September 27, 1996
RISK ASSESSMENT FOR
THE SECTION 403 RULEMAKING
Draft Report
VOLUME I
CHAPTERS 1 TO 8
DO NOT CITE OR QUOTE
Prepared
by
Battelle
505 King Avenue
Columbus, Ohio 43201
for
Chemical Management Division
Office of Pollution Prevention and Toxics
Office of Prevention, Pesticides, and Toxic Substances
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
DISCLAIMER
The material in this document has not been subject to Agency
technical and policy review. Views expressed by the authors are their own
and do not necessarily reflect those of the U.S. Environmental Protection
Agency. Mention of trade names, products, or services does not convey,
and should not be interpreted as conveying, official EPA approval,
endorsement, or recommendation. Do not quote or cite this document.
This report is copied on recycled paper.
Draft -Do Note Cite or Quote ii September 27, 1996
-------
TABLE OF CONTENTS
Page
1.0 INTRODUCTION, BACKGROUND, AND OVERVIEW 1
1.1 Requirements of §403 3
1.1.1 Statutory Requirements 3
1.1.2 EPA's Response to Statutory Requirements 4
1.2 Objectives of Risk Assessment 6
1.3 Organization of Report 7
.3.1 Chapter One - Introduction, Background, and Summary 7
.3.2 Chapter Two - Hazard Identification 8
.3.3 Chapter Three - Exposure Assessment 8
.3.4 Chapter Four - Dose-Response Assessment 9
.3.5 Chapter Five - Integrated Risk Assessment 10
1.3.6 Chapter Six - Validation Studies for the IEUBK Model 11
1.3.7 Chapter Seven - Risk Summary 11
1.4 Overview of Methodology 6
1.4.1 General Approach 12
1.4.2 Current Blood-Lead Distribution 14
1.4.3 Baseline Environmental Lead Conditions 15
1.4.4 Post-§403 Environmental Lead Conditions 15
1.4.5 Post-§403 Blood-Lead Distribution 16
1.4.6 Projected Population of Children 17
1.4.7 Reductions in Childhood Blood Lead and Health Effects 17
1.5 Data Sources, Analysis Tools and Limitations 6
1.5.1 Hazard Identification 18
1.5.2 Exposure Assessment 18
1.5.3 Dose-Response Assessment 20
1.5.4 Integrated Risk Assessment 21
2.0 HAZARD IDENTIFICATION 24
2.1 Measures of Body-lead Burden 26
2.2 Mechanisms of Lead Toxicity 28
2.2.1 Physiological Mechanisms 30
2.2.2 Neurotoxic Effects of Lead 31
2.2.3 Hematologic Effects of Lead 34
2.3 Health Effects of Lead Exposure 36
2.3.1 Neurological Effects of Lead 37
Draft -Do Note Cite or Quote
iii
September 27, 1996
-------
TABLE OF CONTENTS
(Continued)
Page
2.3.2 Hematological Effects of Lead 43
2.3.3 Other Effects of Lead 44
2.4 Representative Health Effects 48
2.4.1 Elevated Blood-Lead Concentration 49
2.4.2 IQ Point Deficits 49
3.0 EXPOSURE ASSESSMENT 51
3.1 Sources and Pathways of Lead 52
3.2 Supporting Evidence in Epidemiologic Studies '. 59
3.2.1 Baltimore Repair and Maintenance (R&M) Study 62
3.2.2 Rochester Study 63
3.2.3 Urban Soil Lead Abatement Demonstration Project (USLADP) 66
3.2.4 Birmingham Urban Lead Uptake Study 68
3.2.5 Cincinnati Longitudinal Study 68
3.2.6 Brigham and Women's Hospital Longitudinal Study 69
3.3 Lead in Dust, Soil, and Paint in the Nation's Housing 70
3.3.1 The Distribution of Lead Levels in Household Dust, Soil,
and Paint 71
3.3.1.1 HUD National Survey 71
3.3.1.2 The Baltimore Repair and Maintenance (R&M) Study 79
3.3.1.3 The Rochester Lead-in-Dust Study 83
3.3.2 Characterizing the Population of Children in the Nation's Housing Stock 88
3.4 Distribution of Childhood Blood-lead 89
3.4.1 NHANES III 90
3.4.2 Baltimore Repair and Maintenance (R&M) Study 94
3.4.3 Rochester Study 95
4.0 DOSE-RESPONSE ASSESSMENT 97
4.1 Estimation of Mean Blood Lead Concentration 100
4.1.1 IEUBK Model 100
4.1.2 Epidemiological Model 103
4.2 Utilizing Dust Lead Loadings 107
4.2.1 Wipe versus Blue Nozzle (BN) Vacuum Conversions 107
4.2.2 Wipe versus Baltimore Repair and Maintenance (BRM)
Vacuum Conversions 108
Draft -Do Note Cite or Quote iv September 27, 1996
-------
TABLE OF CONTENTS
(Continued)
Page
4.3 Estimating the Effect of Pica for Paint on Childhood Blood-lead Levels 109
4.3.1 IEUBK Model 110
4.3.2 EPI Model 111
4.4 Health Outcomes 111
4.4.1 Decrements in IQ Scores 112
4.4.2 Increased Incidence of IQ Scores Less Than 70 116
4.4.3 Incidence of Elevated Blood-Lead Levels 121
4.5 Options for Standards 125
4.5.1 Analysis Methods 126
4.5.2 Assumptions 128
4.5.3 Results 128
5.0 INTEGRATED RISK ANALYSIS 132
5.1 Baseline Characterization of Children's Blood-lead Concentrations
and Health Effects 134
5.2 Intervention Activities 136
5.2.1 Interventions 140
5.2.2 Reductions in Environmental Lead Levels Following
Interventions 142
5.2.3 Intervention Triggers 145
5.2.4 Reductions in Blood-Lead Levels Following Interventions 146
5.3 Characterizing the Risks Following Intervention 147
5.3.1 Characterization of Risks for Various Sets of Standards 152
5.3.1.1 Varying Dust Standard Options 156
5.3.1.2 Varying Soil Standard Options 160
5.3.1.3 Varying Paint Standard Options 166
5.3.1.4 Varying All Standard Options 171
5.4 Sensitivity and Uncertainty Analyses 180
5.4.1 Components of the Sensitivity Analysis 180
5.4.1.1 Alternative Age Range of Children 182
5.4.1.2 Alternative Assumptions on Average IQ Score Decline
Per Unit Increase in Blood-Lead Concentration 182
5.4.1.3 Alternative Approach to Characterizing a Baseline
Blood-Lead Distribution from NHANES III Data 183
Draft -Do Note Cite or Quote v September 27, 1996
-------
TABLE OF CONTENTS
(Continued)
Page
5.4.1.4 Uncertainty in Converting Dust-Lead Loadings for
Comparison to Standards 184
5.4.1.5 Alternative Assumptions on Post-Intervention
Environmental-Lead Levels 185
5.4.1.6 Alternative Methods to Observing Differences in
Health Effects Between Pre- and Post-Intervention 186
5.4.2 Results of the Sensitivity Analysis 190
5.4.2.1 Alternative Age Range of Children 190
5.4.2.2 Alternative Assumptions on Average IQ Score Decline
Per Unit Increase in Blood-Lead Concentration 192
5.4.2.3 Alternative Approach to Characterizing a Baseline
Blood-Lead Distribution from NHANES III Data 193
5.4.2.4 Uncertainty in Converting Dust-Lead Loadings for
Comparison to Standards 194
5.4.2.5 Alternative Assumptions on Post-Intervention
Environmental-Lead Levels 195
5.4.2.6 Alternative Methods to Observing Differences in
Health Effects Between Pre- and Post-Intervention 198
6.0 CONFIRMATION STUDIES FOR IEUBK MODEL 205
6.1 Methods 208
6.1.1 Descriptive Measures 208
6.1.2 Input Data Selection 209
6.2 Results 213
6.3 Conclusions 221
7.0 RISK SUMMARY 222
7.1 Scientific Basis for §403 and Tools for the Risk Assessment 223
7.2 Health Risk Reductions and Numbers of Children and Housing Units Effected . 225
7.3 Robustness of Risk Assessment Data Sources and Methodology 228
7.4 Conclusions of Risk Assessment 232
8.0 REFERENCES 233
LIST OF TABLES
Table 2-1. Interpretation of Blood-Lead Concentrations and Follow-Up Actions
Recommended by CDC 50
•\
Draft-Do Note Cite or Quote vi September 27, 1996
-------
TABLE OF CONTENTS
(Continued)
Page
Table 3-1. Childhood Lead Exposure Studies Conducted in Urban Communities That
Present Evidence of the Positive Relationship Between Environmental-Lead
Levels and Blood-Lead Concentrations 60
Table 3-2. Childhood Lead Exposure Studies Conducted in Smelter Communities That
Present Evidence of the Positive Relationship Between Environmental-Lead
Levels and Blood-Lead Concentrations 61
Table 3-3. Estimated Total Number of Occupied Housing Units in the National Housing
Stock in 1997 According to Year-Built Category 73
Table 3-4. Summary of the Distribution of Lead Loadings in Floor-Dust Samples Within
Housing Units in the HUD National Survey, Weighted to Reflect the Predicted
1997 Housing Stock 75
Table 3-5. Summary of the Distribution of Lead Concentrations in Floor-Dust Samples
Within Housing Units in the HUD National Survey, Weighted to Reflect the
Predicted 1997 Housing Stock 76
Table 3-6. Summary of the Distribution of Lead Loadings in Window Sill-Dust Samples
Within Housing Units in the HUD National Survey, Weighted to Reflect the
Predicted 1997 Housing Stock 76
Table 3-7. Summary of the Distribution of Lead Concentrations in Window Sill-Dust
Samples Within Housing Units in the HUD National Survey, Weighted to
Reflect the Predicted 1997 Housing Stock 77
Table 3-8. Summary of the Distribution of Soil-Lead Concentrations for Housing Units in the
HUD National Survey, Weighted to Reflect the Predicted 1997 Housing Stock ... 77
Table 3-9. Summary of the Distribution of Observed Maximum XRF Lead Levels in Paint for
Housing Units in the HUD National Survey, Weighted to Reflect the Predicted
1997 Housing Stock 78
Table 3-10. Predicted Numbers and Percentages of Units Having Lead-Based Paint in the 1997
Occupied Housing Stock, Based on Information from the HUD National Survey ... 78
Table 3-11. Summary of Average Pre-lntervention Floor Dust-Lead Loading for Housing
Units in the Baltimore R&M Study 80
Table 3-12. Summary of Average Pre-lntervention Floor Dust-Lead Concentrations for
Housing Units in the Baltimore R&M Study 80
Table 3-13. Summary of Average Pre-lntervention Window Sill Dust-Lead Loading for
Housing Units in the Baltimore R&M Study 81
Table 3-14. Summary of Average Pre-lntervention Window Sill Dust-Lead Concentrations for
Housing Units in the Baltimore R&M Study 81
Table 3-15. Summary of Average Pre-lntervention Dripline Soil-Lead Concentrations for
Housing Units in the Baltimore R&M Study 83
Table 3-16. Summary of Observed Maximum XRF Paint-Lead Concentration at Pre-lntervention
for Housing Units Slated for R&M Intervention in the Baltimore R&M Study 83
Table 3-17. Summary of Average Pre-lntervention Floor Dust-Lead Loading for Housing
Units in the Rochester Study 85
Table 3-18. Summary of Average Pre-lntervention Floor Dust-Lead Concentrations for
Housing Units in the Rochester Study 85
Table 3-19. Summary of Average Pre-lntervention Window Sill Dust-Lead Loading for
Housing Units in the Rochester Study 86
Draft -Do Note Cite or Quote
VII
September 27, 1996
-------
TABLE OF CONTENTS
(Continued)
Page
Table 3-20. Summary of Average Pre-lntervention Window Sill Dust-Lead Concentrations
for Housing Units in the Rochester Study 86
Table 3-21. Summary of Average Pre-lntervention Dripline Soil-Lead Concentrations for
Housing Units in the Rochester Study 87
Table 3-22. Summary of Average Pre-lntervention Soil-Lead Concentrations from Play
Areas for Housing Units in the Rochester Study 87
Table 3-23. Summary of Observed Maximum XRF Paint-Lead Concentration at
Pre-lntervention for Housing Units in the Rochester Study 88
Table 3-24. Estimated Number of Children in the 1997 National Housing Stock, by Age of
Child and Year-Built Category 89
Table 3-25. Summary of Blood-Lead Concentration Data for Children Aged 1-2 Years and
Aged 1-5 Years, Based on NHANES III (Phase 1) 92
Table 3-26. Estimated Probabilities of Elevated Blood-Lead Concentrations in Children
Aged 1-2 Years and Aged 1-5 Years, Based on NHANES 111 (Phase 1) 92
Table 3-27. Estimated Percentage of Children With Blood-Lead Concentrations Exceeding 10
fjg/dL, and the Geometric Mean and Geometric Standard Deviation of Blood-Lead
Concentration, for Children Aged 1 to 2 Years Within Selected Subgroups 93
Table 3-28. Summary Statistics on Blood-Lead Concentration Measured Prior to Intervention
in the Baltimore Repair and Maintenance Study 95
Table 3-29. Summary Statistics on Blood-Lead Concentration Measured in the Rochester
Lead-in-Dust Study 96
Table 4-1. Summary of Default Parameter Values Used in the IEUBK Model (Version 0.99D). 102
Table 4-2. Parameter Estimates and Associated Standard Errors for the Multimedia
Exposure Model Based on Data from the Rochester Lead-in Dust Study 105
Table 4-3. Summary Information for Studies Included in the Schwartz (1994) Meta-Analysis. 113
Table 4-4. Experts Who Participated in the Assessment of the Relationship Between IQ
Scores and Blood-lead Levels by Wallsten and Whitfield 117
Table 4-5. Piecewise Linear Function for Estimating the Increased Percentage of Children
Having IQ Scores less than 70 Due to Lead Exposure 118
Table 4-6. Definitions of Performance Characteristics Used to Characterize the Performance
of Options for the §403 Standards Based on Empirical Data from Lead Exposure
Studies 127
Table 4-7. Summary of Estimated Standards Which Achieved a Negative Predictive Value of
95% or an Estimated 95% Probability of a Child's Blood-Lead Concentration
Below 10/;g/dL in a Dwelling that is at or Below the Standard 129
Table 4-8. Proposed Options for §403 Standards To Be Evaluated in the Risk Assessment
and Economic Analysis 131
Table 5-1. Estimated Baseline (1997, Pre-lntervention) Number and Percentage of Children
Aged 1 to 2 Years Having Specific Health Effects 138
Table 5-2. Interventions Defined for the §403 Risk Assessment Effort 141
Table 5-3. Expected Post-Intervention Lead Levels Associated With Performing §403
Interventions 144
Table 5-4. Intervention Triggers Defined for the Risk Assessment of §403 146
Table 5-5. Projected Impact of §403 on House 1011501 in the National Survey 149
Table 5-6. Ranges of Standards Considered 150
Draft -Do Note Cite or Quote
viii
September 27, 1996
-------
TABLE OF CONTENTS
(Continued)
Page
Table 5-7. Characterization of Impact of Various Options for Dust Standards: Soil and
Paint Standards fixed (400 yug/g for Soil Cover, 3000 i/g/g for Soil Removal,
5 ft2 damaged LBP for Paint Maintenance, 20 ft2 damaged LBP for
Paint Abatement) 154
Table 5-8. Characterization of Impact of Various Options for Soil Standards: Dust and
Paint Standards fixed (100/yg/ft2 for Dust Lead Loading, 500//g/ft2 for
Window Sill Dust Lead Loading, 5 ft2 damaged LBP for Paint Maintenance,
20 ft2 damaged LBP for Paint Abatement) 162
Table 5-9. Characterization of Impact of Various Options for Paint Standards: Dust and
Soil Standards fixed (100 jyg/ft2 for Dust Lead Loading, 500/yg/ft2 for Window
Sill Dust Lead Loading, 400 //g/g for Soil Covering, 3000 //g/g for Soil
Removal) 170
Table 5-10. Characterization of Impact of Various Sets of Dust, Soil, and Paint Standards. . 173
Table 5-11. Comparison of Blood-Lead Concentrations Before and After §403 179
Table 5-12. Procedures and Their Alternatives that were Included in the Sensitivity
Analysis 181
Table 5-13a. Estimated Baseline1 Number and Percentage of Children Having Specific Health
Effects, for Two Age Groups of Children and Under Three Assumptions on
Average Decline in IQ Score per Unit Increase in Blood-Lead Concentration ... 191
Table 5-13b. Estimated Baseline1 Average (and Standard Deviation) IQ Loss, for Two Age
Groups of Children and Under Three Assumptions on Average Decline in IQ
Score per Unit Increase in Blood-Lead Concentration 192
Table 5-14. Estimated Baseline Health Effects, As Calculated Under Two Approaches to
Calculating the Baseline Distribution of Blood-Lead Concentration Using
NHANES III Data 193
Table 5-15. Number (and Percentage) of Units in the 1997 National Housing Stock
Projected to Exceed Various Combinations of Environmental-Lead Standards
Under Section 403 Rules, As Determined from Three Different Sets of
Converted Dust-Lead Loadings 195
Table 5-16a. Estimated Percentages of Children Aged 12-35 Months Having Specific Health
and Blood-Lead Effects, Based on the IEUBK Model, forVarious Options for
Post-Intervention Environmental-Lead Levels 196
Table 5-16b. Estimated Percentages of Children Aged 12-35 Months Having Specific Health
and Blood-Lead Effects, Based on the EPI Model, for Various Options for Post-
Intervention Environmental-Lead Levels 197
Table 5-17a. Estimated Percentages of Children Aged 12-35 Months Having Specific Health
and Blood-Lead Effects, Based on the IEUBK Model Under the Approach Used
in the Risk Assessment and an Alternative Approach 200
Table 5-17b. Estimated Percentages of Children Aged 12-35 Months Having Specific Health
Effects, Based on Blood-Lead Concentrations at Pre- and Post-Intervention as
Determined from the EPI Model Under the Approach Used in the Risk
Assessment and an Alternative Approach 201
Table 5-18. Estimated Distribution of Post-§403 Blood-Lead Concentrations for Children
1-2 Years Old Based on the IEUBK Model for Both the Risk Assessment and
the Adjusted Blood-Lead Effects Model Approach 203
Table 6-1. Comparison of Observed and IEUBK Model Predicted Blood-Lead Levels 214
Draft -Do Note Cite or Quote
IX
September 27, 1996
-------
TABLE OF CONTENTS
(Continued)
Page
Table 7-1. Percent of Housing Units Requiring §403 Interventions and Estimated
Reduction from Baseline Health Risks Due to §403 for Three Sets of
Standards Options 227
Table 7-2. Reductions From Baseline Health Risks (in percent) Obtained by Varying
Intervention Effectiveness Due to §403 Intermediate Standards For Selected
Health Endpoints and Risk Assessment Methodology 231
LIST OF FIGURES
Figure 1-1. §403 Risk Assessment Steps 13
Figure 4-1. Methodology Associated With Characterizing the Relationship Between a
National Distribution of Environmental Lead Levels and a National Distribution
of Blood-Lead Concentrations 99
Figure 4-2. IEUBK Model Predicted Blood-Lead Concentration for Children Two Years Old
Plotted Separately Versus Soil-Lead Concentration and Dust-Lead
Concentration for Fixed Default Values of the Remaining Model Parameters ... 103
Figure 4-3. EPI Model Predicted Blood-Lead Concentration Plotted Separately Against
Floor Dust-Lead Loading, Sill Dust-Lead Loading and Soil Lead Concentration
for Fixed Values of the Remaining Model Inputs 106
Figure 4-4. Estimated IQ Point Loss Due to Lead Exposure Plotted Against Blood-Lead
Concentration 114
Figure 4-5. Estimated IQ Point Loss Due to Lead Exposure Plotted Against Concentration
of Lead in Soil and Dust, Utilizing IEUBK Model Predictions 115
Figure 4-6. Estimated IQ Point Loss Due to Lead Exposure Plotted Against Concentration
of Lead in Soil, Utilizing EPI Model 115
Figure 4-7. Increase in Percentage of Children with IQ Below 70 Due to Lead Exposure
Plotted Against Blood-Lead Concentration 119
Figure 4-8. Increase in Percentage of Children with IQ Below 70 Due to Lead Exposure
Plotted Against Concentration of Lead in Soil and Dust, Utilizing IEUBK Model
Predictions 119
Figure 4-9. Increase in Percentage of Children with IQ Below 70 Due to Lead Exposure
Plotted Against Concentration of Lead in Soil, Utilizing EPI Model 120
Figure 4-10. Percentage of Children with Blood-Lead Concentration Greater than 25 //g/dL
Due to Lead Exposure Plotted Against Geometric Mean Blood-Lead
Concentration, Assuming a GSD of 1.6 122
Figure 4-11. Percentage of Children with Blood-Lead Concentration Greater than 25 //g/dL
Due to Lead Exposure Plotted Against Concentration of Lead in Soil and Dust,
Utilizing IEUBK Model Predictions 123
Figure 4-12. Percentage of Children with Blood-Lead Concentration Greater than 25 //g/dL
Due to Lead Exposure Plotted Against Concentration of Lead in Soil, Utilizing
EPI Model 124
Figure 5-1. Baseline Distribution of Blood-Lead Levels Based on NHANES III, Phase 1 (0.2
Percent of Children Had Blood-Lead Concentration Greater than 32 //g/dL) ... 135
Draft -Do Note Cite or Quote
September 27, 1996
-------
TABLE OF CONTENTS
(Continued)
Page
Figure 5-2. Baseline Distribution of IQ Decrements Due to Elevated Blood-Lead
Concentration Based on NHANES III, Phase 1 (0.03 Percent of Children Had in
Excess of 10 IQ Points Lost) 137
Figure 5-3. Post-§403 Risk Characterization Process 148
Figure 5-3a. Projected Health Endpoints Based on Various Options for Dust Standards,
Part 1; Soil Cover 400 fjglg, Soil Removal 3000 fjg/g, Paint Maintenance 5 ft2
Paint Abatement 20 ft2. (Dashed reference line represents baseline risk.) .... 158
Figure 5-3b. Projected Health Endpoints Based on Various Options for Dust Standards,
Part 2; Soil Cover 400 fjglg, Soil Removal 3000 /jg/g, Paint Maintenance 5 ft2,
Paint Abatement 20 ft2. (Dashed reference line represents baseline risk.) .... 159
Figure 5-4a. Projected Health Endpoints Based on Various Options for Soil Standards,
Part 1; Floor Dust 100/yg/ft2, Window Sill Dust 500//g/ft2, Paint Maintenance
5 ft2, Paint Abatement 20 ft2. (Dashed reference line represents
baseline risk.) 164
Figure 5-4b. Projected Health Endpoints Based on Various Options for Soil Standards,
Part 2; Floor Dust 100 jug/ft2, Window Sill Dust 500//g/ft2, Paint Maintenance
5 ft2, Paint Abatement 20 ft2. (Dashed reference line represents
baseline risk.) 165
Figure 5-5a. Projected Health Endpoints Based on Various Options for Paint Standards,
Part 1; Floor Dust 100//g/ft2, Window Sill Dust 500//g/ft2, Soil Cover 400
fjglg, Soil Removal 3000 //g/g. (Dashed reference line represents baseline
risk.) 169
Figure 5-5b. Projected Health Endpoints Based on Various Options for Paint Standards,
Part 2; Floor Dust 100//g/ft2, Window Sill Dust 500//g/ft2, Soil Cover 400
fjg/g, Soil Removal 3000 //g/g. (Dashed reference line represents baseline
risk.) 168
Figure 5-6a. Projected Health and Blood-Lead Endpoints Based on Various Sets of Options
for Dust, Soil, and Paint, Part 1. (Dashed reference line represents baseline
risk.) 174
Figure 5-6b. Projected Health Endpoints Based on Various Sets of Options for Dust, Soil,
and Paint, Part 2. (Dashed reference line represents baseline risk.) 175
Figure 5-7. Projected Post-1403 Blood-Lead Concentration Distributions Based on EPI and
IEUBK Models at Standards of Floor Dust-Lead -200 //g/ft2; Window Sill Dust-
Lead - 500 //g/ft2; Soil Cover - 400 //g/g; Soil Removal - 3000 fjglg; Paint
Maintenance - 5 ft2; Damaged LBP, and Paint Abatement - 20 ft2 Damaged
LBP 178
Figure 5-8. Comparison of NHANES III Blood-Lead Concentration Distribution to
Distributions Estimated Using the Adjusted Blood Lead Effects Model and the
Post-1403 Risk Assessment Method 204
Draft -Do Note Cite or Quote xi September 27. 1996
-------
1.0 INTRODUCTION, BACKGROUND, AND OVERVIEW
CHAPTER 1 SUMMARY
This introductory chapter provides the background, purpose, and objectives of
this report. %403 of Title IV, as amended in Title X, requires EPA to define
standards for lead-based paint hazards, lead-contaminated dust, and lead-
contaminated soil. This report
• documents the scientific basis for %403.
* characterizes the hearth risks to young children from exposures to
lead
• estimates the reductions in childhood health risks and blood-lead
concentrations expected to results from various options for $403,
and
• estimates the numbers of children and housing units affected by
various options for §403.
This information is provided to help the risk managers evaluate and compare
various regulatory options for §403.
Title X of the Housing and Community Development Act, known as the Residential
Lead-Based Paint Hazard Reduction Act of 1992, contains legislation designed to evaluate and
reduce exposures to lead in paint, dust, and soil in the nation's housing. This act provides the
framework for developing a national strategy for reducing and preventing lead exposures to
children. Consistent implementation of this strategy by federal, state, local and private agencies
requires a uniform definition of lead hazards. Title X includes legislation that requires the U.S.
Environmental Protection Agency (EPA) to define standards for lead in paint, dust, and soil.
More specifically, §403 of Title IV of the Toxic Substances Control Act (TSCA), as amended in
Title X, requires EPA to "promulgate regulations which shall identify, for purposes of this title
and the Residential Lead-Based Paint Hazard Reduction Act of 1992, lead-based paint hazards,
lead-contaminated dust, and lead-contaminated soil."
§403 will set standards (condition and location of lead-based paint, levels of lead in dust
and soil) against which to compare a residential environment when evaluating the presence and
magnitude of lead-based paint hazards. Federal, state, local, and private agencies will use these
standards to determine which homes require actions be taken to reduce or prevent the threat of
Draft - Do Not Cite or Quote 1 September 27, 1996
-------
childhood lead poisoning. Following the conduct of these actions health risks associated with
childhood lead poisoning will be reduced for children currently residing in the residence as well
as those that may later live in the residence. Proper selection of the standards requires both an
understanding of the health risks associated residential exposures to lead, the amount by which
these risks can be reduced through interventions and the numbers of homes and children affected
by the standards.
This report was prepared to support the §403 rulemaking, and to help EPA to document
the scientific basis for §403. In addition, the report provides information to the risk managers on
the relative efficacy of various options for the §403 standards for reducing the health risks
associated with lead exposures. First, the report summarizes EPA's assessment of the health
risks to young children from exposures to lead-based paint hazards, lead-contaminated dust, and
lead-contaminated soil in the nation's housing. Second, the report documents the approach
developed by EPA to estimate the reductions in these risks following implementation of §403,
and applies this methodology to evaluate several regulatory options. The benefits of a regulatory
option for §403 are expressed in terms of the reduction in health risks attained from the passage
of an option for the §403 standards. Finally, the report provides estimates of the numbers of
homes and children that will be affected by various options for the standards.
Information presented in this Risk Assessment will ultimately be used as input to the
Regulatory Impacts Analysis (RIA) for the proposed rule, as well as any interim cost-benefit
analyses. While the Risk Assessment quantifies the impact of §403 in terms of health risks and
blood-lead levels and documents the scientific basis for the rule, the RIA expresses the impact of
§403 in terms of costs: monetary costs of implementing the regulation, monetary benefits
associated with reductions in health risks and blood lead concentrations for various options for
the regulation, and the estimated economic impacts of the regulations. The RIA also examines
the likelihood of interventions actually taking place. Finally, the RIA summarizes other
regulatory actions designed to reduce risks from lead, and presents environmental equity analyses
for adults and children.
This report documents the critical decisions on risk-assessment-related tools and data that
are being relied upon in the RIA analyses, which relate environmental levels of lead to children's
blood-lead levels and, ultimately, potential health effects. This report includes a description of
Draft - Do Not Cite or Quote 2 September 27, 1996
-------
the data used, an assessment of the strengths and weaknesses of that data, and discussions of any
additional uncertainties which result from using these particular data sets and tools to create
estimates of risk reduction on a national basis. Section 1.1 discusses the statutory requirements of
§403 and EPA's response to the statutory requirements. The objectives of the risk assessment
are presented in Section 1.2. Summaries of each component of the risk assessment are given in
Section 1.3. Section 1.4 provides an overview of the methodology utilized to conduct the
integrated risk assessment. Finally, Section 1.5 describes the major data sources and analysis
tools employed in each component of the risk assessment along with a discussion of their
limitations.
1.1 REQUIREMENTS OF §403
1.1.1 Statutory Requirements
On October 29,1992, the United States Congress enacted the Residential Lead-Based
Paint Hazard Reduction Act (Title X of HR 5334). This includes an amendment to the Toxic
Substances Control Act (Title IV: Lead Exposure Reduction) that requires the EPA
Administrator to identify lead-based paint hazards. Specifically, §403 of TSCA Title IV states:
"The Administrator shall promulgate regulations which shall identify, for
purposes of this title and the Residential Lead-Based Paint Hazard Reduction Act
of 1992, lead-based paint hazards, lead-contaminated dust, and lead-contaminated
soil."
This statute requires EPA to define standards for lead in dust and soil and to define what
constitutes a lead-based paint hazard. The statute defines lead-based paint to be dried paint film
with a lead content exceeding 1.0 mg/cm2 or 0.5 percent (5,000 parts per million (ppm)) by
weight. The statute requires EPA to identify lead-based paint hazards, i.e., the condition,
location and amount of lead-based paint that causes exposures to lead in paint, lead-contaminated
dust and lead-contaminated soil that would result in adverse health risks.
The Title X statute provides definitions for lead-based paint, lead-based paint hazard,
lead-contaminated soil, lead-contaminated dust and other relevant terms. Definitions given in the
statute and utilized in the risk assessment are provided in the glossary in Appendix A.
Draft - Do Not Cite or Quote 3 September 27. 1996
-------
1.1.2 EPA's Response to Statutory Requirements
The Agency's approach to the §403 requirements is to establish quantitative standards for
levels of lead in residential soil and dust. For soil, the standard will be defined in units of mass
of lead per mass of soil (ug Pb/g soil). A tiered standard was utilized for the soil standard
specified in the Interim 403 Guidance (EPA, 1994) and the Agency has decided to continue to
use a two tiered standard for soil. The lower level recognizes levels of lead in soil that may be of
a health concern and thereby warrant interim controls to reduce this concern. Because of the
limited data available on the efficacy of interim controls for soil, the Agency felt more extreme
actions may be required for high levels of lead in soil. The higher level for the soil standard
identifies soil levels that present a larger health concern and require more intensive action to
eliminate this concern, such as soil abatement.
There are two methods for measuring the amount of lead in dust. Dust-lead loading
measures the mass lead collected per surface area sampled and is usually expressed in terms of
micrograms of lead collected per square foot sampled (ug Pb/ft2'. Dust-lead concentration
measures the mass of lead collected per mass of dust collected and is usually stated in terms of
micrograms of lead collected per gram of dust collected (ug Pb/g dust). Both are useful
measures for evaluating exposures to lead in dust. Dust-lead loading measures the amount of
lead available to the child and dust-lead concentration measures the source strength of lead in the
dust. A high dust-lead loading might represent a surface containing a large amount of dust at a
low lead concentration or a surface containing a small amount of dust at a high lead
concentration. Both measures have been used to predict blood-lead concentrations and, there is
currently no consensus on which may be the better predictor.
There are two approaches for collecting samples of dust from a surface: wipe and vacuum
sampling. Although dust-lead loadings can be measured using both wipe or vacuum sampling,
dust-lead concentrations can be measured only via vacuum sampling. The interim standards
provided for dust lead in the Interim 403 Guidance Document (EPA, 1994) were defined in terms
of lead loading. The decision there was based, in part, on the wider availability, familiarity, and
lower cost of wipe sampling compared to vacuum sampling. The agency made a policy decision
to define the §403 dust standards in terms of lead loading.
Draft - Do Not Cite or Quote 4 September 27, 1996
-------
Dust can accumulate on multiple surfaces in the home: floors, furniture, window sills,
and window troughs. Levels of lead accumulated in dust on the window sill may not represent
the same level of health concern as lead in floor dust. The Interim 403 Guidance Document
(EPA, 1994) provided standards for lead in dust on floors, window sills, and window troughs.
Technical analyses conducted to support the Risk Assessment for §403 indicated that there is
mixed evidence on whether levels of lead in window trough dust provide additional information
on the impact of lead on childhood blood lead beyond the information contained in floor and
window sill dust. Window troughs are also more difficult to sample than window sills. EPA's
approach for the §403 dust standard is to provide two standards, one for lead in floor dust and
one for lead in window sill dust.
With respect to paint, the Agency recognizes that a principal pathway of childhood
exposure to lead is through the contamination of dust and soil. Activities designed to reduce
exposure to lead in dust and soil will therefore mitigate the contribution of lead-based paint to
childhood health effects and blood lead by the elimination or reduction of the pathway from paint
to dust and soil. By statutory definition, intact lead-based paint is not considered a hazard unless
it is present on accessible, friction, or impact surfaces. However, in cases where painted surfaces
are deteriorated, there is an increased potential for both direct ingestion of paint chips containing
lead and for contamination of the residential dust and soil. As a result, the standard for the lead-
based paint hazard is defined in terms of the condition, location and amount of painted surfaces
that contain lead-based paint.
Although the standards defined by this rule will not specifically require the conduct of
any lead exposure reduction activities, EPA recognizes that they will be used by federal, state,
local, and private entities in their efforts to manage the hazards of lead in paint, dust, and soil.
Therefore, EPA's approach to the Risk Assessment is to document the scientific basis for placing
standards on lead-based paint hazards, lead-contaminated dust and lead-contaminated soil, to
estimate the number of interventions performed to meet the standards defined by this rule and to
characterize the reductions in childhood health risks and blood-lead concentrations achieved
from conducting these interventions.
Draft - Do Not Cite or Quote 5 September 27, 1996
-------
1.2 OBJECTIVES OF RISK ASSESSMENT
The overall objectives of this report are to
1. Document the scientific basis for §403.
This report assesses the risks of childhood exposure, to lead-based paint hazards, lead-
contaminated dust, and lead-contaminated soil. Each component of the risk
assessment is documented hi this report: hazard identification, exposure assessment,
and dose-response assessment.
These individual characterizations are integrated to assess the risks of lead exposures.
to young children and the reductions in these risks expected to take place as a result of
§403. The objectives and activities of each component of the risk assessment are
described further in Section 1.3.
2. Characterize the health risks to young children from specific residential exposures to
lead.
This document estimates risks to young children from specific residential sources of
lead. These sources are: (1) interior and exterior lead-based paint; (2) lead-
contaminated dust, which may contain lead derived from deteriorated interior paint
and tracked- or.blown-in exterior soil; and (3) lead-contaminated soil, which may
contain lead from deteriorated exterior paint, from past leaded-gasoline vehicle
emissions, or from other sources.
This risk assessment focuses on risks to children aged 1-2 years old. Other
populations also certainly face risks from lead exposure including childrenof
other ages, pregnant women, and the general adult population. Characterization
of risks and risk reduction for 1-2 year old children was chosen as representative
of total risk and risk reduction in the interest of keeping a manageable scope and
time frame for the risk assessment. Also, as discussed in Sections 2.3 and 2.4 this
may be the subgroup of children most appropriate for estimation of health effects.
3. Characterize the reduction in risk expected to result from implementation of the §403
rule for a variety of options for the §403 standards.
This report characterizes the incremental risk reductions expected to result from
specific environmental interventions hi specific types of housing units. Because the
§403 rule does not mandate specific action, it was not possible to analyze the risk
reductions associated with specific interventions required by the regulation. Instead,
the Agency's approach in this risk assessment is to characterize the risk reduction
consequences that might occur if broadly defined interventions are undertaken to
reduce exposures to lead in dust, soil, and paint. Intervention activities considered in
Draft-- Do Not Cite or Quote 6 . • September 27, 1996
-------
this report are: cleaning of house dust, maintenance of interior or exterior paint,
encapsulation/abatement of interior or exterior paint, soil cover, and soil removal.
4. Estimate the numbers of children and housing units affected by various options for the
§403 standards.
This report estimates numbers of children and numbers of homes in the nation's
housing stock that would be affected by the rulemaking. The time frame of the risk
assessment is 1997, with the assumption that all actions resulting from the §403 rule
occur within that time frame.
Please note that the objectives of this report do not include the actual selection of the
§403 standards. Actual selection of the standards is a policy decision to be made by the risk
managers. The purpose of this report is to provide relevant information on the scientific basis for
setting the standards and the comparative risk reductions that are expected to result for various
options for the §403 standards.
1.3 ORGANIZATION OF REPORT
The report has seven chapters, including this introduction, background, and summary.
This section summarizes the remaining chapters.
1.3.1 Chapter One - Introduction, Background, and Summary
The questions addressed in Chapter One are: What health risks are addressed in this
report? What is the basis and requirements for addressing these risks? What is EPA's response to
addressing these requirements? What are the objectives of the risk assessment? What are the data
sources and tools used to assess the risks? What are the limitations of the data sources and tools?
The health risks addressed in this report are introduced in Section 1: health risks to
young children from exposures to lead-based paint hazards, lead-contaminated dust, and lead-
contaminated soil. The regulatory basis for addressing these risks, the regulatory requirements
and EPA's response to the requirements are described in Section 1.1. The overall objectives of
the risk assessment are presented in Section 1.2, and the objectives of each component of the risk
assessment are discussed in Section 1.3. Section 1.5 describes the data sources and analysis tools
employed in this risk assessment and discusses their limitations. Section 1.4 provides an
Draft - Do Not Cite or Quote 7 September 27, 1996
-------
overview of the methodology and approach utilized to assess the risks of childhood exposures to
lead-based paint hazards, lead-contaminated dust, and lead-contaminated soil.
1.3.2 Chapter Two - Hazard Identification
The goal of the hazard identification was to answer the following questions: What
measure of body burden to lead should be used in this risk assessment? What are the biologic
effects of exposures to lead? What are the adverse health effects linked to body burdens of lead?
