 (Output Choices), respond "Yes" to the
query "Send to Overlay File", respond "7" to the query "Number of Runs for Range".
This will produce output runs at 7 equally spaced levels of soil lead from 0 to 1500 /*g/g,
namely at 0, 250, 500, 750, 1000, 1250, and 1500 /tig/g. The user who is not familiar
with this option may also wish to respond "Yes" to the query "Display summary
outputs". Return to the Range Selection Menu.
5. Run the Multiple Runs Analysis by selecting Option 3 on the Range Selection Menu.
The user should see the message that the data sets RANGE#.LAY and RANGE#.TXT
have been saved. The data set RANGE#.LAY is needed to obtain the probability plot
values. The data set RANGE#.TXT is needed to document the input parameters for the
run.
6. In order evaluate the range-finding runs, exit from the Range Selection Menu to the
Computation Menu, then to the Main Menu. Select Option 3: Output Menu, then Option
2: Plot menu, then select the GSD and the blood lead level of concern (LOG). The
default values GSD = 1.60 and LOG = 10 jug/dl are used here, so no selection is
necessary; otherwise, use Option 6. Then use Option 5: Plot Overlay File (probability
density functions). Tab down and select the appropriate RANGE#.LAY file, then select
the age range "H", ages 0-84 months, or any other range, as needed. The probability of
exceeding blood lead 10 /xg/dL for each soil lead concentration from 0 to 1500 ^tg/g by
steps of 250 fjig/g is shown in Table 5-13.
5-29
-------
TABLE 5-13. RANGE FINDING RUN FOR TARGET SOIL LEAD
CONCENTRATION
OVERLAY PLOT
1
2
3
4
5
6
7
SOIL LEAD CONCENTRATION
0*g/g)
0
250
500
750
1000
1250
1500
PROBABILITY OF
EXCEEDING 10 pig/dL, percent
0.00
1.99
12.03
26.86
42.68
55.50
64.01
7. As a result of the range-finding runs shown in Table 5-13, the soil lead target
concentration is between 250 /*g/g (ROE = 1.99 %) and 500 /xg/g(ROE = 12.03%).
In order to narrow the list of possible values, repeat steps 4, 5, and 6 with a smaller
range of values. We selected Start = 320 jug/g and End = 420 /xg/g in Option 2 (Range)
of the Range Selection Menu, and selected 6 runs in Option 4 of the; Range Selection
menu. Run the Multiple Runs Analysis with Option 3. This produces an output data set
RANGE#+1.LAY. Plot the results in RANGE#+1.LAY for soil lead concentrations of
320, 340, 360, 380, 400, and 420 jtg/g. The results are shown in Table 5-14.
TABLE 5-14. FOCUSED RUN FOR TARGET SOIL LEAD CONCENTRATION
OVERLAY PLOT
1
2
3
4
5
6
SOIL LEAD CONCENTRATION
Og/g)
320
340
360
380
400
420
PROBABILITY OF
EXCEEDING 10 /tg/dL, percent
3.24
3.45
3.90
4.15
4.70
5.00
5-30
-------
8. Table 5-14 shows that the highest value of 420 jug/g appears to produce ROE = 5.00%.
To confirm this, repeat Step 7 with a much smaller range of values. We selected
Start = 400 /*§/§ and End = 430 yug/g in Option 2 (Range) of the Range Selection
Menu, and selected 4 runs in Option 4 of the Range Selection menu. Run the Multiple
Runs Analysis with Option 3. This produces an output data set RANGED+2.LAY. Plot
the results in RANGE#+2.LAY for soil lead concentrations of 400, 410, 420, and
430 /Ag/g. The results are shown in Table 5-15. This procedure has identified a soil
lead concentration of 410 pg/g as the target level.
TABLE 5-15. VERIFICATION RUN FOR TARGET SOIL LEAD CONCENTRATION
OVERLAY PLOT
1
2
3
4
SOIL LEAD
CONCENTRATION
Og/g)
400
410
420
430
PROBABILITY OF
EXCEEDING
10 ftg/dL, percent
4.70
5.00
5.00
5.32
DUST LEAD
CONCENTRATION
0*g/g)
290
297
304
311
9. The user may wish to view the dust lead concentrations corresponding to this procedure.
In order to view RANGE#+2.TXT, return to the Main Menu, then the Computation
Menu and select Option 4: Batch Mode. Select Batch Mode Option 2: View TXT File,
the RANGE#+2.TXT. The dust lead concentrations are shown in the last column of
Table 5-15.
5-31
-------
6. REFERENCES
Agency for Toxic Substances and Disease Registry (1988). Nature and extent of childhood lead poisoning in the
United States: A Report to Congress. U.S. Department of Health and Human Services, Public Health
Service, Atlanta, GA.
Alexander, F. W.; Clayton, B. E.; Delves, H. T. (1974). Mineral and trace-metal balances in children receiving
normal and synthetic diets. Q.J. Med. 43; 89.
Alexander, F. W. (1974). The uptake of lead by children in differing environments. Environ. Health Perspectives
73: 155-159.
Alexander, F. W.; Delves, H. T.; Clayton, B. E. (1973). The uptake and excretion by children of lead and other
contaminants. In: Earth, D.; Berlin, A.; Engel, R.; Recht, P.; Smeets, J. eds. Environmental health
aspects of lead: proceedings, international symposium; October 1972; Amsterdam, The Netherlands.
Luxembourg: Commissiion of the European Communities; pp. 319-331.
Allott, R. W.; Hewitt, C. N.; Kelly, M. R. (1990). The environmental half-lives and mean residence times of
contaminants in dust for an urban environment: Barrow-in-Furness. Sci. Total Environ. 93: 403-410.
Altman, P. L.; Dittmer, D. S. (eds) (1972). Biological Data Book, 2nd Ed. pp 195-201. Bethesda,
MD: Federation of American Societies for Experimental Biology.
American Academy of Pediatrics (1972). Lead content of paint applied to surfaces accessible to young children.
Pediatrics 49(6): 918-921.
American Water Works Association Research Foundation (AWWARF) (1988). Review of the Biological Basis of
the Proposed Drinking Water Lead Standard. Prepared by Karch and Assoc., Inc. Washington DC.
Amitai, Y.; Brown, M. J.; Graef, J. W.; Cosgrove E. (1991). Residential deleading: Effects on the blood lead
levels of lead-poisoned children. Pediatrics 88: 893-897.
Aungst, B. J.; Fung, H. (1981). Kinetic characterization of in vitro lead transport across the rat small intestine.
Toxicol., Appl. Pharmacol. 61: 39-57.
Baes, C. F.; Sharp, R. D.; Sjoreen, A. L.; Shor, R. W. (1984). A Review of Parameters for Assessing
Transport of Environmentally Released Radionuclides through Agriculture. Report prepared by Oak
Ridge National Laboratory (Martin Marietta) for U.S. Dept. of Energy under U.S. EPA Contract #EOA-
78-D-X0394 (AD-89-F-2-A106).
Bander, L. K.; Morgan, K. J.; Zabik, M. E. (1983). Dietary lead intake of preschool children. Am. J. Public
Health 73(7): 789-794.
Barltrop, D.; Meek, F. (1979). Effect of particle size on lead absorption from the gut. Arch. Environ. Health
July/Aug. 1979.
Barltrop, D.; Khoo, H. E. (1975). The influence of nutritional factors on lead absorption. Postgrad
Med, J. 51: 795-800.
Barltrop, D.; Meek, F. (1975). Absorption of different lead compounds. Postgrad. Med. J. 51: 805-809.
6-1
-------
Barltrop, D.; Thornton, I.; Strehlow, C. D.; Webb, J. S. (1975). Absorption of lead from dust and soil.
Postgrad Med. J. 51: 801-804.
Barry, P. S. I. (1981). Concentrations of lead in the tissues of children. Br. J. Ind. Med 38: 61-71.
Barton, J. C.; Conrad, M. E.; Harrison, L.; Nuby, S. (1978). Effects of iron on the absorption and retention of
lead. J. Lab. Clin. Med. 92: 536-547.
Barton, J. C.; Conrad, M. E.; Harrison, L.; Nuby, S. (1980). Effects of vitamin D on the absorption and
retention of lead. Am. J. Physiol. 238; G124-G130.
Bentler, P. M., EQS Structural Equations Program Manual. BMDP Statistical Software, Los Angeles CA,
1989.
Binder, S.; Sokal, D.; Maughan, D. (1986). Estimating soil ingestion: The use of tracer elements in estimating
the amount of soil ingested by young children. Arch. Environ. Health 41: 341-345.
Blake, K. C. H.; Barbezat, G. O.; Mann, M (1983a). Effect of dietary constituents on the gastrointestinal
absorption of Pb in man. Environ. Res. 30: 182-187.
Blake, K. C. H.; Mann, M. (1983b). Effect of calcium and phosphorus on the gastrointestinal absorption of
203Pb in man. Environ. Res. 30:188-194.
Bollen, K. A. (1990). Structural equations with latent variables. New York: John Wiley and Sons.
Bornschein, R. L.; Clark, C. S.; Pan, U. W.; Succop, P. A. et al. (1990). Midvale Community Lead Study.
Department of Environ. Health, University Cincinnati Medical Center. July 1990.
Bornschein, R. L.; Clark, C. S.; Grote, J.; Peace, B.; Roda, S.; Succop, P. A. (1988). Soil lead—Blood lead
relationship in a former lead mining town. In: Environmental Geochemistry and Health, Monograph
Series 4, Lead in Soil: Issues and Guidelines. (Eds) B. E. Davies and B. G. Wixson. Science Review
Limited, Northwood, England, pp. 149-160.