What health endpoints should be quantified in the risk assessment? For what age groups of
children should the health endpoint be evaluated?
Blood-lead concentration, as discussed in Section 2.1, is the most commonly used
measure of lead body burden for linking exposures to lead toxicity and is utilized throughout this
risk assessment. A summary on the mechanisms of lead toxicity is presented in Section 2.2 For
a more comprehensive assessment, the reader is referred to the evaluations conducted by EPA
(EPA, 1986) and ATSDR (ATSDR, 1993) and other literature. Several health effects are linked
to childhood lead body burdens hi Section 2.3, focusing on the adverse neurotoxic effects of
decreased intelligence, developmental delays, behavioral problems, seizures and even coma.
These effects are of particular significance because of the developing nervous system of the
young child and because the central nervous system is the primary target organ for lead toxicity
in children. (ATSDR, 1988). The health risks associated with childhood lead poisoning
encompasses a wide range of exposure levels. IQ point deficit and elevated blood-lead
concentration are selected in Section 2.4 to represent the spectrum of health effects of lead
exposure. This report assess the health risks and blood-lead concentrations of children ages 1-2.
The reasons for selecting this age group of children are discussed in Section 2.4
1.3.3 Chapter Three - Exposure Assessment
Chapter Three examines the following questions: What measures of exposure should be
assessed? Are lead-based paint hazards, lead-contaminated dust and lead-contaminated soil
sources and environmental pathways of lead exposures? Is there evidence of a relationship
between exposures to lead in the environment and blood-lead concentrations? What is the
Draft - Do Not Cite or Quote 8 September 27, 1996
-------
distribution of lead-based paint hazards, lead contaminated dust, and lead-contaminated soil in
the nation's housing? What is the distribution of children's blood-lead concentrations?
Both environmental levels of lead and childhood blood lead are used in this report to
characterize exposures to lead. Pathways and sources of environmental lead exposure, including
lead-based paint, lead contaminated dust, and lead-contaminated soil, are summarized in Section
3.1. The rather extensive evidence on the relationship between childhood blood-lead
concentrations and environmental-lead levels is summarized in Section 3.2. The distribution of
lead in residential dust, soil, and paint in the nation's housing is estimated in Section 3.3. Tables
of results contained in that section show that residential levels of lead can be very high for some
homes and that older homes (built before 1940) tend to have higher levels of residential lead
compared to newer homes (homes built after 1960). Estimating the national distribution of lead
exposures required estimating the number of homes in the nation's housing stock and the number
of children in 1997. The methods employed to determine these two quantities are discussed in
Appendix C. Finally, the national distribution of blood-lead concentrations for children of ages
1-2 is presented in Section 3.4
1.3.4 Chapter Four - Dose-Response Assessment
The primary questions explored in Chapter Four are: What is the dose-response
relationship between exposures to environmental lead and the health effects evaluated in this risk
assessment? What are reasonable ranges of options for the standards to be further examined in
the risk assessment?
The linkage between IQ deficits and lead exposures is usually provided in terms of blood-
lead concentrations rather than environmental-lead levels. Therefore, the solution to the first
question prompted further inspection of two additional questions: What is the dose-response
relationship between environmental lead and childhood blood-lead concentration? What is the
dose-response relationship between childhood blood-lead concentrations and health effects?
Two dose-response models are used in this report to estimate the relationship between
environmental lead and childhood blood-lead concentration. First, a mechanistic model
developed by EPA, EPA's Integrated Exposure, Uptake and Biokinetic Model for Lead (IEUBK
Model), is discussed in Section 4.11. Second, a regression model constructed using the data
Draft - Do Not Cite or Quote 9 September 27, 1996
-------
from the Rochester Lead-in Dust Study (EPI model) is described in Section 4.12. The IEUBK
model does not account for the contribution of lead-based paint on childhood blood-lead
concentrations. The approach implemented for estimating the effect of pica for paint on
childhood blood lead is presented in Section 4.3. Because EPA's approach is to define the dust-
lead standard in terms of a wipe dust-lead loading and because dust samples in the HUD National
Survey were collected via vacuum sampling, an additional question was examined: Can dust-lead
loadings collected using a vacuum sampler be converted to a wipe equivalent dust-lead loading?
Section 4.2 presents equations for converting a dust-lead loading collected via a vacuum sampler
to a wipe equivalent dust-lead loading. Finally, the methods for computing the childhood blood
lead and health risks evaluated in this report from blood-lead concentrations and from
environmental lead exposures are detailed in Section 4.4.
Two approaches were employed to estimate ranges of options for the §403 standards:
regression models and sensitivity/specificity analyses. Table 4.8 in Section 4.5 presents the
results of these analyses.
1.3.5 Chapter Five - Integrated Risk Assessment
This chapter integrates the characterizations from the hazard identification, exposure
assessment, and the dose-response assessment to answer the following questions: What are the
blood-lead concentrations and health risks to children ages 1-2 resulting from residential
exposures to lead? What interventions or abatements might be performed to comply with the
§403 standards? What are the expected reductions in these endpoints resulting from the
rulemaking for various options for the §403 standards. How many children and housing units
will be affected by various options for the §403 standards? What assumptions and data inputs
are likely to have the largest impact on the risk assessment? How sensitive are the results of the
risk assessment to the assumptions and data inputs?
An overview of the methodology employed in the integrated risk assessment is provided
in Section 1.4. Blood-lead concentrations and health risks to children ages 1-2 estimated to exist
in 1997 prior to the passage of §403 (pre-§403) are shown in Table 5-1 in Section 5.1.
Intervention options and environmental levels of residential lead expected after conduct of these
interventions are presented in Tables 5.2 and 5.3, respectively, in Section 5-2. Tables 5.7 to 5.10
Draft - Do Not Cite or Quote 10 September 27, 1996
-------
in Section 5.3 display the numbers of homes affected by various options for the §403 standards
and characterize the childhood blood-lead concentrations and health risks expected to occur in
the post-§403 environment for these options. Figures 5-3 to 5-6 in Section 5.3 graphically
illustrate the impact of changing the levels of lead in the §403 standards on both the number of
homes requiring an intervention and the post-§403 childhood blood-lead concentrations and
health risks.
Table 5-11 in Section 5-4 lists the assumptions and inputs considered to have the greatest
impact on the risk assessment and Tables 5-12 to 5-15 in Section 5-4 present the outcomes of the
sensitivity analyses.
1.3.6 Chapter Six - Validation Studies for the IEUBK Model
This chapter considered one question: Is the IEUBK model an appropriate tool for
predicting a national distribution of blood-lead concentrations based on a national distribution of
environmental-lead levels?
The analyses summarized in Chapter 6 did not find any evidence to conclude that the
IEUBK model is not an appropriate tool for predicting a national distribution of blood lead
levels based on a national distribution of environmental lead levels.
1.3.7 Chapter Seven - Risk Summary
On an overall basis, after each of the activities described in the previous chapters have
been completed, this chapter summarizes the conclusions of the risk assessment.
• The health risks to young children from exposure to lead-based paint hazards, lead-
contaminated dust, and lead-contaminated soil are unacceptably high.
• The health risks to our nation's children can be reduced.
• The standards defined by §403 will help reduce the health risks to our nation's
children.
Draft - Do Not Cite or Quote 11 September 27, 1996
-------
1.4 OVERVIEW OF METHODOLOGY
This section presents the conceptual approach taken by the Agency in the integrated risk
assessment, including the basis for choosing the specific risk assessment tools and data that were
used, as well as the major assumptions that were made in implementing those tools.
1.4.1 General Approach
The general approach taken in this risk assessment is to estimate the health effects
benefits derived from intervention activities expected to be performed in response to candidate
standards, which can then be compared to the costs of those interventions. As with most risk
assessments addressing the health impacts attributable to reduced environmental exposures to
lead, the level of lead hi blood is used as the index of exposure. The predicted decrease in
children's blood lead concentrations is then related to the decrease in expected adverse health
effects.
This assessment focuses on two routes of childhood lead exposure: ingestion of dust and
soil through normal hand-to-mouth activity, and ingestion of paint chips. These are believed to
be the major exposure routes that will be affected by the Title X Lead-Based Paint Program.
(The sources and pathways of lead exposure are summarized in Section 3.1.)
In initially formulating the details of its approach, the Agency decided to rely for the most
part upon existing data on the prevalence of lead in residential environments and on the
relationship between environmental lead and adverse health impacts to children.
The Agency has also used measured empirical data whenever such data existed and was
deemed to be of sufficient quality to provide reasonable estimates of the phenomena of interest
(e.g., the national blood-lead distribution in children, paint conditions and lead levels in
residential dust and soil, etc.). Best technical judgement and predictive models were used to
estimate changes in environmental-lead levels due to intervention activities and consequent
changes in the distribution of blood-lead concentrations in children. Where assumptions are
believed to be tenuous or thought to be critical to any of the conclusions, analyses have been
conducted to demonstrate the sensitivity of the results to those assumptions.
The key concern with predictive models used to relate environmental-lead levels to
children's blood lead is whether the models provide plausible predictions of blood-leads over the
Draft - Do Not Cite or Quote 12 September 27, 1996
-------
range of environmental-lead levels of interest. Two different types of predictive models are
utilized for determining the relationship between environmental lead and blood lead: the IEUBK
model and an EPI model. To help ensure that the IEUBK and EPI models provide plausible
predictions of blood-lead concentrations, EPA developed an approach to calibrate reductions in
blood-lead concentrations predicted from these models to the baseline blood-lead concentrations
developed from empirical data. This approach is summarized as follows:
1. NHANES III was employed to characterize the baseline (pre-§403) distribution of
blood-lead concentration for children aged 1-2 years.
2. Use the environmental-lead levels for HUD National Survey units as input to either
the IEUBK or EPI model to predict the decline in the distribution of blood-lead
concentration for children aged 1-2 expected to result from the §403 rulemaking.
3. Use the baseline distribution of blood-lead concentration and the estimated decline in
the distribution of blood-lead to derive a final post-intervention blood-lead
distribution (post-§403).
Additionally, the models are used only to estimate the "primary prevention" benefits of the
standards (i.e., benefits to children who are bom into units where interventions have already
occurred and, therefore, who have not been previously exposed to higher levels of environmental
lead). The Agency believes that use of the models in these limited circumstances avoids or
minimizes many of the confounding factors that are often encountered when attempting to
predict blood-lead responses related to a wide range of environmental conditions, exposure
situations, and intervention methods.
The process of assessing the risk reductions associated with the §403 standards consists
of a number of discrete steps, as depicted in Figure 1-1. The Agency's approach to each of these
steps is summarized below, and described in detail in the chapters that follow.
Draft - Do Not Cite or Quote 13 September 27, 1996
-------
Current
Environmental
Conditions
Current
Blood-Lead
Distribution
Intervention
Activities
Health
Effects
Assessment
403-Related
Blood-Lead
Reduction
Post-403
Environmental
Conditions
Post-403
Blood-Lead
Distribution
Figure 1-1. §403 Risk Assessment Steps
1.4.2 Current Blood-Lead Distribution
The current "pre-§403" distribution of children's blood lead has been assumed to be that
described by the third National Health and Nutrition Examination Survey (NHANES HI). The
Agency is assuming 1997 as the "pre-§403" time frame since the rule is expected to be
promulgated in that year. NHANES HI reported blood-lead results for children aged 1-2 years
and 3-5 years. Although the results reported were for the years 1988-1989, the Agency is using
these results to represent the current blood-lead distribution. The Agency recognizes that
children's blood leads may be lower today (levels reported in previous NHANES surveys have
shown significant declines over time), but has no information upon which to project additional
declines from the 1988-1989 time frame to the present. The Agency notes that much of the
reduction in children's blood-lead concentrations that occurred from the 1960s through the 1970s
was due to the phase-out of leaded gasoline and activities to halt the use of leaded solder in food
cans. It is likely that the benefits from these types of activities have already been realized and
that the sharp declines in blood-lead concentrations seen in the 1970s and 1980s would not
continue in the period from NHANES in to the present. The results of the NHANES survey are
discussed in Section 3.4.
Draft - Do Not Cite or Quote
14
September 27, 1996
-------
1.4.3 Baseline Environmental Lead Conditions
An estimated baseline distribution of current environmental lead conditions is necessary
in order to predict the post-§403 conditions. The Agency's estimate of current conditions is
based upon the Department of Housing and Urban Development's national survey, which was
conducted in 1989-1990 (HUD National Survey). The HUD National Survey contains data on
age of housing, condition and location of deteriorated lead-based paint, and levels of lead in
interior dust and exterior soil. More recent data from the 1993 American Housing Survey (AHS)
were used to update the information on numbers of housing units obtained from the HUDH
Survey. This survey, conducted every two years by HUD, collects data on the nation's housing
characteristics such as age of home, number of rooms, number and ages of residences, and
income levels of residences. The AHS survey does not collect information on environmental-
lead levels.
Additional methods were used to project the numbers of units in the current housing
stock from the 1993 AHS. Together, this information allowed the Agency to estimate the current
distribution of environmental lead conditions in the nation's housing according to housing age
(pre-1940,1940-1959,1960-1979, and post 1979). This categorization of the housing stock
allows the Agency to account for differences in numbers, types, and costs of intervention
activities that may be associated with the various categories. The details of the Agency's
approach are described in Section 3.3 and Appendix C.
1.4.4 Post-§403 Environmental Lead Conditions
In order to project the distribution of environmental lead conditions that would result
from the promulgation of the §403 standards, the Agency identified a number of specific
intervention activities that were assumed to occur at housing units that would be identified as
posing a hazard under the standards. These intervention activities include both interim controls
and more permanent, abatement measures. For each of the intervention activities, post-
intervention environmental lead conditions were assumed. In general, these intervention
activities are assumed to be media-specific. For example, if a unit was identified as a hazard
based only upon the presence of deteriorated exterior paint, the intervention activity would
address only the exterior paint. An exception is that both interventions dealing with either
Draft - Do Not Cite or Quote 15 September 27, 1996
-------
interior paint or activities involving soil excavation. These two interventions are assumed to be
followed by cleaning of interior dust to HUD clearance levels. This is due to the expectation that
these activities would create high levels of interior leaded dust that would warrant special
cleaning.
Finally, the expected duration of the reduction in environmental-lead levels that result
from each intervention was also estimated. The assumed effectiveness and duration of
interventions are presented in Section 5.2.
1.4.5 Post-§403 Blood-Lead Distribution
For projecting the estimated blood-lead distributions expected to result from the post-
§403 environmental conditions, two modeling approaches have been used. In one, the Agency's
IEUBK model was used, with dust and soil input values taken from the post-§403 environmental
levels described above. The remaining environmental input parameter values were selected to
represent national average levels of exposure to other media. The lEUBK-based approach does
not account for exposures due to the direct ingestion of chips of lead-based paint through the
model itself. Rather, for homes with damaged lead-based paint (where paint chips were assumed
to be available for ingestion), a set percentage of children were assumed to have a specified
increase in blood-lead concentration due to this route of exposure. Where paint stabilization or
removal was the intervention expected to occur in these units, the avoidance of the increase in
blood lead concentration from ingestion of paint chips was assumed.
In the other approach, an empirical model (EPI model) was developed using the data from
the Rochester Lead-in-Dust Study. This model directly accounts for ingestion of paint chips as
well as dust and soil via hand-to-mouth activity.
In combination, these models are used to estimate a range of predicted blood-lead
concentrations that are expected in children exposed to the post-§403 conditions. The details of
the Agency's approach are described in Section 4.1.
The estimated post-§403 blood-lead distribution is used to predict the reductions in health
effects and blood-lead concentrations that are expected to result from implementation of §403.
The predicted post-§403 blood-lead distributions are based on distributions of environmental-
lead levels developed from the HUD National Survey data, predicted declines in the distributions
Draft - Do Not Cite or Quote 16 September 27, 1996
-------
of environmental lead that are expected to result from intervention activities performed under
§403, and either the IEUBK or EPI model. The predicted post-§403 blood-lead distributions are
used herein for the purposes of comparing various regulatory options. The predicted post-§403
blood-lead distributions and associated health effects are NOT MEANT to be an accurate
reflection of what childhood blood-lead concentrations and health effects will be in future years
after §403.
1.4.6 Projected Population of Children
The number of children estimated to occupy the housing categories was calculated based
upon published literature predicting the numbers of children expected to reside in the United
States in 1997 by age group. Numbers of children of ages less than one year, 1 to 2 years, and 3
to 5 years were calculated for each of the four housing group age categories. As discussed in
Section 2.4, the risk assessment is evaluating the health risks and blood-lead concentrations for
children aged 1-2. In addition, the statute defines target housings to be housing constructed
before 1978 that has the possibility of housing a child less than six years of age. Analyses are
conducted in the sensitivity analyses to determine the impact of basing the risk assessments on
children ages 1-2.
The numbers of children occupying housing categories is necessary to estimate the
numbers of these children who will be affected by interventions that are expected to occur at each
housing type. The details of the Agency's approach are described in Section 3.3.2 and
Appendix C.
1.4.7 Reductions in Childhood Blood Lead and Health Effects
Benefits from the estimated blood-lead reductions were quantified for three health
outcomes: decrease in IQ scores, incidence of IQ scores less than 70, and incidence of blood-lead
concentrations greater than 25 |ig/dl. Low IQ scores are associated with lower levels of
educational attainment and lower lifetime earnings. IQ scores less than 70 are indicative of costs
ranging from special education to life-long institutional care. Blood-lead concentrations greater
than 25 ng/dL represent levels at which medical intervention may be necessary. The details of
the Agency's approach are described in Section 4.4.
Draft - Do Not Cite or Quote 17 September 27, 1996
-------
1.5 DATA SOURCES. ANALYSIS TOOLS AND LIMITATIONS
All data sources and research studies referred to in this report are listed in the references
at the end of the report. Data sources and analysis tools are described in more detail in the
individual sections of the report where they are applied. This section highlights the major data
sources, and analysis tools implemented in the risk assessment and points out possible limitations
and data gaps in the risk assessment.
1.5.1 Hazard Identification
There is a wealth of information on adverse biological responses to lead. An exhaustive
compilation or review of all available data on the effects of lead is not within the scope of this
report. For a comprehensive review, the reader is referred to EPA's report, Air Quality Criteria
for Lead (EPA, 1986), and the two ATSDR reports, Toxicological Profile for Lead
(ATSDR.1993) and The Nature and Extent of Lead Poisoning in Children in the United States: A
Report to Congress (ATSDR, 1988). The documented evidence on the adverse biological
responses to lead is one of the major strengths of this risk assessment.
Quantifying health risks to children from exposures to lead requires the selection of
specific endpoints. The neurotoxic and blood lead endpoints selected for this risk assessment
have been used to support previous regulatory decisions. It is possible that if other endpoints
were selected, the baseline risks to lead exposures would be larger and the potential risk
reduction associated with various options might be larger.
1.5.2 Exposure Assessment
The primary data set used in this study to estimate a national distribution of exposures to
lead-based paint hazards, lead-contaminated dust, and lead-contaminated soil is the HUD
National Survey. This study was designed to be a nationally representative study of
environmental-lead in the nation's housing built prior to 1980. Possible limitations of this
dataset are: 1) environmental levels may have declined since the study was conducted in 1988-
1991,2) blood samples were not taken in the study, 3) at most three floor dust samples were
collected in each dwelling unit, 4) dust samples may not have been collected from areas
frequented by children, 5) considerable measurement error may be present in some of the
Draft - Do Not Cite or Quote 18 September 27, 1996
-------
measured dust-lead concentrations due to the small amounts of dust collected for some samples,
and 6) the sample size of 284 homes is a small number for characterizing environmental-lead
levels in our nation's housing stock. A laboratory study was conducted, methodology developed,
and dust-lead concentrations adjusted (revised downwards) to correct for the small weights of
dust collected for some samples. These results are presented in Appendix Z. For comparison
purposes, exposures to lead in paint, dust, and soil from two epidemiologic studies, Baltimore
Repair and Maintenance (R&M) study and the Rochester Lead-in-Dust (Rochester) Study, are
also presented in Section 3.
The HUD National Survey characterized residential environmental-lead levels only in
homes built prior to 1980. Therefore, an additional limitation of the exposure assessment is that
environmental-lead levels in houses built after 1980 had to be inferred.
The primary data set used in the risk assessment to estimate a national distribution of
blood-lead concentration is NHANES III. This is an ongoing survey with a statistical sampling
plan to insure representativeness. Some possible limitations of NHANES m are 1) blood-lead
concentrations may have declined since the study was conducted in 1988-1991, and 2) seasonal
rhythms in blood-lead concentrations are not accounted for in the database. If blood-lead
concentrations have declined since the conduct of NHANES in due to the activities of federal,
state, local, and private agencies, then the baseline risks due to childhood lead exposures will be
overestimated and the reductions in those risks resulting from §403 will also be overestimated.
Because of the serious implication of this data limitation, the sensitivity analysis examines the
impact of a 1% annual decline in childhood blood-lead concentrations on the estimated baseline
risks and the reduction in risks resulting from §403.
The evidence in the literature supporting the existence of a positive relationship between
environmental lead and blood-lead concentration is one of the strengths of this risk assessment.
Eight epidemiological studies were selected to help demonstrate that relationship. However,
quantification of this relationship, as discussed below in Section 1.5.3, is more problematic
The 1993 American Housing Survey (AHS) is the principal data source used for
estimating the nations's housing stock. A possible limitation with the 1993 AHS survey is that it
characterizes the nation's housing in 1993. Other datasets and assumptions, as described in
Section 3.4, were required to extrapolate this information to 1997. Estimates of the number of
Draft - Do Not Cite or Quote 19 September 27, 1996
-------
children who reside in each home were based on information in the 1993 AHS on the average
number of residents per home and Census projections of the number of children per resident
(Day, 1993). A possible limitation in the analyses is the assumption that the average number of
residents per home is the same for all types of homes.
1.5.3 Dose-Response Assessment
Conversions factors were developed to convert dust-lead loadings based on the vacuum
samplers used in the HUD National Survey and the R&M study to wipe equivalent dust-lead
loadings. However, only limited data were available for constructing these conversion factors and
there is considerable uncertainty associated with predictions generated from these equations.
Therefore, the sensitivity analysis examined the impact of the uncertainty in the conversion
equations on the wipe equivalent dust-lead loadings and the predicted blood-lead concentrations
based on these equations.
Currently, the data from just one study, the Rochester Study, was used for development
of the EPI model. Limitations of the EPI model derived from the Rochester Study data concern
the use of data from a single city, with over 84% of the sampled homes built prior to 1940 and
approximately 40% of the sampled children African American. The IEUBK model is a
biological simulation model for predicting a plausible distribution of blood-lead concentrations
based on available information on children's exposure to lead. Analyses conducted in Section 5
are based on the default values for the intake and uptake parameters recommended in the
guidance manual (EPAa, 1994) and the soil- and dust- lead concentrations are being used.
A major limitation of this Risk Assessment is whether or not the model predicted blood-
lead concentrations based on the soil and dust lead concentrations measured in the HUD National
Survey are representative of blood-lead concentrations in exposed children in the nation's
housing. Because of possible limitation in both the EPI model and the IEUBK model for
predicting blood-lead concentrations from environmental-lead levels, results from both models
are presented in this report. Furthermore, analyses are conducted in Section 6 to assess whether
there is any evidence to show that utilizing the IEUBK model in the Risk Assessment is not
appropriate.
Draft - Do Not Cite or Quote 20 September 27, 1996
-------
Several papers were available in the literature for estimating the relationship between
blood lead and IQ decrements. In addition, there were at least three papers that reviewed and
evaluated the data from multiple studies to develop a model based on more than one study (meta
analysis). The relationship between IQ decrement and blood-lead concentration employed in
this Risk Assessment is based on the results of the meta-analysis in (Schwartz, 1994). Heath
risks associated with childhood lead exposures are very sensitive to the estimated IQ points lost
per one ug/dL change in blood lead. The sensitivity analysis examines the impact of other
alternative estimates of this relationship on the Risk Assessment.
1.5.4 Integrated Risk Assessment
The integrated risks assessment employs the datasets mentioned in Section 1.5.2 for
exposure assessment and the dose-response models referred to in Section 1.4.3. In addition, in
order to estimate the impact of §403, generic interventions that might be implemented by federal,
state, local and private agencies to comply with the §403 standards were developed. These
interventions were not meant to simulate every possible nuance of post-§403 activities, but rather
serve as categories of interventions to capture generic activities conducted to meet the standards
for lead-based paint hazards, lead-contaminated dust and lead-contaminated soil. However,
evaluation of the impact of §403 required specification of lead levels in paint, dust and soil after
each one of the interventions. Only limited data were available for estimating the post-
intervention levels of lead. This constitutes one of the major data gaps and limitations for this
risk assessment. The sensitivity analyses examines the impact of changes in the post-intervention
environmental-lead levels on the risk reductions expected to occur as a results of §403. In
addition, the sensitivity analysis examines an alternative approach that does not require
specification of post-intervention environmental-lead levels.
Tables 5.7 to 5.11 in the Integrated Risk Assessment Chapter (Section 5.3) present
estimates of the health and blood-lead effects predicted to exist after implementation of the §403
rule for a variety options for the §403 standards. There is considerable uncertainty in these
numbers due to uncertainty in the
Draft - Do Not Cite or Quote 21 September 27. 1996
-------
1. estimated distribution of environmental-lead levels in the nation's housing,
2. the measures of lead in dust, soil, and paint considered in this risk assessment,
3. the estimated number of homes that will take action in order to comply with the §403
standards,
4. assumed efficacy of interventions conducted to comply with the §403 standards,
5. dose-response models used to predict average (geometric mean) blood-lead
concentrations from environmental-lead levels,
6. dose-response models used to blood-lead concentrations to relate IQ related health
effects,
7. methodology employed to predict a national distribution of post-§403 blood-lead
concentrations and associated health effects.
The intent of the risk characterization is to estimate risks associated with residential
environmental lead exposure and reductions in risk that will be associated with different sets of
§403 standards. It is well known that there is much uncertainty associated with characterizing
specific health risks associated with environmental lead exposure. There are numerous sources
and pathways of lead exposure, particularly for children. There are other significant factors
affecting children's blood-lead concentration and health risks that are not captured in the risk
assessment methodology nor affected by the §403 rule. Such factors include home and personal
cleaning habits, diet and nutritional status, bio-availability of the lead found in residential
environmental media, non-residential exposures, parent's occupation, water, hobbies, and
children's behavioral patterns. Characterization of environmental levels of lead in paint, dust,
and soil are subject to many sources of uncertainty which include chemical analysis, sampling,
spatial, and temporal variability. Use of available data introduces additional uncertainty,
introduced by conversion factors and locality differences. If one accepts a 1.6 GSD as reflective
of variability in blood-lead concentrations in a population of similarly exposed individuals, then
a perfect model to predict a national distribution of blood-lead concentrations would account for
the additional variability (2.05 GSD) observed in the NHANES national distribution. However,
even in this ideal case, the model would only account for approximately 57% of the variance in
Draft - Do Not Cite or Quote 22 September 27, 1996
-------
the national distribution. (Calculated as: [(log 2.05)2 - (logl.6)2 ] / (log 2.05)2.) Modeling is
particularly difficult for children with moderately high blood-lead concentrations (e.g. 10 ug/dL
to 20 jig/dL) and exposures to multiple media. This uncertainty in characterizing the relationship
between environmental levels of lead and health effects was a major factor in EPA's decision to
approach this rule from a risk management perspective. From this perspective, the results of this
risk assessment should not be taken as precise estimates of total health risks associated with a
particular set of standards. Rather the risk assessment should be taken as providing a reasonable
analysis of:
1. The degree to which implementation of a §403 rule that results in environmental
interventions may be expected to reduce health risks associated with lead exposure;
and
2. The relative change in risk reduction when different options for the §403 standards
are chosen.
The fact that all predictions in the modeling process are in terms of geometric means and
associated distributions lends credibility to the analyses. Nevertheless, the substantial
uncertainty in the modeling effort to relate multi-media lead exposure to a national distribution
of health effects is recognized. The sensitivity analyses help characterize the uncertainties in the
analysis.
Draft - Do Not Cite or Quote 23 September 27, 1996
-------
2.0 HAZARD IDENTIFICATION
CHAPTER 2 SUMMARY
This chapter presents information on the toxicity of lead, through a discussion
of how body-lead burden is measured, how lead works in the body, and the
resulting adverse health effects. Two health effects, elevated blood-lead
concentration and IQ point deficits, are identified to represent the spectrum of
adverse health effects resulting from lead exposure. These representative effects
are used in the integrated risk analysis to assess the potential benefits of the
proposed $403 rule.
Blood-lead concentration is a commonly used measure of body lead burden.
An extensive body of research relates health effects of lead exposure to blood-lead
concentration. For example, lead-related reductions in intelligence, impaired
hearing acuity, and interference with vitamin D metabolism have been documented
in children at blood-lead concentrations as low as 10 to 15 ug/dL, with no
apparent threshold. At higher exposure levels, these effects become more
pronounced and other adverse health effects are observed in a broader range of
body systems. Increased blood pressure, delayed reaction times, anemia, and
kidney disease may become apparent at blood-lead concentrations between 20 and
40 ug/dL. Symptoms of very severe lead poisoning, such as kidney failure,
abdominal pain, nausea and vomiting, and pronounced mental retardation, can
occur at blood-lead levels as low as 60 ug/dL. At even higher levels, convulsions,
coma, and death may result.
Adverse health effects (e.g., IQ deficits, neurological dysfunction) in children have long
been associated with elevated lead exposure. The concentration of lead in whole blood, usually
expressed in micrograms of lead per deciliter of whole blood (ug/dL), is the most common
measure of a person's internal exposure to lead. Blood lead can be measured easily and
accurately as compared to alternative physiological measures such as lead in bone or hair, and is
more directly relevant to the assessment of exposure than are environmental measures (CDC,
1991; EPA, 1986). Centers for Disease Control guidelines on childhood lead poisoning
prevention have traditionally been and currently are defined in terms of blood-lead concentration
(CDC, 1991). While lead exposure in adulthood is a concern, fetuses, infants, and young
children are the population most at risk from exposure to lead (EPA, 1986; ATSDR, 1993). This
intensified risk is due to children's increased oral activity (e.g., hand-to-mouth behavior) and
ability to absorb lead, coupled with the susceptibility of their rapidly developing central nervous
Draft - Do Not Cite or Quote 24 September 27, 1996
-------
systems (Bellinger, 1995; Goyer, 1993). The central nervous system is the primary target organ
of lead, though lead is stored throughout the body (e.g., bones, tissues, blood) (ATSDR, 1993).
The blood-lead concentration at which health professionals express concern has
decreased over time. At high blood-lead concentrations (i.e., lead poisoning), lead exposure can
cause coma, convulsions, and death. At lower concentrations, observed adverse effects from lead
exposure in young children include reduced intelligence, reading and learning disabilities,
impaired hearing, and slowed growth (CDC 1991). The phase-out of leaded gasoline (halting the
fallout from leaded gasoline emissions), the restriction on the residential use of lead-based paints,
and restrictions against lead solder in cans and water systems have greatly reduced blood-lead
concentrations nationwide (EPA, 1986; Brody et al., 1994; Pirkle et al., 1994; Section 3.0).
Simultaneously, however, further research suggested that levels previously thought safe were, in
fact, hazardous. In the first half of the 20th Century, medical care providers were concerned
about childhood blood-lead levels above 80 (ig/dL; by the 1960s the level of concern was
reduced to 60 |ig/dL and above; by the 1970s the level of concern was at 40 ug/dL; and by the
1980s the level was lowered to 25 ng/dL (CDC, 1994). In 1991, the CDC reduced its blood-lead
concentration community level of concern to 10 ug/dL (CDC, 1991). This was in response to
scientific evidence appearing in the immediately preceding years that adverse health effects occur
at blood-lead concentrations at least as low as 10 )ig/dL. hi fact, no discernable threshold in the
relationship between adverse health effects and blood-lead concentrations has been identified
(ATSDR, 1993).
Despite the nationwide reductions hi blood-lead concentrations over time, one effect of
reducing the blood-lead concentration of concern is increasing the number of children expected
to have blood-lead concentrations above that level. The National Health and Nutrition
Examination Surveys (NHANES) trace the health and nutritional status of the U. S. population.
The results of the most recent NHANES (NHANES IE, phase 1,1988-1991) demonstrate that
significant declines in childhood blood-lead concentrations have occurred in recent years, but
that significant numbers of children's blood-lead concentrations remain above 10 |ig/dL. The
geometric mean blood-lead concentration, reported in NHANES m - phase 1, for children
between the ages of one and five years was 3.6 iig/dL (Pirkle et al., 1994). This is 11.4 ug/dL
lower than the estimate reported in NHANES n (conducted between 1976 and 1980) of 15 ug/dL
Draft - Do Not Cite or Quote 25 September 27, 1996
-------
(Pirkle et al., 1994). Despite the lower blood-lead concentration reported in NHANES ffl,
approximately 1.7 million children between one and five years of age are estimated to have
blood-lead concentrations above 10 ug/dL (Brody et al., 1994).
There is an extensive body of literature relating health effects of lead exposure to
measures of body-lead burden. This literature is summarized in several government reports,
including
• Air Quality Criteria for Lead (EPA, 1986)
• The Nature and Extent of Lead Poisoning in Children in the United States: A Report
to Congress (ATSDR, 1988a)
• Air Quality Criteria for Lead: Supplement to the 1986 Addendum (EPA, 1990)
• Comprehensive and Workable Plan for the Abatement of Lead-Based Paint in
Privately Owned Housing (HUD, 1990)
• Preventing Lead Poisoning hi Young Children (CDC, 1991)
• Toxicological Profile for Lead (ATSDR, 1993)
These sources were used extensively in the next sections, although the original sources are cited
for specific results whenever possible. Section 2.1 discusses commonly used measures of body
lead burden. The mechanisms of lead toxicity are described in Section 2.2. The scientific
evidence on resulting health effects is presented in Section 2.3. Finally, specific health effects
are selected in Section 2.4 for use in this risk assessment.
2.1 MEASURES OF BODY-LEAD BURDEN
For purposes of risk assessment, it would be ideal to precisely relate particular health
outcomes, such as decreased learning deficits or increased motor coordination, to environmental
lead levels. Unfortunately, most studies of lead in the environment use measures of body-lead
burden, such as blood-lead concentration, as biomarkers of lead exposure. Similarly, studies that
assess lead hazard interventions tend to use blood-lead concentration to measure intervention
effectiveness (EPA, 1995). There is extensive evidence that body-lead burden is associated with
Draft - Do Not Cite or Quote 26 September 27, 1996
-------
lead levels in environmental media (EPA, 1986; CDC, 1991). In addition, there is an extensive
body of literature relating health effects of lead exposure to measures of body-lead burden.
The most common screening and diagnostic measure of body-lead burden is blood-lead
concentration. Other measures include lead in bones, teeth, and hair. Approximately 94% of the
total body burden of lead in adults (73% in children) is found in the bones (Barry, 1975). Blood-
lead concentration has the advantage of being easily and inexpensively measured. A
disadvantage, however, is that it reflects a mixture of recent and past exposure. Bone-lead levels
are more reflective of cumulative exposure to lead. The half-life of lead in the blood of adults is
approximately one month (Griffin, et al., 1975b; Rabinowitz, et al., 1976), whereas the half-life
of lead in bone is several decades (ATSDR, 1993). Because lead cycles between the blood and
bone, a single blood lead measurement cannot distinguish between low-level chronic exposure
and high-level acute exposure (ATSDR, 1993). Because of recycling from the bone, both types
of exposure could result in the same blood-lead concentration. Despite these limitations, blood-
lead concentration remains the one readily accessible measure that can demonstrate in a relative
way the relationship of various effects to increases in lead exposure (ATSDR, 1993).
Of the other measures, bone and tooth lead may be used to measure cumulative exposure
to lead, while hair lead is an indicator of more recent exposure. Bone-lead content may be
measured by x-ray fluorescence (XRF), although the reliability of this method is questioned by
many researchers, especially at levels below 10 ppm (Wedeen, 1988). Since teeth can store lead
up to the time of shedding or extraction, levels of lead in shed teeth have been used as an
indicator of lead exposure in some studies (Smith, et al., 1983; Pocock, et al., 1989; Bergomi, et
al., 1989; Needleman, et al., 1990). Hair lead has been used as an indicator for intermediate
exposure (2 months) in children (Wilhelm, et al., 1989). However, artificial hair treatments such
as dyeing, bleaching, or permanents, can invalidate metal analysis of hair (Wilhelm, et al., 1989)
and external surface contamination problems are such that it is difficult to differentiate between
externally and internally deposited lead (EPA, 1986). After consideration of the disadvantages of
using bone, tooth, and hair lead as biomarkers of exposure, most researchers in the area of lead
exposure conclude that blood lead remains the most efficient and useful way to assess body lead
burden.
Draft - Do Not Cite or Quote 27 September 27, 1996
-------
Physiological changes, such as alterations in heme (the component of blood that contains
iron) synthesis, that are known to implicate lead exposure may also be used as biomarkers of
exposure. Generally, blood-lead levels are determined concurrently with these physiological
biomarkers. Interference with heme synthesis following lead exposure can lead to a reduction of
hemoglobin concentration in blood (Bernard and Becker, 1988) and an increase in urinary
coproporphyrin (EPA, 1986). A relationship between 8-aminolevulinate-dehydratase (ALAD)
activity measured in erythrocytes and blood-lead levels of 5 to 95 ng/dL has been observed
(Hernberg, et al., 1970). Although 6-aminolevulinate (ALA), a potential neurotoxin that
accumulates from decreased ALAD activity, can be detected in urine when blood-lead levels are
above 35 ^ig/dL in adults (25 to 75 jig/dL) (Roels and Lauwerys, 1987), ALA in urine is not
considered as sensitive a measure of current lead exposure as ALAD activity (Hernberg, et al.,
1970). The concentration of erythrocyte protoporphyrin (EP) rises above background at blood-
lead levels of 25 to 30 ug/dL and there is an association between blood-lead levels and EP (CDC,
1985; Hernberg, et al., 1970). Determination of EP in blood is used as an indicator of past
chronic exposure, since elevated EP reflects average blood-lead levels for about 4 months
following the exposure (Janin, et al., 1985). In the case of each of these physiological measures,
other conditions may produce similar effects, leading to false positive outcomes when these
measures are used alone as biomarkers for body lead burden.