Bornschein, R. L.; Succop, P. A.; Krafft, K. M. et al. (1986). Exterior surface dust lead, interior house dust
lead and childhood lead exposure in an urban environment. In: Trace Substances in Environmental
Health II—A Symposium. (Ed) D. D. Hemphill, University of Missouri, Columbia MO. pp. 322-332.
Bornschein, R. L.; Succop, P.; Dietrich, R. N.; Clark, C. S.; Que Hee, S.; Hammond, P. B. (1985). The
influence of social and environmental factors on dust lead, hand lead, and blood lead levels in young
children. Environ. Res. 38: 108-118.
Bruenger, F. W.; Stevens, W.; Stover, B. J. (1973). The association of Pb with constituents of erythrocytes.
Health Physics 25: 37-42.
Brunekreef, B. (1984). The relationship beyween air lead and blood lead in children: A critical review. Sci. Total
Environ. 38: 79-123.
Brunekreef, B.; Dook, N.; Clausing, P. (1987). Variability of exposure measurements in environmental
epidemiology. Amer. J. Epidem. 125: 892-898.
Buncher, C. R.; Succop, P. A.; Dietrich, K. N. (1991). Structural equation modeling in environmental risk
assessment. Environ. Health Persp. 90: 209-213.
6-2
-------
Bushnell, P. J.; DeLuca, H. F. (1983). The effects of lactose on the absorption and retention of dietary lead.
J. Nutr. 113: 365-378.
Calabrese, E. and Stanek, E. J. (1991). A guide to interpreting soil ingestion studies. II. Qualitative and
quantitive evidence of soil ingestion. Reg Toxicol Pharmacol 13: 278-292.
Calabrese, E.; Stanek, E. J.; Gilbert, C. E. (1991). Evidence of soil-pica behavior and quantification of soil
ingested. Human Experi. Toxicol. 10: 245-249.
Calabrese, E. Qualitative and quantitative evidence of soil ingestion. Presented at the Symposium on the
Bioavailability and Dietary Uptake of Lead. Sept. 24-27, 1990. Chapel Hill, NC.
Calabrese, E.; Barnes, R.; Stanek, E. J. et al. (1989). How much soil do young children ingest:
An epidemiologic study. Reg. Toxicol. Pharmacol. 10: 123-137.
Campbell, B. C.; Meredith, P. A.; Moore, M. R.; Watson, W. S. (1984). Kinetics of lead following intravenous
administration in man. Toxicol. Ltrs. 21: 231-235.
Centers for Disease Control and Prevention (1991). Preventing lead poisoning in young children. October, 1991.
Chamberlain, A. C. (1985). Prediction of response of blood lead to airborne and dietary lead from volunteer
experiments with lead isotopes. Proc. R. Soc. Lond. B. 224: 149-182.
Chamberlain, A. C.; Heard, M. J.; Little, P. et al. (1978). Investigations into lead from motor vehicles. Report
of work at Environmental and Medical Sciences Division, AERE, Harwell. HL78/4122 (C. 10).
Chaney, R. L.; Mielke, H. W.; Sterrett, S. B. (1988). Speciation, Mobility and Bioavailability of Soil Lead.
In: Environmental Geochemistry and Health Monograph Series 4. Lead in Soil: Issues and Guidelines.
(Eds) B. E. Davies and B. G. Wixson. Science Reviews Limited, Northwood, England, pp. 105-129.
Charney, E.; Kessler, B.; Farfel, M.; Jackson, D. (1983). Childhood lead poisoning: A controlled trial of the
effect of dust-control measures on blood lead levels. NE. J. Med. 309(18): 1089-1093.
Chisholm, J. J.; Mellits, E. D.; Quaskey, S. A. (1985). The relationship between the level of lead absorption in
children and the age, type, and condition of housing. Environ. Res. 38: 31-45.
Chisolm, J. J., Jr.; Harrison, H. E. (1956). Quantitative urinary coproporphyrin excretion and its relation to
edathamil calcium disodium administration in children with acute lead intoxication. J. Clin Invest.
35: 1131-1138.
Clark, C. S.; Bornschein, R. L.; Ryan, J. et al. (1988). The Cincinnati Soil-Lead Abatement Demonstration
Project. Presented at Lead in Soil Conference, Chapel Hill, NC, March 7-9, 1988.
Clark, C. S.; Bornschein, R. L.; Succop, P. A. et al. (1985). Condition and type of housing as an indicator of
potential environmental lead exposure and pediatric blood lead levels. Environ. Res. 38: 46-53.
Clarkson, T. W.; Kench, J. E. (1958). Uptake of Lead by Human Erythrocytes in vitro. Bioch 69: 432-439.
Clausing, P.; Brunekreef, B.; van Wijnen, J. H. (1987). A method for estimating soil ingestion by children.
Int. Arch. Occup. Environ. Health 59: 73-82.
Cohen, J.; Marcus, A.; Elias, R. (1990). Estimating childhood multi-media lead exposure: Expanded
exposure/uptake/biokinetic model. Presented at the 83rd Annual Meeting and Exhibition, Air and Water
Management Assoc. Pittsburgh, PA. June 24-29, 1990. Paper 90-12.2.
6-3
-------
Cohen, J.; Marcus, A. H.; Elias, R. W. (1991). Estimating childhood multi-media lead exposure: Expanded
exposure/uptaske/biokinetic model. Chemical Speciation and Bioavailability 3: 179-186.
Cohen, J.; Marcus, A.; Schwartz, J. (1988). Use of probabilistic risk assessment in development of U.S. EPA
regulations on lead. Presented at Air Pollution Control Assoc meeting, New York, June 1988.
Cohen, N.; Kneip, T. J.; Rulon, V.; Goldstein, D. H. (1974). Biochemical and lexicological response of infant
baboons to lead driers in paint. Environ. Health Persp. May 1974: 161-173.
Cohen, N.; Kneip, T. J.; Rulon, V.; Goldstein, D. H. (1974). Biochemical and toxicological response of infant
baboons to lead driers in paint. Environ. Health Perspectives: 73: 161-173.
Cosgrove, E.; Brown, M. J.; Madigan, P. et al. (1989). Childhood lead poisoning: Case study traces source to
drinking water. J. Environ. Health 52: 346-349.
Cotter-Howells, J.; Thornton, I. (1991) Sources and pathways of environmental lead to children in a Derbyshire
mining village. Environ. Geochem. Health 13: 127-135.
Crawford-Brown, D. J. (1983). An age-dependent model for the kinetics of uptake and removal of radionuclides
from the GI tract. Health Physics 44(6): 609-622.
Davies, B. E.; Elwood, P. C.; Gallacher, J.; Ginnever, R. C. (1985). The relationships between heavy metals in
garden soils and house dusts in an old lead mining area of North Wales, Great Britain. Environ. Pollution
(Series B) 9: 255-266.
Davies, D. J. A.; Thornton, I.; Watt, J. M. et al. (1990). Lead intake and blood lead in two-year-old UK urban
children. Sci. Total Environ. 90: 13-29.
Davis, B. K.; Jamall, I. S. (1991). Health risks in Eastern Europe associated with environmental pollution.
Toxicologist 11(1): 191.
Davis, S.; Waller, P.; Buschbom, Ray; Ballou, J.; White, P. (1990). Quantitative estimates of soil ingestion in
normal children between the ages of 2 and 7 years: Population-based estimates using Aluminum, Silicon,
and Titanium as soil tracer elements. Arch Environ. Health 45(2) 112-122.
Dolcourt, J. L.; Finch, C.; Coleman, G. D. et al. (1981). Hazard of lead exposure in the home from recycled
automobile storage batteries. Pediatrics 68(2): 225-228.
Duggan, M. J.; Inskip, M. J.; Rundle, S. A.; Moorcroft, J. S. (1985). Lead in playground dust and on the hands
of schoolchildren. Sci. Total Environ. 44: 65-79.
Duggan, M. J. (1983). The uptake and excretion of lead by young children. Arch. Environ. Health 38: 246-247.
Ershow, A.; and Cantor K. (1989). Total water and tap water intake in the United States: Population-based
estimates of quantities and sources. Report prepared under National Cancer Institute Order
#263-MD-810264. Life Sciences Research Office, Federation of American Societies for Experimental
Biology. Bethesda, MD.
Farfel, M. R.; Chisolm, J. J. (1991). An evaluation of experimental practices for abatement of residential lead-
based paint: Report on a pilot project. Environ. Res. 55: 199-212.
Farfel, M. R.; Chisolm, J. J. (1990). Health and environmental outcomes of traditional and modified practices
for abatement of residential lead-based paint. AJPH 80(10): 1240.
6-4
-------
Forbes, G. B.; Reina, J. C. (1972). Effect of age on gastrointestinal absorption (Fe,Sr,Pb) in the rat. J. Nutr.
102: 647-652.
Freeman, G. B.; Johnson, S. C.; Liao, P.; Feder, P. I.; Killinger, J. M.; Chaney, R. L.; and Bergstrom, P. D.;
(1991). Effect of soil dose on bioavailability of lead from mining waste soil in rats. Chemical Speciation
and Bioavailability3: 121-128.
Fullmer, C. S.; Edelstein, S.; Wasserman, R. H. (1985). Lead-binding properties of intestinal calcium binding
proteins. J. biol. Chem. 260: 6816-6819.
Fullmer, C. S,; Rosen, J. F. (1990). Effect of dietary calcium and lead status on intestinal calcium absorption.
Env. Res. 15: 91-99.
Garber, B. T.; Wei, E. (1974). Influence of dietary factors on the gastrointestinal absorption of lead. Toxicol.
Appl. Pharmacol. 27: 685-691.
Gardels, M. C.; Sorg, T. J. (1989). A laboratory study of the leaching of lead from water faucets. J. AWWA
July 1989: 101-113.
Gartrell, M. J.; Craun, J. C.; Podrebarac, D. S.; Gunderson, E. L. (1986). Pesticides, selected elements, and
other chemicals in infant and toddler total diet samples. J. Assoc. Off Anal Chem. 69(1): 123-145.
Gibaldi, M. (1982). Pharmacokinetics. Marcel Dekker, Inc. New York, New York.
Harley, N. H.; Kneip, T. H. (1985). An integrated metabolic model for lead in humans of all ages. Final report
to U.S. EPA, Contract B44899, New York University Department Environmental Medicine.
Heard, M. J.; Chamberlain, A. C. (1982). Effect of minerals and food on uptake of lead from the gastrointestinal
tract in humans. Hum. Toxic. 1: 411-415.
Jacobs Engineering Group Inc., Environmental Systems Division (1989). Human Health Risk Assessment
Protocol for the Populated Areas of the Bunker Hill Superfund Site. Report prepared for U.S. EPA—
Region X. TES IV Contract No. 68-01-7351. Moscow, Idaho.
James, H. M.; Hilburn, M. E.; Blair, J. A. (1985). Effects of meal and meal times on uptake of lead from the
gastrointestinal tract in humans. Human Toxicol. 4: 401-407.
Karalekas, P. C.; Ryan, C. R.; Taylor, F. B. (1983). Control of lead, copper, and iron pipe corrosion in Boston.
J. AWWA—Res. Tech., Feb 1983: 92-95.
Karam, H. S.; Beck, B. D. (1990). Current issues in determining acceptable lead concentrations in soils.
Comments on Toxicology 3(6): 509-529.
Kootsey, J. M. (1989). Introduction to Computer Simulation (SCoP), National Biomedical Simulation Resource,
Duke University Medical Center, Durham, NC.
Kostial, K.; Kello, D.; Jugo, S.; Rabar, I.; Maljkovic, T. (1978). Influence of age on metal metabolism and
toxicity. Environ. Heakth Oersoect. 25: 81-86.
Kostial, K.; Simonivic, J.; Pisonic M. (1971). Lead absorption from the intestine in newborn rats. Nature
(London) 233: 564.
Lacey, R. F.; Moore, M. R.; Richards, W. N. (1985). Lead in water, infant diet and blood: The Glasgow
duplicate diet study. Sci. Total Environ. 41: 235-257.
6-5
-------
LaVelle, J. M.; Poppenga, R. H.; Thacker, B. J.; Giesy, J. P.; Weis, C. P.; OthoudtR.; Vandervoort C.
(1991). Bioavailability of lead in mining wastes: an oral intubation study in young swine. Chemical
Speciation and Bioavailability 3: 105-111.
Laxen, D. P. H.; Raab, G. M; Fulton, M. (1987). Children's blood lead and exposure to lead in household dust
and wate—a basis for an environmental standard for lead in dust. Sci. Tot. Environ. 66: 235-244.
Leggett, R. W. (1994). A model of the biokinetics of Pb in the human circulation. Submitted for publication.
Maes, E. F.; Swygert, L. A.; Paschal, D. C.; Anderson, B. S. (1991). The contribution of lead in drinking
water to levels of blood lead. I. A cross-sectional study. (Submitted for publication).
Mahaffey-Six, K.; Goyer, R. A. (1972). The influence of iron deficiency on tissue content and toxicity of
ingested lead in the rat. J. Lab. Clin. Med. 79: 128-136.
Mallon, R. P. (1983). A Metabolic Model of Lead Kinetics Based Upon Measured Organ Burdens During
Chronic Exposure Experiments with Infant and Juvenile Baboons. Ph.D. Thesis, Department of Biology,
New York University Medical Center, New York, NY.
Mallon, R. P. (1983). A metabolic model of lead kinetics based upon measured organ burdens during chronic
exposure experiments with infant and juvenile baboons. Doctoral Dissertation, New York University.
Marcus, A. H. (1985a). Multicompartment kinetic model for lead: III. Lead in blood plasma and erythrocytes.
Environ. Res. 36: 473-489.
Marcus, A. H. (1985b). Testing Alternative Nonlinear Kinetic Models in Compartmental Analysis.
In: Mathematics and Computers in Biomedical Applications. J. Eisenfeld and C. DeLisi (Eds). Elsevier
Science Publishers BV, North-Holland (IMACS).
Marcus, A. H. (1988a). Geometric standard deviation of blood lead for children in the vicinity of point sources.
Report from Battelle Columbus Division, Washington Operations to U.S. EPA. Contract No. 68-02-4246.
Marcus, A. H. (1988b) Prediction of blood lead levels in East Helena children: Comparison of aggregate air
lead, disaggregate, and uptake/biokinetic models. Battelle/Columbus Division, Washington Operations,
Washington, D.C.
Marcus, A. H. (1989a). Uptake of lead from formula and food by infants: Reanalysis of the Ryu et al. data.
Report from Battelle Columbus Division, Arlington Operations to U.S. EPA Office of Toxic Substances.
Contract No. 68-02-4294.
Marcus, A. H. (1989b). Contributions to a risk assessment for lead in drinking water. Report from Battelle
Arlington to U.S. EPA Office of Toxic Substances. Contract 68-D8-0115.
Marcus, A. H. (1989c). Statistical reanalyses of relationship of blood lead in Edinburgh Children to lead in dust
and water. Report from Battelle Arlington Office to U.S. EPA Office of Toxic Substances. Contract
No. 68-D8-0115.
Marcus, A. H. (1991a). Estimation of childhood lead burdens from newly installed faucets. Report from Battelle
Columbus to U.S. EPA Office of Toxic Substances. Report No. 68-D8-0115.
Marcus, A. H. (1991b). Inter-site comparisons of environmental lead uptake. Presented at Symposium on the
Bioavailability and Dietary Uptake of Lead, ECAO/USEPA. Chapel Hill, NC, September 24-27, 1990.
Report from Battelle Columbus Division, Arlington Office, to U.S. EPA Office of Toxic Substances.
Contract No. 68-02-4246.
6-6
-------
Marcus, A. H. (1991c). Relationship between soil lead, dust lead, and blood lead over time: A reanalysis of the
Boston lead data. Report from Battelle Columbus Division, Arlington Office, to U.S. EPA Office of
Toxic Substances. Contract No. 68-02-4246.
Marcus, A. H. (1992). Use of site-specific data in models for lead risk assessment and risk management.
In: An Update of Exposure and Effects of Lead, B. Beck (Ed), Fund. Appl. Toxicol. 18: 10-16.
Marcus, A. H.; Bernholc, A. (1990). Variability of household water lead levels in American cities. Report from
Battelle Arlington Operations to U.S. EPA. Contract No. 68-D8-0115.
Marcus, A. H.; Cohen, J. (1988). Modeling the Blood Lead—Soil Lead Relationship. In: Environmental
Geochemistry and Health Monograph Series 4. Lead in Soil: Issues and Guidelines. (Eds) B. E. Davies
and B. G. Wixson. Science Reviews Limited, Northwood, England, pp. 161-174.
Marcus, A. H.; Elias, R. W. (1994). Estimating the contribution of lead-based paint to soil lead, dust lead, and
childhood blood lead, Lead in Paint, Soil, and Dust: Health Risks, Exposure Studies, Control Measures,
Measurement Methods, and Quality Assurance, ASTM STP 1226, Michael E. Beard and S. D. Allen Iske,
Eds., American Society for Testing and Materials, Philadelphia, 1994.
Marcus, A. H.; Holtzman, A. P. (1989). Relationships between infant blood lead concentrations and lead in
water or liquid diet. Report from Battelle Columbus Division to U.S. EPA Office of Drinking Water.
Contract No. 68-02-4246.
Matte, T. D.; Figueroa, J. P.; Ostrowski, S.; Burr, G.; Jackson-Hunt, L.; Baker, E. L. (1991a). Lead exposure
from conventional and cottage lead smelting in Jamaica. Arch. Environ. Contam. Toxicol. 21: 65-71.
Matte, T. D.; Figueroa, J. P.; Ostrowski, S.; Burr, G.; Jackson-Hunt, L.; Baker, E. L. (1991b). Relationship
between soil lead levels and blood levels among childreb living near a lead smelter in Jamaica. Chemical
Speciation and Bioavailability 3: 173-179.
Menton, R. G.; Burgoon, D. A.; Marcus, A. H. (1994). Pathways of lead contamination for the Brigham and
Women's Hospital Longitudinal Lead Study, Lead in Paint, Soil and Dust: Health Risks, Exposure
Studies, Control Measures, Measurement Methods, and Quality Assurance, ASTM STP 1226, Michael E.
Beard and S.D. Allen Iske, Eds., American Society for Testing and Materials, Philadelphia, 1994.
Momcilovic, B.; Kostial, K. (1974). Kinetics of lead retention and distribution in suckling rats. Arch. Environ.
Health 33: 115-117.
Mordenti, J. (1986). Man versus beast: Pharmacokinetic scaling in mammals. J. Pharma. Sci.
75(11): 1028-1040.
Morton, A. P.; Partridge, S.; Blair, J. A. (1985). The intestinal uptake of lead. Chem. Brit. 21: 923-927.
Munro, I. C.; Willes, R. F.; Truelove, J. F. (1975). Absorption and tissue distribution of inorganic lead in the
developing infant monkey (Macaca irus). Toxicol. Appl. Pharmacol. 33: 128-129.
Mushak, P. (1991). Gastro-intestinal absorption of lead in children and adults: overview of biological and
biophysico-chemical aspects. Chemical Speciation and Bioavailability 3: 87-104.
Mushak, P. (1992). Defining lead as the premier environmental health issue for children in America. Environ.
Res. 59: 281-309.
Mykkanen, H. M.; Wasserman, R. H. (1981). Gastrointestinal absorption of lead ( Pb) in chicks: influence of
lead, calcium, and age. Jutr. Ill: 1757-1765.