2.2 MECHANISMS OF LEAD TOXICITY
Lead is a very dynamic compound with a wide spectrum of effects in humans. Its effects
are seen at the subcellular level as well as at the level of general function that encompasses all
systems in the body. The subcellular mechanisms of action, followed by a discussion of the
neurotoxic effects and the heme effects of lead poisoning, are included in this chapter. Wherever
possible, mechanisms included in the subcellular mechanisms section are related to the specific
effects of lead on the nervous system and the blood. There remain many gaps, however, in the
information needed to explain the varied mechanisms of lead in the body in different organs.
Lead has been recognized as a naturally occurring element since the beginning of
civilization. Today, the major environmental sources of lead are paint, auto exhaust, food, dust,
soil, and water. "Inorganic lead" includes the metallic form of lead, its salts and oxides.
Draft - Do Not Cite or Quote 28 September 27, 1996
-------
Inorganic lead primarily enters the body through inhalation (breathing in air) and ingestion
(eating or drinking). Inorganic lead is absorbed, distributed throughout the body, and removed
from the body (excreted). It is not broken down to simpler compounds (metabolized) in the
body. "Organic lead," which is formed by lead in combination with an alkyl group (carbon and
hydrogen), is found primarily in gasoline as tetraethyl lead. Organic lead enters the body through
inhalation and can pass through the skin because of its properties. It is broken down
(metabolized) in the liver and then removed from the body.
The rate at which lead is absorbed into the body depends on the chemical and physical
properties of the form of lead and on the physiological characteristics of the exposed person
(nutritional status, age, etc). When lead is inhaled it becomes deposited in the lower respiratory
tract and is completely absorbed. The amount of lead absorbed from the gastrointestinal tract of
adults is 10-15% of the amount ingested. In pregnant women and children, the amount absorbed
can increase to 50%. The amount absorbed greatly increases during compromised nutritional
status of the individual or during periods of iron or calcium deficiency. Once lead is absorbed it
enters the bloodstream and is dispersed throughout the body where it is distributed between the
blood, the mineralizing tissue (bone and teeth), and soft tissue (kidney, bone marrow, liver, and
brain).
The lead in the mineralizing tissues accumulates in two different areas of the bone: 1) an
area where lead can quickly be exchanged in the blood and 2) a more stable area where lead is
stored long-term. This stable area for storing lead can pose a special risk because, when the body
is under stress such as during pregnancy, lactation, or chronic disease, this lead may be
'mobilized,1 thereby increasing the blood lead. Because of this more stable compartment of lead
in bone, significant declines in blood lead can require months or years to occur after exposure has
been stopped.
Of the lead found in the blood, 99% is associated with the red blood cells (erythrocytes).
The remaining 1% is in the plasma, where it can be released to tissues. The blood lead not
retained is either excreted by the kidneys, or through bile, enters the gastrointestinal tract. In
exposures to a single dose of lead, one-half of the lead from the original exposure remains in the
blood about 25 days after exposure, in soft tissues about 40 days, and in stable bone more than 25
years. Consequently, after a single exposure a person's blood-lead concentration may begin to
Draft - Do Not Cite or Quote 29 September 27. 1996
-------
return to normal, but the total body burden (amount of lead in the body) may still be elevated.
Lead exposure does not need to be acute for lead poisoning to occur. It is the total body burden,
accumulated over a lifetime, that is related to the risk of adverse health effects.
2.2.1 Physiological Mechanisms
Lead affects the cellular organdie structures and processes as well as the general
functioning of the body, which results in neurotoxicity, hematological effects, possible
hypertension, kidney damage, reproductive dysfunction, developmental abnormalities, etc. These
effects are described in Section 2.3.
The biological basis of lead toxicity is its ability to bind (attach) to substances crucial to
various physiological functions. In this way, lead may interfere with cell functions by competing
with native (substances normally found in the body), essential metals for binding sites, inhibiting
enzyme activity, and inhibiting or otherwise altering essential ion transport. These effects are
modulated by the stability of the binding site, how the lead is distributed in the body, and the
differences in biochemical organization of different cells and organs. As a result, there is no
single, well-defined mechanism that explains how lead works in all tissues in humans.
Studies on the mechanism of lead toxicity at the cellular level appear to implicate the
mitochondria (energy powerhouse in the cell) and membranes (both cellular and intracellular) as
primary targets for lead (EPA, 1986).
Lead effects mitochondria in numerous ways. These include structural changes and
marked disturbances in mitochondria! function within the cell, especially energy metabolism and
ion transport. These effects are associated with the accumulation of lead within the
mitochondria. Structural changes include the swelling of mitochondria, and the distortion and
loss of the small inner folds, called cristae, which carry many enzymes. The uncoupling of
energy metabolism, inhibition of cellular respiration, and altered activities of intracellular
calcium in mitochondria due to lead have been demonstrated in many studies. These
investigations have particularly been concerned with the effects of lead on the brain, heme
synthesis, and erythropoiesis (formation and production of red blood cells).
Lead also effects cellular and intracellular membranes and the mitochondria by altering
ion transport, particularly calcium. This leads to the inhibition of enzymes and interferes with
Draft - Do Not Cite or Quote 30 September 27. 1996
-------
normal transport systems. The overall impact of these effects is to disturb the development and
functioning of many of the major organ systems in the body, particularly the central nervous
system, resulting in significant adverse health effects.
2.2.2 Neurotoxic Effects of Lead
The data assessing the neurotoxicological mechanisms of lead provide limited
information about how lead affects the nervous system. For over a decade the hippocampus was
thought to be the principal target of lead in the brain. The hippocampus was selected because:
1) the hippocampus contains relatively high concentrations of zinc, and zinc-dependent functions
may be sensitive to lead, 2) the hippocampus contains a dense plexus of cholinergic fibers that
are affected by lead exposure, and 3) the hippocampus is functionally related to behaviors
involving memory and learning (Petit, 1983). More recent investigations have shown that other
areas, particularly the mesolimbic system (Moresco, 1988, Lasley, 1988) where low levels of
lead have been found, cannot be excluded. Continuing research may help to determine which
areas of the brain have an affinity for lead.
Lead is an ion and the hypothesis for its molecular mechanism has been based on its
interaction with other physiologically important ions like calcium (Pounds, 1984) and zinc
(Fowler, 1989). In addition, the activity of protein kinase C (Markovac, 1988) is effected by lead
due to its zinc binding sites. Calcium and zinc are found in a multitude of sites and reactions
throughout the body. It is thought, however, that there must be other, more specific events that
determine how lead acts at a defined site. How this occurs is unknown at this time.
A number of scientists working on the neurotoxicity of lead discuss the proposed
mechanisms of how lead affects the nervous system. Among these scientists, Silbergeld (1992)
and Bellinger (1995) both discuss possible mechanisms for lead neurotoxicity in the context of
neurodevelopmental (effects occurring during development of the nervous system), and
neuropharmacological (interaction of lead with cells of the brain) effects.
Neurodevelopmental Effects: During development, the central nervous system (the
brain and spinal cord) goes through a number of programmed changes involving the overall
growth in cell numbers, size in the organ, and proliferation and outgrowth of cells that establish
connections between cells. Many factors regulate these processes, including growth factors,
Draft - Do Not Cite or Quote 31 September 27. 1996
-------
neurotransmitters functioning as trophic agents, and glycoprotein cell adhesion molecules
(Jacobson, 1990).
One of the potential mechanisms for lead's effect on the developing brain has been
investigated by Goldstein (1990), who suggests that the immature endothelial cells forming the
capillaries of the developing brain are less resistant to the effects of lead than are capillaries from
mature brains. As a result, blood carrying lead to the brain may easily pass into the newly
forming compartments of the brain and effect many parts of this developing organ. In
comparison, the capillaries of adults are developed and help to prevent the passage of ions like
lead across the blood-brain barrier. It has been suggested that lead may affect the differentiation
of capillary endothelial cells from the fetal brain in a similar way to its effects on neurons
undergoing development (Dressier and Goldstein, 1991). This provides a neurotoxicological
basis for the observed increased risk to pregnant women, infants, and young children of exposure
to lead.
Silbergeld (1990) found that exposure of fetal animals to lead affects both regional
growth and neuron-specific differentiation/synaptogenesis (development of synapses) in the
central nervous system. Of these, synaptogenesis appears to be the more sensitive (Silbergeld,
1991; Regan, 1989), and lead is thought to interfere with the normal development of synapses in
the brain. A synapse is a junction where the axon of one neuronal cell (or neuron) terminates
with the dendrite of another neuron. Nerve impulses move from one nerve cell to another by
traveling through the synapse. The normally functioning brain seems to exhibit a deletion of
synapses that are unused. Those synapses which are frequently used are kept and strengthened.
Goldstein (1990,1992) suggests that lead may disrupt, or delay, this normal synaptic
developmental process and that perhaps the resulting connections in the brain are "poorly
chosen," leading to functional impairment in the brain. Although this hypothesis is speculative,
lead's ability to facilitate the unstimulated release or prevent the stimulated release of
neurotransmitters, which are important for the morphological organization of neurons, may be
related to how neurons are chosen to survive (Audesirk, 1985). This may result in a nervous
system that appears normal but in which cell to cell connections are not normal. These
abnormalities then may be translated into the kind of neurobehavioral deficits which result in
cognitive and behavioral deficits.
Draft - Do Not Cite or Quote 32 September 27. 1996
-------
At the age of two the synaptic density and number of synapses per neuron peaks in the
frontal cortex in humans. It is suggested that this area of the brain is an important target for lead
(Huttenlocher, 1979). The age of two is also important from the standpoint of scientists
evaluating lead toxicity outcome. Many studies have been conducted at this age because blood-
lead concentrations tend to peak at age two, children are more cooperative for assessment at this
age, hand to mouth activity is greatest, and level of cognitive ability is sufficiently developed.
Neuropharmacological Effects of Lead: Lead may also act as a neuropharmacological
toxicant in the brain (Silbergeld, 1992; Bellinger, 1995). Silbergeld (1992) discusses how lead
interferes with the synaptic mechanisms of the release of neurotransmitters and signal
transduction. These effects are due to the presence of lead in the synapse. Theoretically, these
effects are reversible if lead is removed. However, exposure to lead for a long time may result in
permanent alteration in cellular responsiveness at pre- and post-synaptic levels. These
pharmacologic effects may include the effects of lead to facilitate transmitter release, modulate
ion conductance and, as a result, alter the electrophysiological output of the neuron.
Disruption of ion transport at membranes may be the mechanism by which lead produces
its pharmacologic effects in the nervous system. Lead can also substitute for calcium and zinc
in ion transport events at the synapse. These events include sodium channels, calcium
channels, calcium-binding modulators like calmodulin, messengers like adenyl cyclase and
protein kinase C (Dressier, 1991). Lead may affect ion channels by occupying zinc-binding sites
and preventing ion movements (Alkodon, 1990).
At the neuron, lead seems to be more disruptive inside than outside. For example, inside
the neuron, mitochondria! release of calcium is quite sensitive to lead (Silbergeld, 1975). Protein
kinase C, which is very sensitive to lead, modulates receptor currents affecting long-term
potentiation and other forms of synaptic response that may underlie learning and memory
(Markovac, 1988). Dopamine sensitive adenyl cyclase, Na,K-ATPase, is also relatively sensitive
to lead (Ewers, 1980, Fox, 1991).
Outside the neuron, neurotransmitter release or transmitter-gated ion channels are
sensitive to lead at higher concentrations (Alkondon, 1990; Minnema, 1986; Kostial, 1957;
Silbergeld, 19741; Audesirk, 1985). If lead is kept out of the neuron, relatively high levels of
lead are required to affect function. However, lead can enter certain neurons under the right
Draft - Do Not Cite or Quote 33 September 27, 1996
-------
conditions (Sulzer, 1987; Silbergeld, 1977). Under these conditions mitochondria, protein
kinase C, and other intracellular sites become accessible. The differential ability to prevent lead
entry may be an important protective mechanism to prevent neurotoxicity. There has been
speculation of a lead-binding protein in humans (DuVal, 1989) which may serve to concentrate
and transport lead to certain parts of the brain.
Peripheral Neuropathy: Lead induces degeneration of the protective Schwann cells in
the motor neurons of the peripheral nervous system, which causes segmental loss of the myelin
covering of the neuron and possible neuron degeneration (Fullerton, 1966). Dyck (1980) and
Windebank (1980) suggest that lead induces a breakdown in the blood-nerve barrier, allowing
lead and fluid to enter the endoneurium, and disruption of myelin membranes. The degeneration
of sciatic and tibial nerve roots is also possible. Sensory nerves are less sensitive to lead than
motor nerves. Peripheral neuropathy is usually present only after prolonged high exposure to
lead. Peripheral neuropathy may be reversible or permanent depending on the severity of
exposure. Motor nerve dysfunction has been assessed clinically by the electrophysiologic
measurement of nerve conduction velocities and shown to occur at blood-lead levels as low as 40
ug/dL.
To summarize, the mechanisms for lead neurotoxicity are not well understood. Several
mechanisms have been proposed which seek to explain why children are more sensitive than
adults to lead and how lead acts molecularly to effect the nervous system. Because most of the
molecular events in the nervous system are also found throughout the rest of the body, it is
difficult to explain why the nervous system is the most sensitive system in the body to lead.
2.2.3 Hematologic Effects of Lead
Red blood cells, which carry oxygen to body tissues, develop in the bone marrow of the
body. Lead has adverse affects on heme synthesis (the formation of hemoglobin), which can
result in anemia, and red blood cell formation, which can result in decreased life span of these
cells.
Hemoglobin constitutes 90% of red blood cells. Hemoglobin consists of globin protein
and heme, which is a metal complex consisting of an iron atom in the center of a porphyrin
structure and provides the red color to hemoglobin.
Draft - Do Not Cite or Quote 34 September 27, 1996
-------
Effect of Lead on Heme Synthesis: When an individual is exposed, lead quickly
reaches the blood, circulates in the body, and enters different tissues including the bone marrow,
where it can have an impact on various parts of the formation of heme. The process of heme
biosynthesis starts with glycine and succinyl-coenzyme A, proceeds through the formation of a
protein entity called protoporphyrin DC, and culminates with the insertion of iron into the
porphyrin ring to form heme. In addition to being a constituent of hemoglobin, heme is found in
many hemoproteins, such as myoglobin, the P-450 component of the mixed-function oxygenase
system, and the cytochromes of cellular energetics. Therefore, disturbing heme biosynthesis by
exposure to lead poses the potential for multiple-organ toxicity.
Lead's effects on the heme synthesis pathway are: 1) stimulation of mitochondria!
5-aminolevulenic acid synthetase (ALA-S), which mediates formation of 6-aminolevulinic acid
(ALA), 2) direct inhibition of the cytosolic enzyme, 6-aminolevulenic acid dehydrase (ALA-D)
which catalyzes formation of porphobilinogen from ALA, and 3) inhibition of insertion of iron
into protoporphyrin DC to form heme, a process mediated by ferrochelatase.
Lead's mechanism of action seems to be due to its effect on cellular mitochondria. Lead
enters the mitochondria of the cell where it impairs mitochondria! function and exercises many of
its effects on the production of heme. In the mitochondria, lead increases the activity of the
enzyme ALA-S, which increases the amount of ALA that is formed. Lead, in the cytosol of the
cell, also decreases the activity of ALA-D, an enzyme which catalyzes reactions of ALA to form
other molecules in heme biosynthesis. The result is an increase in the level of ALA.
Ferrochelatase, an enzyme also found in the mitochondria, catalyzes the incorporation of
iron into protoporphyrin DC to form heme. Lead tends to inhibit ferrochelatase from
incorporating the iron into the protoporphyrin ring, thereby preventing the formation of heme.
Instead, there is an increase in erythrocyte protoporphyrin in the red blood cells. Erythrocyte
protoporphyrin (EP) can be measured in blood as zinc protoporphyrin (ZPP) or free erythrocyte
protoporphyrin (FEP).
Effect of Lead on Hemoglobin Production and Red Cell Formation: As described
above, heme production is decreased by lead. Heme production regulates globin production so
that globin production is also decreased, resulting in the decreased production of hemoglobin.
These effects can lead to anemia (reduction in circulating red blood cell mass). Lead can induce
Draft - Do Not Cite or Quote 35 September 27, 1996
-------
two types of anemia. Acute high-level lead poisoning has been associated with hemolytic
(excessive red blood cell destruction) anemia. In chronic lead poisoning, lead induces a
hypochromic (light colored red cells) normocytic (normal size red cells) anemia by both
interfering with normal hemoglobin production and by diminishing red cell survival (cells are
incapable of functioning in a normal manner and the life span is shortened).
The molecular mechanism for the diminished red cell life span is thought to be due to
lead's inhibition of the enzymes (Na+, K+)-ATPase and pyrimidine-5-nucleotidase (Py-5-N).
With enzyme inhibition there is irreversible loss of potassium ion from the cell with undisturbed
input of sodium into the cell, resulting in a relative increase in sodium. Because the cells tend to
shrink, there is an increase in sodium concentration which results in increased mechanical
fragility and cell lysis (destruction of cells through rupture of cell membrane) in lead-induced
anemia. It has also been suggested that enzymes (like Py-5-N and glucose-6-phosphate
dehydrogenase) that help stabilize the red cell membrane may be effected by lead.
2.3 HEALTH EFFECTS OF LEAD EXPOSURE
Lead is a powerful toxicant with no known beneficial purpose in the human body
(ATSDR, 1988a). The toxic effects of lead are seen primarily in the central nervous system, but
virtually all parts of the body can be damaged at high exposure levels. Acute lead poisoning,
associated with blood-lead levels above 70 ug/dL, causes abdominal pains, vomiting, and
diarrhea. Without proper treatment, lead poisoning can result in convulsions, coma, and even
death. At lower exposure levels, subtle neurological effects are of most concern.
Although occupational exposure to lead is dangerous and the subsequent health effects
are well-documented, infants and young children are more at risk from lead exposure than are
adults, as their neurological systems are developing and are more vulnerable to damage. At the
same time, their frequent hand-to-mouth activities bring them into greater contact with lead in
the environment and their bodies absorb a larger percentage of ingested lead than do those of
adults. The increased risk of lead exposure appears to be most evident at age 2 (Clark, 1985;
Goyer, 1993), and studies have shown a strong association between blood-lead concentration
measured at age 2 and IQ scores later in life (Bellinger, 1992; Schwartz, 1994; Pocock, 1994).
Because lead is readily transferred across the placenta, a developing fetus is at risk for lead
Draft - Do Not Cite or Quote 36 September 27, 1996
-------
exposure and toxicity. For this reason, women of childbearing age are also a population of
concern.
Over time, the lead exposure levels at which adverse health effects are reported have
declined dramatically. Although cases of severe lead poisoning still occur, recent research has
focused on the effects of chronic, low-level lead exposure. At blood-lead levels as low as 10 to
15 |ig/dL, researchers have documented slower reaction times, reductions in intelligence and
short-term memory, other neurobehavioral deficits, adverse effects on heme (iron in blood)
biosynthesis, and on vitamin D and calcium metabolism. In addition, longitudinal studies have
shown reductions in gestational age and birth weight associated with prenatal blood-lead levels
of 10 to 15 ug/dL. At or above 40 ug/dL, children may experience reduced hemoglobin, the
accumulation of a potential neurotoxicant known as ALA, and mild anemia. Very severe lead
poisoning involving symptoms such as kidney failure, gastrointestinal problems, coma,
convulsions, seizures, encephalopathy, and pronounced mental retardation, can occur at blood-
lead levels of 60 ug/dL or higher. Specific health effects of lead exposure, the blood-lead levels
at which these effects have been observed, and the scientific literature in which the effects were
reported are summarized in Table B-l in Appendix B. This table is reproduced from the
Toxicological Profile for Lead (ATSDR, 1993).
2.3.1 Neurological Effects of Lead
The most severe neurological effect of lead in adults and children is lead encephalopathy,
a general term used to describe various diseases that affect brain function. Early symptoms
include dullness, irritability, poor attention span, headache, muscular tremor, loss of memory,
and hallucinations. The condition may worsen, sometimes abruptly, to delirium, convulsions,
paralysis, coma, and death (Kumar, et al., 1987). While physical symptoms of lead poisoning
can be treated, the effects on the central nervous system may be irreversible. Long-lasting
impacts on intelligence, motor control, hearing, and emotional development of children have
been documented at levels of lead in the body that are not associated with obvious symptoms and
were once thought to be safe.
Effects on Adults: Occupational exposure to lead has often been associated with
subjective signs of neurotoxicity. The literature contains numerous case reports and small cohort
Draft - Do Not Cite or Quote 37 September 27, 1996
-------
studies that describe symptoms such as malaise, forgetfulness, irritability, lethargy, headache,
fatigue, and dizziness at blood-lead levels that range from 40 to 120 jig/dL, following acute,
intermediate, and chronic-duration occupational exposure to lead (Awad, et al., 1986; Baker,
et al., 1979; Haenninen, et al., 1979; Holness and Nethercott, 1988; Marino, et al., 1989; Matte,
et al., 1989; Pagliuca, et al., 1990; Pasternak, et al., 1989; Pollock and Ibels, 1986; Schneitzer,
etal., 1990).
Neurobehavioral testing has revealed effects of lead in adults at blood-lead levels
between 40 and 80 ug/dL, well below the levels that cause encephalopathy (120 |ig/dL - EPA,
1986). Disturbances in oculomotor function (saccadic eye movements) have been observed in
lead workers with mean blood-lead levels of 57 to 61 ng/dL (Baloh, et al., 1979; Spivey, et al.,
1980; Glickman, et al., 1984). Disturbances in reaction time, visual motor performance, hand
dexterity, IQ test and cognitive performance, nervousness, mood, or coping ability were observed
in workers with blood-lead levels of 50 to 80 |ig/dL (Arnvig, et al., 1980; Haenninen, et al.,
1978; Hogstedt, et al., 1983; Mantere, et al., 1982; Valciukas, et al., 1978). However, there is
some evidence to the contrary. No neurobehavioral effects were observed in a study of 288
randomly selected lead workers with mean blood lead of 40.1 U£/dL, compared to 181
demographically similar controls with mean blood lead of 7.2 ug/dL (Ryan, et al., 1987).
Numerous studies measure the conduction velocity of electrically stimulated nerves in the
arm or leg of occupationally exposed workers. Nerve conduction velocity (NCV) is used to
measure slowed reaction times associated with lead exposure and is considered a sensitive
indicator of lead toxicity. Studies indicate that NCV effects occur in adults at blood-lead levels
below 70 fig/dL, possibly as low as 30 ug/dL. Decreased NCV has been observed in both
prospective and cross-sectional studies (Seppalainen, et al., 1983; Rosen, et al., 1983; Treibig, et
al., 1984; Araki, et al., 1980). There is some evidence indicating that changes in NCV associated
with lead exposure may be transient (Araki, et al., 1980; Muijser, et al., 1987).
Effects on Children: High-level lead exposure produces encephalopathy in children,
starting at approximately 80 to 100 ng/dL (NAS, 1972; Bradley and Baumgartner, 1958; Bradley,
et al., 1956; Gant, 1938; Rummo, et al., 1979; Smith, et al., 1983; EPA, 1986). However, low-
level exposure also may result in long-lasting impacts on intelligence, motor control, hearing,
and neurobehavioral development of children.
Draft - Do Not Cite or Quote 38 September 27. 1996
-------
Results are available from four large-scale, longitudal studies conducted in Boston,
Cincinnati, Cleveland, and Port Pine, Australia. These studies indicate that disturbances in early
neurobehavioral development occur at exposure levels that until recently were considered safe, or
even normal. In the Boston study, 4-8 point differences in performance on the Bayley Mental
Development Index (MDI) were reported at 6,12,18, and 24 months, after adjusting for other
covariates, between low (prenatal mean of 1.8 iig/dL) and high (prenatal mean of 14.6 |ig/dL)
exposure infants (Bellinger, et al., 1985a, 1985b, 1986a, 1986b, 1987a). These findings were
confirmed in more recent studies (Bellinger, et al., 1989a, 1989b). Additional follow-up showed
that deficits in McCarthy General Cognitive Index scores at age 5 were significantly correlated
with blood-lead levels at age 24 months, although not with prenatal blood lead measures.
Similar results were reported in the Cincinnati study (Dietrich, et al., 1986,1987a, 1987b). In
addition, study results suggest that the effect of prenatal lead exposure on the MDI was mediated
in part through its effects on birth weight and reduced gestational age, which were each
significantly associated with MDI scores (Dietrich, et al., 1987a). Results reported for the
Cleveland study were mixed, but while the authors tended to conclude that there was not strong
evidence of developmental effects of lead (Ernhart, et al., 1985,1986,1987,1988; Wolf, et al.,
1985; Ernhart and Green, 1990), other reviewers suggest that such effects may be inferred from
the reported results (Davis and Svendsgaard, 1987; EPA, 1986; ATSDR, 1993). In the Port Pine
study, reduced MDI scores at 24 months were associated with postnatal blood-lead levels
measured at age 6 months, but not with prenatal exposure measured through cord and maternal
blood-lead levels (Baghurst, et al., 1987; Vimpani, et al., 1985,1989; Wigg, et al., 1988).
Results of a follow-up neurobehavioral assessment conducted at age 3 to 4 years, using the
McCarthy Scales of Children's Abilities, indicated significant associations between postnatal
blood-lead levels (geometric means of 14 (ig/dL at 6 months and approximately 21 ug/dL at 15
and 24 months) and ability test scores (McMichael, et al., 1988).
In addition, all four studies report lower IQ scores at school-age for children who had
earlier exhibited elevated blood-lead levels. In Boston, slightly elevated blood-lead levels at age
24 months (mean of 6.5 |ig/dL) were associated with intellectual and academic performance
deficits at age 10 years (Bellinger, 1992). In Cincinnati, postnatal blood-lead levels measured
through age 3 years were inversely associated with IQ scores measured at age 5, although the
Draft - Do Not Cite or Quote 39 September 27, 1996
-------
effect was not statistically significant when adjusted for covariates (Deitrich, et al., 1993). In
Cleveland, a significant association was reported between blood-lead concentration at age 2
(mean of 16.7 ^ig/dL) and IQ measured at 5 years (Ernhart, et al., 1989). In Port Pirie,
statistically significant associations were reported between IQ measured at age 7 and blood-lead
levels from birth through age 7, with the strongest associations for blood-lead levels measured at
15 months to 4 years (Baghurst, et al., 1992).
Taken together, these studies provide strong evidence that low-level prenatal or early
postnatal exposure to lead results in neurobehavioral developmental delays through age 5.
Strong relationships between blood-lead concentration in early childhood, age 15 months to 4
years, and IQ scores were also reported, even when only slight elevations in blood-lead levels
were present.
Additional evidence of IQ point loss associated with elevated blood-lead levels in school-
age children is reported in cross-sectional studies throughout the world. A study of Danish
children related tooth-lead concentration to performance on several psychometric tests (Hansen,
et al., 1989). Children with elevated tooth-lead levels (above 18.7 jag/g) were matched by sex
and socioeconomic status with children with lower levels (below 5 ug/g). High lead children
scored lower on the Wechsler Intelligence Scales for Children (WISC) IQ test than children with
lower lead levels, although no difference in scores was observed for the Performance IQ and
several experimental tests. Impaired neuropsychological functioning due to lead exposure was
observed through differences in performance on the Bender Visual Motor Gestalt Test and on a
behavioral rating scale. A study of school children in Edinburgh, Scotland (Fulton, et al., 1987)
found that elevated blood-lead levels (mean of 11.5 ug/dL) were associated with lower scores on
IQ tests and on mathematical and reading attainment tests, after adjusting for covariates. No
threshold in the relationship, below which lead does not have an effect on intelligence and
attainment, was observed even for blood-lead concentrations below 10 ug/dL. A study of
Chinese children (Wang, et al., 1989) also reported a significant dose-response relationship
between blood-lead concentration (above 10 ug/dL) and IQ scores, after adjusting for covariates.
A significant effect of lead on IQ is not uniformly reported, however. Children randomly
selected from birth records in Birmingham, United Kingdom, were assessed using a variety of
cognitive, performance, neuropsychological, and behavioral endpoints (Harvey, et al., 1988).
Draft - Do Not Cite or Quote 40 September 27, 1996
-------
The effect of lead (mean of 13.5 (ig/dL) was not significant for most endpoints, and for none of
the three IQ measures. Both tooth lead and blood lead were examined as predictors of
intelligence in a study of 6 year old children in London (Smith, et al., 1983; Pocock, et al., 1989).
Neither measure of lead exposure was a significant predictor, once social factors were controlled.
No evidence of an association between blood-lead levels (mean of 12.75 ug/dL) and intelligence
was reported in another study of London children that included more middle class families
(Lansdown, et al., 1986).
A possible explanation for these seemingly contradictory results is that the effect of lead
on IQ may be overshadowed by the effects of home and societal factors, such as birth order,
parental IQ and level of education, and socioeconomic status. For example, a study of 104
children under age 7 and of lower socioeconomic status indicated that MDI and IQ scores were
significantly associated with blood-lead levels ranging from 6 to 59 ng/dL, after controlling for
socioeconomic and other factors (Schroeder, et al., 1985). In a five-year follow-up of 50 of these
children, IQ was inversely correlated with initial and concurrent blood-lead levels, but the effect
of lead was not significant when socioeconomic status and other covariates were included in the
analysis (Schroeder and Hawk, 1987). However, in a replication of the study among children of
uniformly low socioeconomic status, the effect of lead was evident at both the initial and five-
year follow-up (Hawk, et al., 1986; Schroeder and Hawk, 1987). These results suggest that the
effects of lead may be more easily detected in groups with similar home and societal
backgrounds.
Both current and long-term indicators of lead-exposure were studied to establish which
indicator was best correlated with psychometric test scores (Bergomi, et al., 1989). Total and
verbal IQ scores were negatively correlated with tooth-lead levels and ALAD activity. Tooth-
lead levels were also negatively correlated with Toulouse Pieron test results, which evaluate
ability figure identification, discrimination, and attention. The most predictive measure of lead
exposure was tooth lead, which is indicative of cumulative lead exposure. Blood lead, which is
indicative of a mix of current and past exposure, and hair lead, which is indicative of short-term
exposure, had little predictive value in this study.
The effect of lead on IQ and other developmental indicators is well-established for
children with markedly elevated blood-lead concentrations. For example, five point IQ
Draft - Do Not Cite or Quote 41 September 27, 1996
-------
decrements, fine motor dysfunction, and altered behavioral profiles were reported among
preschool children exhibiting pica for paint and plaster, whose blood-lead levels were greater
than 40 ng/dL (mean of 58 ug/dL), when compared with matched controls who did not eat paint
and plaster (de la Burde and Choate, 1972). At age 7 to 8, three point IQ decrements and
impairments in learning and behavior were reported for these children, even though blood-lead
levels had declined (de la Burde and Choate, 1975). Blood-lead concentrations for control
children were not reported, but, given the timing of the study, children in the control population
may have had what would now be considered elevated blood-lead levels. A study that included
children who had previously had encephalopathy indicated that these children had increased
incidence of hyperactivity and IQ decrements of approximately 16 points resulting form lead
exposure (Rummo, et al., 1979). In the same study, asymptomatic children with long-term
exposure (means of 51-56 ug/dL) had IQ decrements of 5 iig/dL on average, compared to control
children (mean of 21 ug/dL).
A study of the long-term effects of low-level lead exposure found that children with
higher dentin lead levels were more likely to drop out of high school and have a reading
disability (Needleman, et al., 1990). Higher lead levels were also associated with lower ranking
in high school class and increased absenteeism. Lower scores on vocabulary and grammatical-
reasoning tests were reported, along with poor hand-eye coordination, delayed reaction times,
and slowed finger tapping, compared to children with lower lead exposure. Earlier results
indicated that children with high dentin lead levels had deficits in IQ scores, speech and language
processing, attention, and classroom performance in first and second grades (Needleman, et al.,
1979). IQ deficits continued through the fifth grade. In addition, children with higher lead levels
needed more special academic services, and had a higher failure rate in school (Bellinger, et al.,
1986c).
A lead-related decrease in hearing acuity has been reported in young children, with
hearing thresholds at 2000 Hz increasing with blood-lead levels in the range of 6 to 59 ug/dL
(Robinson, et al., 1985). Analysis of NHANESII data indicated that the probability of increased
hearing thresholds at 500,1000, 2000, and 4000 Hz was associated with increased blood-lead
levels from below 4 ug/dL to over 50 ug/dL. In addition, this study reported increased
Draft - Do Not Cite or Quote 42 September 27, 1996
-------
probability a child was hyperactive and delays in developmental milestones (age at which child
first sat up, walked, and talked) associated with elevated blood lead (Schwartz and Otto, 1987).
2.3.2 Hematological Effects of Lead
The effects of lead on the blood's biochemical functions are interrelated and have
variable biological impact. Heme (the component of blood that contains iron) is critical to the
basic function of many organ systems, including the blood-forming tissue, liver, brain, and
kidneys. As noted earlier, lead can disturb the formation of hemoglobin (red blood cells), which
may cause anemia at high exposure levels. The heme-mediated generation of an important
hormonal metabolite of vitamin D (1,25-dihydroxyvitamin D) may be disturbed by lead. This
hormone serves a number of functions in humans, including the regulation of calcium
metabolism. In addition to the direct effects of lead on heme biosynthesis, there are potentially
significant indirect impacts on the central nervous system, caused by the accumulation of the
potential neurotoxicant, ALA. Lead also inhibits coproporphyrin utilization and the conversion
of zinc erythrocyte protoporphyrin (ZPP) into heme. The effects of lead on heme biosynthesis
are described in detail in Section 2.2.3 and in the Air Quality Criteria for Lead (EPA, 1986).
The threshold blood-lead level for a decrease in hemoglobin is approximately SO ug/dL in
occupationally exposed adults and 40 ug/dL in children (EPA, 1986). However, adverse effects
on hematocrit may occur at even lower blood-lead levels in children. In a cross-sectional study
of 579 children ages 1 to 5 years, a strong association between blood-lead level and the
probability of anemia was observed between 20 and 100 ug/dL, with the strongest effect in the
youngest children. In this study, anemia (defined as hematocrit below 35%) was not observed at
lead levels below 20 |ig/dL (Schwartz, et al., 1990).
Anemia is not usually an early manifestation of lead poisoning and is evident only when
the blood-lead level is significantly elevated for prolonged periods. Some of the hematologic
signs of lead poisoning resemble other diseases or conditions. Two rare diseases, acute
intermittent porphyria (a group of diseases with unusual and characteristic manifestations, which
have in common the excretion of porphyrins) and coproporphyria (high excretion of
coproporphyria), also result in heme abnormalities similar to those of lead poisoning. Nutritional
deficiencies may increase the development of anemia since lack of proper vitamins and minerals
Draft - Do Not Cite or Quote 43 September 27, 1996
-------
may result in iron deficiency anemia. Iron deficiency makes lead induced anemia worse in
children and vice versa. Lead also exacerbates hemolytic anemia associated with vitamin E
deficiency by enhancing mechanical fragility of cells.
Lead-induced disturbances in red blood cell formation and maturation also occur by way
of alterations in pyrimidine metabolism, as described in Section 2.2.3. Erythrocite Py-5'-N
activity is inhibited in lead workers, with the greatest inhibition and marked accumulations of
pyrimidine mucleotides apparent in workers with overt intoxication, including anemia (Paglia, et
al., 1975,1977). Erythrocyte Py-5'-N activity is inhibited in children at very low blood-lead
levels, with no threshold apparent (Angle and Mclntire, 1978; Angle, et al., 1982). The adverse
effects of decreased Py-5'-N activity at low blood-lead levels, in the absence of detectable effects
on hemoglobin levels and erythrocyte function or survival, are not known.
Lead can inhibit ALA-D activity and stimulate ALA-S activity, which results in
accumulation of ALA in the body and excretion. ALA may be neurotoxic at higher levels.
General population studies indicate that ALA-D activity is inhibited at very low blood-lead
levels, with no threshold apparent, in adults (Hernberg and Nikkanen, 1970; Roels, et al., 1976),
children (Chisolm, et al., 1985; Roels and Lauwerys, 1987), and newborns (cord blood) and their
mothers at delivery (Lauwerys, et al., 1978). The adverse effects of the decreased ALA-D
activity observed at low blood-lead levels are not known.
2.3.3 Other Effects of Lead
Death: It is well known that severe lead poisoning can lead to encephalopathy and death.
There is some evidence, too, of higher death rates due to cerebrovascular disease among lead
workers (Fanning, 1988; Malcolm and Barnett, 1982; Michaels, et al., 1991). In infants, high
levels of lead have been suggested as a causative agent in Sudden Infant Death Syndrome (SIDS)
(Drasch, et al., 1988).
Hypertension: There may be a relationship between lead exposure and hypertension.
Increased heart rate and hypertension have been noted in occupationally exposed workers after
exposure to high levels of lead following exposure durations of as short as four weeks (Marino,
et al., 1989). Hypertension has also been associated with lead exposure in the general
populations (Khera, et al., 1980b; Pirkle, et al., 1985), although the evidence is mixed (Pocock,
Draft - Do Not Cite or Quote 44 September 27. 1996
-------
et al., 1984,1985, 1988; Gartside, 1988; Coate and Fowles, 1989). Cardiovascular effects other
than blood pressure changes have been observed in individuals occupationally exposed to lead
(electrocardiographic (ECG) abnormalities - Kosmider and Petelenz, 1962; ischemic ECG
changes - Kirkby and Gyntelberg, 1985).
Gastrointestinal Effects: Colic is a consistent early symptom of lead poisoning in
occupationally exposed cases, in individuals acutely exposed to high levels of lead, such as
occurs during the removal of lead-based paint. Colic is characterized by a combination of the
following symptoms: abdominal pain, constipation, cramps, nausea, vomiting, anorexia, and
weight loss. Although gastrointestinal symptoms typically occur at blood-lead levels of 100 to
200 ng/dL, they have sometimes been noted in workers whose blood-lead levels were as low as
40 to 60 ug/dL (Studies listed in Table B-l). EPA has established a lowest observed adverse
effect limit (LOAEL) of 60 to 100 ug/dL for colic in children (EPA, 1986).
Renal Effects: The characteristics of early, or acute, lead-induced nephropathy (kidney
disease) include nuclear inclusion bodies, mitochondrial changes, and cytomegaly of the
proximal tubular epithelial cells; disfunction of the proximal tubules (Fanconi's syndrome)
manifested as aminoaciduria, glucosuria, and phosphaturia with hypophosphatemia; and
increased sodium and decreased uric acid excretion. These effects appear to be reversible.