6-7
-------
O'Flaherty, E. J. (1986). The rate of decline of blood lead in lead industry workers during medical removal:
The effect of job tenure. Fund Appl. Toxicol. 6: 372-3.
O'Flaherty, E. J. (1993). Physiologically-based models for bone-seeking elements. IV. Kinetics of lead
disposition in humans. Toxicol. Appld. Pharmacol. 118: 16-29.80.
Pennington, J. A. T. (1983). Revision of the total diet study food list and diets. J. Am. Dietetic Assoc.
82(2): 166-173.
Poppenga, R. H.; Thacker, B. J.; Giesy, J. P. (1990). Bioavailability of lead in mining wastess: An oral
intubation study in swine. Report from Michigan State University to U.S. EPA Region VIII.
Portier, C. J.; Kaplan, N. L. (1989). Variability of safe dose estimates when using complicated models of the
carcinogenic process. Fund Appl. Toxicol. 13(3): 533-544.
Proceedings of the Symposium on the Bioavailability and Dietary Exposure of Lead, (1991) September 1990,
Chapel Hill, North Carolina, Chemical Speciation and Bioavailability, Volume 3, No 3-4. ISBN
0-946682-10-0.
Raab, G. M.; Laxen, D. P. H.; Fulton, M. (1987). Lead from dust and water as exposure sources for children.
Environ. Geochem. Health 9(3-4): 80-85.
Rabinowitz, M. B. (1987). Stable isotope mass spectrometry in childhood lead poisoning. Biological Trace
Element Research. 12: 223-229.
Rabinowitz, M. B. (1987). Stable isotope mass spectrometry in childhood lead poisoning. Biological Trace
Element Research. 12: 223-229.
Rabinowitz, M. B.; Kopple, J. D.; Wetherill, G. W. (1976). Kinetic analysis of lead metabolism in healthy
humans. J. Clin. Invest. 58: 260-270.
Rabinowitz, M. B.; Kopple, J. D.; Wetherill, G. W. (1980). Effect of food intake and fasting on gastrointestinal
lead absorption in humans. Am. J. Clin. Nutr. 33: 1784-1788.
Rabinowitz, M.; Leviton, A.; Bellinger, D. (1984). Home refinishing, lead paint, and infant blood lead levels.
Amer. J. Public Health. 75: 403-404.
Rabinowitz, M.; Leviton, A.; Needleman, H.; Bellinger, D.; Waternaux, C. (1985). Environmental correlates of
infant blood lead levels in Boston. Environ. Res. 38: 96-107.
Revicki, D. A.; Elixhauser, A.; Hersey, J.; Marcus, A. (1991). The cost-effectiveness of alternative lead
abatement strategies. Report from Battelle to CDC/CEHIC.
Ryu, J. E.; Ziegler, E. E.; Nelson, S. E.; Fomon, S. J. (1983). Dietary intake of lead and blood lead
concentration in early infancy. Am. J. Dis. Child. 137: 886-891.
Sawyer, M.; Kearney, T.; Spector, S. et al. (1985). Lead intoxication in children—Interdepartmental Conference.
University of California, San Diego (Specialty Conference). West J. Med. 143: 357-364.
Schilling, R. J.; Bain, R. P. (1988). Prediction of children's blood lead levels on the basis of household-specific
soil lead levels. Am. J. Epidemiology 128(1): 197-205.
Schock, M. R.; Neff, C. H. (1988). Trace metal contamination from brass fittings. J. AWWA 80(11): 47-56.
6-8
-------
Sherlock, J. C.; Quinn, M. J. (1986). Relationship between blood lead concentrations and dietary lead intake in
infants: The Glascow Duplicate Diet Study 1979-1980. Food Additives and Contaminants 3: 167-176.
Simmons, C. M. (1989). City of Raleigh's experience in the pilot public education program for lead in drinking
water. Raleigh, North Carolina.
Stark, A. D.; Quah, R. F.; Meigs, J. W.; DeLouise, E. R. (1982). The relationship of environmental lead to
blood-lead levels in children. Environ. Res. 27: 372-383.
Steele. M. J.; Beck, B. D.; Murphy, B. L. (1990). Assessing the contribution from lead in mining wastes to
blood lead. Reg. Toxicol. Pharm., 11: 158-190.
Succop, P. A.; O'Flaherty; Bornschein, R. L. et al. (1987). A kinetic model for estimating changes in the
concentration of lead in the blood of young children. In: International Conference: Heavy Metals in the
Environment, (Eds) Lindberg, S. E.; Hutchinson, T. C.; New Orleans, September 1987. (EP Consultants
Ltd., Edinburgh, pp. 289-291).
Thornton, E.; Daview, D. J. A.; Watt, J. M.; Quinn, M. J. Lead exposure in young children from dust and soil
in the United Kingdom. Environ. Health Persp. 89: 55-60.
Trotter, R. T. (1990). The cultural parameters of lead poisoning: A medical anthropologist's view of intervention
in environmental lead exposure. Environ. Health Persp. 89: 79-84.
Tukey, J. W. (1977) Exploratory Data Analysis. Addison-Wesley Publishing Co., Reading MA.
U.S. Department of Housing and Urban Development (1990). Comprehensive and workable plan for the
abatement of lead-based paint in privately owned housing. Office of Policy Development and Urban
Research. PB91-193953.
U.S. Environmental Protection Agency (1986). Air Quality Criteria for Lead Volume I-IV. Environmental
Criteria and Assessment Office, Office of Research and Development, RTP, NC. EPA 600/8-83-028 a-d.
U.S. Environmental Protection Agency (1989a). Review of the National Ambient Air Quality Standards for Lead:
Exposure Analysis Methodology and Validation. U.S. EPA Office of Air Quality Planning and Standards,
RTP, NC. EPA-450/2-89/011.
U.S. Environmental Protection Agency (1989b). Guidance Concerning Soil Lead Cleanup Levels at Superfund
Sites. Directive #9355.4-02. Office of Solid Waste and Emergency Response. September 1989.
U.S. Environmental Protection Agency (1989c). Exposure Factors Handbook. U.S. EPA Office of Health and
Environmental Assessment, Washington, DC. EPA/600/8-89/043.
U.S. Environmental Protection Agency (1990a). Technical Support Document on Lead. Document prepared for
Office of Solid Waste and Emergency Response by U.S. EPA/ECAO Cincinnati, OH. ECAO-CIN-757.
U.S. Environmental Protection Agency (1990b). Report of the Clean Air Scientific Advisory Committee on its
Review of the OAQPS Lead Staff Paper and the ECAO Air Quality Criteria Document Supplement.
EPA-SAB-CASAC-90-002. January 1990.
U.S. Environmental Protection Agency (1990c). Risk Assessment Data Evaluation Report (RADER), by
TerraGraphics Environmental Engineering under Contract #68-W9-0008, WA #C10012, October 1990.
U.S. Environmental Protection Agency (1991a). Three City Urban Soil-Lead Demonstration Project—Midterm
Project Update. U.S. EPA ECAO/HEARD. 21S2001.
6-9
-------
U.S. Environmental Protection Agency (1991b). National Air Quality and Emissions Trends Report 1989. Report
prepared by U.S. EPA OAQPS, RTP, NC. EPA-450/4-91-003.
U.S. Environmental Protection Agency (1991c). National Primary Drinking Water Regulation for Lead
(NPDWR). Federal Register 56-26460. June 7, 1991.
U.S. Environmental Protection Agency (199Id). Sampling Manual for Site-Specific Data Collection for the Lead
Integrated Biokinetic Uptake Model. Report to U.S. EPA from Cadmus Assoc. (in preparation).
U.S. Environmental Protection Agency (1992a). An SAB Report: Review of the Uptake Biokinetic Model for
Lead. EPA-SAB-IAQC-92016. March 1992,
U.S. Environmental Protection Agency (1992b). Guidance for Data Useability in Risk Assessment (Part A).
Directive #9285.7-09A. Office of Emergency and Remedial Response. April 1992.
U.S. Environmental Protection Agency (1993). Urban Soil Lead Abatement Demonstration Project, Volume I:
Integrated Report, Office of Research and Development. EPA/600/AP-93/001a, July 1993.
van Wijnen, J. H.; Clausing, P.; Brunekreef, B. (1990). Estimated soil ingestion by children. Environ. Res.
51: 147-162.
von Lindern, I. (1991). Projected Blood Lead Distribution for Proposed Soil Cleanup Guidelines—Bunker Hill
Site Residential Soils Feasibility Study (Technical Memorandum from TerraGraphics Environmental
Engineering to Sally Martyn, U.S. EPA Region X).
Walter, S. D.; Yankel, A. J.; von Lindern, I. H. (1980). Age-specific risk factors for lead absorption in
children. Arch. Environ. Health 35:53-58.
Weis, C. P.; Poppenga, R. H.; Thacker, B. J. (1994). Design of pharmocokinetic and bioavailability studies of
lead in an immature swine model. In: Lead in paint, soil, and dust: Health risks, exposure studies,
control measures, measurement methods, and quality assurance. Beard, M.E.; Iske, S.A., Eds. ASTM
1226. Philadelphia: American Society for Testing and Materials.
Weis, C. P.; LaVelle, J. M. (1991). Characteristics to consider when choosing an an animal model for the study
of lead bioavailability. In: Proceedings of the Symposium on the Bioavailability and Dietary Exposure of
Lead, Chemical Speciation and Bioavailability 3: 113-119.
Weitzman, M.; Aschengrau, A.; Bellinger, D.; Jones, R.; Hamlin, J. S.; Beiser, A. (L993). Lead-contaminated
soil abatement and urban children's blood lead levels. J. Amer. Med. Assoc. 269: 1647-1654.
West, J. R.; Smith, H. W.; Chasis, H. (1948). Glomerular filtration rate, effective renal blood flow, and
maximal tubular excretory capacity in infancy. J. Pediatr. 32: 10-18.
Whitfield, R. G.; Cohen, J.; Marcus, A. H.; Zaragoza, L. J. (1989). Using data and judgment to estimate blood-
lead distributions among children from soil-lead levels and other environmental factors. Unpublished
report for Argonne National Laboratory, Energy and Environmental Systems Division.
Woodward-Clyde (1991). Demographics workplan California Gulch Study Area Leadviile, Colorado. Prepared
for Asarco by WoodwardClyde. Project No. 22443E-33441. July 1991.
Yaffe, Y. et al. (1983). Identification of lead sources in California children using the stable isotope ratio
technique. Arch. Environ. Health 38(4): 237-245.
6-10
-------
Yankel, A. J.; von Lindera, I. H.; Walter, S. D. (1977). The Silver Valley lead study: the relationship of
childhood lead poisoning and environmental exposure. J. Air Pollut. Control Assoc. 27: 763-767.
Ziegler, E. E.; Edwards, B. B.; Jensen, R. L.; Mahaffey, K. R.; Fomon, S. J. (1978). Absorption and retention
of lead by infants. Pediatr. Res. 12: 29-34.
6-11
-------
APPENDIX A: HOW TO CALCULATE THE
GEOMETRIC STANDARD DEVIATION FROM
BLOOD LEAD DATA, IF YOU MUST
A.I A DIRECT METHOD FOR CALCULATING THE GEOMETRIC
STANDARD DEVIATION
One of the simplest approaches to calculating a GSD from a sample of blood lead and
environmental lead data is based on the idea that children with similar environmental lead
exposures will have similar geometric mean blood lead levels. For children of a given age
with similar soil lead (denoted PbS), dust lead (denoted PbD), and other lead exposures, we
can reasonably characterize the variability in blood lead level (denoted GSD) calculated with
respect to the actual geometric mean blood lead level of this group of children (denoted
GMB) without modelling blood lead levels. The procedure shown here is the simplest
procedure we have found, but even with this procedure, the user must be prepared to do a
great deal of statistical calculation. We will illustrate how an empirical GSD may be
calculated from data after we describe the procedure:
STEP 1: Divide the data set into subgroups, where each group has children of a
given age, with soil lead levels in a given interval, dust lead levels in a
given interval, and with distinct levels of other important variables.
Each such group corresponds to a "box" or cell of soil and dust lead
levels, and levels of other variables if used.
STEP 2: From each individual blood lead (denoted PbB) in each cell, calculate
In(PbB), where hi denotes the natural logarithm.
STEP 3: Within each cell, calculate the mean and the standard deviation of the
In(PbB) values. Then, for that cell,
GMB = exp(mean of In(PbB) values within the cell)
GSD = exp(standard deviation of In(PbB) values within the cell)
where exp denotes the process of calculating the exponential
function of the indicated quantity. Exponential and natural
logarithm functions are available in most statistical packages for
microcomputers and on most scientific calculators.
STEP 4: Calculate the inter-individual GSD for this neighborhood by finding the
median or middle value of the GSD values in the sample. The median
is found by ordering the GSD values from all cells from smallest to
A-l
-------
largest. If the number of GSD values is odd, the median is the middle
value; if sample size is an even number, find the average of the two
middle values. Since the number of observations in each box or cell is
different, each GSD should be counted a number of times according to
the number of degrees of freedom (cell count minus one) for that GSD.
STEP 5: Users with some statistical background may wish to examine the within-
cell GSD's for patterns based on the data, such as by plotting GSD
against GMB or against the within-cell value of age, PbS, PbD, or other
stratifying variables.
STEP 6: Users with more statistical background may wish to use other
approaches to calculating a "typical" GSD, such as by calculating a
mean or pooled variance of the within-cell variances of In(PbB).
We would caution such users that the data should be carefully evaluated
for outliers, either in raw PbB values or in the calculated GSD. One
convenient approach for visual detection of outliers is a normal
probability of within-cell variances after a variance-stabilizing
transformation such as the cube root of the within-cell variance of
In(PbB).
EXAMPLE: In a sample of 166 children from the Midvale, Utah study of 1989
(Bornschein et al., 1990), we found that the estimation of blood lead
levels could be considerably improved by determining whether or
not the children lived in houses in which paint had recently been
removed. There is substantial evidence that inadequately controlled
lead paint abatement may increase blood levels in resident young
children by 2 to 4 ug/dL on average in the first 6 to 12 months after
paint removal (Rabinowitz et al., 1984; Marcus et al.., 1991;
Menton et al., 1993). Interviews with the family provided such
information for 162 of the 166 children, which was used as an
additional stratifying variable.
The worksheet for determining subgroups are shown in Table A-l. Each table gives
the blood leads of children of a given age, divided by whether or not there was recent paint
removal in the residence. Within each table, the children are divided according as the PbS
and PbD values at their residence. Each cell in the table corresponds to intervals of 250 ppm
of PbS and 250 ppm of PbD. The soil lead levels were averages of non-missing values of
perimeter, bare area, play area, and garden soils. It should be noted that data for most of
the cells are not available with such detailed sub-division of the data set, and that at higher
soil and dust lead concentrations, there is usually only one observation per cell. There was
only one case in which two children from the same family had the same age, in years, and
analyses without this duplication would produce very similar results. Otherwise, all blood
lead levels (denoted PbB) within each cell come from different families. This is believed to
A-2
-------
TABLE A-l. CELLS OF BLOOD LEAD LEVELS IN 165 MIDVALE
CHILDREN, BY PAINT REMOVAL STATUS, AGE, AND INTERVALS
OF 250 Atg/g IN SOIL AND DUST LEAD1
Paint
Removal
.
.
.
,
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Age
0
2
2
3
4
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
2
Soil
Pb
375
.
375
625
125
125
125
125
375
375
375
375
1125
.
125
375
375
375
625
625
875
875
1125
1375
1625
.
Dust
Pb
375
625
125
625
125
125
375
125
375
1125
1375
875
625
375
375
625
875
375
625
625
1125
875
1125
625
375
Blood Lead (jug/dL)
Smallest -»~>Largest
5.5
6.
4.
3.
6.5
3.
1.
0.5
3.
5.5
3.
3.5
13.5
5.5
2.5
4.
4.5
3.5
3.
3.5
3.
6.
1.
6.
3.
4.
.
.
6.
4.5
7.
.
3.
5.5
.
.
7.
6.
.
.
10.5
,
6.
.
.
.
.
.
.
.
.
.
7.
.
.
8.
6.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
10.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
A-3
-------
TABLE A-l (cont'd). CELLS OF BLOOD LEAD LEVELS IN 165 MIDVALE
CHILDREN, BY PAINT REMOVAL STATUS, AGE, AND INTERVALS
OF 250
IN SOIL AND DUST LEAD
1
Paint
Removal
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Age
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
4
4
4
4
4
Soil
Pb
.
125
125
125
375
375
625
625
875
1125
1125
1125
.
125
125
375
375
375
1125
1375
125
125
375
375
Dust
Pb
625
.
125
625
375
1125
375
625
1125
.
625
875
625
125
375
125
375
1375
625
1125
375
125
375
.
625
Blood Lead (jig/dL)
Smallest -^~->Largest
7.
5.
2.5
6.
1.5
4.5
14.5
4.
5.5
13.
9.5
19.
5.
2.5
2.
6.5
3.
13.
16.5
5.
2.
4.
5.5
2.
1.5
.
.
5.5
.
.
.
7.
.
.
,
,
.
.
7.5
.
4.
.
.
.
.
7.5
6.
.
7.
.
.
5.5
.
.
.
.
11.5
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
8.
.
.
.
.
,
.
.
,
,
,
,
.
.
.
.
.
.
.
.
.
.
.
12.
.
.
,
,
,
.
.
.
.
.
,
.
.
.
.
.
.
.
.
.
.
,
,
,
.
.
.
.
.
.
A-4
-------
TABLE A-l (cont'd). CELLS OF BLOOD LEAD LEVELS IN 165 MTOVALE
CHILDREN, BY PAINT REMOVAL STATUS, AGE, AND INTERVALS
OF 250 /*g/g IN SOIL AND DUST LEAD1
Paint
Removal
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Age
4
4
4
4
5
5
5
5
5
5
5
0
0
0
0
1
1
1
1
1
1
1
1
1
1
Soil
Pb
625
875
1125
2125
125
125
375
625
625
625
1125
125
125
375
375
.
125
125
375
375
375
625
625
875
1125
Dust
Pb
375
625
2375
875
125
375
375
375
625
1375
875
125
375
375
875
375
125
375
375
625
875
125
375
625
1375
Blood Lead (jug/dL)
Smallest -^Largest
5.
7.5
5.
8.
2.
2.5
4.
5.
4.
4.5
13.5
3.
0.5
8.5
5.
8.
5.5
5.5
3.5
5.5
16.5
2.5
9.
6.
5.5
.
.
.
4.5
3.5
6.
.
.
.
,
16.5
1.5
.
.
22.5
.
.
.
.
.
.
9.
.
.
.
8.5
10.
.
,
,
.
.
3.5
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5..
.
.
.
.
.
.
•
.
.
.
.
,
.
.
,
.
.
6.
.
.
.
.
,
.
.
.
.
.
.
.
,
.
.
.
,
.
.
.
.
.
.
A-5
-------
TABLE A-l (cont'd). CELLS OF BLOOD LEAD LEVELS IN 165 MIDVALE
CHILDREN, BY PAINT REMOVAL STATUS, AGE, AND INTERVALS
OF 250 /tg/g IN SOIL AND DUST LEAD1
Paint
Removal
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Age
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
Soil
Pb
.