Characteristics of chronic lead nephropathy include progressive interstitial fibrosis, dilation of
tubules and atrophy or hyperplasia of the tubular epithelial cells, and few or no nuclear inclusion
bodies, reduction in glomerular filtration rate, and azotemia. These effects are irreversible. The
acute form is reported in lead-intoxicated children and sometimes in lead workers. The chronic
form is reported mainly in lead workers. A summary of studies reporting acute or chronic
nephropathy may be found in ATSDR, 1993. Additional detail is reported in EPA, 1986.
Vitamin D Metabolism: Lead appears to interfere with the conversion of vitamin D to its
hormonal form, 1,25-dihydroxyvitamin D. Evidence for this effect comes primarily from studies
of children with high lead exposure (Rosen, et al., 1980; Mahaffey, et al., 1982). However, the
effect of lead on vitamin D metabolism may only be apparent in children with chronic nutritional
deficiency and chronically elevated blood-lead levels (Koo, et al., 1991).
Draft - Do Not Cite or Quote 45 September 27, 1996
-------
Thyroid: Limited evidence from occupationally exposed workers suggests that lead may
adversely affect thyroid function (Tuppurainen, et al., 1988). However, no effects of lead on
thyroid function have been found in children (Siegel, et al., 1989).
Growth: A number of epidemiological studies have reported an association between
blood-lead levels and growth in children (Nye, 1929; Johnson and Tenuta, 1979; Lauwers, et al.,
1986; Schwartz, et al., 1986; Lyngbye, et al., 1987; Angle and Kuntzelman, 1989). However, a
study of lead-poisoned subjects and nonexposed sibling controls failed to establish an association
between blood-lead levels and growth or the genetic predisposition for adult height (Sachs and
Moel, 1989). Moreover, a recent longitudinal study in Cleveland found no statistically
significant effect of blood-lead levels on growth (height, weight, and head circumference) from
birth through age 4 years and 10 months (Greene and Ernhart, 1991). Growth rates, measured as
covariate-adjusted increases in stature from 3 and IS months of age, were inversely correlated
with corresponding increases in blood-lead levels in a longitudinal study of 260 infants in
Cincinnati (Shukla, et al., 1987,1989).
Development: Lead-related effects on children's development, such as reduced birth
weight, reduced gestational age, and neurobehavioral developmental deficits, have been reported.
The evidence on birth weight and gestational age is mixed, with some studies reporting
reductions associated with lead exposure (Moore, et al., 1982; McMichael, et al., 1986), while
others report no differences (Needleman, et al., 1984; Factor-Litvak, et al., 1991; Green and
Ernhart, 1991). The evidence on neurobehavioral development is more consistent, with most
studies reporting an association between lead exposure and developmental deficits (Baghurst, et
al., 1987; Vimpani, et al., 1985,1989; Wigg, et al., 1988; Bellinger, et al., 1985a, 1985b, 1986a,
1986b, 1987a, 1989a, 1989b; Dietrich, etal., 1986,1987a, 1987b). A short summary of these
results is included in Section 2.3.1. There is some evidence that early developmental deficits
may not persist until age 4-5 (Bellinger, et al., 1991). Finally, one study demonstrated an
association between cord blood-lead levels and minor congenital anomalies (Needleman, et al.,
1984), although lead was not associated with increased incidence of major congenital anomalies.
Immune System: The data on immunological effects of lead in occupationally exposed
adults are inconsistent, but indicate that while lead may have an effect on the cellular component
of the immune system, the humoral component is relatively unaffected (ATSDR, 1993). The
Draft - Do Not Cite or Quote 46 September 27, 1996
-------
data on immunolo'gical effects of lead on children are very limited, but no effects have been
detected (ATSDR, 1993; Reigart and Graber, 1976).
Reproduction: A large body of literature clearly indicates that high levels of lead cause
adverse effects on both male and female reproductive functions. Women who are exposed to
high levels of lead during pregnancy have experienced an increased rate of miscarriages and
stillbirths (Nordstrom, et al., 1979; Baghurst, et al., 1987; McMichael, et al., 1986; Wibberly, et
al., 1977). In addition, women who were significantly exposed during childhood may be at risk
of spontaneous abortion and stillbirth and their children more likely to experience learning
disabilities (Hu, 1991). Lead-induced effects on male reproductive functions, including reduced
sperm production, have been reported in studies of occupationally exposed males (Chowdhury,
et al., 1986; Assennato, et al., 1987; Lancrajan, et al., 1975; Wildt, et al., 1983). Reproductive
effects of chronic low-level exposure are less known. A recent prospective study found no effect
on the rate of spontaneous abortions among women who resided near a lead smelter (mid-
pregnancy mean blood lead concentration of 15.9 Hg/dL) (Murphey, et al., 1990).
Genotoxic Effects: Results of assays made following in vivo exposure from occupational
sources are contradictory, but do suggest that lead may have an effect on chromosomes. While
increased frequencies of chromosomal aberrations have been observed in occupationally-exposed
workers,(Huang, et al., 1988b; Nordenson, et al., 1978), most of the available data show no such
increase (Bauchinger, et al., 1977; Maki-Paakkanen, et al., 1981; O'Riordan and Evans, 1974;
Schmid, et al., 1972). Sister chromatid exchanges may (Huang, et al., 1988b; Grandjean, et al.,
1983; Leal-Garza, et al., 1986), or may not (Maki-Paakkanen, et al., 1981; Dalpra, et al., 1983)
be increased as a result of lead exposure.
Cancer: The information available that has examined the association of occupational
exposure to lead with increased cancer risk is generally limited in its usefulness because the
actual compound(s) of lead, the route(s) of exposure, and level(s) of lead to which the workers
were exposed were not reported. Furthermore, the potential for exposure to other chemicals,
including arsenic, exists, particularly in lead smelters. Nonetheless, a statistically significant
increase in total malignant neoplasms, largely due to small, statistically nonsignificant increases
in digestive, respiratory, and urinary tract tumors has been observed among lead production
workers (Cooper, 1976; Cooper and Gaffey, 1975; Kang, et al, 1980). In addition, a statistically
Draft - Do Not Cite or Quote 47 September 27, 1996
-------
significant increase in rectal cancer was found in workers exposed to tetraethyl lead
(Fayerweather, et al, 1991).
2.4 REPRESENTATIVE HEALTH EFFECTS
The childhood lead poisoning problem encompasses a wide range of exposure levels,
with varying health effects at different levels of exposure. As described in the previous section,
even low-level exposure to lead can result in adverse health effects. At low levels, the health
effects may not be severe or obvious, but a large number of children are affected. As the
exposure level increases, the severity of the health effects increases, but the number of children
affected decreases.
Both individuals and society as a whole are damaged by adverse health effects associated
with lead exposure. In this section, two representative effects, elevated blood-lead concentration
and IQ point deficit, are identified to represent the spectrum of health effects of lead exposure.
Each representative end point may be used both to estimate the number of children who will
benefit under the proposed rule and also the economic benefit to society. The representative
health effects and blood-lead concentrations are used in the integrated risk analysis in Chapter 5
to estimate the numbers of children who may benefit under the proposed §403 rule. The
estimation of economic benefits was considered in selecting the health endpoints, as economic
benefits resulting from the rule are estimated in the accompanying §403 RIA. Estimation of the
benefits of reducing lead exposure requires selection of an age group for characterizing the health
risks of lead exposure. The selection of the age group in this risk assessment was based on the
most appropriate age of child for the estimation of health effects. The effects of lead exposure
are thought to be strongest for fetuses, infants, and young children, because of their rapidly
developing central nervous system and because the brain is a primary target organ of lead.
Furthermore, the increased risk of lead exposure appears to be most evident at age 2 (Clark,
1985; Goyer, 1993) and studies have shown a strong association between blood-lead
concentration measured at age 2 and IQ scores later in life (Bellinger, 1992; Schwartz, 1994;
Pocock, 1994). Therefore, the health benefits of reducing childhood lead exposure are estimated
at age 2 in this risk assessment.
Draft - Do Not Cite or Quote 48 September 27, 1996
-------
2.4.1 Elevated Blood-Lead Concentration
Although an elevated blood-lead concentration is not a health effect in and of itself, the
relationship between blood-lead concentration and adverse health effects is well-established. In
addition, CDC guidelines on childhood lead poisoning prevention traditionally have been and
currently are defined in terms of blood-lead concentrations. Table 2-1 summarizes CDC's
recommended actions for children with elevated blood-lead concentrations (CDC, 1991). Based
on these guidelines, two levels of elevated blood-lead concentration are used to estimate benefits
under the proposed rule:
Incidence of blood-lead levels greater than 10 //g/dL: This level is the lowest blood-
lead level at which a child is considered lead poisoned by CDC. While extensive interventions
are not always recommended by CDC, children with blood-lead concentrations at or above 10
ug/dL require more frequent rescreening at minimum, and may require environmental or more
extensive medical interventions.
Incidence of blood-lead levels greater than 25 /ig/dL: This level is the blood-lead level
at which extensive medical intervention may be necessary. Current CDC guidelines (CDC, 1991)
recommend a provocative chelation test for children with blood-lead concentrations between
25 |ig/dL and 44 |ig/dL. It is further recommended that children with positive provocative
chelation tests, and all children with blood-lead concentrations at or above 45 ng/dL, should
receive one or more courses of chelation therapy. For these children, the medical intervention
accompanies an environmental assessment, remediation of sources of lead, and parental
education on ways to reduce lead exposure.
2.4.2 IQ Point Deficits
In this section, two IQ based endpoints are identified to represent the spectrum of
neurotoxicological effects of lead. While tests that focus on a specific neurological effect might
be more sensitive to the effects of lead than IQ tests, the selection of a representative effect is
difficult. Differences in the level, timing, and route of exposure for individuals may result in
differing effects of lead. For example, early exposure to lead (before age 2) may affect language
skills, while later exposure is more likely to affect spatial-symbolic skills (Shaheen, 1984). In the
absence of details of the exposure scenario, which are rarely available, exposure-related
Draft - Do Not Cite or Quote 49 September 27, 1996
-------
Table 2-1. Interpretation of Blood-Lead Concentrations and Foliow-Up Actions
Recommended by CDC
Class
I
MA
MB
III
IV
V
Blood-Lead
Concentration
U/g/dL)
S 9
10- 14
15- 19
20-44
45-69
2 70
Recommended Action
A child in Class 1 is not considered to be lead-poisoned. No action is
recommended.
Many children (or a large proportion of children) with blood-lead levels in the
range should trigger communitywide childhood lead poisoning prevention
activities. Children in this range may need to be rescreened more frequently.
A child in Class MB should receive nutritional and educational interventions and
more frequent screening. If the blood-lead level persists in this range,
environmental investigation and intervention should be done.
A child in Class III should receive environmental evaluation and remediation and
a medical examination. Such a child may need pharmacologic treatment of
lead poisoning. A provocative chelation test is recommended for children with
blood-lead levels between 25 and 44 |/g/dL.
A child in Class IV will need both medical and environmental interventions,
including chelation therapy.
A child with Class V lead poisoning is a medical emergency. Medical and
environmental management must begin immediately.
differences will be most apparent on tests, such as IQ tests, that measure performance over a
range of neurological functions (Bellinger, 1995). The relationship between blood-lead
concentration and IQ scores has been reported consistently in the literature and is quantified by
meta-analysis (Needleman and Gatsonis, 1990; Schwartz, 1993; Schwartz, 1994; Pocock, et al.,
1994; Section 4.4; Appendix D). The following IQ-based health endpoints are used in the risk
assessment to represent the neurotoxicological effects of lead exposure:
IQ Points Lost: This health effect is used to represent the neurological loss due to low
level lead exposure. Lower IQ scores are associated with a lower level of educational attainment
and lower life-time earnings.
Increased Incidence of IQ scores less than 70: This health effect is selected to
represent the increased likelihood of mental retardation resulting from lead exposure. An IQ of
70 is two standard deviations below the population mean and is used as an indicator of mental
retardation. Children who are mildly mentally retarded require special education classes in
school. Children who are severely mentally retarded may require life-long institutional care.
Draft - Do Not Cite or Quote
50
September 27, 1996
-------
3.0 EXPOSURE ASSESSMENT
CHAPTER 3 SUMMARY
The goal of this exposure assessment is to document the important sources of
lead in the environment, to document the ma/or pathways by which lead is
exposed to children, to characterize the current distribution of environmental-lead
levels in the nation's housing stock, and to characterize the current distribution of
average blood-lead concentration among the nation's children. Information from
the exposure assessment is used with the findings of hazard identification
(Chapter 2) and dose response assessment (Chapter 41 to provide input to the risk
characterization (Chapter 5).
The 1997 national housing stock is predicted to contain 99,272,000 occupied
housing units, containing nearly eight million children aged 1 to 2 years. Of these
units, approximately 62% are forecast to contain lead-based paint, and 14% to
contain more than 5 ft2 of deteriorated lead-based paint. In this risk assessment,
the HUD National Survey is used to characterize the distribution of environmental-
lead levels in these units. Data from this survey and other epidemiological studies
suggest that lead levels in dust and soil tend to decrease with the age of the unit.
Data from the Baltimore Repair and Maintenance Study provide strong evidence
that environmental-lead levels are high in the presence of deter/orated lead-based
paint. Data from the Rochester Lead-in-Dust Study indicate that lead levels can be
high within urban environments and in older units.
Data from Phase I of the Third National Health and Nutrition Examination
Survey (NHANES III) were used to characterize a pre-intervention distribution of
children's blood-lead levels in the nation's housing stock. The geometric mean
blood-lead concentration for children aged 1-2 years is 4.05 ug/dL, with a
geometric standard deviation of 2.06. Slightly over 10% of these children are
estimated to have blood-lead concentrations greater than 10 ug/dL. Blood-lead
concentration data from the Baltimore Repair and Maintenance Study and the
Rochester Lead-in-Dust Study indicate that blood-lead concentrations are higher in
the presence of lead-based paint and high environmental-lead levels.
The goal of this exposure assessment is to document the important sources of lead in the
environment, to document the major pathways by which lead is exposed to children, to
characterize the current distribution of environmental-lead levels in the nation's housing stock,
and to characterize the current distribution of average blood-lead concentration among the
nation's children. Information from the exposure assessment is used with the findings of hazard
identification (Chapter 2) and dose response assessment (Chapter 4) to provide input to the
integrated risk analysis (Chapter 5).
Draft - Do Not Cite or Quote 51 September 27, 1996
-------
Section 3.1 provides documented sources and pathways of lead exposure in the nation's.
residential environment. A number of epidemiological studies have investigated the extent to
which lead is present in certain residential environments and how this lead exposure affects
blood-lead concentration in children. These studies are introduced and summarized in
Section 3.2.
It is of interest to assess lead exposure in that portion of the national housing stock in
which children reside or can potentially reside (hereafter referred to simply as the "national
housing stock")- Section 3.3 characterizes the extent to which lead exposure is present in the
national housing and Section 3.4 characterizes the distribution of childhood blood-lead
concentrations. These characterizations are provided for 1997, the year in which regulations
developed in response to §403 are expected to be promulgated.
This chapter provides several sources of data on housing stock characteristics, population
estimates, and environmental-lead levels in housing units. In the exposure assessment, these data
have been used to make inferences on residential lead exposure to children in the United States.
The extent to which any exposure assessment accurately portrays the exposure scenario of
interest depends on the relevance and representativeness of the data used in the analyses and hi
the methods applied to these data to meet the objectives of the exposure assessment. Therefore,
an effort has been made in this chapter to present the methods used, to identify assumptions and
approximations made in the analysis and when they were made, and to identify uncertainties and
limitations in the data. Supporting information and detailed results to accompany the
information in this chapter are presented in Appendix C.
3.1 SOURCES AND PATHWAYS OF LEAD
Lead is a heavy, stable element occurring naturally in the earth's crust. Through natural
activity such as crustal weathering and human activity such as mining, this metal has been
distributed throughout the human environment. Lead's historic use as an ingredient in various
manufactured and refined products has increased its introduction into the environment. As a
result, lead has been detected in water, soil, air, plants, animals, and humans. As lead does not
naturally biodegrade, its exposure potential tends to accumulate over time as more and more lead
is deposited in the environment.
Draft - Do Not Cite or Quote 52 September 27, 1996
-------
Research has identified a variety of environmental sources and reservoirs of lead which
can contribute to overall lead exposure in a child. Figure 3-1 illustrates the major sources and
reservoirs of lead, how lead is introduced into the human environment, and various pathways of
human exposure. According to this figure, both natural sources (e.g., crustal weathering) and
manufactured sources (e.g., auto and industrial emissions, paint and industrial dusts, solder, lead
glazes) have contributed lead to various components of the human environment. These
components act as lead reservoirs. Lead is exchanged among these reservoirs by various
environmental pathways (e.g., ambient air to soil to dust). Lead in such media as inhaled air,
dusts, food, or drinking water contributes to human lead exposure via direct pathways between
these reservoirs and man. As data supporting the dangers of lead exposure have been identified,
a combination of state and Federal action has curtailed the impact of certain sources and
reservoirs of lead in the environment, resulting in a change in the predominance of historically
significant sources.
In the scientific literature (e.g., Bornschein et al., 1986), quantitative exposure models, or
pathways models, have been applied to data from environmental-lead studies to identify the most
significant pathways by which residential, childhood environmental-lead exposure occurs and to
provide quantitative estimates of the relative contributions of the numerous hypothesized
exposure paths. These pathways models support Figure 3-1 by identifying direct and indirect
effects of lead levels in the residential environment on lead concentrations in children's blood.
Examples of direct effects on a child's blood-lead concentration include the effect of refinishing
painted surfaces within a housing unit, and the effect of a child's pica habits or mouthing
behavior. An indirect effect occurs when lead in one residential environmental medium (e.g.,
soil) contaminates another medium (e.g., interior dust), which in turn contributes directly to
elevated blood-lead concentration.
The information that follows provides the current status of the sources of lead included in
Figure 3-1 that have historically been recognized in the scientific literature as most associated
with elevated blood-lead concentrations in children. Most of the information comes from
detailed investigations on sources of lead documented in EPA (1986), CDC (1991), and ATSDR
(1993).
Draft - Do Not Cite or Quote 53 September 27, 1996
-------
SURFACE AND
GROUND
WATER
PAINT,
INDUSTRIAL
DUSTS
FECES
Figure 3-1. Pathways of Lead from the Environment to Humans, Main Organs of
Absorbsion and Retention, and Main Routes of Excretion
(Sources: EPA, 1986; EPA, 1996)
Draft - Do Not Cite or Quote
54
September 27, 1996
-------
Airborne Lead
Historically, emissions from lead smelters, battery manufacturing plants, solid waste
incinerators, and automobiles have made major contributions to airborne lead levels. Fallout of
atmospheric lead contributes to lead levels in soil, household dust, and street dust. Lead is
deposited on soil, plants, and animals, which thereby is incorporated into the food chain.
Until recently, leaded gasoline emissions was one of the primary sources of lead exposure
in the United States. Under Title H of the Clean Air Act (FR 1973 December 6), EPA specified
lead as a pollutant compound of concern, and instituted a controlled phase-out of leaded gasoline
by December 31,1995. As a result, there was a 73% reduction in lead consumed in gasoline
from 1975 to 1984 (EPA, 1986), and a 64% reduction in national lead emissions from 1985 to
1989 (ATSDR, 1993). This reduction has corresponded to a similarly dramatic decrease in
average lead concentration in children's blood (CDC, 1991; Annest, 1983). The phase-out of
leaded gasoline has contributed to airborne lead becoming only a minor lead-exposure pathway
for children not exposed to specific point-emitting lead sources (CDC, 1991).
Indoor air may be considered an important indirect lead-exposure pathway when lead-
based paint or lead-contaminated dust or soil is disturbed during renovation and remodeling
activities. Inadequate dust control or use of paint stripping techniques that vaporize lead in paint
are ways that lead is introduced into the air during renovation and remodeling activities (EPA,
1994d).
EPA has set a National Ambient Air Quality Standard of 1.5 jig of lead per cubic meter of
air averaged over three months (ATSDR, 1993; 40 CFR 50.12).
Drinking and Cooking Water
Detectable levels of lead are rare in surface and ground water that serve as sources of
drinking water in this country. Typically, lead contamination of drinking water occurs after the
water leaves the treatment plant (CDC, 1991). By traveling within service lines and household
plumbing, drinking water can become contaminated upon encounter with lead pipes, connectors,
and solder. At a residence, water can also become contaminated within lead-containing water
fountains, coolers, faucets, and other fixtures. Through the authority of the 1986 Safe Drinking
Water Act and its amendments, EPA banned the use of lead materials and solders in new
Draft - Do Not Cite or Quote 55 September 27. 1996
-------
plumbing and plumbing repairs, required that public water suppliers notify the public about lead
presence in drinking water, and encouraged local government measures to test and remediate
lead-contaminated drinking water in schools and day-care centers. As a result, lead in drinking
and cooking water is generally not a predominant source of lead exposure among lead-poisoned
children (CDC, 1991).
Analysis of environmental-lead data from several studies, including the Baltimore R&M
Study and the Rochester Study (Section 3.2), concluded that lead levels in drinking water
generally do not have a statistically significant effect oh blood-lead concentrations. However,
lead in drinking water is still considered an important exposure source when present due to the
high absorption rate of lead in water (CDC, 1991).
The Safe Drinking Water Act set an action level for lead in drinking water of 15 ppb.
Those systems that exceed the action level must inform the public, while taking measures to
reduce lead levels and continue monitoring procedures^ In 1991, EPA promulgated maximum
contaminant level goals (MCLGs) and national primary drinking water regulations (NPDWRs)
for lead and copper (56 FR 26460, June 7,1991). This rule set the MCLG for lead within
drinking water at the tap to be 0 ppb (ATSDR, 1993; 40 CFR 141,142).
Food ;
Many studies have shown that children's dietary intake of lead has receded over recent
years. For example, data from the U.S. Food and Drug Administration (FDA) indicate that
dietary lead intake in two-year-old children has declined from an approximate average of 30
Hg/day in 1982 to 5 ug/day in the period 1986-1988 (CDC, 1991). U.S. FDA intervention and
outreach activities, along with reduced lead entering the food chain due to the phase-out of
leaded gasoline, have contributed to this decline. The phase-out of lead-soldered food cans
(1.4% of the U.S.-produced food and soft drink cans in 1989, compared to 47% of such cans
produced in 1980), along with public education on proper food storage and cooking techniques,
have made large contributions to reducing the amount of lead ingested with food (CDC, 1991).
Education is especially important in those areas of the country with traditions of using lead-
containing pottery in cooking and preparing folk remedies containing lead.
Draft - Do Not Cite or Quote 56 September 27, 1996
-------
While production of lead-soldered food and soft drink cans have been virtually eliminated
in the U.S., such cans may still be used by other countries who import food to the U.S. In
addition, lead can be introduced to food grown in lead-contaminated soil. Improper handling of
food in the home (e.g., storing food in containers such as lead-soldered cans and lead-glazed
pottery) can cause food to be a source of lead exposure. Thus, while lead exposures through food
ingestion have declined considerably in recent years, these exposures can still occur if proper
precautions are not addressed.
Lead-Based Paint
Lead-based paint is currently considered the most significant high-dose source of lead
exposure in pre-school children (CDC, 1991). From the turn of the century through the 1940's,
paint manufacturers used lead as a primary ingredient in many oil-based interior and exterior
house paints. Usage gradually decreased through the 1950s and 1960s, as largely lead-free latex
paints and exterior paint with lower lead concentrations were manufactured. Although the
Consumer Product Safety Commission (CPSC) banned lead-based paints from residential use in
1978 (currently, paints may not have greater than 0.06% lead by weight), the presence of lead-
based paint in the nation's housing stock remains high. An estimated 64 million (or 83% of)
privately-owned, occupied housing units built prior to 1980 .contain some components covered
with lead-based paint (EPA, 1995a). Approximately 12 million of these units contain at least one
child under the age of seven years. The estimated percentage of public housing units built prior
to 1980 and containing lead-based paint is even higher: 86% (EPA, 199Sa).
The exposure to lead from lead-based paint is considerably higher when the paint is in a
deteriorated state or is found on accessible, chewable, impact, or friction surfaces (EPA, 1986;
CDC, 1991). Thus, young children are especially susceptible to lead poisoning from lead-based
paint, as they may ingest lead-based paint chips or come into contact with dust or soil that has
been contaminated by deteriorated lead-based paint (see below). Both adults and children can be
exposed to hazardous levels of lead by inhaling the fine dust or by ingesting paint-dust during
hand-to-mouth activities. The U.S. Department of Housing and Urban Development (HUD) has
prepared guidelines on controlling lead-based paint hazards, as improper control procedures can
actually increase the threat of lead-based paint exposure by dispersing fine lead dust particles in
Draft - Do Not Cite or Quote 57 September 27, 1996
-------
the air and over accessible household surfaces (HUD, 1995b; Farfel and Chisolm, 1990). The
potential for lead-based paint to contaminate a variety of environmental media within a
household makes lead-based paint the greatest source of public health concern regarding lead
exposure (CDC, 1991).
Contaminated Dust and Soil
While enforcement of national air quality standards continues to reduce the threat of lead
exposure via air from point sources, the fallout of atmospheric lead over time has resulted in a
continued exposure route through soil (CDC, 1991). In addition, soil can become contaminated
by deteriorated lead-based paint or by the improper removal of lead-based paint from a housing
unit. The same soil, once tracked indoors, can become a component of household dust causing
yet another source of lead exposure. Children are exposed to lead from soil or dust in their
homes during typical hand-to-mouth activities.
Lead-contaminated soil and dust are thought to be the major pathway by which young
children are exposed to lead from lead-based paint hazards (EPA, 1986). Exterior house paint
can flake off or leach into the soil around the outside of a home, contaminating children's playing
areas. Indoors, normal wear of lead-based paint (especially around windows and doors) and
contaminated soil tracked into the house can contaminate interior dust. When lead takes the form
of small particles, as it typically does when found within household dust (Que Hee et al., 1985),
it is more easily absorbed into the body (Mahaffey, 1977).
A number of studies have assessed the effect of dust- and soil-lead levels on childhood
blood-lead concentrations. A few studies have concluded that the effect of residential lead-based
paint on blood-lead levels occurs via the pathway of dust- and soil-lead to blood. For example,
analysis of data from the Cincinnati Longitudinal Study (Section 3.2.5) identified a significant
lead pathway from exterior dust to interior dust to hands to blood, with lead in paint and soil
contributing to lead in exterior dust (Bornschein et al., 1986). Analysis of data from the Brigham
and Women's Hospital Longitudinal Study (Section 3.2.6) concluded a significant pathway from
soil to window sill-dust to floor-dust to blood (Menton et al., 1995). It is likely that exposure of
young children to lead in dust and soil is primarily due to their propensity to mouth fingers, toys,
and other nonfood items that contain contaminated dust. In unpublished, EPA-supported
Draft - Do Not Cite or Quote 58 September 27. 1996
-------
pathways analyses of data from the Baltimore R&M Study (Section 3.2.1) and the Rochester
Study (Section 3.2.2), mouthing tendencies were found to be an important contributor to
childhood blood-lead concentrations.
3.2 SUPPORTING EVIDENCE IN EPIDEMIOLOG1C STUDIES
Extensive evidence of the relationship between childhood blood-lead concentrations and
environmental-lead levels is offered in the scientific literature. Evidence from two types of
studies is available. Epidemiological studies investigate the association between elevated blood-
lead concentrations and elevated levels of lead in a child's residential environment. Intervention
studies investigate the impact on children's blood-lead concentrations of reducing childhood lead
exposure via a range of intervention strategies. Epidemiological studies have demonstrated that
elevated blood-lead concentrations are associated with elevated lead levels in the dust, paint, and
soil of the surrounding environment. Causation, however, is better demonstrated by intervention
studies. If children receiving an intervention strategy that targets a particular lead exposure
source (e.g., paint, dust, or soil) exhibit greater reductions in blood-lead concentrations than
those reported for a suitable control population, then the targeted source may be at least partially
responsible for the prior exposure.
A review of intervention studies (EPA, 1995b) concluded that reductions in blood-lead
concentrations have occurred following interventions of lead in paint, dust, and soil. While such
studies suggest causation, their results are not necessarily indicative of the magnitude of the
association between the levels of lead in targeted environmental media and blood-lead
concentrations. This is because intervention studies typically examine children already exposed
to environmental lead. Exposed children retain a store of lead in their tissues that routinely
mobilizes into the blood. In fact, this mobilization is heightened following an intervention
(Schroeder and Tipton, 1968; Rabinowitz, 1991) as the change in exposure caused by the
intervention disrupts the body's equilibrium. Blood-lead concentrations following the
intervention, therefore, represent a combination of the now reduced environmental lead exposure
and the increased (at least temporarily) mobilized lead stores.
During the past 25 years, studies have been conducted to investigate the sources
responsible for lead exposure in children. Many of these studies are limited, small, or not
Draft - Do Not Cite or Quote 59 September 27. 1996
-------
relevant to the current exposure situation. These studies include investigations of the sources and
extent of lead exposure in both urban and smelter communities. The studies listed in Tables 3-1
and 3-2 provide evidence regarding associations between childhood blood-lead concentrations
and environmental-lead levels in urban and smelter communities, respectively. The results of
these studies are qualitatively similar in that the association between environmental lead and
blood lead is consistently positive and, when considered without the confounding from additional
variables, usually found to be statistically significant. However, any effort to combine the
disparate results from multiple studies into a single set of coefficients that provide one
representative, quantitative measure of the relationship between blood-lead concentration and
soil- and dust-lead levels is complicated by the qualitative dissimilarity among studies (e.g.,
differences in sampling and analysis methods, sampling locations, studied populations, and types
of communities).
Table 3-1. Childhood Lead Exposure Studies Conducted in Urban Communities That
Present Evidence of the Positive Relationship Between Environmental-Lead
Levels and Blood-Lead Concentrations
Study/Community
Baltimore (MD) Repair and
Maintenance Study
Rochester (NY) Lead-in-Dust
Study
Baltimore (MD) Urban Soil Lead
Abatement Demonstration Project
(USLADP)
Boston (MA) USLADP
Cincinnati (OH) USLADP
Birmingham (UK) Urban Lead
Uptake Study
Cincinnati (OH) Longitudinal
Brigham and Women's Hospital
Longitudinal Study (Boston. MA)
New Haven, CT
Omaha, NE
Study Duration
1992-1997
1993
1988-1991
1989-1991
1989-1991
1984-1985
1980-1987
1980-1983
1977
1970-1977
Study Type
Abatement
Efficacy
Health
Assessment
Soil Abatement
Efficacy
Health
Assessment
Health
Assessment
Health
Assessment
Health
Assessment
Health
Assessment
Reference(s)
Farfel and LJm, 1995
Rochester School of Medicine and
NCLSH, 1995; Lanphear et al., 1995
EPA, 1996;
Weitzman et al., 1993;
Aschengrau et al., 1994
Davieset al., 1990;
Thornton et al., 1990;
Davieset al., 1987
Bornschein et al., 1985a; due Hee et al.,
1985; Bornschein et al., 1985b;
Bornschein et al., 1986
Bellinger et al., 1986; Rabinowitz et al.,
1985a; Rabinowitz et al., 1985b;
Rabinowitz et al., 1 984a; Rabinowitz et
al., 1984b; Rabinowitz et al., 1982
Stark et al., 1982; Stark et al., 1978
Angle and Mclntire, 1979; Angle et al.,
1974; Angle et al., 1984
Draft - Do Not Cite or Quote
60
September 27, 1996
-------
Table 3-2. Childhood Lead Exposure Studies Conducted in Smelter Communities That
Present Evidence of the Positive Relationship Between Environmental-Lead
Levels and Blood-Lead Concentrations
Study/Community
Granite City (IL) Educational Intervention
Butte-Silver Bow (MT) Environmental
Health
Clear Creek/Central City (CO) Mine
Waste Exposure
Midvale (UT) Community
Child Lead Exposure Study (Leeds, AL)
Philadelphia (PA) Neighborhood Lead
Leadville (CO) Metals Exposure
Silver Creek Mine Tailings Exposure
(Park City, UT)
Telluride, ID
Kellogg (ID) Revisited
Helena Valley (MT) Child Lead
El Paso. TX
Study
Duration
1991
1990
1990
1989
1989
1989
1988
1987
1986
1983
1983
1971-1973
Reference(s)
Kimbrough et al., 1 994
Butte-Silver Bow Dept. of Health, et
al., 1991
ATSDR, 1992
Bornschein et al., 1 990;
Que Hee et al., 1985
ATSDR, 1991 a
ATSDR, 1991b
Colorado Dept. Of Health, et al.,
1990
ATSDR, 1988
Bornschein et al., 1 989;
QueHeeetal., 1985
Panhandle District Health Dept. et
al., 1986
Lewis and Clark County Health
Dept. etal., 1986
Landrigan et al., 1975
Early childhood lead exposure studies emphasized exposure to lead in paint, leaded
gasoline emissions, and emissions from industrial sources. These studies, therefore, measured
lead levels in these media and sought to relate them directly to resident children's blood-lead
concentrations. Due to the assessment by many researchers in childhood lead exposure that
ingestion of dust and soil via hand-to-mouth behavior represents the principal mechanism of lead
exposure in young children today (CDC, 1991), more recent studies have focused principally on
lead exposure from residential soil and dust. As indicated in Figure 3-1, residential soil and dust
are assumed to have been contaminated by these same original sources: lead-based paint,
industrial emissions or tailings, and leaded gasoline emissions.
Draft - Do Not Cite or Quote
61
September 27, 1996
-------
Due to the reduction in lead sources such as gasoline emissions over time, the most recent
epidemiologic studies provide a more accurate picture of the relationship between child blood-
lead concentrations and lead-based paint hazards. Consequently, the §403 risk assessment has
relied primarily on information from recent studies conducted in urban areas in the absence of
specific point emission sources. The remainder of this section summarizes key objectives and
conclusions on the effects of childhood lead exposure for eight urban studies in Table 3-1 that
were conducted in the 1980s and 1990s. Environmental-lead data and blood-lead concentration
data from two of these studies, the Baltimore Repair and Maintenance Study and the Rochester
Lead-in-Dust Study, are summarized in Sections 3.3 and 3.4. These data were also used to
develop statistical models to relate blood-lead concentration to environmental-lead levels.
Necessary information on data from the three USLADP studies was not made available hi tune to
allow the data to be included in this risk assessment. Data for the three other studies
(Birmingham Urban Lead Uptake Study, Cincinnati Longitudinal Study, and the Brigham and
Women's Hospital Longitudinal Study) were either not available to the risk assessment effort or
were not considered for specific reasons.
3.2.1 Baltimore Repair and Maintenance (R&M) Study
The objectives of the Baltimore R&M Study were to characterize the efficacy of
comprehensive lead-paint abatement up to six years post-abatement and to characterize the
efficacy and costs of three levels (low, medium and high) of less costly Repair and Maintenance
interventions. While published analyses of the R&M data are not yet available, these data were
available for use in the §403 risk assessment.
In 1992, three types of housing units were recruited for this study. In the first group, IS
previously-abated dwellings were chosen from 90 low-income housing units that were abated
between May, 1988, and April, 1992, by Baltimore City and Kennedy Krieger Institute Pilot
Abatement Projects. The second group, slated to receive R&M interventions in this study,
consisted of 75 older (mostly pre-1940), low-income dwellings in Baltimore City. Finally, 15
modern urban dwellings free of lead-based paint were chosen to represent control units. These
units were chosen from an urban renewal area and included units that were fully gut-rehabilitated
Draft - Do Not Cite or Quote 62 September 27, 1996
-------
since 1980. All units in the study had to include at least one eligible child aged 6 to 48 months
who spent most of his/her time at the unit.
Prior to any intervention in this study, blood-lead concentrations were measured for 115
children that lived in 87 of the housing units. Although environmental and blood samples were
collected both before and after interventions, only the results and data of the pre-intervention
samples are considered in this report. The BRM vacuum method, consisting of a modified HVS3
cyclone collector, was the primary dust sampling method used in the R&M Study (EPA, 1995c).
Within each housing unit, rooms were divided into three groups: rooms with windows on the
first floor, rooms with windows on the second floor, and rooms with no windows. A composite
sample of floor dust was collected from each of the three groups. Two additional composite
samples, one of window sill dust and one of window well dust, were collected from the first two
groups of rooms.
Lead levels in paint were determined through in situ x-ray fluorescence (XRF)
measurement. Only those components suspected of being covered with lead-contaminated paint
were measured, in order to identify whether a unit contained lead-based paint. As a result, paint-
lead measurements in this study were high and do not represent a random sampling of painted
surfaces in a housing unit.
Soil samples were taken from (1A inch) soil cores collected at the foundation and property
boundaries. However, few units had available soil at these locations to sample. Dust and soil
samples were analyzed for lead using inductively coupled plasma-atomic emission spectrometry
or graphite furnace atomic absorption spectroscopy. Two-hour stagnation drinking water
samples were also collected. A structured questionnaire collected information on study children
and the households.
Summaries of pre-intervention environmental-lead levels observed in the Baltimore
R&M Study are presented in Section 3.3 and observed blood-lead concentrations are summarized
in Section 3.4.
3.2.2 Rochester Study
The Rochester study, conducted in 1993, was a cross-sectional design study whose
primary objective was to obtain information on the association between lead levels in house dust
Draft - Do Not Cite or Quote 63 September 27, 1996
-------
and blood-lead concentrations of resident children (HUD, 1995a) Children between the ages of
12 and 31 months (considered by the study group as the age of greatest risk for lead exposure)
and living in the city of Rochester, NY, were eligible for this study, provided they did not satisfy
any of the following:
• they or their environment had underwent recent interventions that were likely to alter
blood or dust lead (e.g., major renovation, recent ingestion of prescribed iron
products, or any medical or environmental intervention for an elevated blood-lead
level),
• they spent more than 20 hours per week away from home, or
• they lived with an adult exposed to lead from an occupational or recreational activity.
Random sampling techniques were used to recruit children bom from March 1,1991, to
September 30,1992, at either Rochester General Hospital, Strong Memorial Hospital, or St.
Mary's Hospital. Data for 205 families and children were included in the analysis.
During visits to the home of each study participant, an environmental health team
obtained a venipuncture blood sample from the eligible child, completed a behavioral
questionnaire for the household regarding lead exposure, collected environmental samples
(interior dust, exterior soil, water), and took in situ measurements of lead in paint. The dust
samples were collected from floors, window sills, and window wells within rooms in which the
child was frequently present. When considering dust-lead loadings, the §403 risk assessment
effort considered only those dust results collected using wipe techniques ("Little Ones" baby
wipes). However, because a secondary objective of the Rochester study was to evaluate various
dust sampling methods relative to predicting children's blood lead levels, dust samples were also
collected using the University of Cincinnati Dust Vacuum Method (DVM) and the BRM
vacuum. Lead concentrations of dust samples collected using the BRM vacuum are also
summarized. Side-by-side dust samples were collected at specific locations, with each sample
corresponding to a particular collection method and the wipe sample being the first to be
collected. Dust samples were analyzed using either flame or graphite furnace atomic absorption
Draft - Do Not Cite or Quote 64 September 27, 1996
-------
spectroscopy. Soil samples, taken at the play area and dripline, were analyzed using flame
atomic absorption spectroscopy.
Children enrolled in the Rochester study were not specifically recruited because of
elevated blood-lead concentrations. However, a disproportionate percentage of these children
exhibited two risk factors associated with elevated blood-lead concentrations: residing in older
housing (84% of the homes were built prior to 1940) and belonging to low-income families (55%
of households had incomes below $15,500).
The geometric mean blood-lead concentration for the 205 children in the Rochester study
was 6.38 ug/dL (geometric standard deviation, 1.85). Twenty-three percent of the children had
blood-lead concentrations above 10 |ig/dL, 8% above 15 ug/dL, and 3% above 20 (ig/dL.
Further summaries of blood-lead concentrations in this study are presented in Section 3.4.
A multiple regression approach using backward selection techniques (Neter and
Wasserman, 1974) was used to determine those environmental variables and questionnaire
variables most important to predicting blood-lead concentration in children. In addition to wipe
dust-lead loading, the following factors were significantly associated with increased blood-lead
concentrations among children: African-American race, children engaging in soil pica, single
parent household, and having a high ferritin level. Adjusting for these factors, wipe dust-lead
loading accounted for 10.1% of the variation in blood-lead concentrations (HUD, 1995a).
The Rochester Study also investigated the relationship between soil-lead concentration
and children's blood-lead concentration. One composite soil sample was obtained from a
maximum of 12 core samples (3 per side of house) taken two feet away from the foundation, and
a second composite sample was obtained from 8-10 samples taken where the child frequently
played. A coring device was used to take samples at a depth of Vz inches only where bare soil
was present. Twenty-seven percent of the children in this study exhibited soil pica. Soil-lead
concentration was a significant (positive) predictor of blood-lead concentration, even when
adjusting for dust-lead loading.
The Rochester Study concluded that lead-contaminated dust affects children's blood-lead
levels, even when those levels are in the low to moderate range (< 25 ug/dL). This relationship
differs according to the dust sampling method and the type of surface sampled. At the relatively
low levels of dust-lead in this study, dust-lead loadings were found to be a better predictor of
Draft - Do Not Cite or Quote 65 September 27, 1996
-------
blood-lead concentration than were dust-lead concentrations. The study suggests that of the three
dust collection methods considered, dust-lead results from samples collected using either wipe or
BRM-vacuum methods should be used when making inferences on children's blood-lead
concentrations.
Summaries of environmental-lead levels observed in the Rochester Study are presented in
Section 3.3. Data from the Rochester study were employed to develop the epidemiologic model
used in this risk assessment (Section 4.3.2).
3.2.3 Urban Soil Lead Abatement Demonstration Project (USLADP)
The USLADP, authorized hi 1986 under the Superfund Amendments and Reauthorization
Act, was conducted to determine whether reducing lead levels in soil accessible to children
decreases their blood-lead concentration. While other observational studies of childhood lead
exposure such as the Rochester Study have shown that differences in soil lead exposure are
associated with differences in blood-lead concentration, this project specifically addressed
whether controlled reductions in external soil lead exposure were associated with reductions in
blood-lead concentrations. The USLADP consisted of three studies conducted in Baltimore,
MD, Boston, MA, and Cincinnati, OH. This project considered soil abatements hi urban areas
and focused on inner-city children.
In Baltimore, data were analyzed for 185 children aged 6 to 72 months. These children
resided in either the study area (expectation of moderate risk of lead poisoning) or a control area.
The Boston study included 149 children aged 6 to 48 months, considered to be at risk for lead
exposure and residing hi one of the study areas (history of high incidence of lead poisoning).
Only children with blood-lead concentrations ranging from 7 to 24 ug/dL were included in the
Boston study. In Cincinnati, families with children under five years of age and residing in one of
the study areas (selected as having similar socioeconomic, housing type, etc. characteristics) were
enrolled in the study. Data for 206 children were analyzed from the Cincinnati study.
Within each city, a series of neighborhoods were considered in the study from which the
participating households were selected. Selected units within certain neighborhoods were to
have interventions performed, while units in other neighborhoods were selected as control units.
For purposes of data summary and analysis, study units were grouped according to intervention
Draft - Do Not Cite or Quote 66 September 27, 1996
-------
strategy. Environmental media sampled included dust, soil, drinking water, and paint.
Household interviews were also conducted to obtain information on such factors as household
behavior and socioeconomic status.
The following two main conclusions were drawn from the USLADP (EPA, 1996):
1. "When soil is a significant source of lead in the child's environment, under certain
conditions, the abatement of that soil will result in a reduction in exposure that will
cause a reduction in childhood blood lead concentrations."
2. "Although these conditions for a reduction in blood are not fully understood, it is
likely that five factors are important in determining the magnitude of any possible
reduction: (1) the past history of exposure of the child to lead, as reflected in the pre-
abatement blood lead; (2) the initial soil lead concentration and the magnitude of the
reduction in soil lead concentrations; (3) the initial interior house dust lead loading
and the magnitude of reduction in house dust lead loading; (4) the magnitude of other
sources of lead exposure, relative to soil; and (5) the strength of the exposure
pathway between soil and the child relative to other lead exposure pathways in the
child's environment."
In taking soil samples, the entire soil region surrounding the residence was partitioned
into distinct areas (e.g., front, back), and samples were taken from each partition. At each core
sample, the top 2" and bottom 2" of the sample core were retained. A single core sample was
taken when less than two meters in either direction were available for sampling. Larger areas had
core samples taken at the foundation and at the boundary.
Dust samples were collected by vacuum methods in all three cities. In Baltimore, the
Sirchee-Spittler dust buster vacuum sampler without a frame (a 4' x 4' sample area is demarcated
with tape) was used. A minimum of three areas were sampled: the main entrance to the
household and two areas often frequented by the child when playing. In Boston, the same
sampler and sampling sites were used, but a plastic 25 cm x 25 cm frame was used instead of
tape. In Cincinnati, a personal air monitoring vacuum pump was used with a plastic 25 cm x 25
cm frame. Dust was sampled from a floor area adjacent to the main entrance; a composite of
dust samples from at least 3 floors areas including the child's bedroom and a high traffic area in
the main living area; a composite of dust samples from at least 3 window well and sill areas
including from within the child's bedroom and the main living area; dust from an entryway
Draft - Do Not Cite or Quote 67 September 27, 1996
-------
floormat placed by sample collection personnel. Dustfall and exterior surface dust were also
measured.
Data from the USLADP studies became available for analysis and summary in 1996,
when this risk assessment was being performed. As a result, these data were not summarized in
this report, and the data have not been used in the analysis efforts of the risk assessment.
3.2.4 Birmingham Urban Lead Uptake Study
This study was conducted in Birmingham, England, from 1984-1985, and consisted of
183 randomly-selected children, aged 24 months (± 2 months), born in and still residing in urban
Birmingham. A stratified subset of 106 children were selected for the study, of which 97
completed the study. The objective of the study was to simultaneously examine lead uptake via
all identified environmental pathways for young children hi an urban environment.
Soil samples were collected using a stainless-steel trowel surface scrape (0-5 cm). One
composite soil sample was obtained from 25 core samples. A specially adapted vacuum was
used to collect dust samples from the child's main play area, the child's bedroom and under the
doormat. All exposed floor space was sampled. Samples were also taken from the bag of the
vacuum cleaner most often used by the household.
The main conclusion from this study was that childhood blood-lead concentration was
found to be significantly associated with a combination of dust-lead loading, the rate of touching
objects, water-lead concentration, and smoking habits of the parents. Only an estimated 3% of a
child's average total uptake of lead per day was attributed to breathable air; the remainder was
attributed to dust, food, and water ingestion.
Because this study was conducted outside of the United States over ten years ago, it was
considered less representative of current childhood lead exposure in the United States than more
recent studies. Therefore, the data from this study were not used in the §403 risk assessment.
3.2.5 Cincinnati Longitudinal Study
Objectives of the Cincinnati Longitudinal study, conducted from 1980-1987, were to
provide a complete picture of a child's lead exposure history and to investigate the factors
responsible for excessive lead exposure. Approximately 250 expectant mothers residing within a
Draft - Do Not Cite or Quote 68 September 27, 1996
-------
prespecified set of census tracts in the Cincinnati (OH) area were enrolled for this study. These
census tracts were identified as having a long history of producing children with elevated blood
lead levels. The mothers were patients at one of three prenatal clinics. Once these mothers
delivered, blood-lead concentrations were measured in the children from birth through 5 years of
age.
Soil samples were collected by surface scrapings. Surface scrapings were collected from
the child's play area outside, if one existed. Interior dust samples were collected using a personal
sized vacuum within 484 cm2 plastic frame area wherever child frequents. A maximum of five
sites were sampled within the home. Each sample entails three sweeps of the vacuum within the
frame. Exterior dust samples were collected via scraping exterior surfaces with a stainless steel
spatula. Paint-lead levels from a maximum of IS surfaces were measured using XRF techniques.
Dust collection from children's hands was performed via repeated wiping with multiple pre-
moistened wipes.
This study observed high levels of lead contamination in the residential environments,
with most contamination occurring in areas immediately outside of the unit and within the
entranceways. Statistical analyses indicated that the pathway from exterior dust to interior dust
to hands to blood was of most significance in this study.
The data for this study were not available to the §403 risk assessment effort.
3.2.6 Brigham and Women's Hospital Longitudinal Study
The objective of this early study was to examine the relationship between children's
blood-lead levels and various environmental factors from late pregnancy to two years of age.
Children were selected from births occurring between April 1979 and April 1981 at Brigham and
Women's Hospital in Boston, MA. Births were categorized into the highest, lowest, and middle
deciles of umbilical cord blood lead. The 249 infants selected were nearly equally drawn from
three distinct categories of cord blood levels. All families resided within a 12 mile radius of
hospital, spoke English as their primary language, and the infants had no serious illness.
Umbilical cord blood was collected, as was blood at 6,12,18, and 24 months of age. The
sample was predominantly white, middle- to upper-middle class families in an urban
environment.
Draft - Do Not Cite or Quote 69 September 27, 1996
-------
Soil samples were collected at a distance of 3 meters from the road. Dust samples were
collected at 1,6,18, and 24 months using wipe techniques from a living room surface (floor or
furniture top) and from a window sill. Samples were collected from within a plastic frame with a
930 cm2 opening (a 465 cm2 opening for window sills). Lead levels in paint were measured by a
PGT model XE-3 XRF instrument. Air samples were collected from personal air monitors, and
drinking water samples were collected from the kitchen tap after a 4-liter flush.
Mean blood-lead concentrations at 24 months was 6.8 ug/dL. At 24 months, blood-lead
concentration was found to be significantly associated with soil-lead levels, dust-lead levels, the
presence of deteriorated paint, and the occurrence of recent refinishing activities at the residence.
Water-lead and airborne-lead levels were not significant factors. While none of the children
were classified as having excessive pica tendencies, evidence existed between mouthing
tendencies and increased blood-lead concentration. These findings agreed with earlier studies
which considered children with higher blood-lead concentrations. In addition, blood-lead
concentrations were found to be approximately 44% higher within specimens collected in
summer months, indicating a possible seasonality factor associated with blood-lead
concentration.
Due to the age of the study and its focus on a specific area of the country, data from this
study were not used in the §403 risk assessment.
3.3 LEAD IN DUST. SOIL. AND PAINT IN THE NATION'S HOUSING
This section provides information on the distribution of environmental-lead levels in the
nation's housing stock, with a focus on lead in residential dust, soil, and paint. The §403 risk
assessment uses data from the HUD National Survey of Lead-Based Paint in Housing to
characterize the distribution of environmental-lead levels in the nation's occupied housing stock
in 1997. These environmental-lead data are summarized in Section 3.3.1. To provide supporting
information on environmental-lead levels in occupied housing, environmental-lead data from the
Baltimore R&M Study and the Rochester Study are also summarized in this section. Section
3.3.1 also includes estimated numbers of occupied housing units in the 1997 national housing
stock.
Draft - Do Not Cite or Quote 70 September 27. 1996
-------
To provide a link between childhood and residential environmental lead exposures,
Section 3.3.2 presents estimated numbers of children of specific age groups in 1997 residing
within housing units of specific ages.
3.3.1 The Distribution of Lead Levels in Household Dust, Soil, and Paint
In this section, environmental-lead levels in residences are summarized for three studies.
The first study presented, the HUD National Survey, is used to characterize environmental-lead
levels in the nation's occupied housing stock in 1997, prior to §403 interventions. The other two
studies, the Baltimore R&M Study and the Rochester Study, provide supporting information on
environmental-lead levels for specific housing groups or exposure conditions.
3.3.1.1 HUD National Survey
For the §403 risk assessment effort, the primary source of information on environmental-
lead levels in the national housing stock was the National Survey of Lead-Based Paint in
Housing (EPA, 1995a). This survey was sponsored by the U.S. Department of HUD, in response
to a mandate in the 1987 amendments to the Lead-Based Paint Poisoning Prevention Act to
obtain "an estimate of the amount, characteristics and regional distribution of housing in the
United States that contains lead-based paint hazards at differing levels of contamination."
Conducted in 1989-1990, the privately-owned unit portion of the survey (cited as the "HUD
National Survey" in this document) measured lead levels in paint, dust, and soil within 284
privately-owned, occupied housing units. The units were selected via a statistically-based
sampling design to represent the national housing stock built prior to 1980. Units built in 1980
or later were not included in the survey, as they were assumed to be free of lead-based paint
because of the Consumer Product Safety Commission's 1978 ban on the sale of lead-based paint
and its use in residences.
In the HUD National Survey, lead loadings (fig of lead per square-feet of area sampled)
and lead concentrations (ug of lead per gram of sample) were measured from dust samples
collected on floors, window sills, and window wells. Dust samples were collected using the Blue
Nozzle vacuum method. Lead concentrations in the soil at each unit were measured by collecting
soil samples along the foundation, the entryway to the unit, and from remote areas in the yard,
Draft - Do Not Cite or Quote 71 September 27, 1996
-------
using a soil corer with plunger. Lead concentrations in paint (milligrams of lead per square-
centimeter of painted surface) were measured using in situ XRF techniques in selected rooms as
well as on the exterior of the unit.
In the §403 risk assessment effort, data from the 284 privately-owned units in the HUD
National Survey were used to characterize environmental-lead levels in the nation's occupied
housing. Table C-7 of Appendix C contains the following summary of environmental-lead
levels for each of these units:
• two weighted arithmetic averages of dust-lead loading: one for floors and one for
window sills (where sample results were weighted according to area of sample)
• two weighted arithmetic averages of dust-lead concentration: one for floors and one
for window sills (where sample results were weighted according to mass of sample)
• the weighted arithmetic average soil-lead concentration (where remote sample results
were weighted twice that of the entryway and dripline results)
• the maximum observed paint-lead concentration for the ulterior and the exterior, as
measured by XRF techniques.
Note that the last bullet indicates the maximum observed (or measured) paint-lead concentration
in a unit. To identify whether a unit was suspected of containing LBP, statistical modeling was
performed in the HUD National Survey to obtain a predicted maximum XRF paint-lead
concentration for each unit. If the predicted maximum XRF value for a unit was at least 1.0
mg/cm2, the unit was considered to contain LBP. In the §403 risk assessment effort, the
predicted maximum XRF value was used only to identify the presence of LBP for each unit.
In the HUD National Survey, each unit was assigned a sampling weight equal to the
number of pre-1980 privately-owned, occupied units in the national housing stock that were
represented by the given unit in the survey. The sum of all 284 sampling weights equaled the
number of pre-1980 privately-owned, occupied units in the national housing stock at the time of
the survey. Sampling weights in the HUD National Survey were determined according to four
demographic variables associated with the units:
Draft - Do Not Cite or Quote 72 September 27, 1996
-------
• Age category of unit
• Number of units in the building
• Census region
• Presence of a child under age 7 years
In order to use the environmental-lead levels from the HUD National Survey to
characterize environmental-lead levels in the 1997 national housing stock, it was necessary to
revise the sampling weights of the HUD National Survey units to represent the 1997 occupied
housing stock, both publicly-owned and privately-owned. The method for revising the sampling
weights is documented in Section C.I. 1.2 of Appendix C; Table C-7 of Appendix C lists the
revised weights for each unit. The revised weights, therefore, indicate the number of units in the
1997 national housing stock that are represented by the given HUD National Survey unit, and
therefore, represented by its environmental-lead levels. The estimated numbers of units in the
1997 national housing stock are presented in Table 3-3, within four age categories.
Table 3-3. Estimated Total Number of Occupied Housing Units in the National Housing
Stock in 1997 According to Year-Built Category
Year In Which
the Unit Was
Built
Pre-1940
1940-1959
1960-1979
Post- 1979
Number of
National Survey
Units
77
87
120
281
Estimated Total:
Estimated Numbers of
Units in the 1997 National
Housing Stock
19,676,000
19,718,000
34,985,000
24,893,000
99,272,000
1 Units built from 1960-1979 and containing no lead-based paint were
also placed in this category.
The HUD National Survey did not consider units built after 1979, as all such units were
assumed to be free of lead-based paint due to the Consumer Product Safety Commission's 1978
ban on the sale of lead-based paint and its use in residences. Therefore, in characterizing the
1997 national housing stock from the HUD National Survey, post-1979 housing was represented
by the 28 units built between 1960 and 1979 and containing no lead-based paint (i.e., the
Draft - Do Not Cite or Quote
73
September 27, 1996
-------
predicted maximum amount of lead in paint within the unit was less than 1.0 mg/cm2).
Therefore, the revised sampling weight for these 28 units was subdivided into two parts: one
part representing 1960-1979 units, and the other representing post-1979 units. See Section
C.I. 1.3 of Appendix C on the rationale for selecting these 28 units to represent the post-1979
housing stock and on the method for obtaining the portion of the sampling weight representing
post-1979 units.
Using the environmental-lead levels and the updated 1997 sampling weights for the HUD
National Survey units from Appendix C, Tables 3-4 and 3-5 are a predicted summary of lead
loadings and concentrations, respectively, in floor-dust samples across units in the 1997 housing
stock. Tables 3-6 and 3-7 summarize lead loadings and concentrations, respectively, in window
sill-dust samples. Table 3-8 summarizes lead concentrations in soil. Table 3-9 presents
summaries of the observed maximum XRF value from the National Survey units, weighted by
the updated 1997 sampling weights. The percentages of units in the 1997 housing stock having
lead-based paint, as well as the percentages having damaged lead-based paint, are estimated in
Table 3-10.
The tables that summarize dust-lead loadings and dust-lead concentrations (Tables 3-4
through 3-7) indicate that the geometric means and the medians decrease with the age of the unit.
This finding is consistent with the hypothesis that the potential for dust contamination by lead is
higher in older units, due to their propensity to contain lead-based paint and to be located in older
neighborhoods with lead-contaminated soil. Window sill dust-lead levels in units built prior to
1940 were considerably higher than those of the other units. These tables also indicate that lead
levels in dust tend to be higher on window sills than floors, especially in older units. The same
trends were observed in soil-lead concentration (Table 3-8), whose geometric mean and median
decreased with the age of the unit, and whose levels were considerably higher in pre-1940 units
than in the other units.
The components tested for lead-based paint in the HUD National Survey were selected
based on a predetermined sample design that did not target only components suspected of having
lead-based paint. Therefore, the summarized values in Table 3-9 represent both lead-
contaminated and lead-free painted surfaces. The relationship between lead levels in paint and
age of unit is strongest for the median and upper percentiles, indicating that while low paint-lead
Draft - Do Not Cite or Quote 74 September 27, 1996
-------
measurements are likely to be observed in all housing regardless of age, large paint-lead
measurements are more likely to be observed in older units.
This risk assessment predicts that approximately 62% of the 1997 occupied housing stock
will contain LBP (Table 3-10), based on information from the HUD National Survey and under
the assumption that no units built after 1979 contain LBP. This percentage is less than 83%, the
percentage of pre-1980 occupied housing predicted to contain LBP according to the HUD
National Survey (EPA, 1995a). The estimate of 62% is relative to all occupied housing, even
units built after 1979. The percentages of units with LBP within the three pre-1980 year-built
categories match those reported in the National Survey report (EPA, 1995a). Table 3-10 also
indicates that approximately 14% of units are predicted to contain more than five square feet of
deteriorated LBP, with over half of these units built prior to 1940.
Table 3-4. Summary of the Distribution of Lead Loadings in Floor-Dust Samples Within
Housing Units in the HUD National Survey, Weighted to Reflect the Predicted
1997 Housing Stock
Year Unit Was Built
Before 1940
1940-1959
1960-1979
1960- 1979 units
with no LBP2
Floor Dust-Lead Loadings (fig/ft1)1
Geometric
Mean
43.6
22.5
13.8
10.3
Geometric
Standard
Deviation
3.13
2.93
2.52
2.04
5th
Percentile
7.19
3.65
3.79
4.17
25th
Percentile
17.9
11.3
7.05
5.77
Median
32.9
22.1
12.9
9.15
75th
Percentile
127.
52.9
26.6
18.1
95th
220.
147.
57.8
42.1
1 Data summarized in this table are area-weighted arithmetic mean dust-lead loadings from floors for the 284 privately-
owned, occupied National Survey units (see Appendix C). These loadings are converted to represent loadings from dust
samples obtained from wipe collection techniques. In the summaries, each unit is weighted by its 1997 weight, which is
presented in Appendix C.
2 Units with no LBP have a predicted maximum XRF value (interior and exterior) less than 1.0 mg/cm2. These units represent
post-1979 units in the S403 risk assessment effort.
Draft - Do Not Cite or Quote
75
September 27. 1996
-------
Table 3-5. Summary of the Distribution of Lead Concentrations in Floor-Dust Samples
Within Housing Units in the HUD National Survey, Weighted to Reflect the
Predicted 1997 Housing Stock
Year Unit Was Built
Before 1 940
1940-1959
1960-1979
1960-1 979 units
with no LBP2
Floor Dust-L
Geometric
Mean
511.
204.
123.
93.
Geometric
Standard
Deviation
4.0
2.6
3.0
2.2
5th
PBrCBfltllB
96.9
44.7
20.2
22.2
Bad Concentrations (f/g/g)1
25th
Percentile
297.
102.
68.2
53.8
Median
589.
215.
134.
90.5
75th
P6rc0ntil6
831.
322.
207.
164.
95th
P6rc6ntilo
6320.
1240.
487.
458.
1 Data summarized in this table are mass-weighted arithmetic mean dust-lead concentrations from floors for the 284
privately-owned, occupied National Survey units (see Appendix C). These concentrations were adjusted to reflect the
weight of the entire dust sample, not just the tap weight. In the summaries, each unit is weighted by its 1997 weight,
which is presented in Appendix C.
2 Units with no LBP have a predicted maximum XRF value (interior and exterior) less than 1.0 mg/cm2. These units
represent post-1979 units in the §403 risk assessment effort.
Table 3-6. Summary of the Distribution of Lead Loadings in Window Sill-Dust Samples
Within Housing Units in the HUD National Survey, Weighted to Reflect the
Predicted 1997 Housing Stock
Year Unit Was Built
Before 1940
1940-1959
1960-1979
1960- 1979 units
with no LBP2
Window Sill Dust-Lead Loadings (pg/ft2)1
Geometric
Mean
208.
33.1
24.9
13.5
Geometric
Standard
Deviation
12.2
8.33
11.0
8.02
5th
Porcontite
2.36
0.568
0.474
0.315
25th
PwCBfltllS
50.0
9.74
4.10
5.04
Median
243.
29.6
26.8
12.8
75th
Psrcontilft
1440.
143.
227.
73.9
95th
Percentile
9640.
795.
753.
180.
1 Data summarized in this table are area-weighted arithmetic mean dust-lead loadings from window sills for the 284
privately-owned, occupied National Survey units (see Appendix C). These loadings are converted to represent loadings
from dust samples obtained from wipe collection techniques. In the summaries, each unit is weighted by its 1997
weight, which is presented in Appendix C.
2 Units with no LBP have a predicted maximum XRF value (interior and exterior) less than 1.0 mg/cm2. These units
represent post-1979 units in the §403 risk assessment effort.
Draft - Do Not Cite or Quote
76
September 27, 1996
-------
Table 3-7. Summary of the Distribution of Lead Concentrations in Window Sill-Dust
Samples Within Housing Units in the HUD National Survey, Weighted to
Reflect the Predicted 1997 Housing Stock
Year Unit Was Built
Before 1 940
1940-1959
1960-1979
1960-1 979 units
with no LBP2
Window Sill Dust-Lead Concentrations (frg/g)1
Geometric
Mean
1730.
487.
388.
247.
Geometric
Standard
Deviation
5.2
4.0
4.9
3.3
5th
Percentila
103.
48.1
28.7
26.0
25th
Percentile
698.
244.
138.
129.
Median
1740.
378.
519.
274.
75th
P0rc6ntilo
6700.
1360.
1540.
503.
95th
Percentile
56500.
3230.
1650.
1250.
1 Data summarized in this table are mass-weighted arithmetic mean dust-lead concentrations from window sills for the 284
privately-owned, occupied National Survey units (see Appendix C). These concentrations were adjusted to reflect the
weight of the entire dust sample, not just the tap weight. In the summaries, each unit is weighted by its 1997 weight,
which is presented in Appendix C.
2 Units with no LBP have a predicted maximum XRF value (interior and exterior) less than 1.0 mg/cm2. These units
represent post-1979 units in the §403 risk assessment effort.
Table 3-8. Summary of the Distribution of Soil-Lead Concentrations for Housing Units in
the HUD National Survey, Weighted to Reflect the Predicted 1997 Housing
Stock
Year Unit Was Built
Before 1940
1940-1959
1960-1979
1960- 1979 units
with no LBP2
Soil-Lead
Geometric
Mean
464.
91.7
32.7
22.4
Geometric
Standard
Deviation
3.1
3.2
2.6
2.3
5th
Percentile
39.5
22.0
6.11
5.58
concentrations Vfl'fli
25th
Percentile
258.
44.3
19.7
13.6
Median
569.
75.8
28.6
21.2
75th
Percentile
1160.
146.
58.3
45.0
95th
Percentile
2020.
485.
186.
82.5
1 Data summarized in this table are weighted arithmetic mean soil-lead concentrations for the 284 privately-owned,
occupied National Survey units (see Appendix C). Within each unit's average, remote sample results were weighted
twice that of the entryway and dripline results. In the summaries, each unit was weighted by its 1997 weight, which is
presented in Appendix C.
2 Units with no LBP have a predicted maximum XRF value (interior and exterior) less than 1.0 mg/cm2. These units
represent post-1979 units in the §403 risk assessment effort.
Draft - Do Not Cite or Quote
77
September 27, 1996
-------
Table 3-9. Summary of the Distribution of Observed Maximum XRF Lead Levels in Paint
for Housing Units in the HUD National Survey, Weighted to Reflect the
Predicted 1997 Housing Stock
Year Unit
Was Built1
#
National
Survey
Units3
Observed Maximum XRF Paint-Lead Levels (rug/cm*)'
Geometric
Mean
Geometric
Standard
Deviation
5th
Percentile
25th
B «MkdkM*SI Jk.
r8rc6fTui9
Median
75th
Percentile
95th
Percentile
Interior of Unit
Before 1 940
1940-1959
1960-1979
72
83
116
1.86
1.02
0.71
3.78
2.42
1.79
0.30
0.40
0.30
0.60
0.60
0.60
0.90
0.70
0.60
6.40
1.60
0.90
20.2
8.00
1.60
Exterior of Unit
Before 1940
1940-1959
1960-1979
60
76
103
3.14
1.45
0.72
3.75
3.05
2.43
0.30
0.20
0.00
0.70
0.60
0.50
4.00
1.50
0.60
7.10
2.60
0.90
26.9
10.3
3.60
Data summarized in this table are observed maximum XRF paint-lead level for National Survey units across both interior and
exterior painted surfaces (see Appendix C). Each unit's observed maximum XRF paint-lead level was weighted by the 1997
weight for the unit, which is presented in Appendix C.
2 No units built after 1979 were included in the HUD National Survey. In the S403 risk assessment effort, these units are
assumed to be free of LBP.
3 Number of privately-owned units in the HUD National Survey in which an observed maximum XRF paint-lead level was
available (for either the interior or exterior).
Table 3-10. Predicted Numbers and Percentages of Units Having Lead-Based Paint in the
1997 Occupied Housing Stock, Based on Information from the HUD National
Survey1
Year Unit Was Built
Before 1940
1940-1959
1960-1979
After 1979
All Housing
Number <%» of Units with
Lead-Based Paint
17,248,000(87.7%)
18,047,000(91.5%)
26,452,000 (75.6%)
0(0%)
61,747,000(62.2%)
Number (%) of Units with More
Than 5 ft2 of Lead-Based Paint
7,755,000 (39.4%)
3,065,000(15.5%)
2,651,000(7.6%)
0 (0%)
13,470,000(13.6%)
1 A unit in the HUD National Survey is labeled as containing LBP if its predicted
maximum XRF value in either the interior or the exterior is at least 1.0 mg/cm2.
Results are weighted using the 1997 weights presented in Appendix C.
Draft - Do Not Cite or Quote
78
September 27, 1996
-------
In preparing summaries of dust concentration data in Tables 3-5 and 3-7, these data were
adjusted for the effect of underestimated sample weights. "Tap weight" is the portion of a dust
sample that was tapped out of the sample collection filter. In the HUD National Survey, the dust-
lead concentration equaled the amount of lead in the entire dust sample, divided by the tap
weight. The adjustment to dust-lead concentration data attempted to correct for the weight of the
entire dust sample, not just the tap weight. Appendix Z contains details on the adjustment
method. Lead concentrations for dust samples with a tap weight of less than 0.7 mg were set to
missing.
As discussed in Section C.I.3, missing values for dust-lead loading, dust-lead
concentration, or soil-lead concentration for a National Survey unit were replaced by nonmissing
values prior to the data summaries in Tables 3-5 through 3-8. For a particular data parameter, the
value assigned to a unit having missing values equaled the average value across units within the
same category of year built and having the same lead-based paint status (i.e., presence or absence
of a maximum XRF value in the interior or exterior at 1.0 mg/cm2 or above).
The §403 dust lead standard will be defined as a lead loading of a wipe sample. Dust
samples in the HUD National Survey were collected using the Blue Nozzle vacuum sampler.
The Blue Nozzle vacuum dust-lead loadings were converted to wipe equivalent dust-lead
loadings using the conversion equations presented in Section 4.2.
3.3.1.2 The Baltimore Repair and Maintenance (R&M) Study
In the Baltimore R&M Study (Section 3.2.1), the BRM vacuum was used to collect dust
samples. Dust-lead loadings were converted to wipe equivalent dust-lead loadings using the
conversion equations presented in Section 4.2. Tables 3-11 and 3-12 summarize pre-intervention
lead loadings and concentrations, respectively, in floor-dust samples across study units. Tables
3-13 and 3-14 summarize lead loadings and concentrations, respectively, in window sill-dust
samples. Table 3-15 summarizes lead concentrations in soil samples taken at the dripline. Table
3-16 presents summaries of the observed maximum XRF value within study units slated for
R&M interventions in the study, for the interior only and the exterior only, as well as for the
entire unit. XRF measurements were not made in the previously abated and modem urban
homes.
Draft - Do Not Cite or Quote 79 September 27, 1996
-------
Table 3-11. Summary of Average Pre-lntervention Floor Dust-Lead Loading for Housing
Units in the Baltimore R&M Study
Unit Category
All Study Units
Previously Abated Units
Units Slated for R&M
Intervention
Modern Urban Units
# Units
90
16
58
16
Floor Dust-Lead Loading (/ig/ft2)1
Geometric
Mean
31.89
35.62
44.41
8.60
Geometric
Standard
Deviation
2.11
1.50
1.54
1.54
Minimum
3.78
17.67
18.63
3.78
Maximum
157.48
85.13
157.48
15.28
1 Data summarized in this table are area-weighted arithmetic mean dust-lead loadings from floors for each unit. These
loadings have been converted to represent loadings from dust samples obtained from wipe collection techniques.
Table 3-12. Summary of Average Pre-lntervention Floor Dust-Lead Concentrations for
Housing Units in the Baltimore R&M Study
Unit Category
All Study Units
Previously Abated Units
Units Slated for R&M
Intervention
Modern Urban Units
# Units
90
16
58
16
Floor Dust-Lead Concentration f/ig/g)1
Geometric
Mean
1303.22
1214.02
2436.50
144.79
Geometric
Standard
Deviation
4.06
2.55
2.61
2.19
Minimum
48.85
331.86
425.62
48.85
Maximum
60304.19
7357.76
60304.19
704.20
1 Data summarized in this table are area-weighted arithmetic mean dust-lead concentrations from floors for each unit.
Draft - Do Not Cite or Quote
80
September 27, 1996
-------
Table 3-13. Summary of Average Pre-lntervention Window Sill Dust-Lead Loading for
Housing Units in the Baltimore R&M Study
Unit Category
All Study Units
Previously Abated Units
Units Slated for R&M
Intervention
Modern Urban Units
# Units
90
16
58
16
Window Sill Dust-Lead Loading (/ig/ft2)'
Geometric
Mean
320.64
153.34
668.48
46.75
Geometric
Standard
Deviation
3.27
2.21
1.62
1.59
Minimum
22.80
45.52
201.82
22.80
Maximum
1869.49
840.35
1869.49
80.74
1 Data summarized in this table are area-weighted arithmetic mean dust-lead loadings from window sills for each unit.
These loadings have been converted to represent loadings from dust samples obtained from wipe collection techniques.
Table 3-14. Summary of Average Pre-lntervention Window Sill Dust-Lead Concentrations
for Housing Units in the Baltimore R&M Study
Unit Category
All Study Units
Previously Abated Units
Units Slated for R&M
Intervention
Modern Urban Units
# Units
90
16
58
16
Window Sill Dust-Lead Concentration (pg/g)1
Geometric
Mean
5601.53
1881.89
20148.29
160.98
Geometric
Standard
Deviation
8.54
4.61
2.42
2.66
Minimum
7.25
254.63
2809.83
7.25
Maximum
141056.96
31497.47
141056.96
447.12
1 Data summarized in this table are area-weighted arithmetic mean dust-lead concentrations from window sills for each
unit.
Draft - Do Not Cite or Quote
81
September 27, 1996
-------
Tables 3-11 through 3-14 indicate that geometric mean levels are highest for units slated
for R&M intervention, while modern urban units have geometric mean levels that are an order of
magnitude lower than the other two housing groups. Units slated for R&M interventions should
not be considered representative of inner city homes. Prior to the interventions, many of them
were in poor condition. They might be considered to represent the worst case of residential
environmental-lead levels. The floor dust-lead loadings for previously-abated units and units
slated for R&M interventions are comparable to those reported for pre-1940 housing in the
National Survey, while these units have considerably higher window sill dust-lead levels
compared to the National Survey. Units slated for R&M interventions have very high dust-lead
concentrations and window sill dust-lead loadings, due to the deteriorated condition of most of
these units. Modern urban units have dust-lead levels that are similar to the National Survey
units built from 1960-1979 and containing no LBP.
Soil-lead concentrations summarized in Table 3-IS are based on small numbers of units,
due to the lack of available soil to sample for many of the study units. Geometric mean soil-lead
concentrations presented in Table 3-15 are high compared to those in the HUD National Survey.
The paint-lead measurements summarized in Table 3-16 are extremely high, as the data
represent only units slated for R&M interventions. Paint-lead measurements were taken
primarily from components suspected of containing LBP, in order to identify and prioritize
surfaces requiring LBP intervention. Thus, the data summarized hi Table 3-16 reflect a LBP-
contaminated environment and are not typical of all painted surfaces in occupied housing.
Draft - Do Not Cite or Quote 82 September 27, 1996
-------
Table 3-15. Summary of Average Pre-lntervention Dripline Soil-Lead Concentrations for
Housing Units in the Baltimore R&M Study
Unit Category
All Study Units
Previously Abated Units
Units Slated for R&M
Intervention
Modern Urban Units
# Units
28
2
16
10
Soil-Lead Concentration (/ig/g)1
Geometric
Mean
444.48
2192.33
1258.08
61.13
Geometric
Standard
Deviation
5.06
1.60
1.95
1.65
Minimum
28.85
1570.20
334.73
28.85
Maximum
3539.16
3060.94
3539.16
153.69
1 Data summarized in this table are area-weighted arithmetic mean dripline soil-lead concentrations for each unit.
Table 3-16. Summary of Observed Maximum XRF Paint-Lead Concentration at Pre-
lntervention for Housing Units Slated for R&M Intervention in the Baltimore
R&M Study1
Location Within a
Unit
Entire Unit
Exterior Only
Interior Only
# Units
36
35
36
Observed Maximum XRF Paint-Lead Concentration (mg/cm2)
Geometric
Mean
38.38
24.75
28.22
Geometric
Standard
Deviation
1.68
2.63
1.80
Minimum
9.30
0.60
7.40
Maximum
98.10
86.30
98.10
1 XRF data were not available for previously-abated units and modern urban units in the study, as they were assumed to be
free of LBP.
3.3.1.3 The Rochester Lead-in-Dust Study
Tables 3-17 and 3-18 summarize lead loadings and concentrations, respectively, in floor-
dust samples across study units. Tables 3-19 and 3-20 summarize lead loadings and
concentrations, respectively, in window sill-dust samples. Table 3-21 summarizes lead
concentrations in soil samples taken at the dripline, while Table 3-22 summarizes lead
concentrations in soil samples taken at the child's play area. Table 3-23 presents summaries of
the observed maximum XRF value within the study units, for the interior only and the exterior
only, as well as for the entire unit. All of these tables summarize results across all units, as well
as within the four age categories in which the HUD National Survey units were categorized in
Section 3.3.1.1.
Draft - Do Not Cite or Quote
83
September 27, 1996
-------
Note from Table 3-17 that approximately 84% of the study units were built prior to 1940.
Therefore, while the Rochester study considers units in an urban environment and does not
attempt to target a particular lead exposure environment in recruiting the units, most of the units
are older units and contain families with low income levels.