125
125
375
375
375
625
1875
,
.
125
375
375
625
875
1625
1875
.
.
125
125
375
875
875
1625
Dust
Pb
125
375
.
375
625
625
625
.
625
375
625
875
875
3625
1375
625
375
625
125
375
625
1125
1375
Blood Lead (^g/dL)
Smallest -»-»Largest
8.5
3.
2.5
6.
3.
8.5
6.5
10.5
8.5
4.
4.5
5.5
4.
2.
2.
15.5
7.5
3.5
3.5
4.5
2.
7.
9.
7.5
9.5
,
4.
3.5
.
10.
.
.
.
5.5
.
.
.
.
.
18.
.
5.
3.5
.
.
.
.
.
4.5
5.
.
-
.
,
.
,
.
.
8.
.
.
.
.
.
.
5.
4.
.
.
.
.
5.5
5.
.
•
.
.
.
.
.
.
,
.
.
,
.
,
5.
8.
.
.
,
.
.
.
5.5
.
-
.
.
.
.
.
.
.
.
,
.
.
.
5.
.
.
.
.
.
.
19.5
.
-
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
,
.
„
.
A-6
-------
TABLE A-l (cont'd). CELLS OF BLOOD LEAD LEVELS IN 165 MTOVALE
CHILDREN, BY PAINT REMOVAL STATUS, AGE, AND INTERVALS
OF 250 iig/g IN SOIL AND DUST LEAD1
Paint
Removal
1
1
1
1
1
1
1
Age
4
5
5
5
5
5
5
Soil
Pb
3125
.
125
125
375
625
1875
Dust
Pb
1625
.
125
375
375
375
625
Blood Lead (jug/dL)
Smallest -»-»Largest
13.
4.5
4.
4.
1.5
6.5
5.5
.
.
5.
7.5
.
.
.
.
•
.
.
.
.
.
.
.
.
.
.
-
.
.
.
,
.
.
•
An isolated decimal point denotes a missing value.
give a much more valid estimate of variability than within-family GSD's for children of
different ages, but similar genetic and non-lead environmental factors and similar family
behavior patterns.
The statistics for GMB and GSD were calculated as described in Step 3, for each cell
where enough data were available (at least 2 PbB values in order to calculate GSD). The
results are shown in Table A-2. The PbS and PbD values are the cell midpoints, and
provide convenient plot points. Some of the GSD values are very high, as for the cell whose
two values are PbB = 1.5 and 10 /xg/dL.
The distribution of GSD values for all cells is shown in Table A-3, in the form of
a "stem-and-leaf" plot (Tukey, 1977). No weighting scheme has been applied. Many users
would prefer a weighted GSD where the number of observations in each cell is taken into
account. This can be done by counting each GSD estimate as representing the number of
degrees of freedom (denoted DF) in the GSD estimate. In this application, DF = N — 1,
where N is the number of PbB values in the cell. A DF-weighted stem-and-leaf plot is
shown in Table A-4. In the unweighted case, the median GSD = 1.694 may be taken as a
representative value for this community. In the weighted DF case, a somewhat larger
median GSD = 1.768 may be used.
A-7
-------
TABLE A-2. GEOMETRIC MEAN AND GEOMETRIC STANDARD DEVIATION
OF BLOOD LEADS IN CELLS OR GROUPS, BY PAINT REMOVAL STATUS,
AGE, AND INTERVALS OF 250 pglg IN SOIL AND DUST LEAD1
Paint
Removal
.
.
.
.
.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Age
(Years)
0
2
2
3
4
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
Soil Lead
G*g/g)
375
375
625
125
125
125
125
375
375
375
375
1125
.
125
375
375
375
625
625
875
875
1125
1375
Dust
Lead
G*g/g)
375
625
125
625
125
.
125
375
125
375
1125
1375
875
625
375
375
625
875
375
625
625
1125
875
1125
N
1
1
1
1
1
2
1
1
2
1
2
1
1
1
2
3
1
1
4
3
1
1
2
1
Geometric
Mean Blood
Lead
(/xg/dL)
5.5
6.
4.
3.
6.5
4.243
1.
0.5
3.674
5.5
4.583
3.5
13.5
5.5
2.739
5.360
4.5
3.5
6.402
5.013
3.
6.
3.240
6.
GSD
.
.
.
.
.
1.633
.
,
1.332
,
1.821
.
.
.
1.138
1.324
.
1.693
1.365
.
5.273
A-8
-------
TABLE A-2 (cont'd). GEOMETRIC MEAN AND GEOMETRIC STANDARD
DEVIATION OF BLOOD LEADS IN CELLS OR GROUPS, BY PAINT
REMOVAL STATUS, AGE, AND INTERVALS OF 250 jig/g IN
SOIL AND DUST LEAD1
Paint
Removal
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Age
(Years)
1
2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
4
Soil Lead
G*g/g)
1625
.
.
125
125
125
375
375
625
625
875
1125
1125
1125
.
125
125
375
375
375
1125
1375
.
Dust
Lead
(Mg/g)
625
375
625
.
125
625
375
1125
375
625
1125
.
625
875
625
125
375
125
375
1375
625
1125
375
N
1
2
1
1
5
1
1
1
1
3
1
1
1
1
1
1
2
1
3
1
1
1
1
Geometric
Mean Blood
Lead
0*g/dL)
3.
4.899
7.
5.
5.055
6.
1.5
4.5
14.5
6.854
5.5
13.
9.5
19.
5.
2.5
3.873
6.5
4.762
13.
16.5
5.
2.
GSD
.
1.332
,
.
2.013
.
.
.
1.696
.
,
.
.
.
.
2.546
,
1.768
.
,
.
A-9
-------
TABLE A-2 (cont'd). GEOMETRIC MEAN AND GEOMETRIC STANDARD
DEVIATION OF BLOOD LEADS IN CELLS OR GROUPS, BY PAINT
REMOVAL STATUS, AGE, AND INTERVALS OF 250 jig/g IN
SOIL AND DUST LEAD1
Paint
Removal
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
Age
(Years)
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
0
0
0
0
1
1
1
1
Soil Lead
G*g/g)
125
125
375
375
625
875
1125
2125
125
125
375
625
625
625
1125
125
125
375
375
.
125
125
375
Dust
Lead
0*g/g)
125
375
.
625
375
625
2375
875
125
375
375
375
625
1375
875
125
375
375
875
375
125
375
375
N
2
2
1
2
1
1
1
1
3
3
2
1
1
1
1
2
5
1
1
2
1
1
1
Geometric
Mean Blood
Lead
0*g/dL)
5.477
5.745
2.
3.240
5.
7.5
5.
8.
4.245
4.440
4.899
5.
4.
4.5
13.5
3.
0.5
8.5
5.
8.
5.5
5.5
3.5
GSD
1.560
1.063
.
2.972
.
,
.
.
2.065
2.061
1.332
.
.
.
16.5
1.5
,
22.5
.
A-10
-------
TABLE A-2 (cont'd). GEOMETRIC MEAN AND GEOMETRIC STANDARD
DEVIATION OF BLOOD LEADS IN CELLS OR GROUPS, BY PAINT
REMOVAL STATUS, AGE, AND INTERVALS OF 250 jtg/g IN
SOIL AND DUST LEAD
1
Paint
Removal
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Age
(Years)
1
1
1
1
1
1
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
Soil Lead
0*g/g)
375
375
625
625
875
1125
,
125
125
375
375
375
625
1875
.
.
125
375
375
625
875
1625
1875
Dust
Lead
0*g/g)
625
875
125
375
625
1375
125
375
.
375
625
625
625
.
625
375
625
875
875
3625
1375
625
N
1
1
1
1
2
1
1
4
6
1
2
1
1
1
1
1
1
3
1
1
1
1
1
Geometric
Mean Blood
Lead
(/*g/dL)
5.5
16.5
2.5
9.
6.
5.5
8.5
3.
2.5
6.
3.
8.5
6.5
10.5
8.5
4.
4.5
5.5
4.
2.
2.
15.5
7.5
GSD
.
.
9.
.
.
.
4.
3.5
10.
.
.
.
.
.
5.5
.
.
.
.
A-ll
-------
TABLE A-2 (cont'd). GEOMETRIC MEAN AND GEOMETRIC STANDARD
DEVIATION OF BLOOD LEADS IN CELLS OR GROUPS, BY PAINT
REMOVAL STATUS, AGE, AND INTERVALS OF 250 jtg/g IN
SOIL AND DUST LEAD1
Paint
Removal
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Age
(Years)
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
Soil Lead
0*g/g)
.
.
125
125
375
875
875
1625
3125
,
125
125
375
625
1875
Dust
Lead
0*g/g)
375
625
125
375
.
625
1125
1375
1625
.
125
375
375
375
625
N
2
1
5
4
1
1
1
1
1
1
2
2
1
1
1
Geometric
Mean Blood
Lead
Oig/dL)
3.5
3.5
4.5
2.
7.
9.
7.5
9.5
13.
4.5
4.
4.
1.5
6.5
5.5
GSD
18.
5.
3.5
.
.
.
,
,
,
5.
7.5
.
•
A.2 A MORE SOPHISTICATED STATISTICAL METHOD FOR
ESTIMATING THE GEOMETRIC STANDARD DEVIATION
The GSD actually represents the residual variability in the logarithm of the predicted
blood lead level. A direct regression method that is an overly simplified approximation to
the IEUBK model at steady state exposure may be useful in deriving a residual GSD from a
blood lead and environmental lead study. The method is based on the concepts that: (1) the
IEUBK model at low to moderate steady-state exposure yields predicted blood leads that are
A-12
-------
TABLE A-3. STEM AND LEAF PLOT OF
GEOMETRIC STANDARD DEVIATION FOR MTOVALE CHILDREN1'2
MINIMUM:
LOWER QUARTILE:
MEDIAN:
UPPER QUARTILE:
MAXIMUM:
1
1
1
1
1
2
2
2
2
2
1.048
1.332
1.694
2.071
5.273
H
M
H
0011
22333333
55
6667
88
00000
3
5
7
9
***OUTSIDE VALUES***
3
5
13
2
N = 32 groups, unweighted.