As seen in the HUD National Survey, dust-lead loadings and concentrations are highest
among the units built prior to 1940 (Tables 3-17 through 3-20). For units built prior to 1980, the
geometric mean floor dust-lead levels were often lower in the Rochester study than in the HUD
National Survey. However, the ten units built after 1979 had higher geometric mean dust-lead
levels than for the 1940-1959 and 1960-1979 year-built categories. It is unclear why these ten
units would have such high dust-lead levels, other than existing within an urban environment.
Geometric mean soil-lead concentrations were higher than those observed in the HUD
National Survey. As seen with dust-lead levels, units built after 1979 had a higher geometric
mean soil-lead concentration compared to the 1960-1979 housing group. Less than half of the
units had soil samples taken from the play area (Table 3-22), where geometric mean
concentrations were generally lower than at the dripline for older units.
Table 3-23 indicates that at least 40% of the units within each age category (even houses
built after 1979) contained LBP as measured by XRF. Units built prior to 1940 had the highest
geometric mean paint-lead levels and the highest percentage of units with LBP in the study.
However, of most interest is the frequency to which LBP was detected in the ten units built after
1979. Two of these units contained LBP in the interior, while four contained LBP in the exterior.
It is possible that LBP exists within some of these units, as high lead levels were also observed in
dust and soil. As a result, one may not wish to exclude post-1979 housing in
urban settings when identifying LBP hazards in the nation's housing. However, interpretation of
XRF results should take into account imprecisions associated with the measurements. For
example, units with a maximum XRF value slightly greater than 1.0 mg/cm2 are classified as
having LBP, but the maximum value may be statistically equivalent to a value less than 1.0
mg/cm2.
Draft - Do Not Cite or Quote 84 September 27, 1996
-------
Table 3-17. Summary of Average Pre-lntervention Floor Dust-Lead Loading for Housing
Units in the Rochester Study
Year Unit Was Built
All Units
Before 1940
1940-1959
1960-1979
After 1979
# Units
205
172
19
4
10
Floor Dust-Lead Loading (//g/ft2)1
Geometric
Mean
17.70
19.79
8.36
7.84
15.00
Geometric
Standard
Deviation
3.20
3.18
2.61
2.40
3.34
Minimum
1.21
1.66
1.21
2.13
3.48
Maximum
8663.53
8663.53
26.93
13.21
250.30
1 Data summarized in this table are area-weighted arithmetic mean dust-lead loadings from floors for each unit. Results
included in the summaries are only for dust samples collected using wipe techniques.
Table 3-18. Summary of Average Pre-lntervention Floor Dust-Lead Concentrations for
Housing Units in the Rochester Study
Year Unit Was Built
All Units
Before 1940
1940-1959
1960-1979
After 1979
# Units
204
172
18
4
10
Floor Dust-Lead Concentration (/ig/g)1
Geometric
Mean
350.84
395.68
209.43
60.80
226.14
Geometric
Standard
Deviation
3.74
3.58
4.61
2.68
2.99
Minimum
8.25
8.25
16.52
16.93
57.02
Maximum
40717.31
40717.31
7899.91
163.88
1 1 24.86
1 Data summarized in this table are area-weighted arithmetic mean dust-lead concentrations from floors for each unit.
Results included in the summaries are only for dust samples collected using BRM vacuum techniques.
Draft - Do Not Cite or Quote
85
September 27, 1996
-------
Table 3-19. Summary of Average Pre-lntervention Window Sill Dust-Lead Loading for
Housing Units in the Rochester Study
Year Unit Was Built
All Units
Before 1940
1940-1959
1960-1979
After 1979
# Units
196
164
18
4
10
Window Sill Dust-Lead Loading (//g/ft2)1
Geometric
Mean
196.03
233.68
72.01
52.31
113.09
Geometric
Standard
Deviation
3.96
3.67
6.16
1.38
1.95
Minimum
2.83
2.85
2.83
36.23
26.88
Maximum
14901.36
14901.36
4393.03
70.75
320.40
1 Data summarized in this table are area-weighted arithmetic mean dust-lead loadings from window sills for each unit.
Results included in the summaries are only for dust samples collected using wipe techniques.
Table 3-20. Summary of Average Pre-lntervention Window Sill Dust-Lead Concentrations
for Housing Units in the Rochester Study
Year Unit Was Built
All Units
Before 1940
1940-1959
1960-1979
After 1 979
# Units
199
166
19
4
10
Window Sill Dust-Lead Concentration (//g/g)1
Geometric
Mean
2787.03
3859.20
497.25
473.44
674.30
Geometric
Standard
Deviation
8.44
7.33
9.90
2.92
8.56
Minimum
3.15
15.85
5.31
159.77
3.15
Maximum
368111.11
368111.11
15017.51
1900.67
8625.00
1 Data summarized in this table are area-weighted arithmetic mean dust-lead concentrations from window sills for each
unit. Results included in the summaries are only for dust samples collected using BRM vacuum techniques.
Draft - Do Not Cite or Quote
86
September 27, 1996
-------
Table 3-21. Summary of Average Pre-lntervention Dripline Soil-Lead Concentrations for
Housing Units in the Rochester Study
Year Unit Was Built
All Units
Before 1940
1940-1959
1960-1979
After 1979
# Units
186
158
14
4
10
Soil-Lead Concentration (//g/g)1
Geometric
Mean
731.03
937.83
291.14
66.35
135.26
Geometric
Standard
Deviation
3.68
3.17
3.30
1.79
3.10
Minimum
12.31
12.31
29.70
29.00
26.00
Maximum
21049.00
21049.00
1788.00
1 1 1 .00
876.00
Data summarized in this table are area-weighted arithmetic mean dripline soil-lead concentrations for each unit.
Table 3-22. Summary of Average Pre-lntervention Soil-Lead Concentrations from Play
Areas for Housing Units in the Rochester Study
Year Unit Was Built
All Units
Before 1940
1940-1959
1960-1979
After 1979
# Units
87
79
6
1
1
Soil-Lead Concentration U/g/g}1
Geometric
Mean
266.90
277.54
184.51
138.00
215.00
Geometric
Standard
Deviation
2.78
2.79
3.09
.
•
Minimum
28.00
28.00
55.40
138.00
215.00
Maximum
7300.00
7300.00
767.00
138.00
215.00
Data summarized in this table are area-weighted arithmetic mean soil-lead concentrations from play areas for each unit.
Draft - Do Not Cite or Quote
87
September 27, 1996
-------
Table 3-23. Summary of Observed Maximum XRF Paint-Lead Concentration at Pre-
Intervention for Housing Units in the Rochester Study
Year the Unit
Was Built
#
Units
% of Units
with LBP1
Observed Maximum XRF Paint-Lead Levels (mg/cm*)
Geometric
Mean
Geometric
Standard
Deviation
Minimum
Maximum
Entire Unit (interior and exterior)
All Units
Before 1940
1940-1959
1960-1979
After 1 979
All Units
Before 1940
1940-1959
1960-1979
After 1979
205
172
19
4
10
89%
95%
68%
50%
40%
205
172
19
4
10
83%
91%
63%
0%
20%
12.80
16.63
5.50
1.04
1.93
3.88
3.05
4.97
1.89
5.92
Interior of Unit
7.58
9.90
2.86
0.61
1.32
4.38
3.67
4.71
1.12
5.72
0.50
0.50
0.50
0.57
0.50
59.77
59.77
37.57
1.93
39.43
0.50
0.50
0.50
0.57
0.50
57.57
57.57
32.87
0.73
39.43
Exterior of Unit
All Units
Before 1940
1940-1959
1960-1979
After 1979
204
171
19
4
10
79%
84%
63%
50%
40%
8.14
10.33
3.47
1.00
1.63
4.91
4.40
5.22
1.97
4.95
0.50
0.50
0.50
0.50
0.50
59.77
59.77
37.57
1.93
37.43
1 LBP is defined as the maximum XRF value exceeding 1.0 mg/cm*.
3.3.2 Characterizing the Population of Children in the Nation's Housing Stock
Tables 3-4 through 3-10 in Section 3.3.1.1 contained summaries of estimated
environmental-lead levels in the nation's occupied housing stock prior to implementing
interventions that would occur under the proposed §403 rule. These summaries were based on
data from the HUD National Survey with sampling weights revised to represent the 1997
national occupied housing stock. To characterize the extent to which these environmental-lead
levels provide exposures to children and to characterize the benefits associated with §403, it was
Draft - Do Not Cite or Quote
88
September 27, 1996
-------
necessary to estimate numbers of children of specific age groups who reside within the housing
units represented in Table 3-3 of Section 3.3.1.1.
Methods used to obtain numbers of children in the national housing stock are presented
in Section C.I .2. These methods used estimates of the 1997 birth rate, average number of
children per 1,000 people, and average number of residents per housing unit. While this risk
assessment focused on characterizing the blood-lead concentrations and associated health effects
for children aged 12-35 months (i.e., 1-2 years), the sensitivity analysis (Section 5.4) also
considered children aged 12 to 71 months (1 to 5 years). Therefore, the methods in Appendix C
were applied to both age groups.
Table 3-24 provides the number of children residing in the 1997 housing stock according
to age of unit and age of child. Numbers of children associated with the 1997 sampling weights
for each HUD National Survey unit are displayed in Table C-7 of Appendix C.
Table 3-24. Estimated Number of Children in the 1997 National Housing Stock, by Age of
Child and Year-Built Category
Year-Built
Category
Prior to 1 940
1940-1959
1960-1979
After 1979
Entire Nation1
Age of Child
12-35 Months
1,578,000
1,581,000
2,805,000
1 ,996,000
7,961 ,000
12-71 Months
4,043,000
4,051,000
7,188,000
5,115,000
20,397,000
1 Value may differ from sum of previous rows due to rounding.
3.4 DISTRIBUTION OF CHILDHOOD BLOOD-LEAD
The §403 risk assessment, as described in Section 2.4, characterizes health effects and
blood-lead concentrations for children aged 1 to 2 years (i.e., 12 to 35 months). The national
distribution of blood-lead concentration for children aged 1-2 years is based on NHANES ffl.
This information is summarized in Section 3.4.1. Supporting information on children's blood-
lead concentrations is provided through summary statistics from the Baltimore R&M Study and
Draft - Do Not Cite or Quote
89
September 27, 1996
-------
the Rochester Study. These summaries are presented in Section 3.4.2. While blood-lead
concentrations in these two epidemiological studies are not representative of lead exposure on a
national scale, they provide additional evidence on the prevalence of elevated childhood blood-
lead concentrations.
3.4.1 NHANES III
The National Health and Nutrition Examination Surveys, conducted by the CDC's
National Center for Health Statistics (NCHS), trace the health and nutritional status of the
noninstitutionalized, civilian U. S. population. The surveys consist of adult, youth, and family
questionnaires, followed by standardized physical examinations.
The Third National Health and Nutrition Examination Survey (NHANES HI), conducted
from 1988 to 1994, was the seventh in a series of national examination studies conducted by
NCHS since 1960. The target population for NHANES HI included the civilian
noninstitutionalized population 2 months of age and older. The primary objectives of NHANES
III were the following (CDC, 1992):
"To produce national population health parameters; to estimate the national prevalence
of selected diseases and disease risk factors; to investigate secular trends in selected
diseases and risk factors; to contribute to the understanding of disease etiology; and to
investigate the natural history of selected diseases."
Approximately 40,000 persons were sampled in NHANES HI, including approximately 3,000
children aged 1 to 2 years. Phase 1 of NHANES ffl, conducted from 1988-1991, provided the
most recently-collected data on blood-lead concentrations that were available for analysis in the
§403 risk assessment. These data included 1341 children aged 1 to 2 years, of which 924 had
blood-lead concentrations reported.
Study participants in NHANES ffl were subjected to a physical examination conducted by
a physician, a dentist, and health technicians. For participants aged 12 months and older, these
examinations included taking a blood sample via venipuncture. This sample was analyzed for
f
lead content by graphite furnace atomic absorption spectrophotometry.
Draft - Do Not Cite or Quote 90 September 27. 1996
-------
To provide for a nationally representative sample, a complex survey design was employed
in NHANES III (CDC, 1992; CDC, 1994). Although estimates of national population health and
nutrition parameters were the primary objectives of the survey, suitably precise estimates for
certain age and race groups were obtained through oversampling. As a result, the NHANES IE
provides national and subpopulation estimates of the distribution of childhood blood-lead
concentrations.
In NHANES III, each subject was assigned a series of sampling weights. Each weight,
determined from the 1990 Current Population Survey (CPS), indicates the total number in the
U.S. population represented by a subject at a given stage of the survey. Because the risk
assessment is using the blood-lead concentrations, the sampling weight assigned to subjects at
the time of the physical examination is utilized in the risk assessment.
Tables 3-25 and 3-26 summarize the blood-lead concentrations from NHANES in for
children aged 1-2 years (12-35 months) and aged 1-5 years (12-59 months). According to Table
3-25, the estimated geometric mean blood-lead concentration for children aged 1-2 years is 4.046
Hg/dL, with a geometric standard deviation of 2.057. The geometric mean declines to 3.571
jig/dL for children aged 1 to 5 years. Table 3-26 contains probabilities of observing elevated
blood-lead concentrations within these two age groups. Slightly over 10% of U.S. children aged
1-2 years are estimated to have blood-lead concentrations greater than 10 |ig/dL (the current
action level established by the CDC). Blood-lead concentrations greater than 25 ug/dL are
relatively rare. The probability of elevated blood-lead concentration and the geometric mean
blood-lead concentration declines as children aged 3-5 years are also considered, supporting the
hypothesis that blood-lead concentration tends to peak at some age between 1 and 2 years.
For children aged 1 to 2 years, Table 3-27 presents geometric mean blood-lead
concentration and percentage of children with a blood-lead concentration above 10 ng/dL, for
selected subgroups of the U.S. population (family income, race, and urban status). These results
illustrate that socioeconomic status is an important factor in the incidence rate of elevated lead
exposure. Low-income families, especially those containing non-Hispanic African-Americans,
have the highest percentage of children with blood-lead concentration exceeding 10 ug/dL.
Urbanicity is also an important factor, with children residing in urban centers having the highest
probability of exhibiting elevated blood-lead concentrations. Urban centers are usually
Draft - Do Not Cite or Quote 91 September 27, 1996
-------
associated with high environmental-lead exposure, due to the high density of older buildings
containing lead-based paint and remaining fallout of leaded gasoline emissions from urban
traffic. These results confirm that while childhood blood-lead concentrations may have declined
over the years, childhood lead exposure remains a real public health threat for certain subgroups
of the U.S. population.
Table 3-25. Summary of Blood-Lead Concentration Data for Children Aged 1-2 Years and
Aged 1-5 Years, Based on NHANES III (Phase 1)
Age
. Range
(years)
1-2
1-5
# Children
with Blood-
Lead Cone.
Reported
924
2232
Sum of
NHANES
Sample
Weights1
7,812,631
19,206,122
Blood-Lead Concentration (pg/dL)
Minimum
0.70
0.70
Maximum
56.6
72.0
Geometric
Mean
4.046
3.571
Geometric
Standard
Deviation
2.057
2.084
1 Weights assigned at the time of the physical examination. Included in this sum are the weights for all children who were
examined, including those who did not have a blood-lead concentration reported.
Table 3-26. Estimated Probabilities of Elevated Blood-Lead Concentrations in Children
Aged 1-2 Years and Aged 1-5 Years, Based on NHANES III (Phase 1)
Age
Range
(years)
1-2
1-5
» Children
with Blood-
Lead Cone.
Reported
924
2232
Sum of
NHANES
Sample
Weights1
7,812,631
19,206,122
Percentages with Elevated Blood-Lead Concentration
;>10//g/dL
10.5%
8.0%
2l5pg/dL
3.5%
2.5%
*20//g/dL
1.3%
0.9%
225/ig/dL
0.6%
0.4%
1 Weights assigned at the time of the physical examination. Included in this sum are the weights for all children who were
examined, including those who did not have a blood-lead concentration reported.
Draft - Do Not Cite or Quote
92
September 27. 1996
-------
Table 3-27. Estimated Percentage of Children With Blood-Lead Concentrations Exceeding
10 /vg/dL, and the Geometric Mean and Geometric Standard Deviation of
Blood-Lead Concentration, for Children Aged 1 to 2 Years Within Selected
Subgroups
Percentage
Geometric Mean (GSD)
Family
Income Level2
Urban Status3
Low
Mid
High
Central City,
i 1.000.000
Persons
Central City,
< 1.000,000
Persons
Non-Central City
All Children1
17.4%
5.00 (2.04)
8.0%
3.73 (2.01)
7.1%
3.31 (2.02)
10.7%
4.04 (2.08)
20.4%
4.92 (2.05)
3.7%
3.46(1.93)
Selected Race Groups
Non-Hispanic
White
12.2%
4.17(2.02)
7.7%
3.58 (2.00)
7.9%
3.33 (2.06)
6.5%
3.60(1.96)
18.6%
4.21 (2.12)
3.4%
3.25(1.97)
Non-Hispanic
African-American
30.0%
6.86 (2.05)
7.5%
4.27 (2.07)
2.6%
3.54(1.69)
26.0%
5.53 (2.44)
16.9%
5.76(1.78)
5.2%
5.23(1.66)
Mexican-
American
7.4%
4.39(1.81)
6.9%
3.94(1.87)
< 1%
2.69(1.63)
9.5%
4.00(1.95)
9.5%
4.70(1.88)
6.6%
4.35(1.78)
1 Includes race/ethnicity groups not shown separately.
2 Income level was defined by poverty income ratio (PIR) categorized as low (0< PIR< 1.30), mid (1.30sPIR< 3.0), and high
(PIRzS.O). Persons with missing information on income level are not included in summaries by income level.
3 Persons with missing information on urban status are not included in summaries by urban status.
Source: National Center for Health Statistics. NHANES III, 1988-91.
The blood-lead concentration summaries in Tables 3-25 through 3-27 may differ from
what is reported in published literature on NHANES HI results (Pirkle et al., 1994; Brody et al.,
1994). The summaries in these three tables were calculated using the public-use dataset for
Phase 1 of NHANES III. Reasons why these summaries would differ from published results
include:
When subjects received two physical examinations in Phase I of the NHANES m, the
public-use dataset reported results only for the first examination, regardless of whether
a blood-lead concentration was available for the subject from this examination. In
contrast, published results included blood-lead information from the second
examination when information from the first examination was not reported.
Draft - Do Not Cite or Quote
93
September 27, 1996
-------
• If a subject did not have a detectable blood-lead concentration, the NHANES HI set the
blood-lead concentration equal to a small, positive value. This value differed between
the dataset used to produce the published results and the public-use dataset.
• The sampling weights were refined after the published results were produced. The
refined weights are included in the public-use dataset.
Data from Phase 1 of NHANES in are used to characterize the distribution of blood-lead
concentration for children aged 1 to 2 years in 1997, prior to any intervention activities that occur
as a result of implementing the proposed §403 rules. More details on this distribution are
presented in Section S.I.
3.4.2 Baltimore Repair and Maintenance (R&M) Study
Childhood blood-lead concentrations collected prior to any interventions performed in the
Baltimore R&M Study (Section 3.2.1) provide evidence of elevated blood-lead concentrations
for children in high-exposure environments. Units slated for R&M interventions were
documented to contain lead-based paint and elevated lead levels in household dust. Modern
urban units acted as negative controls, being free of lead-based paint. Previously-abated units
were abated for lead-based paint previous to this study, and therefore reflect a post-abatement
environment. The urban setting of this study indicates an increased potential for lead exposures.
Table 3-28 summarizes the pre-intervention blood-lead concentrations. The overall
geometric mean blood-lead concentration for children aged 1-2 years is 9.94 ug/dL, which is
over twice the value of 4.05 ug/dL for NHANES IK In particular, geometric mean blood-lead
concentrations for children aged 1-2 years in previously-abated units and units slated for R&M
intervention (11.86 ug/dL and 10.56 ug/dL, respectively) are high, while for 1-2 year old
children residing among the modem urban units, where potential for lead exposure was reduced,
the geometric mean was 2.82 ug/dL.
The probabilities that children's blood-lead concentrations exceed 10,15,20, or 25 |ig/dL
are high for all but the modem urban units (Table 3-28). Approximately 58% of 1-2 year olds in
the R&M Study had blood-lead concentrations greater than 10 ug/dL, compared to the NHANES
III estimate of 10.5% (Table 3-26). Again, this is likely due to the increased lead exposure
associated with these children compared to the national population. Of the different unit groups
Draft - Do Not Cite or Quote 94 September 27, 1996
-------
Table 3-28. Summary Statistics on Blood-Lead Concentration Measured Prior to
Intervention in the Baltimore Repair and Maintenance Study
Age
Range
(years)
Sample
Size
Minimum
Blood-Lead concentration (i/yaLt
Maximum
Geometric
Mean
Qsonwtnc
Standard
Deviation
Percentages with Elevated Blood-Lead
Concentration
*10
/rg/dl
zlS
ligltU.
220
j/g/dL
a 25
jig/dL
All Children with Measured Pro-Intervention Blood-Lead Concentration1
All2
1-2
163
93
0.9
0.9
65.5
65.5
10.42
9.94
2.12
2.29
58.3
53.8
35.0
33.3
14.7
16.1
10.4
12.9
Children in Previously-Abated Units
All3
1-2
23
12
3.65
3.65
28.8
24.15
12.70
11.86
1.60
1.71
73.9
66.7
43.5
50.0
13.0
8.3
4.3
0.0
Children in Units Slated for Repair and Maintenance
All2
1-2
69
41
1.75
1.75
65.5
65.5
10.20
10.56
1.87
1.99
49.3
51.2
27.5
26.8
10.1
14.6
5.8
9.8
Children in Modem Urban (control) Units
All4
1-2
19
14
0.9
0.9
10.15
5.8
3.04
2.82
1.74
1.67
5.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1 Includes children who did not complete the study and children in non-study units targeted to move into vacant study units
after intervention. As a result, numbers of children within the four specified housing groups at pro-intervention do not total
the numbers of children with pre-intervention blood-lead concentrations.
2 Children aged 6-57 months.
1 Children aged 10-57 months.
4 Children aged 16-43 months.
in the study, the highest percentage of children with blood-lead concentrations above 10 ug/dL
occurred for previously-abated units (67%). Recall from Section 3.1.3.2 that environmental-lead
levels were high in these units as well, suggesting that lead exposures were not removed as a
result of the abatements performed previous to this study. No children aged 1-2 years who
resided in modem urban units had elevated blood-lead concentrations.
3.4.3 Rochester Study
Blood-lead concentration data were collected in the Rochester Study (Section 3.2.2) for
205 children aged 12-30 months of age. While units having recent major renovations or the
potential for lead contamination from exterior sources were not considered in this study, no
Draft - Do Not Cite or Quote
95
September 27, 1996
-------
attempt was made to include units based on environmental-lead levels or the presence of lead-
based paint. Therefore, blood-lead concentrations in this study likely reflect typical lead
exposure conditions in an urban setting.
Table 3-29 summarizes the blood-lead concentrations from the Rochester study. The
geometric mean blood-lead concentration was 6.38 ug/dL (geometric standard deviation, 1.85).
Twenty-three percent of the children had blood-lead concentrations above 10 ug/dL, 8% above
IS |ig/dL, and 3% above 20 ug/dL. These numbers are higher than those for NHANES ffl, but
lower than those for the Baltimore R&M study.
Table 3-29. Summary Statistics on Blood-Lead Concentration Measured in the Rochester
Lead-in-Dust Study
Number of
Children
205
Minimum
1.4
Maximum
31.7
Geometric
Mean
6.38
Geometric ~
Standard
Dovifttlon
1.85
PttTTAntP'»*t «uM« Ctauntaul DlMMt-l «ta*4
210
WJ/dL
23.4
Concentration
^S
W/dL
7.8
*20
ttgldL
2.9
225
//g/dL
1.5
Draft - Do Not Cite or Quote
96
September 27, 1996
-------
4.0 DOSE-RESPONSE ASSESSMENT
CHAPTER 4 SUMMARY
Chapter 4 presents the approach for characterizing the relationship between
environmental lead exposure and the resulting adverse health effects. The
relationship is established in two stages. First, the relationship between
environmental lead levels and blood-lead concentration is characterized by two
models, the IEUBK and EPI models. Then the relationship between blood-lead
concentration and representative health effects is established. This relationship is
applied in the integrated risk assessment, using environmental data from the HUD
National Survey to estimate the number of children who will benefit from the S403
rule. Additional topics covered in this chapter describe assumptions required for
this application, including 1) methods for converting environmental lead levels
measured by different sampling methods; 2) methods for incorporating the effect
of paint pica, and 3) development of a range of options for the standards.
This chapter presents the approach for characterizing the relationship between children's
exposure to lead in dust, soil, and paint and the resulting adverse health effects. It would be ideal
to relate particular health outcomes directly to environmental lead levels. However, most studies
of lead in the environment relate residential measures of lead exposure (e.g., floor or window sill
dust-lead loadings) to measures of body lead burden (e.g. blood-lead concentration), rather than
directly to health effects. In addition, most studies of health effects of lead exposure relate
specific health outcomes (e.g. IQ point loss) to measures of body lead burden, rather than directly
to environmental lead levels.
Thus, the dose-response relationship between environmental lead levels and health
outcomes for this risk assessment will be established in two stages. First, the relation between
environmental lead levels and blood-lead concentration will be estimated via quantitative
models. Then, the relationships between blood-lead concentration and health effects documented
in the scientific literature will be used to establish the link between environmental lead exposure
and health outcomes.
Sections 4.1 through 4.3 discuss the methodology associated with characterizing the dose-
response relationship between environmental lead levels and blood-lead concentration.
Section 4.1 describes two models used to relate environmental lead levels to blood-lead
concentrations. While the goal of each model is similar, the modeling approaches are markedly
Draft - Do Not Cite or Quote 97 September 27, 1996
-------
distinct. The first, described in Section 4.1.1, is a biokinetics-inspired model that attempts to
simulate how lead is absorbed, processed, and eliminated by the human body. The second
model, described in Section 4.1.2, is a regression model that seeks to relate observed
environmental lead levels to observed blood-lead concentrations.
Each of these models is used to predict a national distribution of blood-lead
concentrations in both a pre- and post-§403 environment. Environmental lead levels in the HUD
National Survey homes are used as input to the models to predict the geometric mean blood-lead
concentration of children exposed to environmental lead conditions evidenced in HUD National
Survey homes. The national distribution of blood-lead concentrations is then developed by
allowing each home to represent a proportion of the total number of 1-2 year old children
(Section 3.3).
Environmental dust-lead levels hi the HUD National Survey were measured by the blue
nozzle vacuum sampling method. However, §403 dust-lead standards will be specified in terms
of wipe sampling, and the EPI model dust-lead parameters are based on wipe sampling. Section
4.2 discusses conversion equations developed to relate different dust sampling methods that will
be used, for example, to convert blue nozzle vacuum dust loadings to wipe dust loadings.
Section 4.3 describes the methodology used to an account for the effect of a child's pica
for paint on the distribution of blood-lead concentrations developed using each model.
Methodology discussed in this chapter related to characterizing the relationship between a
national distribution of environmental lead levels and a national distribution of blood-lead levels
is illustrated in Figure 4-1 Additional details on methodology related to predicting the national
distribution of blood-lead concentrations is provided in Chapter 5.
Section 4.4 presents the approach for relating blood-lead concentration to the
representative health effects identified in Chapter 2: IQ deficits and elevated blood-lead
concentrations. Lead-related IQ deficits are measured directly as decrements hi IQ scores and
through the increased incidence of IQ scores less than 70. Incidence of elevated blood-lead
concentrations is estimated at 10 ug/dL and 25 ug/dL, based on CDC guidelines (CDC, 1991), as
described in Section 2.4.1.
Draft - Do Not Cite or Quote 98 September 27, 1996
-------
HUD National Survey of
Environmental Levels
Dust, Soil, Paint
Determine Homes for Intervention
Conversion of BN Loading to Wipe Loading
Input to EPI Model
Conversion of BN Loading
to Wipe Loading
(pre- and post-intervention)
EPI Model Applied
Pica Effect Incorporated
into Model Prediction
HUD/EPI Prediction of
National Distribution of
Blood-Lead Concentrations
Input to IEUBK Model
Conversion of Wipe Loading
to BN Concentration
(post-intervention)
IEUBK Model Applied
Pica Effect Applied
Subsequent to Model
Prediction
HUD/IEUBK Prediction of
National Distribution of
Blood-Lead Concentrations
Figure 4-1. Methodology Associated With Characterizing the Relationship Between a
National Distribution of Environmental Lead Levels and a National Distribution
of Blood-Lead Concentrations
Draft - Do Not Cite or Quote
99
September 27, 1996
-------
The tools and methodology presented in Sections 4.1 through 4.4 are used in Chapter 5 to
help characterize the reduction in childhood health risks and blood-lead concentrations for
various options for the §403 standards. Because standards are set for multiple media it was not
possible to look at all sets of standards. Section 4.5 presents a range of options for standards that
will be evaluated in the integrated risk analysis in Chapter 5 and discusses the methodology used
to specify the ranges.
4.1 ESTIMATION OF MEAN BLOOD LEAD CONCENTRATION
This section describes the first stage of the dose-response relationship, estimation of the
relationship between environmental lead levels and blood-lead concentration. Two modeling
approaches will be used to estimate blood-lead concentrations from environmental lead exposure.
In the first approach, blood-lead concentrations are estimated using EPA's Integrated Exposure,
Uptake, and Biokinetic Model for Lead (version 0.99D), hereafter referred to as the IEUBK
model (EPA, 1994a; 1995e). The second approach estimates blood-lead concentrations using a
multiple regression model fitted to epidemiological data.
4.1.1 IEUBK Model
The IEUBK model is a biokinetics-inspired simulation model designed to predict the
probability of elevated blood-lead levels in children given a set of environmental lead levels.
The model addresses three components of environmental risk assessment: 1) multi-media nature
of exposures to lead, 2) lead pharmacokinetics, and 3) inter-individual variability in blood-lead
levels, through the estimation of the probability distribution of blood-lead levels for children
exposed to similar environmental lead concentrations.
Specifically, the model uses lead concentrations measured in dust, soil, air, water, diet,
and other ingested media as inputs to estimate a longitudinal exposure pattern from birth to seven
years of age (EPA, 1995e). The model then estimates a distribution of blood lead levels for a
population of children exposed to that exposure pattern. The center of this distribution, the
geometric mean, is predicted by the model. A constant empirical estimate is used by the model
to represent the variability about the geometric mean. In statistical terminology, this variation is
referred to as the geometric standard deviation (GSD). A value of 1.6 for the GSD was estimated
Draft - Do Not Cite or Quote 100 September 27, 1996
-------
from residential community blood-lead studies and is assumed by the model. The GSD pertains
to the inter-individual and biological variability in blood-lead levels of individual children
exposed to similar environmental lead levels. It must be recognized here that the IEUBK model
is not intended to predict the blood-lead level of an individual child, and therefore cannot
substitute for a medical evaluation of an individual child.
It is beyond the scope of this document to describe the IEUBK model in detail. Very
briefly, the model has three distinct functional components that work together in series:
exposure, uptake, and biokinetic components. Each model component is a set of complex
equations and parameters. The Technical Support Document (EPA, 1995e) provides the
scientific basis of the parameters and equations used in the model and the Guidance Manual
(EPA, 1994a) includes a detailed description of the exposure pathways, absorption mechanism,
and biokinetic compartments and associated compartmented transfers of lead.
The IEUBK model can be implemented by running a PC-compatible software program.
The program provides the user access to input values for the exposure and uptake parameters.
The biokinetic parameter values, however, are not accessible. The Guidance Manual (EPA,
1994a) recommends using the input values specified by the software for the exposure and uptake
parameters whenever more representative data are not available. In the absence of better
information, these default input values are used in the risk characterization to predict blood-lead
levels via the IEUBK model. In addition, blood-lead concentrations are predicted at age 24
months. Table 4-1 presents the default input values for the exposure and uptake parameters.
For purposes of illustration, Figure 4-2 displays the relationship between blood-lead
concentration predicted by the IEUBK model and soil- or dust-lead concentration for children
aged 2 years.
The dust-lead concentration was set to 200 ppm for the soil curve as the HUD National
Survey data report a geometric mean dust-lead concentration of 192 ppm. Similarly, the soil-
lead concentration was set to 100 ppm for the dust curve because the HUD National Survey data
estimated a geometric mean soil-lead concentration of 78 ppm. Air, water, diet, and other
ingested media were set to model default values. From the dust curve in Figure 4-2, the
predicted blood-lead concentration is 3 ug/dL for a dust-lead concentration of 100 ppm and a
Draft - Do Not Cite or Quote 101 September 27. 1996
-------
Table 4-1. Summary of Default Parameter Values Used in the IEUBK Model (Version
0.99D).
Parameter
Setting*
Parameter
Setting*
Parameter
Setting*
Parameter
Setting*
Air Parameters
Vary air concentration by
year?
No
Outdoor air lead concentration
0.10//g/m3
Indoor air lead concentration
30% of outdoor value
* All air parameters use default values
Diet Intake Parameters
Lead intake in diet, by age of child
0-1 yrs
5.53
//g/day
1-2 yrs
5.78
//g/day
2-3 yrs
6.49
//g/day
3-4 yrs
6.24
//g/day
4-5 yrs
6.01
//g/day
5-6 yrs
6.34
//g/day
6-7 yrs
7.00
//g/day
* All diet intake parameters use default values
Water Intake Parameters
Lead
Cone, in
Water
4//g/L
Drinking water consumption, by age of child
0-1 yrs
0.20
L/day
1-2 yrs
0.50
L/day
2-3 yrs
0.52
L/day
3-4 yrs
0.53
L/day
4-5 yrs
0.55
L/day
5-6 yrs
0.58
L/day
6-7 yrs
0.59 L/day
* All water intake parameters use default values
Soil and Dust Intake Parameters
Soil/Dust
Ingestion
Weighting
Factor
45% soil;
55% dust
Total soil + dust intake, by age of child
0-1 yrs
0.085
g/day
1-2 yrs
0.135
g/day
2-3 yrs
0.135
g/day
3-4 yrs
0.135
g/day
4-5 yrs
0.1
g/day
5-6 yrs
0.09
g/day
6-7 yrs
0.085
g/day
* Soil and dust lead concentrations are input. All other parameters use default values.
Absorption Method Parameters
Parameter
Setting
Half
Saturation
Level
100 //g/day
Total Absorption
Soil
30%
Dust
30%
Water
50%
Diet
50%
Alt.
0%
Fraction of Total Assumed Passive
Absorption
Soil
0.20
Dust
0.20
Water
0.20
Diet
0.20
Alt.
0.20
Blood Lead Parameter
Parameter
Setting
Geometric Standard Deviation (GSD)
1.6
Draft - Do Not Cite or Quote
102
September 27, 1996
-------
1 10 100 100O 10000
Lead Concentration In Soil or Dust (ppm)
Duct Soil
Dust-lead concentration was set to 200 ppm for the soil curve and soil-lead concentration was
set to 100 ppm for the dust curve.
Figure 4-2. IEUBK Model Predicted Blood-Lead Concentration for Children Two Years Old
Plotted Separately Versus Soil-Lead Concentration and Dust-Lead
Concentration for Fixed Default Values of the Remaining Model Parameters
soil-lead concentration of 100 ppm. Similarly, from the soil curve, the predicted mean blood-
lead concentration for 200 ppm for lead concentrations of both soil and dust is 4.5 ng/dL. It is
important to recognize that each point on the predicted curve represents a geometric mean blood-
lead level for children exposed to similar environmental lead levels. The blood-lead levels for
individual children will vary about the predicted geometric mean. The variation about the
geometric mean is captured by the GSD.
4.1.2 Epidemiological Model
An epidemiological (EPI) model is another model to estimate blood-lead levels hi young
children stemming from exposure to environmental lead. The EPI model was developed from
the Rochester Study (Lanphear et al., 1995; HUD, 1995a) data. Measures that were in both the
Draft - Do Not Cite or Quote 103 September 27, 1996
-------
Rochester data and the HUD National Survey data or measures in the Rochester data that could
be closely approximated by measures in the HUD National Survey data were considered for
inclusion in the model. Variables whose definition provided a convenient translation when
applied to the National Survey, whose predictive power in Rochester was high, and whose spread
in the National Survey population covered a wide range of values, were used in the EPI model.
Appendix W provides further details on development of the EPI model.
The EPI model addresses three components of environmental risk assessment: 1) multi-
media nature of exposures to lead, 2) intercorrelations among the environmental lead exposure
variables as they relate to blood-lead levels, and 3) inter-individual variability in blood-lead
levels. As with the IEUBK model, the EPI model is used to estimate a probability distribution of
blood-lead levels for children exposed to similar environmental lead levels. The center of the
distribution, the geometric mean, is predicted by the model. As with the IEUBK, a GSD of 1 .6 is
used to represent the variability about the geometric mean.
The main difference between the EPI model and the IEUBK model is that the EPI model
relates blood-lead concentration to dust-lead loading (rather than dust-lead concentration) and
two other environmental variables based on a multiple regression equation fitted to data from a
single locality. In contrast, the IEUBK model relates blood-lead concentration to dust-lead
concentration, was not developed for data form a single locality, and is based on a simulation of
the lead biokinetics within a child's body.
The EPI model applies to children aged 12-30 months. The model may be written as:
log(PbB) =po+p,-log(DnipS'o//) +^PaintPi
where DripSoil is the dripline soil-lead concentration at the house; PaintPica is an indicator
variable equal to 1 .5 if there is a child with paint pica in the house and the house has deteriorated
lead based paint, and zero otherwise; FloorLoad is the area-weighted floor dust-lead loading for
all surfaces in the house; and SillLoad is the area-weighted average of dust-lead loading from
window sills in the house. In the equation PO, Pl9 P2, P3, and P4 are model parameters, € is a
random error term that represents the residual error left unexplained by the model and log is the
natural logarithm function. In the Rochester data, the extent of paint pica was coded as a 0, 1
Draft - Do Not Cite or Quote 104 September 27, 1996
-------
or 2, where 0 represented no paint pica, 1 was for children who exhibited paint pica only rarely,
and 2 was for children who exhibited paint pica at least sometimes. A value of 1.5 was used for
the EPI model when applied to the HUD National Survey data, since it was the average of the
two values used for children who exhibited any amount of paint pica in the Rochester data.
The model was fit to the Rochester data using the S AS* System. Table 4-2 presents the
resulting parameter estimates and associated standard errors. For example, the EPI model
predicts a geometric mean blood-lead concentration of 5.3 ng/dL for children with no paint pica
or damaged lead-based paint exposure but who are exposed to a floor dust-lead loading of 40
ug/ft2, a drip line soil-lead concentration of 200 ppm, and a window sill dust-lead loading of 100
ug/ft2. It is important to note that the EPI model predicts the geometric mean blood-lead level
for children in similar exposure environments; it should not be used to predict the blood-lead
level for an individual child. The variation in blood-lead concentration about the geometric
mean is captured by the GSD. A GSD of 1.6 is assumed by the EPI model, to be consistent with
the IEUBK model.
Table 4-2. Parameter Estimates and Associated Standard Errors for the Multimedia
Exposure Model Based on Data from the Rochester Lead-in Dust Study.
Parameter
Po
Pi
P2
P3
P.
Predictor Variable
Intercept
log(DripSoil)
PaintPica
log(DustLoad)
log(SillLoad)
Estimated Effect
0.418
0.114
0.248
0.066
0.087
Standard Error
0.240
0.035
0.100
0.040
0.036
For purposes of illustration, Figure 4-3 displays the blood-lead concentration, predicted
by the EPI model, as a function of drip line soil-lead concentration, floor dust-lead loading, and
window sill dust-lead loading. Each medium is represented by a different line on the plot. For
the prediction of blood-lead concentration as a function of soil-lead concentration, the floor dust-
lead loading is held constant at 25 ug/ft2, window sill dust-lead loading is held constant at 50
Draft - Do Not Cite or Quote
105
September 27, 1996
-------
ug/ft2, and paint pica is set equal to zero. For predicting blood-lead concentration as a function
of floor dust-lead loading, soil-lead concentration is held constant at 100 ppm, window sill dust-
lead loading is held constant at 50 u-g/ft2, and again paint pica is set equal to zero. Finally, for
the sill dust-lead loading curve, soil-lead concentration was held constant at 100 ppm, floor dust-
lead loading was held constant at 25 ug/ft2, and paint pica was set equal to zero. The values at
which variables were held constant in Figure 4-2 were chosen to be approximately equal to the
variable's weighted geometric mean estimated from the HUD National Survey data. The blood-
lead levels for individual children will vary about the predicted geometric mean. This variation
is captured by the assumed GSD of 1.6.
Floor (Mg/sq ft)
1OO
Dust or Soil Lead Level
SIM ((ug/sq ft) —
Soil (ppm)
Sill dust-lead loading was set to 50 /sg/ft2 and soil-lead concentration was set to 100 ppm for the floor dust curve.
Floor dust-lead loading was set to 25 /yg/ft2 and soil-lead concentration was set to 100 ppm for the sill dust curve.
Floor dust-lead loading was set to 25 //g/ftj and sill dust-lead loading was set to 50 ^g/ft2 for the soil lead curve.
Figure 4-3. EPI Model Predicted Blood-Lead Concentration Plotted Separately Against Floor
Dust-Lead Loading, Sill Dust-Lead Loading and Soil Lead Concentration for
Fixed Values of the Remaining Model Inputs
As any of the environmental lead levels increase, the geometric mean blood-lead
concentration increases gradually. This is similar to the same plot for the IEUBK model except
that there is no point at which blood-lead concentration begins to rise rapidly as environmental
lead increases.
Draft - Do Not Cite or Quote
106
September 27, 1996
-------
4.2 UTILIZING DUST LEAD LOADINGS
The HUD National Survey is the only national survey of environmental lead levels and
therefore.will be used for prediction of a national distribution of blood-lead concentrations. Dust
lead measurements in the HUD National Survey were collected by the blue nozzle vacuum
method. However, §403 standards for dust will be expressed as a measured lead loading
collected by a dust wipe sample. This discrepancy leads to several instances where conversion
factors will be used in the risk assessment methodology. These include:
Convert blue nozzle loading to wipe loading hi the HUD National survey to
determine the extent to which homes in the United States are impacted by various
options for the §403 dust-lead standard.
Convert blue nozzle loadings to wipe loadings in the HUD National Survey data
before input into the EPI model.
Convert post-intervention wipe loadings to blue nozzle concentrations for input to the
IEUBK model.
In addition, conversion factors were used to convert different sampling methods to wipe
sampling for production of prevalence tables in Chapter 3 and for sensitivity/specificity analyses
in Section 4.5 of this chapter.
This section presents the equations developed to perform the conversions described
above. The conversion equations are presented for samples collected from floors and window
sills, since §403 will set standards for those housing components. Appendix X provides detailed
information concerning the development of all the conversion equations discussed hi this section.
4.2.1 Wipe versus Blue Nozzle (BN) Vacuum Conversions
Three studies reported side-by-side wipe and BN vacuum dust-lead measurements:
1. CAPS Pilot Study [EPA, 1995]
2. National Center for Lead-Safe Housing (NCLSH)/Westat Study [Westat, 1995]
3. Baltimore Repair and Maintenance (R&M) Pilot Study [Battelle, 1992]
Draft - Do Not Cite or Quote 107 September 27, 1996
-------
Log-linear regression models were fitted to each data set separately. Weighted averages
of the parameter estimates from each model were used to obtain the following conversion
equations (written in the scale of the original data) for predicting a wipe dust-lead loading from a
BN vacuum dust-lead loading:
Uncarpeted Floors: Wipeload = 11.4 (BNload)°'69°
Window Sills: Wipeload = 5.79 (BNlMd)' °8
For example, a BN dust-lead loading of 100 ug/ft2 on an uncarpeted floor would be converted to
a wipe dust-lead loading of 273 ug/ft2 (confidence intervals and prediction intervals are provided
in Appendix X to describe the uncertainty associated with these conversions).
The same statistical procedure produced the following equations for predicting a BN
vacuum dust-lead concentration from a wipe dust-lead loading:
Uncarpeted Floors: BN^ - 34.2 (WipeloJ°6'3
Window Sills: BN^ =115 (Wipeload)°4S1
Thus, a wipe dust-lead loading of 100 ug/ft2 on an uncarpeted floor would be converted to a BN
dust-lead concentration of 576 ug/g.
4.2.2 Wipe versus Baltimore Repair and Maintenance (BRM) Vacuum Conversions
Four studies reported side-by-side wipe and BRM vacuum dust-lead measurements:
1. R&M Mini Study
2. Rochester Lead-in-Dust Study
3. NCLSH 5-Method Comparison Study
4. Milwaukee Low-Cost Intervention Study
These studies are described in Appendix X.
Draft - Do Not Cite or Quote 108 September 27, 1996
-------
An analogous approach to that presented in Section 4.2.1 for developing the conversions
resulted in the following equations for predicting a wipe dust-lead loading from a BRM vacuum
dust-lead loading:
Uncarpeted Floors: Wipe = 8.79 BRM °313
Carpeted Floors: Wipe = 2.21 BRM °271
Window Sills: Wipe = 17.0 BRM °421
For instance, a BRM dust-lead loading of 100 ng/ft2 on an uncarpeted floor would be converted
to a wipe dust-lead loading of 37.1 fig/ft2.
Note that the floor dust-lead samples in the Baltimore R&M Study were collected as
composite samples, which eliminates the ability to distinguish uncarpeted floor samples from
carpeted floor samples. However, the number of uncarpeted and carpeted subsamples within
each composite sample can be determined. Therefore, floor samples from the Baltimore R&M
Study were converted using the following heuristic approach to obtain the summary statistics
provided in Chapter 3, and the sensitivity specificity analyses results in Section 4.5:
Wipe = p • 8.79 BRM0313 + (1-p) • 2.21 BRM0271,
where p represents the proportion of the composite sample obtained from uncarpeted floors, and
BRM represents the dust-lead loading measured with the BRM sampler. For example, a
composite floor BRM dust-lead loading of 100 ug/ft2 made up of 3 uncarpeted and 2 carpeted
subsamples would be converted to a floor wipe dust-lead loading of 25.4 fig/ft2 from (0.6 • 8.79 •
1000313) + (0.4 • 2.21 • 1000271). Confidence and prediction intervals were not derived for the
conversion of composite floor samples.
4.3 ESTIMATING THE EFFECT OF PICA FOR PAINT ON CHILDHOOD BLOOD-LEAD
LEVELS
The exposure pathway from lead-based paint to childhood blood-lead concentration can
be both direct and indirect. Indirect exposure takes place when deteriorated lead-based paint
Draft - Do Not Cite or Quote 109 September 27, 1996
-------
contaminates residential dust or soil, which is then ingested by the child. Direct exposure takes
place through the ingestion of paint chips. The two models described in Section 4.1 differ in
their handling of direct and indirect exposure to lead-based paint. The IEUBK model estimates
the geometric mean blood-lead concentration for children receiving indirect exposure to lead-
based paint, through the soil- and dust-lead concentrations used as model inputs. The IEUBK
model does not include a direct mechanism for estimating the contribution of paint chip ingestion
to childhood blood lead. The EPI model does include a mechanism for estimating the effect of
lead-based paint ingestion, as well as the effects of indirect exposure. This section describes how
the effect of pica for paint is applied to the HUD National Survey homes for each model.
4.3.1 IEUBK Model
As described in Section 4.1, environmental conditions observed in the HUD National
Survey are used as input to the IEUBK model. For each home in the HUD National Survey, the
IEUBK model is used to predict the geometric mean blood-lead concentration of children
exposed to those environmental conditions at age 24 months. The distribution of blood-lead
levels in the population of children is then developed by allowing each home in the HUD
National Survey to represent a proportion of the total number of children in the country. For
homes without damaged lead-based paint, the IEUBK model predicted geometric mean blood-
lead concentration and the assumed geometric standard deviation of 1.6 ug/dL are used to model
the distribution of blood-lead levels in children represented by each home.
For each home with damaged lead-based paint (defined as greater than 0 ft2 of interior or
exterior deteriorated lead-based paint), the children represented by that home are divided into
three groups: 1) children who have recently ingested paint chips (0.03%), 2) children who
ingested paint chips at some time (8.97%), and 3) children without pica for paint (91%). The
distribution of blood-lead levels for children in the three groups is estimated as follows:
1. Children who have recently ingested paint chips (0.03%) - Blood-lead concentration
is assigned the value 63 ug/dL with no variation.
2. Children who ingested paint chips at some time (8.97%) - Geometric mean blood-
lead concentration is 3.0 ug/dL greater than the geometric mean blood-lead
Draft - Do Not Cite or Quote 110 September 27, 1996
-------
concentration predicted by the IEUBK model. The adjusted geometric mean blood-
lead concentration and the assumed geometric standard deviation of 1.6 |ag/dL is used
to model the distribution of blood-lead levels for these children.
3. Children without pica for paint (91.0%) - The IEUBK model predicted geometric
mean blood-lead concentration and the assumed geometric standard deviation of 1.6
ug/dL is used to model the distribution of blood-lead levels for these children.
The scientific evidence and assumptions used to select percentages of children assigned
to each group and the adjustments to blood-lead concentrations for children who have ingested
paint chips are described in Appendix Dl.
4.3.2 EPI Model
Because the EPI model incorporates the effect of pica for paint, EPI model predicted
values are used to estimate the distribution of blood-lead concentrations both for children who do
and do not ingest paint chips, as described in Section 4.1.2. For HUD National Survey homes
with no damaged lead-based paint, the predicted geometric mean blood-lead concentration for
children who do not ingest paint chips and the assumed geometric standard deviation of 1.6
Hg/dL are used to model the distribution of blood-lead levels for all children represented by each
home. For homes with damaged lead-based paint, children represented by that home are divided
into two groups: 1) children who have ingested paint chips (9%), and 2) children who do not
ingest paint chips (91%). EPI model predicted geometric mean blood-lead concentrations and
the assumed geometric standard deviation of 1.6 ug/dL are used to estimate the distribution of
blood-lead concentrations for both groups.
4.4 HEALTH OUTCOMES
This section presents the approach for determining the incidence of adverse health
outcomes resulting from lead exposure in young children. Two representative health effects
were identified in Chapter 2: IQ deficits and elevated blood-lead concentrations. These effects
are measured through: 1) decrements in IQ scores, 2) increased incidence of IQ scores less than
70, and 3) incidence of blood-lead concentrations greater than 10 ug/dL, and 4) incidence of
Draft - Do Not Cite or Quote 111 September 27. 1996
-------
blood-lead concentrations greater than 25 jig/dL. These effects were chosen to represent the
spectrum of health effects of lead exposure. In this section, the relationship of each
representative health effect to blood-lead concentration is characterized. These relationships are
applied to the IEUBK and EPI model predicted blood-lead concentrations to relate environmental
lead exposure to health effects. The integrated risk analysis estimates specific health outcomes.
For example, the relationship between blood-lead concentration and IQ scores is used to estimate
the average IQ point loss due to lead exposure and the percentage of children with IQ point
decrements greater than one, two, or three IQ points. Total health risks were not obtained by
summing over the specific health outcomes. While the estimation of economic benefits was
considered when the primary health effects were selected, this report does not convert health
outcomes to monetary values. Economic benefits associated with health outcomes are estimated
in the §403 Regulatory Impacts Analysis.
4.4.1 Decrements in IQ Scores
The IQ point loss health effect represents the neurological loss due to low level lead
exposure. The association between blood-lead levels and IQ scores has been consistently
reported in the scientific literature, as described in Section 2.3.1. Lower IQ scores are associated
with a lower level of educational attainment and lower life-time earnings.
Schwartz (1994) conducted a meta-analysis to combine the findings of multiple studies in
determining the effect that blood-lead concentration has on full-scale IQ score in primary school
age children. The results from seven studies were employed to characterize the decrease in IQ
score associated with a 1 ^g/dL increase in blood-lead concentration. The three longitudinal and
four cross-sectional studies included in the meta-analysis are summarized in Table 4-3.
Additional details are provided in Appendix D2, Tables D2-1 and D2-2. A summary of the
Schwartz (1994) article and a comparison of the results to those reported in other, similar papers
(Schwartz, 1993; Pocock, et al, 1994) are also presented in Appendix D2.
Draft - Do Not Cite or Quote 112 September 27, 1996
-------
Table 4-3. Summary Information for Studies Included in the Schwartz (1994) Meta-
Analysis.
Study
Hawk, et al
(1986)
Hatzakis, et al
(1987)
Fulton, et al
(1987)
Yule, et al
(1981)
Bellinger, et al
(1992)
Dietrich, et al
(1993)
Baghurst, et al
(1992)
Number
of
Children
75
509
501
166
147
231
494
Blood-Lead Concentration
Uig/dL)
Range
6.2 - 47.4
7.4 - 63.9
3.3 - 34.0
7.0 - 33.0
na
na
<12.2-
>28.2
Mean (SO)
20.9 (9.7)
23.7 (9.2)
11.52
13.5(4.1)
6.5 (4.9)
15.2(11.3)
20 (na)
EstinfiBtBcl
Effect1
(SE>
2.55(1.5)
2.66 (0.7)
2.56 (0.9)
5.6 (3.2)
5.8(2.1)
1 .3 (0.9)
3.33(1.5)
Other Study Information
Cross-sectional study of children age 3-7
in Lenoir and New Hanover counties, NC
Cross-sectional study of primary school
age children in a lead smelter community
(Lavrion, Greece)
Cross-sectional study of primary school
age children in Edinburgh, Scotland
Cross-sectional study of primary school
age children in London, England
Longitudinal study in Boston. MA; Blood
lead at age 2; IQ measured at school age
Longitudinal study in Cincinnati, OH;
Integrated blood lead up to age 3; IQ
measured at school age
Longitudinal study in Port Pirie, Australia;
Integrated blood lead up to age 3; IQ
measured at school age
1 Effect was estimated for a doubling of blood-lead concentrations from 10 j/g/dL to 20 /ig/dL.
1 Geometric Mean was reported for this study.
The seven studies used linear, or log-linear, regression models to model the relationship
between IQ scores and childhood blood-lead levels, along with other potentially important
covariates. A log-linear regression model is a regression model fitted to the logarithm of the
independent variables; in this application the independent variable is blood-lead concentration.
The modeled relationships reported for each study were used to estimate that a 1 ug/dL increase
in blood-lead concentration results in a loss of 0.257 IQ points, on average. This relationship is
most applicable for blood-lead concentrations between 10 and 20 ug/dL, due to the modeling
assumptions for those studies that used log-linear models. However, the relationship is applied
over a much broader range of blood-lead concentrations in the risk assessment. This use is
justified because a similar effect was observed in the two studies (Hawk, et al., 1986; Hatzakis, et
al., 1987) that employed linear models, where the interpretation of model parameters is similar to
the manner in which the relationship is applied in the risk assessment. The blood-lead
concentrations in these two studies ranges from 6.2 ug/dL to 63.9 ug/dL, as shown in Table 4-3.
Draft - Do Not Cite or Quote
113
September 27, 1996
-------
The relationship between blood lead and IQ point loss is illustrated in Figure 4-4. For
example, children with blood-lead concentrations of 4 ug/dL are expected to have IQ scores
approximately one point lower, on average, compared to children who are not exposed to lead.
For an individual child, a greater or lesser IQ point loss may be observed.
9
a
7
6 -
5 -
4 -
3 -
2 -
10 20
Blood — Lead Concentration
30
40
Figure 4-4. Estimated IQ Point Loss Due to Lead Exposure Plotted Against Blood-Lead
Concentration
The relationship between IQ scores and blood lead was used in the risk assessment to
estimate the average IQ point decrement for children exposed to environmental lead and to
estimate the benefit, measured as IQ points not lost, following promulgation of the §403 rule. In
addition, the percentage of children with decrements of >1, >2, and >3 IQ points due to lead
exposure were calculated. Both the EPI and the IEUBK models were used to estimate the
distribution of blood-lead concentrations of children exposed to a given set of environmental
conditions, based on homes in the HUD National Survey. The blood-lead levels estimated across
all children were multiplied by 0.257 to estimate the IQ point loss due to lead exposure.
The relationship between environmental lead levels and IQ point loss are presented in
Figures 4-5 and 4-6 for the IEUBK and EPI models, respectively. For each curve, the soil- or
Draft - Do Not Cite or Quote
114
September 27, 1996
-------
dust-lead levels were varied over a range of values, while all other parameters were held fixed as
described in Section 4.1. The predicted blood-lead level was used to estimate the average IQ
point loss in the manner described above. For example, for a soil or floor dust-lead concentration
of 1000 ppm, the IEUBK model predicts that 2 IQ points will be lost. The EPI model predicts
that 1.5 IQ points will be lost for a dripline soil-lead concentration of 1100 ppm. Note that for
Figure 4-5 and subsequent figures in this section illustrating predictions from the EPI model only
the soil-lead concentration curve is illustrated.
6-
— 5
t/t
o
2 H
1O 1OO 10OO
Lead Concentration in Soil or Dust (ppm)
Dust Soil
The IEUBK model was used to relate environmental lead to blood lead.
Dust-lead concentration was set to 100 ppm for the soil curve and soil-lead concentration was
set to 200 ppm for the dust curve.
Figure 4-5. Estimated IQ Point Loss Due to Lead Exposure Plotted Against Concentration
of Lead in Soil and Dust, Utilizing IEUBK Model Predictions
Draft - Do Not Cite or Quote
115
September 27, 1996
-------
O
Q-
2 -
0 -
1O 100 10OO
Soil —Lead Concentration (ppm)
100OO
The EPI model was used to relate environmental lead to blood lead. Floor dust-lead loading was set to 25 //g/ft2,
window sill dust-lead loading was set to 50 //g/ft2 and it was assumed that there was no paint pica.
Figure 4-6. Estimated IQ Point Loss Due to Lead Exposure Plotted Against Concentration
of Lead in Soil, Utilizing EPI Model
4.4.2 Increased Incidence of IQ Scores Less Than 70
The increased incidence of IQ scores less than 70 represents the increased likelihood of
mental retardation resulting from lead exposure. An IQ of 70 is two standard deviations below
the population mean IQ of 100 and can be used as an indicator of mental retardation. Children
who are mildly mentally retarded require special education classes in school. Children who are
severely mentally retarded may require life-long institutional care. This health effect may be
used to estimate the number of children who will benefit under the proposed rule and the societal
benefit through reduced costs of caring for the mentally retarded.
There is limited data available to estimate the increased likelihood of mental retardation
resulting from lead exposure. Because of the lack of data, Wallsten and Whitfield (1986) used
judgmental probability encoding methods to assess health risks due to lead exposure, particularly
in the area of lower IQ scores. The results of their analyses are worth summarizing. As part of
Draft - Do Not Cite or Quote
116
September 27, 1996
-------
this assessment, the increased percentage of children having IQ scores less than 70 was estimated
for populations of children with elevated blood-lead levels.
In the Wallsten and Whitfield study, care was taken to select experts whose opinions
spanned the range of respected opinion. The six experts who participated in the assessment of
the relationship between IQ scores and blood-lead levels are listed in Table 4-4. These experts
were asked to consider a hypothetical experiment in which a large number of children were
randomly assigned at birth to either a control group, or one of six lead-exposure groups. Lead
exposure was to remain fixed until the children reached age seven, at which time the Wechsler
Intelligence Scale for Children - Revised (WISC-R) IQ test would be administered. Blood-lead
levels were to be measured at age three. The lead exposure levels were such that at age three,
members of each of the lead-exposure groups had blood-lead levels of 5, IS, 25,35,45, and 55
ug/dL. The experts were asked to estimate the mean and standard deviation of IQ scores in the
control group. The experts also estimated the expected mean IQ differences between the control
group and each exposure group. Each expert assumed that the IQ standard deviation in exposure
groups was the same as that of the control group. This information was used to estimate the
increased percentage, due to lead exposure, of children having IQ scores less than 70.
Table 4-4. Experts Who Participated in the Assessment of the Relationship Between
IQ Scores and Blood-lead Levels by Wallsten and Whitfield.
Expert
Kim Dietrich
Claire Ernhart
Herbert Needleman
Michael Putter
Gerhard Winneka
William Yule
Affiliation
University of Cincinnati
Cleveland Metropolitan General Hospital
University of Pittsburgh
Institute of Psychiatry, London, UK
University of Dusseldorf, Dusseldorf , West Germany
Institute of Psychiatry, London, UK
If the expert thought it necessary, separate judgements were made for children in low and
high socioeconomic (SES) groups. For this purpose, the low SES group was defined as children
living in households with incomes at, or below, the fifteenth percentile; the high SES group was
Draft - Do Not Cite or Quote
117
September 27, 1996
-------
defined as children living in households with incomes above the fifteenth percentile. Five of the
six experts chose to make separate judgements based on socioeconomic status.
At blood-lead levels ranging from 2.5 to 27.5 ng/dL, the distribution of increased
percentage of children having IQ scores less than 70 was reported by Wallsten and Whitfield
(Table D3-4), for each expert and SES (low and high). These distributions were combined
by calculating the weighted average of the low income median (15% of weight) and high
income SES medians (85%). The increased percentage of children having IQ scores less than
70, due to lead exposure, was estimated from the weighted average of the medians as a
piecewise linear function of blood-lead concentration. This function is reported in Table 4-5
and illustrated in Figure 4-7, over a range of blood-lead concentrations. For example,
1.06% = -0.193 + 0.1044 x 12 of children with blood-lead concentrations of 12 ug/dL are
expected to have IQ scores less than 70 due to lead exposure above and beyond those whose
IQ would naturally fall below that level.
Table 4-5. Piecewise Linear Function for Estimating the Increased Percentage of Children
Having IQ Scores less than 70 Due to Lead Exposure.
Range of Blood-Lead (PbB)
Levels 25.0
Function for Estimating Increased Percentage of Children
Having IQ Scores less than 70 (IQ < 70)
IQ<70 = 0.360 + 0.0204 PbB
IQ<70 = 0.218 + 0.0488 PbB
IQ<70 = 0.217 + 0.1068 PbB
IQ<70 = 0.193 + 0.1044 PbB
IQ < 70 = 0.108 + 0.0976 PbB
IQ<70 = 0.534 + 0.126 PbB
IQ<70 = 0.653 + 0.1328 PbB
IQ<70 = 1.112 + 0.1532 PbB
IQ<70 = 0.942 + 0.1464 PbB
Note to Reader: The piecewise linear function displayed in Table 4-5 is being revised for the
next draft of this report.
Draft - Do Not Cite or Quote 118 September 27. 1996
-------
5-
4 -
3
0-
10 20
Blood—Lead Concentration (/ug/dL)
3O
40
EPI model predictions for a fixed dust-lead loading of 100 ^g/sq.ft, no pica or paint exposure, and
African American race were used to relate soil-lead concentrations to blood lead.
Figure 4-7. Increase in Percentage of Children with IQ Below 70 Due to Lead Exposure
Plotted Against Blood-Lead Concentration
The relationships between environmental lead levels and the increased percentage of
children having IQ scores less than 70 are presented hi Figures 4-8 and 4-9 for the IEUBK
and EPI models, respectively. For each curve, the soil or dust-lead levels were varied over a
range of values, while all other parameters were held fixed. The predicted blood-lead level was
used to estimate the increased percentage of children with IQ scores less than 70 due to lead
exposure. For example, if the soil- or floor dust-lead concentration were 1000 ppm, the IEUBK
model predicts that an additional 0.8% of children will have IQS less than 70. For a dripline
soil-lead concentration of 1000 ppm, the EPI model predicts that 0.5% more children will have
IQS less than 70.
Draft - Do Not Cite or Quote
119
September 27, 1996
-------
CD
o
5
c
CD
cn
D
!O
Q-
3 -
2 -
0 -I
10 100 1000
Lead Concentration in Soil or Dust (ppm)
Dusi Soil
The IEUBK model was used to relate environmental lead to blood lead.
Dust-lead concentration was set to 100 ppm for the soil curve and soil-lead concentration was
set to 200 ppm for the dust curve.
Figure 4-8. Increase in Percentage of Children with IQ Below 70 Due to Lead Exposure
Plotted Against Concentration of Lead in Soil and Dust, Utilizing IEUBK Model
Predictions
Draft - Do Not Cite or Quote
120
September 27, 1996
-------
4 •
O
s
~o>
CD
— 3
.£
"i
c
o
-------
Figure 4-10 illustrates the relationship between geometric mean blood-lead concentration
and the percentage of children with a blood-lead level greater than 25 ug/dL, over a range of
geometric mean blood-lead levels. This relationship was computed using a geometric standard
deviation of 1.6 ug/dL, assuming that blood-lead concentrations may be characterized by using a
log-normal distribution. The same assumptions were applied in the risk characterization to
calculate the percentage of children having a blood-lead concentration greater than 10 (ig/dL.
Relationships between environmental lead levels and the incidence of blood-lead levels greater
than 25 ng/dL are illustrated hi Figures 4-11 and 4-12. For each curve, the soil- or dust-lead
levels were varied over a range of values, while all other parameters were held fixed. The
predicted blood-lead concentration was used to calculate the percentage of children having a
blood-lead concentration greater than 25 ug/dL for the environmental conditions. For a soil-lead
concentration of 2000 ppm the IEUBK model predicts that 10% of children will have blood-lead
concentrations greater than 25 ng/dL. The EPI model, by contrast, predicts that less than 1% of
children will have blood-lead concentrations greater than 25 iig/dL for a dripline soil-lead
concentration of 2000 ppm.
Draft - Do Not Cite or Quote 122 September 27. 1996
-------
90 •
80
70 •
60 •
50 •
40 •
30-
20
r 10 -
0-
10 20 30
Blood Lead Concentration (/ug/dL)
Figure 4-10. Percentage of Children with Blood-Lead Concentration Greater than 25 //g/dL
Due to Lead Exposure Plotted Against Geometric Mean Blood-Lead
Concentration, Assuming a GSD of 1.6
Draft - Do Not Cite or Quote
123
September 27, 1996
-------
80-
y 70-
50 -
"5 40-
30 -
20 J
10 100 1000
Lead Concentration in Soil or Dust (ppm)
10000
Dust
Soil
The IEUBK model was used to relate environmental lead to blood lead.
Dust-lead concentration was set to 100 ppm for the soil curve and soil-lead concentration was
set to 200 ppm for the dust curve.
Figure 4-11. Percentage of Children with Blood-Lead Concentration Greater than 25 //g/dL
Due to Lead Exposure Plotted Against Concentration of Lead in Soil and
Dust, Utilizing IEUBK Model Predictions
Draft - Do Not Cite or Quote
124
September 27, 1996
-------
80-
fe 60
o
50
o>
o 10
c
0 •
10 100
Soil —Lead Concentration (ppm)
1000
10000
The EPI model was used to relate environmental lead to blood lead. Floor dust-lead loading was set to 25 //g/ft2,
window sill dust-lead loading was set to 50 //g/ft2 and it was assumed that there was no paint pica.
Figure 4-12. Percentage of Children with Blood-Lead Concentration Greater than 25 fjg/dL
Due to Lead Exposure Plotted Against Concentration of Lead in Soil, Utilizing
EPI Model
4.5 OPTIONS FOR STANDARDS
The risk characterization in Chapter 5 will characterize risk reduction associated with
implementing different options for the §403 standards. Because multiple standards will be set
for different media it is not possible to evaluate all possible combinations of standards. The
purpose of this section is to provide a range of options for the standards which can be further
evaluated in the risk assessment and economic analysis.
Options for the standards were developed from modeling analyses of two epidemiological
studies, the Baltimore R&M and the Rochester Study. The Rochester study was the primary
source for defining the options for §403 standards in this section.
Draft - Do Not Cite or Quote
125
September 27, 1996
-------
4.5.1 Analysis Methods
Two main analysis methods were used to estimate options for the §403 standards from the
epidemiological data: 1) regression models; and 2) sensitivity/specificity analyses. Each method
was applied to individual standards separately (single media analysis) and to a combination of
standards jointly (multi-media analysis). Summarized final results from the statistical analyses
are presented in this section, while specific details on the statistical analyses are provided in
Appendix Y.
Regression Models
Both single and multi-media regression models were explored for estimating options for
the §403 standards based on data from the Rochester Study. The single media models related
childhood blood-lead concentrations to measures of lead from each media (floor dust, window
sill dust, soil, and paint) separately as follows:
where PbBj represents the blood-lead level of the child living in the ith home, PbEj represents a
measure of environmental lead (either on the original scale or log transformed) from the ith
home, Po and P, are intercept and slope parameters which describe the modeled relationship, and
Sj is the residual error in ln(PbBj) left unexplained by the model. For each model, an estimate of
the environmental lead level at which 95 percent of the population of children would be expected
to be below 10 (ig/dL was provided to help develop ranges of options for the §403 Standards.
The multi-media regression model related childhood blood-lead concentrations to
measures of lead from each media (dust, soil and paint) simultaneously as follows:
ln(PbBj) = po +p,*ln(Dusti) + p2*ln(Soili) +p3*Paintj + Ss
For this model, a joint estimate of environmental lead levels in paint, dust, and soil at which 95
percent of the population of children would be expected to be below 10 |ig/dL was provided.
Draft - Do Not Cite or Quote 1 26 September 27, 1996
-------
Sensitivity/Specificity Analyses
Table 4-6 describes the performance characteristics that were estimated as part of the
sensitivity/specificity analyses. The single media analysis focused on media standards that each
corresponded to a single measure of lead at the primary residence, whereas in the multi-media
analysis, the media standard corresponded to measures of environmental lead in several media
(dust, soil and paint). Thus, if any single measure of environmental lead was above a standard in
the multi-media analysis, the residence was categorized as being above the standard.
Table 4-6. Definitions of Performance Characteristics Used to Characterize the
Performance of Options for the §403 Standards Based on Empirical Data from
Lead Exposure Studies
Blood Lead
Concentration
Standard
In the above table, the lent
above a given blood-lead s
standard for that environm
counts the following perfo
Performance
Characteristic
Sensitivity
(or True Positive Rate)
Specificity
(or True Negative Rate)
Positive Predictive Value
(PPV)
Negative Predictive
Value (NPV)
Above
Below
Media Standard
Below
a
c
Above
b
d
ar 'a' represents the number of children which have a blood lead concentration
tandard and who live in a residence with an environmental lead level below a
ental medium. Letters 'b', 'c', and 'd' represent similar counts. From these
rmance characteristics are calculated
Definition
Probability of a dwelling being above the soil lead
standard given that there is a resident child with an
elevated blood concentration.
Probability of a dwelling being below the soil lead
standard given that a resident child has a low blood
lead concentration.
Probability of a resident child having an elevated
blood lead concentration given that the observed soil
lead in the dwelling is above the standard.
Probability of a resident child having a low blood
lead concentration given that the observed soil lead
in the dwelling is below the standard.
Calculation
b/(a + b)
c/(c+d)
b/(b + d)
c/(a+c)
Draft - Do Not Cite or Quote
127
September 27, 1996
-------
4.5.2 Assumptions
It was assumed that individual standards will be,set for each media, location, or surface
addressed in the rulemaking, and that action will be recommended if any individual standard is
exceeded. Therefore, ranges are provided for each media, location, or surface separately. It is
also assumed that dependencies between media will be addressed in the guidance provided to
risk assessors and in the recommended actions associated with each standard.
Ranges for the options of the standards presented in this report are based on estimated
effects of lead exposure on blood-lead concentration. The high and low values for each range
result from different analyses. The low values are generally taken from the conservative, single
media analyses and the high values from the less conservative, joint analyses. Still, all options
proposed are estimated to provide a minimum 95% probability that a child living in a home with
environmental levels below each of the §403 standards will have a blood-lead concentration less
than 10 ug/dL.
4.5.3 Results
Table 4-7 presents a summary of estimates of standard levels which achieved (for homes
that meet the standard) either:
1. A negative predictive value of 95% from a single media or multi-media
sensitivity/specificity analysis, or
2. An estimated 95% probability that a child's blood-lead concentration is below 10
Ug/dL from the single media or multi-media regression analyses.
The above two criteria were the current target health effects to be considered in choosing the
§403 standards.
As seen in Table 4-7, based on the single media analyses, i.e., each §403 standard
considered separately, options for standards must be set very low to meet the target health
effects. However, Table 4-7 also indicates that if the target health effects are associated with
meeting all §403 standards simultaneously, then the options for standard levels can rise
significantly, even well above the levels proposed in the Interim Guidance (EPA, 1995h). In the
Draft - Do Not Cite or Quote 128 September 27, 1996
-------
Table 4-7. Summary of Estimated Standards Which Achieved a Negative Predictive Value of B5% or an Estimated 95%
Probability of a Child's Blood-Lead Concentration Below 10 /fg/dL in a Dwelling that is at or Below the Standard.
Media
Dripline Soil (pg/g)
Play Area Soil O/g/0)
Uncarpeted Floor big/ft1)
Carpeted Floor (/ig/ft1)
Carpeted and Uncarpeted
Floor big/ft1)
Window Sills b/g/ft')
Window Troughs li/g/ft1)
Percent of Interior Components
with Deteriorated LBP
Average Percent of Deteriorated
LBP per Interior Component
Percent of Exterior Components
with Deteriorated LBP
Average Percent of Deteriorated
LBP per Exterior Component
Maximum of Interior/ Exterior
Percent of Components with
Deteriorated LBP
Maximum of Interior/ Exterior
Average Percent of Deteriorated
LBP per Component
Single Media Analyses1
Regression
Models
nOCnBSt6f
<50
<50
<25
<25
<25
<26
<2S
0%
0%
0%
0%
SensltMtv/Specifidtv
R&M
-------
Rochester study for any single media/surface/location standard alone there were homes with
children with blood-lead concentrations above 10 iig/dL, even at very low media lead levels.
However, when all media/surface/location standards were considered jointly, even at relatively
high levels, all homes with a child with a blood-lead concentration above 10 ng/dL were
identified. In other words, any home with a child with a blood-lead concentration above 10
Hg/dL exhibited a relatively high lead level in at least one media/surface/location being
considered for a §403 standard. This is a significant finding related to the §403 rulemaking as it
illustrates the effectiveness of using multiple standards applied in a single risk assessment to
identify a home with a lead-based paint hazard.
It should be noted that the above conclusion is based upon the multi-media
sensitivity/specificity analyses from Rochester alone. The multi-media regression analyses do
not show the same effect. There could be a number of reasons for this including the fact that the
form of the regression model fitted is not sensitive to interactions between the effect of the
different environmental media. In addition, many blood-lead concentrations are close to the
sensitivity/specificity target value of 10 fig/dL.
Based on Table 4-7, a range of options for a standard for each media/surface/location can
be specified. Table 4-8 below lists the proposed range of options for each standard. For each
standard the upper limit of the range is specified as the maximum estimated level in one of the
multi-media analyses.
The lower limit of the standard is more problematic. Very low levels of standards (e.g.
less than 25 ng/ft2 for dust and less than SO ug/g for soil) were estimated by the single media
analyses to be associated with the target health effect. From a practical standpoint, these
standards are not necessarily achievable. Therefore, the lower limits presented in Table 4-8 for
the range of options for a standard are based on practicality.
For dust-lead standards, this lower limit is defined as 25 ug/ft2 because a dust-lead
standard below 25 (ig/ft2 may be problematic due to laboratory analysis issues. Many currently
accredited laboratories have levels of quantification between 10 fig/ft2 and 25 ug/ft2.
Draft - Do Not Cite or Quote 130 September 27, 1996
-------
Table 4-8. Proposed Options for §403 Standards To Be Evaluated in the Risk Assessment
and Economic Analysis.
Standard
Uncarpeted Floor Dust-Lead Loading
(fjgm2 )
Window Sill Dust-Lead Loading (jug/ft2 )
Dripline Soil-Lead Concentration (/yg/g)
Play Area Soil-Lead Concentration
(pg/g)
Maximum of Percent of Interior
Components with Deteriorated Lead-
Based Paint and Percent of Exterior
Components with Deteriorated Lead-
Based Paint
Range
Low Limit
25
25
50
50
0%
High Limit
400
800
1500
1000
20%
Note to EPA: The amount of damaged lead-based paint is currently being expressed in terms
of square footage rather than percentage of components in the integrated risk
assessment.
The lower limit for a soil-lead standard(s) is defined as 50 ug/g. This level was chosen
since national background levels of lead in soil have been estimated in the neighborhood of 20
ug/g to 35 ug/g.
The lower limit for a paint-lead standard was chosen as zero percent deteriorated lead-
based paint, the level suggested by the single media analyses. Single media analyses indicate
that no single standard for paint can achieve the target health criteria of a 95% probability that a
child's blood-lead concentration when exposed to paint lead will be below 10 u.g/dL. There are
no practical difficulties associated with setting a lower limit of zero percent for deteriorated
lead-based paint.
The ranges hi Table 4-8 were used as bounds on options for standards that were
evaluated in Chapter 5.