76 groups with missing values excluded from plot.
approximately linear functions of PbS and PbD, with age-dependent regression coefficients;
(2) the linear model should be fitted in a logarithmic form so as to estimate relative
variability. In order to use the model, it is necessary to create indicator variables for the age
of the child in the study. These are:
A-13
-------
TABLE A-4. STEM AND LEAF PLOT OF
GEOMETRIC STANDARD DEVIATION FOR MID VALE CHILDREN
(Weighted by Degrees of Freedom)
1
MINIMUM:
LOWER QUARTILE:
MEDIAN:
UPPER QUARTILE:
MAXIMUM:
1
1
1
1
1
2
2
2
2
2
1.048
1.332
1.768
2.061
5.273
H
M
H
0000011
2222233333333
55
66666677
8888
00000000000000
3
5
7777
9
***OUTSIDE VALUES***
3
5
13
2
N = 58 groups, weighted by degrees of freedom.
AGEO = 1 if the child is age 0 to 11 months; AGEO = 0 if not;
AGE1 = 1 if the child is age 12 to 23 months; AGE1 = 0 if not;
AGE2 = 1 if the child is age 24 to 35 months; AGE2 = 0 if not;
AGE3 = 1 if the child is age 36 to 47 months; AGE3 = 0 if not;
AGE4 = 1 if the child is age 48 to 59 months; AGE4 = 0 if not;
A-14
-------
AGE6 = 1 if the child is age 72 to 83 months; AGE6 = 0 if not;
and so on. Then the model that may be fitted, using all of the children in the data set for
which observed or imputed PbS and PbD values are available, using a nonlinear regression
program for parameter estimation, is given by
In(PbB) = ln(AO*AGEO + A1*AGE1 + A2*AGE2 + A3*AGE3 + ...
+ PbS * (BO*AGEO + B1*AGE1 + B2*AGE2 + B3*AGE3 + ...) +
+ PbD * (CO*AGEO + C1*AGE1 + C2*AGE2 + C3*AGE3 + ...) +
+ X * (DO*AGEO + D1*AGE1 + D2*AGE2 + D3*AGE3 + ...))
Here, X represents other predictive covariates for blood lead. In the Midvale example,
X = RMVPAINT = 1 if paint has recently been removed from the premises, and X = 0 if
not. In other applications, it may be useful to use water lead or air lead levels as an
additional predictor. X may be omitted if necessary. In many applications, the regression
parameters may be set equal for some ages. For example, if blood leads stabilize for ages
3 to 5 years, we may set A3 = A4 = A5, B3 = B4 = B5, C3 = C4 = C5, etc. Since the
age-dependence of soil and dust lead exposure may differ somewhat from one site to another,
depending on climate or other factors, no general prescription for how to carry out such
analyses may be given.
When a non-linear regression model is fitted to the data by use of a program that
estimates non-linear parameters, it is then possible to calculate the residual standard deviation
S for the model, so that
S = standard deviation of ln(observed PbB / predicted PbB)
GSD = exp(S).
For the Midvale example described in this Appendix, we find that for N = 143 children with
no missing data for PbD, PbS, or RMVPAINT, S = 0.5701 on the natural log scale, thus
GSD = 1.768. The approximate relative standard error of S2 is (2 / (N - p))°' , where p is
the number of nonlinear parameters estimated from the data. With p = 12 parameters (ages
2 to 5 years were grouped), we have (2 / (143 - 12))°'5 = 0.1236 relative standard
deviation for the variance. An approximate 95 % confidence interval for the true value of
S2 has a lower bound (1 - 2 * (2 / (N - p))°'5), and an upper bound (1 + 2 * (2 / (N -
p))0' ), times S2. For Midvale, the limits are
( 1 - 2 * 0.1236) * (0.5701)2 = 0.2447
A-15
-------
( 1 + 2 * 0.1236) * (0.5701)2 = 0.4053
Thus the lower and upper bounds for S are 0.4947 to 0.6366, and for GSD = exp(S) the
confidence limits are 1.640 to 1.890.
A-16
-------
APPENDIX B: SUMMARY OF REVISIONS TO
LEAD UPTAKE BIOKINETIC MODEL
SOFTWARE VERSIONS
B-l
-------
VISIONS TO LEAD UPTAKE/BIOKINETIC
FROM LEAD 0.2 TO LEAD 0.4
P^"
•*H ^n
gyj ,^X
i**^ ^3
fe ^
Ol^
^
>H E
§o
2 cc
a ^
s w
S Q
!/2 O
^S
a
S
5
H
Lead 0.4 (September, 1990)
o\
oo
ON
i-T
4)
J3
4)
p
&
_>
CA
03
O,
o;
>
o
CO
c
1
G
O
C
to
I
1)
Q
"8
g
"iu
>
Cfl
CO
ex.
£
l_
Default is linea
o>
•o
o
S
g
.2
&
o
JD
CO
8
2
3
a
-1
1
1
Q
Composite = 4 jug/L
First flush = 4 jtxg/L
Flushed = 1 /ig/L
Fountain = 10 /ng/L
% First flush = 50
% Flushed = 35%
% Fountain =15%
(New defaults were provided EPA/ODW)
Dietary lead intakes are based on FDA surveys completed in 1988.
i 5.8-7.5 jug/day for ages 0-7.
4>
3
< "5 i-^
Q * T
£3°
G <£ 4)
O 4) Ofl
•0 Q "
JJ U
w . o
M *G «-"
•° M >,
Z ~ -5
j j , : .s »
Jf'^lToO 1^ S
O\ o Iff -S- II O •goii
" I' «»?* " G -tS
•!ii-l!il i'J
fllflll III
u E ft ft SR *R «£ Q 3 2
w
1
4)
•o
a
« <2
i* 3
W) ,g
.S 4)
J< -0
-.3 S
C „—
Q Q
>,
••Q ^ "^ t- .5 —"
I S -S 8 £ ->« -S
^ - S "" n £" c ,-2
81. is 11 - § " i 1
1 ! 1 I § -S 5! 1 | I .2
ife t. 1 l.^o "fc
i > S a "2^3.*.-'-4)0
f S « £ i 1 5 1 -a s | §
-2-2-S,X«13 =8«Sg's5S
T-,0S (S«T3 -SttC^SS
III .d 3 ! 1 -| 1 1 1 s 1
§-s i J 8-3 1" s !l-s
P(x) = (l/(2.51)(x)(InGSD))iexp((-l/2)((lnx-InGM)/lnGSD)2)
Ordinate of "bell-shaped plot" labelled "probability density function
removed from ordinate. Probability is computed as the integral of tf
ranged. This is graphically illustrated in the "S-shaped" probability
User option: 1) As in Lead 0.2 and option to select maternal blood
which maternal blood lead is calculated based on user-defined exposi
biokinetic model, and newborn blood lead is calculated from biokine
Menu revised to "Change GI Method/Bioavailability". If "yes" is se
the following message appears over the secondary data entry screen:
may vary depending on the lead source. For example, lead from mil
bioavailabilit)' than lead uom smellers. Differences in bioavaiiabiiiti
differences in gastrointestinal absorption of specific lead species and
depending on the source. The following data entry screen allows the
the gastrointestinal absorption coefficients to account for site-specific
•o o
4) O
~ ! &
ts a o
S r £
i-j f-~ 4>
r\ ^. &< x S
•- .-s -2 b
G ~ -° «
Sj JD t+H P
•o " o .g
* 2 § -
— ex -a »
3 •*- 2, ^ 3
CO O 3 C Q
•8 ^ £ —
O o ".2 CH •— •
-------
3
1
LH
f.t,i
o
t/2
^
P
i
u
HH
H
tj
I J
MH
UPTAKE BIO
'O LEAD 0.5
^^
^f ^t
W o
1-1 Q
O ^
H W
i/~t ^J
EVISION
FROM
»
O
j*
§
cJ
«
n3
5
H
f
O\
(December
v»
o
T3
CO
j
1
1
4)
Q.
«
i-V
Lead 0.'
£
g
u.
o
S
S
2
£
j
s*
8
3
*
<8
4J
^3
33
>
"3
•S
Q
tL7
o
»
0
t
II
t t
o
i
CO
3
o
s
2
o
u
6
3
"o
>
«3 ^
•S 1Z "^
» 0 ff'S
« -2 '-3 D
,- o i 8
2 "^ -S- u 9"
« 2 S 'S S
3 5 S |2 J
s s, I ^ -s 1
4_1 /— V C (0 t-
o PL, 3 ^ o o
g O,"o g ^ 2
0 | S 1 V J
"S > S S-LL, ^
^ A "? O *"
6 " u ci ^ ^
3 HH -^- W r2
> > ^ -5
CO
4)
e*
CO
T3
D
C
Wj
"w
t)
TL5
>-.
G
cO
e-
'§
a.
M
o
4J
^O
G
O
O
1
2?
"55
D
'O
D
D
G
.2
Ji
(A
1
3
5
S
G
O
JS
W
4>
cuo
G
cO
4>
<
o !S
3 >
•°. s
1^ eC
1 s
S .2
t» "2
C c
^
•s*
-o S
4i p,
13 ;
.ij «
5- c
2 "i
•Bl
e-o
n) d]
»
G
"° 5
S _
ser to input a range of li
1 make a series of model
specified range.
o 03 ^
•2 S S
5 S §
*i
"3 _
•^ ^
i are made manv
es input for eac
Results Out file
§ ^ .s
s a -
•3 8- jL
"9 t-< 3
c ^ o
c 3 w
.2 » §
.B--S «
ii -
G
J>
.&•
*^
2
CL
2
•g
4)
a
4>
O.
33
.0
2
&i
"^
C
CO
G
o
1
a
£,
*s
G
4)
T3
£
£
CO
x>
2
CM
O
fa
'J3
33
X
o
.s1
CO
S
s
3
tf}
s
^0
<
ly program
j_>
03
e
w
.2
33
X
I
G
1
0.
3
g
S
c
ensity functio
e., probability d
vi
to save graph output
i
3
Allows
1
s
—
3
12
E
•5
G
<+-
O
(A
x;
ex
2
W)
j)"
3
_M
1
o
Z
^
ii
X!
cL,
S
U
>i
.-••»
s
>
o
2
33
Itiple runs on
1
^<
o
"3
G-
1
'o.
e runs to an file and
.&•
~s
1
o
•o
2
G
O,
O
1
CO
"O
4)
£*
^cO
"H-
-3
4)
Xi
c
CQ
u
C
2
0.
3
.S
G
a
13
"8
JD
"H,
1
of individual or
3
to save the data outp
4)
3
CO
1
<
X)
G
cO
O
G
a
^^
i
6
es
3
TJ
'E
•o
.^-
S
£
3
oO
Q
EC
r^
UH
•H
H
CO
1
u=:
01
0
>
S
S
£
rogram after exiting
o.
e*
G
processi
1
•c
•§
T3
3
fr
•c
Q*
. values for one i
•o
i
4-1
O
1
CO
l_<
CO
1
2
1
3
m
1
5
_u
3
03
'5
>
03
O
Z
"3
1
G
O
1
1
Q
03
in
§
to
>
•o
CQ
«J
3
(C
4>
"«
>
~
*T3
1
JS
u
1
f model runs in
o
W
, and produce a serie
i=-
i
dust or
5
ti
O
0.
oo
o
s""'
G
3
a:
ed in an X-Y plot of mean
3
ex
cd
U
S2
c
e output of each
iS
he specified range. '
u.
4)
>•
1
«
>
s. medium lead.
>
T3
«
4)
i
_o
_o
3
1
ecified medium
f
D
3
1
1
CO
w
«
«
ji ^
CO «
U ^
_4J
2
_«
'«
^
o
Z
li-
G
•a
2
G
O
C
o
U
CO
"5
s
99
•Q
g
O
s
B-3
-------
fd
2
3
H
o
cc
•s
MODEL
U
^H
H
[^
n
o
a
n
t^3
|
§
a
o
H
t
!/2
F REVISIO*
O
>
^J
^j
1
«5
TO
tt
M
-
jtA
5
H
^rt
^
OS
ON
•
o
1
H-l
g
«0
•
o
9
w
J
FROM
<
4)
u
ia
<
-v
oj
a
&
Accepts a properly
4)
1
?
>
^
O
Z
c
S
D
"8
1>
T3
O
X
0
OB
CQ
3
f
Js "H
P |
'cO •"
§ bo
3 -s
- "S
1 °
if
1*
^ s
t> i-4
•rt c
11
If, ^*
I -a
M ^
1 1
_o S
S'S
ll
e «
° ts
11
3 «
g-s
G- r,
0) C
fr
g*
n
« w
M 1
=2"S
J2 >
«> S
> ™
1) w
•s
>,
1
CO
U
S
CA
1
^H
"flS
w
<2
ito PBSTAT
i
o
EX
g
4)
JD
C
U
to
4)
<£
U
CO
#
CO
processor), and as
graphic display.
3 >»
tl
o C
0 S
•% ><
•g. «
s,s.
= 1
• J
CO C4_i
H o
to 3
•*3 G
cq D
1 S
£ S
i'l
~S u
•° ^
•8s
*l
H
u S
to o
*"? «
* ,j
XI
«: s
•C^ o
2 t->
o 3
•SH
CO CO
IS
o
« 3
•O G
§ s
S i
1 2
^ 1
•s*
Accepts the output
selected by the usei
_4)
3
cO
'ec
>
sB
O
z
5
1
e
Q.
S
CM
tg
^
w
O
•S
«j
cO
4)
S
4>
£
ad
; in PBSTAT
c
^o
o<
o
l-l
4>
S
'CJ
o.
0)
fS
p
<
H
S
IX
1
O
O
«
93
•S
'S
i-s
*n 3
IS
o 1
SJ'C
OS &
&H J-
c^
P w
?\ ir
Q S
>-) 0
(X 3
PQ ^3
fe|
1 '*
i DOS comm
^ad, however
1 =
s
S *s
y ^
U c
» 1
« t
4)
Ti ^
« »
S .2
J ^
II
J2 «
G «
Produces graph pri
the same functions
J£
3
CO
'cO
>
eO
O
z
§
E.
CQ
c^
5
a
#
•o
c
93
>
<
_i
*
1
g
03
H
S
Q*
CO
a*
1
2
&o
L*
«
C
•c
c.
o.
•c
u
«
«
£
*o
G
1
o
S
OH
S -d
60 S
0 J
higher resolutions (
files produced with
Restructured.
3
4)
'5
G
.2
'1
u
4)
3
£
i
>,
33
•o
1>
O
no user option.
•S
'5
•B
o
c
4)
c
O
"S
•c
£
c
o
'u
2
4)
1!
•o
o
s
a J 'K
21s
_4) 4) •— Ofl
^ — H3 £
eO oO o ^
.•^S > «o
oO ^_,
o *flj « T
Z Q ^^
•o
g
o * —
33, 1 1
j a s 2 s •*-
3 to « u X
3 -5 - -9- -a "S
^3 « "S ^ S 3
o c -o S a ^o
« g 2 M j2 S
^" o 8 ^ o
JNl's §
Hill I
— «3 .2 5 ±1 5
.0 OJ J3 -D _ 03
— T) „ - ^ >
2 i 03 c S —
C « E ° — "3
2 S s •§ g S «
« p « K — ^) o
^ 5 C 2 03 r\ 60
-£ o c c o Q s
"a)
T3
0 0
s 1
1 f
H. 0
= M
""• O
<«-, 5
lows the use
S ,-S
•c
D.
1
B-4
-------
_ •»
s
PN
K^
^
pH
0
t/2
_]
W
Q
i
u
HH
H
i
0
hH
pa
w
CI
s
^^
5k
p^
Q
3
S
o
H
UMMARY OF REVISIONS
cc
T3
^^
C!
O
U
fl
05
S
M
§
H
•o
ON
ON
•
O
^•S
ri
H
^
7=!
^
hJ
O
H
«n
•
0
FROM LEAD
£
ON
b
CO
3
C
eO
^
•O
S
o
•s
oS
J
Lead 0.5 (December, 1990)
3
U.
O
k*
4)
1>
E
S
£
1
<«-
O
-S?
M
3.
bri
a
S
•o
t»
fi
Ofi
a
g
r^
O
w
C
Q
o
C
a
D
3
1
3
"§
O
Default value function is 0.28 fig Pb/g
dust per fig Pb/g of soil.
^
3
*O
O
o
•o
c
o
T3
_^
'o
^4^
O
c
.2
5
,0
'C
•y _,
G "O
0 oj
u JS
mp
c
&D
^
^H
O
(fl
<1>
^3
33
>
3
•s
Q
Default value is 0.2 jig/m3.
c
o
"g
1
u
c
o
o
•a
I
U
"ed
fc^
0
o
3
6
S
•Q
oo
CO
ON
<
Q
li
<*-<
O
an
^i
CO
G
03
1
3
«
T3
£
O
s
EO
G
O
t)
x
cO
Xi
<0
EO
U
4)
_3
00
>
3
<5
u
Q
Default values were based on a
preliminary analysis of 1988 FDA survey
data.
V
^A
-o
!
£>
93
s
3
1
o
g
o
S
a
•§
u
o
4)
a
CO
•g
c
1
1
£
U
o
D
a
'•o
o
U
«
4J
a>
o
'•5
c
'Si
0
'JD
3
<2
o
T3
<4-<
O
M
C
i
RECSUM time constants (e.g., TPLUR)
CQ
^
4)
^
3
0)
V
CO
w
CO
1
§•
y
fj
03
1"
5
X)
1
appeared to be insensitive to changes in
TPLUR.
1
"R
c °
.2 "°
ca tu
a) U
•- 5
4> ^
•5 °>
* *
a o
§ S
% §
1 "S
•5.3
Is
•3 S
1 S
1
tt 3
03 C
t's
O "~
0 S
qj D
eO *-* *^
M .S .S
Forward Euler algorithm calculates the
increase compartmental lead mass over
the iteration interval as the total lead
inflow to the compartment, minus the
total lead outflow at the beginning of the
interval.
£
•c
o
_M>
*(,
tu
1
J)
!B
OB
3
03
4>
£
V-"
o
c
o
'Z3
J5
3
jj
**
o
9)
.S •
13 •^i'
8 g
Ig
3 H
§^
0 ES
£H
« 5
.2 «
11
=3 1
I!
S -S
S
^ 0
<4-l
Js ^2
o fi
03 D
2 'o
CL 1*3
11
O
s O
1 §
•P 2 's
I g> &
's si
o< " -3
"Non-integrated" approach: Saturable
absorption coefficients are calculated for
each medium (soil, diet, drinking water,
etc.) based on intakes from each medium,
and are used to calculate media-specific
uptakes which are summed to yield total
uptakes.
£ -3
a -o
C 0
2 S
f-1 c
& .2
o -g.
S &
^ o
U tfl
x .0
Wjf
<,
la 3
£ S S
^ "3 c
»
------- |