Draft - Do Not Cite or Quote
131
September 27, 1996
-------
5.0 INTEGRATED RISK ANALYSIS
CHAPTER 5 SUMMARY
This chapter provides the estimated health risks and blood-lead
concentrations for children aged 1-2 years associated with current residential
lead exposures (pre-%403). These risks are then compared to the projected
health risks and blood-lead concentration associated with the predicted
residential lead exposures in the post-1403 environment. Baseline risk is
computed based on NHANES survey data; post-%403 risk is estimated
separately using the IEUBK model and an EPI model applied to environmental-
lead levels observed in the HUD National Survey. Post-%403 environmental-lead
levels are adjusted for the assumed effects of intervention initiated by §403,
under various options of standards for lead in dust, soil, and paint. The results
of this chapter are projected childhood health risks and blood-lead
concentrations for a wide range of options for standards.
A sensitivity analysis was performed to characterize how the estimated risk
reductions results differ when alternatives are considered for the most important
assumptions and approaches in the procedure. Alternative procedures were
considered for characterizing baseline and post-intervention blood-lead
distributions. Alternative assumptions were considered on post-intervention
environmental-lead levels, the IQ decrement associated with increased blood-
lead concentration, and the age group of interest. The largest differences in
results, especially those representing the most extreme health effects, tended to
appear when making alternative assumptions on post-intervention
environmental-lead levels and on the IQ decrement associated with increased
blood-lead concentration.
The goals of the integrated risk analysis are to
• identify the baseline distribution of blood-lead concentrations in the nation's
children,
• use this distribution to characterize adverse health effects associated with elevated
blood-lead concentration in children,
• characterize the post-intervention distribution of children's blood-lead
concentrations and associated adverse health effects adjusted for the assumed
effects of performing interventions in response to various options for the §403
standards,
• compare the post-§403 distributions of childhood health effects and blood-lead
concentrations to their respective pre-§403 baselines, noting any risk reduction that
may result for the §403 standards, and
Draft - Do Not Cite or Quote 132 September 27, 1996
-------
• Characterize how sensitive the estimated risk reductions are to the uncertainty
present in key assumptions, parameters, data sources, and analysis tools.
The risk assessment targets children aged 12 to 35 months. As discussed in Section 2.3 and 4.4,
the health effect endpoints of interest are:
• The number and percentage of children with blood-lead concentrations at least
10 ug/dL
• The number and percentage of children with blood-lead concentrations at least
25 ug/dL
• The average of IQ points lost per child, due to exposure to lead-based paint (LBP)
hazards v
• The number and percentage of children with IQ scores less than 70
• The number and percentage of children with IQ decrements of at least 1
• The number and percentage of children with IQ decrements of at least 2
• The number and percentage of children with IQ decrements of at least 3
In addition, in order to understand the impact of §403 on the nation's housing, the number and
percentage of housing units in which some action might be required is predicted for various
options of the standard for each medium.
Section 5.1 presents a characterization of the baseline (i.e., pre-§403) distribution of
children's blood-lead concentrations and associated health effect endpoints. The intervention
activities, expected reductions in environmental-lead levels, and intervention triggers are
discussed in Section 5.2. Section 5.3 characterizes the childhood health risks and blood-lead
concentrations predicted to exist after performing the interventions in response to §403. Finally,
Section 5.4 presents the results of a sensitivity analysis on the effects of the uncertainty present
in key assumptions, parameters, data sources, and analysis tools on the risk assessment.
Draft - Do Not Cite or Quote 133 September 27. 1996
-------
5.1 BASELINE CHARACTERIZATION OF CHILDREN'S
BLOOD-LEAD CONCENTRATIONS AND HEALTH EFFECTS
For purposes of this Risk Assessment, interventions in response to the proposed §403
rules are assumed to begin in 1997. In this section, the national distribution of children's blood-
lead concentrations in 1997 is estimated to characterize the distribution prior to the enactment of
§403. This distribution is then used to estimate selected health endpoints. These
characterizations serve as the baseline for evaluating the risk of environmental-lead exposure to
children. Post-intervention distributions are compared to these baselines to assess risk reduction
resulting from §403.
As discussed in Section 2.4, the target age for characterizing a baseline distribution of
blood-lead concentrations and estimating health effects is from 1 to 2 years (12 to 35 months).
Information from NHANES ID, phase I was used to characterize the 1997 baseline distribution
of blood-lead concentrations in children aged 1 to 2 years. The NHANES m weights, for
children aged 1 to 2 years with non-missing blood-lead concentrations, add up to 5,272,000,
which is less than 7,961,000, the total number estimated for 1997 in Section 3.3. Therefore, the
sampling weights in NHANES ffl were scaled by a factor of 7,961/5,272 so that the distribution
determined by the NHANES ffl represents the same total number of children projected for 1997
as the subsequent post-§403 estimates.
There are many ongoing local, state, and federal initiatives to reduce childhood blood-
lead concentrations. Therefore, the actual distribution of childhood blood-lead concentrations in
1997 may differ from that reported in NHANES m which was conducted between 1988 and
1991. Specifically, if the government strategies already in place are effective, the distribution
reported may assign slightly higher probabilities to elevated blood-lead concentrations than will
actually occur in 1997. Nevertheless, the distribution of childhood blood-lead concentrations in
NHANES III is the best available data for estimating the baseline distribution of blood-lead
concentrations in 1997.
Figure 5-1 contains two plots presenting the estimated baseline distribution of blood-
lead concentrations in 1997 for children aged 1-2 years. The top plot presents the estimated
relative frequency distribution based on a lognormal model. The bottom plot presents the
estimated cumulative frequency distribution. There are two curves in this plot. The jagged
Draft - Do Not Cite or Quote 134 September 27, 1996
-------
25 4
O 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
Blood — Lead Concentration
80
40 •
0 1 2 3 4 5 6 7 8 9 1O 11 12 13 14 IS 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
Blood —Lead Concentration (,ug/dL)
Figure 5-1. Baseline Distribution of Blood-Lead Levels Based on NHANES III, Phase 1
(0.2 Percent of Children Had Blood-Lead Concentration Greater than 32
fjg/dl)
Draft - Do Not Cite or Quote
135
September 27, 1996
-------
curve represents the empirical distribution observed in NHANES ffl; the smooth curve
represents the lognormal model approximation to the data. The plot illustrates that the
lognormal model fits the data well..
This graph is useful for understanding the percentage of children aged 1-2 years having a
blood-lead concentration greater than.a specified value. For example, 10.5% = 100% - 89.5% of
children are estimated to have a blood-lead concentration greater than 10 ng/dL, and 0.6% =
100% - 99.4% are estimated to have a blood-lead concentration greater than 25 jig/dL.
The estimated geometric mean blood-lead concentration was 4.1 iig/dL, and the estimated
geometric standard deviation was 2.1 ug/dL. The distributions in Figure 5-1 were translated to
the distribution of number of IQ points lost due to blood-lead concentration using the methods
discussed in Section 4.4.1. Figure 5-2 displays this translated distribution using the same format
as in Figure 5-1. The lower plot presents the percentage of children with anticipated IQ point
losses within particular ranges. For example, the percentage of children with at least a 2 point
decrement in IQ, due to blood-lead concentration, is estimated to be about 18 percent.
Table 5-1 displays estimated 1997 baseline probabilities for the various adverse health
effects in children aged 1-2 years. These estimates were calculated from the information
summarized in Figures 5-1 and 5-2 and are employed to characterize the pre-§403 risk of lead
exposure in 1997. Thus, it is assumed that this risk will prevail if §403 is not implemented.
The method for estimating the probability of children having IQ less than 70 is described in
Section 4.4.2. Each of these endpoints is estimated from the geometric mean and geometric
standard deviation, assuming a lognormal distribution. The mathematical approach used to
make these inferences is described in Step (5) of Appendix E2.
5.2 INTERVENTION ACTIVITIES
Once promulgated, §403 will prompt intervention activities targeting residential lead
hazards. These interventions will be conducted on behalf of children already exposed to the
targeted lead hazards, as well as children who would otherwise be exposed if the hazards are
not abated or controlled. For the purposes of the Risk Assessment, a lead hazard intervention is
defined as any non-medical activity that seeks to prevent a child from being exposed to the
Draft - Do. Not Cite or Quote 136 September^?, 1996
-------
25
0)
a.
15
10-
3456
Number of IQ Points Lost
10
_0
£
60-
40-
20-
0 ->
3456
Number of 10 Points Lost
10
Figure 5-2. Baseline Distribution of IQ Decrements Due to Elevated Blood-Lead
Concentration Based on NHANES III, Phase 1 (0.03 Percent of Children Had
in Excess of 10 IQ Points Lost)
Draft - Do Not Cite or Quote
137
September 27, 1996
-------
Table 5-1. Estimated Baseline (1997, Pre-lntervention) Number and Percentage of
Children Aged 1 to 2 Years Having Specific Health Effects.
Health Effect
Blood-lead concentration of at least 25 /vg/dL1
Blood-lead concentration of at least 1 0 i/g/dL1
IQ score less than 702
IQ score decrement of greater than 1 3
IQ score decrement of greater than 23
IQ score decrement of greater than 33
IQ decrement
Estimated
Baseline
Number of
Children
46,000
834,000
45,000
4,152,000
1,451,000
564,000
Average
1.35
Estimated
Baseline
Percentage of
Children
0.58
10.5
0.57
52.2
18.2
7.09
SD
1.11
1 Determined from Figure 5-1.
2 Determined from methods in Section 4.4.2
3 Determined from Figure 5-3.
lead in his or her surrounding environment. An intervention, therefore, may range from the in-
home education of parents regarding the dangers of a young child's hand-to-mouth activity, to
the abatement of lead-based paint.
An intervention conducted on behalf of children already exposed to the targeted hazard is
termed secondary prevention (e.g., paint abatement in the home of a child who has a blood-lead
concentration exceeding 20 ug/dL). A primary prevention intervention prevents exposure
before it occurs (e.g., paint abatement in a home before a new family with children moves in).
The distinction between primary and secondary prevention efforts is one of the population
targeted rather than the activity conducted. In fact, a given intervention can have primary and
secondary prevention benefits.
One objective of §403 is to prompt primary prevention interventions targeting lead
hazards in residential soil, dust, and paint. (Secondary prevention will, of course, also take
place.) As the risk assessment needs to model the expected benefits following promulgation of
§403, measures of the effectiveness of these lead hazard interventions are required.
Draft - Do Not Cite or Quote
138
September 27, 1996
-------
Unfortunately, there is no information currently available in the scientific literature regarding
the efficacy (as measured by either health outcomes or by changes in children's blood-lead
concentrations) of primary prevention interventions targeting paint, dust, or soil. There are
limited data on the effectiveness of secondary prevention interventions (EPA, 1995b).
Research suggests that primary prevention interventions will produce greater efficacy
than secondary prevention interventions (Gulson et al., 1995). Bone-lead stores accumulated
by exposed children continue to mobilize into the blood following an intervention and may
mask the intervention's full effectiveness. The effectiveness of interventions studied in the
literature, therefore, have shortcomings as estimates of the efficacy stemming from primary
interventions. Thus, these blood lead declines from a secondary prevention situation are not
used as our primary mechanism for assessing health benefits of the §403 rule. However, a
method which uses the changes in blood-lead concentrations from secondary prevention
settings with a modeled effect of the bone lead stores is examined in the Section 5.4 sensitivity
analysis.
Data on the effectiveness of lead hazard interventions in terms of change in
environmental lead levels following interventions targeting paint, dust and soil were used to
estimate environmental-lead levels following interventions conducted as a result of §403. It is
important to note, however, that only some of the interventions considered viable under current
standards have been studied in the literature. Where available, the reported post-intervention
environmental-lead levels may then be translated into blood-lead concentrations representing
the benefit of primary prevention interventions. The translation is accomplished using both
epidemiological models and the IEUBK Lead Model. Where little or no environmental
effectiveness information is available about a particular intervention, EPA has used its current
understanding of the intervention to develop an estimated effectiveness.
Fully characterizing an intervention requires addressing four questions:
1. What "triggers" the intervention?
2. What procedures are conducted during the intervention?
Draft - Do Not Cite or Quote 139 September 27, 1996
-------
3. How effective is the intervention at reducing environmental lead levels?
4. How effective is the intervention at reducing blood-lead concentrations (for both
primary and secondary prevention)?
The interventions and their associated procedures utilized in the risk assessment are
discussed in Section 5.2.1. The effectiveness of these methods in reducing environmental lead
levels and blood-lead concentrations are documented in Section 5.2.2 and 5.2.3, respectively.
Finally, Section 5.2.4 discusses the circumstances under which each of the defined
interventions are triggered.
5.2.1 Interventions
For the purposes of the Risk Assessment, a total of seven interventions were defined for
lead in paint, dust, and soil. The seven interventions are dust cleaning, exterior LBP
maintenance, exterior LBP abatement, ulterior LBP maintenance and abatement, soil cover, and
soil removal. For interior paint, exterior paint, and soil, two intervention approaches were
defined. These two approaches are intended to reflect the viable range in scope achieved by
interventions of the targeted media. For residential dust, only a dust-cleaning method was
included as a one-time activity to follow LBP interventions. Table 5-2 presents these seven
interventions by defining the procedures conducted and the expected duration of the
intervention's benefits.
The procedures defined in Table 5-2 for each of the interventions are consistent with
intervention practices currently recommended by EPA (and mandated hi some communities).
For example, paint removal must be conducted using appropriate precautions, and LBP
encapsulation must utilize materials approved as encapsulants (i.e., remain effective for 20
years). The procedures exclude interventions previously utilized but now considered hazardous,
such as open-flame burning or abrasive sanding of lead-based paint.
Draft - Do Not Cite or Quote 140 September 27. 1996
-------
Table 5-2. Interventions Defined for the §403 Risk Assessment Effort
Intervention
Dust Cleaning
Exterior
LBP
Interior
LBP
Soil
Maintenance
Encapsulation/
Abatement
Maintenance
Encapsulation/
Abatement
Cover
Removal
Procedures Defining the Intervention
Cleaning the unit using HEPA vacuums and wet
mopping.
Painted surfaces with deteriorated LBP are repaired
by feathering the edges of deteriorating paint and
repainting with new, lead-free paint.
Painted surfaces with deteriorated LBP are
encapsulated, enclosed, or removed using currently
acceptable practices and materials.
Painted surfaces with deteriorated LBP are repaired
by feathering the edges of deteriorating paint and
repainting with new, lead-free paint. Window sills
are covered with permanent barrier. A Dust
Cleaning of the affected area follows the
intervention.
Painted surfaces with deteriorated LBP are
encapsulated, enclosed, or removed using currently
acceptable practices and materials. A Dust
Cleaning of the housing unit follows the
intervention.
Areas of bare soil are reseeded, resodded or
covered with mulch, gravel, etc.
Soil from areas with elevated lead concentrations
are removed and replaced with clean soil, or the
areas are permanently covered. A Dust Cleaning of
the housing unit follows the intervention.
Expected
Duration1
5 years
5 years for paint
20 years for paint
5 years for paint,
5 years for dust
20 years for paint,
8 years for dust
5 years
Permanent
1 Duration is defined as the length of time before the lead levels in the targeted medium or conditions of the medium
require further intervention.
The specified durations of the interventions reflect the length of time before the targeted
media returns to levels or conditions requiring further intervention. For example, the duration
of a paint intervention represents the estimated period of time before formerly intact or repaired
surfaces deteriorate. When defining the duration of interior lead-based paint abatements, the
duration of reduced interior residential dust-lead levels is also defined. Since paint
interventions target only deteriorated lead-based paint, it would be unrealistic to assume that
dust-lead levels remain low permanently. The once intact lead-based paint could, over time,
deteriorate and produce elevated lead levels in residential house dust.
Draft - Do Not Cite or Quote
141
September 27, 1996
-------
Unfortunately, there were only limited data available for estimating the duration of the
methods defined in Table 5-2. The HUD Regulatory Impact Analysis (pages 3-21 through 3-
22) utilized 4 and 8 years as the duration of reduced dust-lead levels following interim paint
controls and paint abatements, respectively. These durations were based on estimates of the
rate of increased dust-lead loading fag/ft2 per year) stemming from residential recontamination
reported in studies of LBP interventions conducted in Baltimore and Cincinnati (page 3-22).
These durations were the starting point for determining the dust durations for the paint
interventions reported in Table 5-2. The efficacy duration for paint in the paint
encapsulation/abatement intervention is consistent with HUD's definition of a LBP abatement
practice: requiring the abatement to be effective for at least 20 years in order to be called an
abatement. The five-year duration for paint maintenance intervention was intentionally set
equal to the estimated dust duration, as the dust intervention is designed to follow paint
interventions. For the soil cover intervention, a 5 year duration was utilized. This duration is
reasonable given the duration of exterior paint repair (which would presumably reduce soil-
lead levels) cited in the HUD RIA (page 3-23). Finally, the soil removal intervention was
assumed to have permanent effectiveness in that the soil exhibiting elevated lead
concentrations had been either removed or permanently covered.
5.2.2 Reductions in Environmental Lead Levels Following Interventions
The effectiveness of the interventions outlined in Table 5-2 is defined in terms of
reductions in environmental-lead levels following conduct of the intervention. More
particularly, the post-intervention environmental-lead levels may be specified for each of the
interventions. Table 5-3 presents the post-intervention environmental-lead levels for each of
the interventions described in Table 5-2. For each intervention, the post-intervention lead
levels are defined for those media expected to be affected by the intervention. For example,
interior paint abatement can be expected to prompt reductions in ulterior dust-lead loadings as
well as in paint-lead loadings. Where relevant, additional details are provided regarding the
effectiveness of the interventions.
The interventions outlined in Tables 5-2 and 5-3 are intended to include state-of-the-art
practices. As a result, defining the effectiveness of these interventions as measured by reduced
Draft - Do Not Cite or Quote 142 September 27, 1996
-------
environmental-lead levels is difficult. Though numerous intervention studies are documented
in the literature, many utilized methods that today would be considered inappropriate. The
available information on intervention effectiveness too often is of little relevance. Where
possible, however, the available data were utilized.
Encapsulation/abatement of interior paint is assumed to reduce residential floor and
window sill dust-lead loadings to 40 and 100 fig/ft2, respectively, while effectively eliminating
(for the duration outlined in Table 5-2) the hazard from deteriorated lead-based paint. The
same degree of effectiveness with regard to residential dust was assumed for maintenance of
interior paint (but for a shorter duration), soil removal, and one-time dust cleaning. This value
was selected after considering the efficacy reported for housing units in the Denver
Comprehensive Abatement Performance (CAP) Study and in the Baltimore Experimental Paint
Abatement Study. The geometric mean floor vacuum dust-lead loading measured in abated
units studied by the Denver CAP Study was 29.0 fig/ft2 approximately two years following
extensive paint abatements; the geometric mean window sill vacuum dust-lead loading was
91.6 iig/ft2 among the same units (page 34 of EPA, 199Sf). Similarly, the Baltimore
Experimental Paint Abatement Study reported a geometric mean floor wipe dust-lead loading
of 40.9 ug/ft2 among 13 housing units 18-42 months following complete paint abatements; a
geometric mean of 103 fig/ft2 was reported for the unit's window sill wipe dust-lead loadings at
the same time (page 62 of EPA, 1995c).
The complete effectiveness in terms of paint levels assumed for the four paint
interventions is consistent with the procedures defined for the interventions and their assumed
durations. These interventions are defined as utilizing practices consistent with ensuring that
the surfaces with deteriorated paint remain intact following the intervention for the specified
duration. Recall that the durations were defined to recognize the potential for paint, intact at
the time of the intervention, becoming deteriorated over tune. Thus, the potential hazard posed
by deteriorated paint is assumed to be completely eliminated by each of the interventions (both
interior and exterior) for the durations specified in Table 5-2.
Draft - Do Not Cite or Quote 143 September 27, 1996
-------
Table 5-3. Expected Post-Intervention Lead Levels Associated With Performing §403
Interventions.
Intervention
Dust Cleaning
Exterior
LBP
Interior
LBP
Soil
Maintenance
Encapsulation
/ Abatement
Maintenance
Encapsulation
/ Abatement
Cover
Removal
Post-Intervention Lead Level
Dust-lead loading = min{40 //g/ft2, pre} for floors
min{100//g/ft2, pre} for window
sills
Dust-lead concentration = min{328 /ig/g, pre} for floors
Paint = 0
Paint = 0
Paint = 0
Dust-lead loading = min{40 //g/ft2. pre} for floors
min{ 100 //g/ft2, pre} for window
sills
Dust-lead concentration = min{328 //g/g, pre} for floors
Paint = 0
Dust-lead loading = min{40 //g/ft2, pre} for floors
minj 100 //g/ft2, pre} for window
sills
Dust-lead concentration = min{328 //g/g, pre} for floors
Soil = 50% of Pre-lntervention Levels
Soil = 1 50 ppm
Dust-lead loading = min{40 //g/ft2, pre} for floors
min{ 100 //g/ft2, pre} for window
sills
Dust-lead concentration = min{328 //g/g, pre} for floors
Comments on Performing the
Intervention
It is assumed that this
intervention would occur only
if dust-lead levels were above
the standard, and if no sources
of lead exposure remain in the
housing unit.
Deteriorated LBP is eliminated
as a potential exposure source
for the duration specified in
Table 5-2.
Deteriorated LBP is eliminated
as a potential exposure source
for the duration specified in
Table 5-2.
Deteriorated LBP is eliminated
as a potential exposure source
for the duration specified in
Table 5-2.
Deteriorated LBP is eliminated
as a potential exposure source
for the duration specified in
Table 5-2.
Residential dust is assumed
unaffected by the intervention.
Residential dust is not
recontaminated by the
intervention
The post-intervention soil-lead concentration assumed following soil removal was
derived from the HUD National Survey and the Boston 3-City data. Among 257 HUD
National Survey housing units with interior and exterior LBP in good condition (e.g., no
deteriorated paint), a geometric mean soil-lead concentration of 61 ppm was reported. The
Boston 3-City Soil Abatement Project reported an average soil-lead concentration of 160 ppm
among 34 housing units 20 months following soil abatements (page 62 of EPA, 1995c). The
same study documented arithmetic average soil-lead concentrations of 171 ppm and 180 ppm
among two other groups (sample sizes of 32 and 26, respectively) of housing units 6 to 12
months following soil abatements (page A-24 of EPA, 1995c). Unfortunately, limited data
Draft - Do Not Cite or Quote
144
September 27, 1996
-------
were available regarding the effectiveness of soil cover. The one available study that examined
soil cover as an intervention strategy (Mielke et al., 1994) also included paint stabilization and
interior dust control in the interventions performed at each residence. Foundation and mid-yard
soil-lead levels, as measured by 2.5 cm core samples pre- and post-intervention, were reduced
by approximately a factor of 3 to 4, four months following the interventions (as noted in Table
5-3 a factor of 2 was assumed in the analyses).
5.2.3 Intervention Triggers
An intervention is triggered if the housing unit exhibits environmental-lead levels in
excess of those defined in §403. The findings of the risk assessment will not depend upon
when the intervention occurs, only on its effectiveness once the intervention is conducted. It is
assumed that the specific interventions conducted are those targeting the environmental media
exhibiting the elevated levels. If either dust, soil, or paint exhibit levels in excess of those
specified by the §403 standards, selecting appropriate interventions that target the problematic
media are assumed. However, since two approaches to intervention are defined for paint and
soil, a question remains as to which of the two is to be selected when the relevant
environmental medium exceeds the §403 standard.
Table 5-4 summarizes the circumstances under which each of the defined interventions in
Table 5-2 would be conducted. The choice of a soil-removal intervention versus a soil-cover
intervention is made strictly on the measured lead concentration for specific areas in the yard.
In contrast, the choice of an encapsulation/abatement approach versus a maintenance approach
to paint intervention is based on the extent to which deteriorated lead-based paint is present.
As noted earlier, dust cleaning is only prompted as a clean-up activity following an interior
paint intervention or soil removal, or as a one-time activity where elevated dust-lead levels are
observed despite the absence of residential sources of lead exposure (e.g., soil or paint). In
such a case, it is assumed that the source of the lead has been abated due to activities conducted
under §403 in the residence, neighboring residences, and the neighborhood in general.
Draft - Do Not Cite or Quote 145 September 27, J996
-------
Table 5-4. Intervention Triggers Defined for the Risk Assessment of §403.
Intervention
Dust Cleaning
Exterior
LBP
Interior
LBP
Soil
Maintenance
Encapsulation/
Abatement
Maintenance
Encapsulation/
Abatement
Cover
Removal
Circumstances Prompting Conduct of the Intervention
Follows any interior paint intervention or soil removal, and when
dust-lead loadings are elevated despite absence of residential
sources of lead exposure (e.g., no deteriorated LBP or elevated
soil lead).
When deteriorated exterior LBP is present, but not extensive (e.g.,
confined to a limited area).
When deteriorated exterior LBP is present and extensive (e.g., not
confined to a limited area).
When deteriorated interior LBP is present, but not extensive (e.g.,
confined to one area of the housing unit).
When deteriorated interior LBP is present and extensive (e.g.,
greater than one area of the housing unit).
When residential soil-lead concentrations exceed lower soil
standard, but do not exceed the higher, emergency standard.
When residential soil-lead concentrations exceed the higher,
emergency soil standard. It is assumed this degree of
intervention would only be warranted in specific areas of the
yard.
5.2.4 Reductions in Blood-Lead Levels Following Interventions
For each home in the National Survey, the post-intervention environmental-lead levels
presented in Table 5-3 were employed to predict the blood-lead concentrations of resident
children. If an intervention was triggered by one or more of the standards, then the
environmental-lead levels in the post-intervention time frame were set equal to those displayed
in Table 5-3. The dose-response models (EEUBK and EPI) for predicting blood-lead
concentration discussed in Section 4.1 were employed to predict childhood blood lead in the
residence following the intervention activity. In this manner, the Risk Assessment estimated
the impact of various options for the §403 standards on environmental-lead levels and
childhood blood-lead concentrations in the nation's housing and children.
Draft - Do Not Cite or Quote
146
September 27, 1996
-------
5.3 CHARACTERIZING THE RISKS FOLLOWING INTERVENTION
This section describes the predicted distributions of blood-lead concentrations and health
effects following promulgation of §403. To enable evaluation of various options for standards,
risks are characterized for different sets of standards, each set affecting a different number of
houses nationwide. A four-step process was employed to characterize these risks for each set
of standards:
1. §403 Intervention: Predict post-§403 environmental-lead levels
2. IEUBK/EPI PbB Prediction Model: Apply IEUBK and EPI models to
environmental-lead levels to predict post-§403 blood-lead concentrations
3. Calibrate Predicted Blood-Lead Concentrations Using NHANES ID: Estimate the
change in modeled pre- and post-§403 blood-lead concentrations and apply that
change to the NHANES in blood-lead concentration distribution
4. Summarize Risk: Predict health effects and blood lead endpoints for children aged
1-2 years.
This process is illustrated in Figure 5-3. In this figure, boxes with rectangular corners
represent datasets or tables of results. Boxes with rounded corners represents steps in the
process that transform the data being fed to it ~ either through a predictive model (e.g., the
IEUBK and EPI models are used to predict blood-lead concentrations from environmental-lead
levels) or computation.
The four numbered steps in the process are illustrated by the four boxes with rounded
corners in Figure 5-3. The remaining text in this section describes each of these steps in more
detail.
Step 1: Predict Post-§403 Environmental Lead Levels. The HUD National Survey data
were used to predict health risks after §403 in the following manner.
1. Environmental-lead levels observed at each home were compared to various
options for the §403 standards.
2. These levels are projected to be reduced as a result of an intervention required by
the rule or to remain unchanged if an intervention is not required as described in
Section 5.2.
Draft - Do Not Cite or Quote 147 September 27, 1996
-------
NHANES
III
1
HUD National Survey Environmental Levels
(Pre-§403 Environmental Levels)
Dust
Soil
Paint
§403
Intervention
Post-§403
Environmental
Levels
IEUBK/EPI
PbB Prediction Model
Modeled
Pre-§403
Blood-Lead
Distribution
Modeled
Post-§403
Blood-Lead
Distribution
J
Calculate Change In Modeled Blood-
Lead Distribution (Post Minus Pro)
and
Derive Post-§403 Blood-Lead
Distribution Based on Modeled
Change and NHANES III
Risk Summarization
(Estimation of Health Effects)
Impacts
Standards
9 Houses Affected
Health Endpolnts
. PbB>25
. PbB> 10
. IQ < 70
• Average, Std Dev IQ decrement
• IQ decrement > 1,2,3
Dust
Soil
Paint
3
Figure 5-3. Post-§403 Risk Characterization Process
Draft - Do Not Cite or Quote
148
September 27, 1996
-------
For example, Table 5-5 illustrates the assumed impact of a §403 intervention on
environmental-lead levels at a particular house in the HUD National Survey (ID 1011501) for
the following set of standards:
floor dust-lead loading:
window sill dust-lead loading:
soil (cover):
soil (removal):
paint (maintenance):
paint (abatement):
200 ng/ft2
500 ug/ft2
400 ug/g
3000 ug/g
Greater than 5 ft2 of damaged LBP
Greater than 20 ft2 of damaged LBP
Table 5-5. Projected Impact of §403 on House 1011501 in the National Survey
Environmental Lead Level
Floor dust-lead loading U/g/ft2)
Floor dust-lead concentration (//g/g)
Window sill dust-lead loading (/t/g/ft2)
Soil-lead concentration (//g/g)
Maximum XRF (mg/cm2)
Damaged LBP (ft2)
Pre-§403
32.4
623
65.3
4,619
1.4
0
Post-§403
32.4
328
65.2
150
1.4
0
Notice that the soil-lead concentration (4619 u,g/g) was above the 3000 (ig/g standard for soil
removal. This would have triggered soil removal. As pointed out in Table 5-3, this causes
post-§403 soil-lead concentration to go down to 150 u.g/g, and dust-lead concentrations to be
reduced to the minimum of pre-intervention levels and 328 ug/g. Because the pre-§403 floor
dust-lead concentration was 623 u,g/g, levels are reduced to 328 u.g/g. The floor dust-lead
loading is not affected because pre-intervention loading was less than 40 ug/ft2. Similarly, the
Draft - Do Not Cite or Quote
149
September 27. 1996
-------
window sill dust-lead loading is unchanged because the pre-intervention loading was below
100 ug/ft2.
Selection of ranges for the proposed standards was discussed in Section 4.5. The ranges
assessed in the risk characterization for lead in dust and soil and lead-based paint are displayed
in Table 5-6. The anticipated effects of different types of interventions were documented in
Section 5.2.
Table 5-6. Ranges of Standards Considered.
Medium
Dust
Soil
Paint
Range of Standards
Floors: 25-400 //g/fl2
Window Sills: 25-800 j/g/ft2
Cover: 50-1 500 //g/g
Removal : 1 000-5000 //g/g
Maintenance: 0-10 ft2 of damaged LBP
Abatement: 5-100 ft2 of damaged LBP
Step 2: Use IEUBK/EPI models to predict blood-lead concentrations. The second step in
the process is the translation of post-intervention environmental-lead levels and pica tendencies
into blood-lead concentration in children. Post-§403 blood-lead levels are being inferred using
two different blood-lead models applied to projected environmental-lead levels anticipated as a
result of promulgating §403 (Step 1 above).
The first of these models is an epidemiologically-based (EPI) model introduced in
Section 4.1. EPA's IEUBK model, is also used to project blood-lead levels following
promulgation of the §403 ruling. The IEUBK model is being used to predict blood-lead levels
for children aged 24 months, as discussed in Section 4.1. For example, for the case illustrated
under Step 1, the geometric mean blood-lead concentration estimated from pre-§403
environmental-lead levels is 22.9 ug/dL and the post-§403 geometric mean blood-lead level is
predicted to be 5.0 ug/dL based on the IEUBK model.
Both the EPI model and the IEUBK model predict geometric mean blood-lead levels for
the subpopulation of children exposed to the specified environmental conditions. However,
Draft - Do Not Cite or Quote 150 September 27, 1996
-------
these geometric means are not sufficient to characterize the national distribution of children's
blood-lead levels. Not every child exposed to floor dust-lead loading of 100 ug/ft2, soil-lead
concentration of 250 ug/g, and maximum paint lead loading of 1.1 mg/cm2 with 5 ft2 of
damaged LBP will have the same blood-lead concentration. The model-predicted geometric
means need to be supplemented by some measure of variability that reflects differences in
blood-lead levels observed under the same measured environmental-lead levels. This is
accomplished by use of the GSD of 1.6 ug/dL presented in the guidance manual for the IEUBK
model (EPA, 1994a) (see Section 4.3.1). Appendix El describes the approach taken to
characterize the variability in blood-lead levels about the estimated geometric means and how
this information is used to determine a distribution of PbB over a population of children.
These additional steps are required to infer the arithmetic average blood-lead concentration or
the proportion of children with a blood-lead concentration are above a specified concentration.
information. Step 2 in this process estimates the pre- and post-§403 distribution of blood-lead
concentrations from environmental-lead levels. Step 3 determines the change in blood-lead
concentrations resulting from the intervention (post-§403 minus pre-§403), and applies this
change to the distribution of blood-lead concentrations inferred from NHANES ffi. This step is
necessary because the NHANES results are regarded as the most reliable baseline
characterization of children's blood-lead concentration available. The IEUBK and EPI models
applied in Step 2, however, are the best tools available for estimating the change in PbB
associated with an intervention. Thus, there are three inputs to this step in the process:
1. A model-predicted, pre-§403 distribution of PbB
2. A model-predicted, post-§403 distribution of PbB
3. A baseline distribution of PbB from NHANES
In this step, the difference between pre-§403 modeled PbB and post-§403 modeled PbB is
applied to the baseline distribution of PbB inferred from NHANES ffl. The details of this step
are described in Steps (1) through (4) of Appendix E2. The result is an estimate of the
Draft - Do Not Cite or Quote 151 September 27, 1996
-------
geometric mean and the geometric standard deviation of blood-lead levels in the nation
following §403. This information is used to predict health risks to children in the next step.
Step 4: Predict health effects and blood lead endpoints for children 1-2 years old. The
last step in the process is the summarization of health risks associated with the baseline and the
predicted post-§403 distributions of blood-lead concentrations. This step estimates the
proportion of children with blood-lead levels above specified thresholds, the proportion of
children anticipated to experience IQ decrements of specified amounts due to elevated blood
lead concentrations, the proportion of children with IQ levels below 70 due to elevated blood
lead concentrations, and the average and standard deviation of IQ point losses, due to elevated
blood lead concentrations. Each of these endpoints is estimated from the geometric mean and
geometric standard deviation, assuming a lognormal distribution. The mathematical approach
used to make these inferences is described in Step (5) of Appendix E2.
The remainder of this section presents the estimated health risks due to lead-based paint
following the rule-making for various options of the standard. Section 5.3.1 summarizes the
specific endpoints listed in Section 5.0 for various options considered for the §403 standards
(see Section 4.5). Section 5.3.2 displays in more detail the estimated risks associated with a
particular "central" set of standards and environmental conditions observed hi the HUD
National Survey. Please note that Section 5.3.2 does not promote a set of standards as being
the best to implement; it is merely provided to present a more in-depth summary of the
predicted reduction in risks associated with an option for the standard.
5.3.1 Characterization of Risks for Various Sets of Standards
Chapter 3 of this report presented numbers of children aged 1-2 years associated with
each housing unit in the HUD National Survey in 1997. These numbers reflect the estimated
numbers of children who will reside hi housing units with similar dust-, soil-, and paint-lead
levels, and demographic variables in 1997. The HUD National Survey environmental-lead
levels as modified by the proposed §403 Interventions, are being used to predict blood-lead
concentrations for the children assigned to each house. Predicted blood-lead concentrations are
in turn used to characterize the national distribution of blood-lead concentrations as a result of
§403.
Draft - Do Not Cite or Quote 152 September 27, 1996
-------
For each medium, several different standards were considered. The ranges of different
standards for each medium were introduced in Section 4.5 based on an analysis of health
effects. Table 5-6 (above) presents a list of these ranges for the standards. There is a
tremendous number of combinations of standards options that could be evaluated for the
various media being considered. To reduce the complexity and reduce the number of options
to consider, dust options are evaluated first with soil and paint standards fixed at central values
within their specified ranges. Then soil options are considered with dust and paint standards
fixed at central values. Options for paint standards are presented in an analogous fashion.
For example, Table 5-7 presents a range of options for floor and window sill dust
standards. In this table, the soil cover standard is set at 400 ug/g, the soil removal standard is
set at 3000 ng/g, the paint maintenance standard is set at 5 ft2 of damaged LBP, and the paint
abatement standard is set at 20 ft2 of damaged LBP. The options for floor dust-lead loading
standards range from 25 to 400 ug/ft2 (in reverse order), and from 25 to 800 fig/ft2 for window
sills. Each column is devoted to a specific pair of standards for floor and window sill dust-
lead loading. For each set of standards, the top part of Table 5-7 indicates the percentage of
homes that would be affected specifically by each of the floor and window sill dust-lead
loading standards, the percentage of homes that would be affected by either the floor or
window sill dust-lead loading standards and the percentage of homes that would be affected by
any one of the standards for dust, soil, or paint specified in this table.
For example, in the first column of standards in Table 5-7, we see that only 0.30 percent
of houses in the nation would be expected to exceed a floor dust lead loading standard of 400
ug/ft2. 10.6 percent of the nation's homes would be expected to have window sill dust-lead
loading exceeding 800 ug/ft2. 10.9 percent of the nation's homes would be expected to exceed
either of these two standards. Note that in this case, the percentage exceeding the floor dust-
lead standard (0.3) and the percentage exceeding the window sill dust-lead standard (10.6)
added to equal the percentage of homes exceeding either dust-lead standard (10.9).
Draft - Do Not Cite or Quote 153 September 27, 1996
-------
Table 5-7. Characterization of Impact of Various Options for Dust Standards: Soil and
Paint Standards fixed (400 //g/g for Soil Cover, 3000 //g/g for Soil Removal,
5 ft2 damaged LBP for Paint Maintenance, 20 ft2 damaged LBP for Paint
Abatement).
Options for Dust Lead Loading Standard (fig/ft2)
Floors
Window Sills
ercantaga of Homes
Exceeding Floor Dust
Standard
Percentage of Homes
Exceeding Window
SHI Dust Standard
Percentage of Homes
Exceeding Any Dust
Standard
Percentage of Homes
Exceeding Any
Standard
400
800
0.297
10.6
10.9
24.3
200
500
1.98
14.1
15.4
26.1
100
500
8.94
14.1
18.0
28.2
100
200
8.94
26.9
29.0
35.7
50
100
16.7
37.0
43.1
47.1
25
25
33.4
54.6
62.9
64.6
Health Effects Projected by EPI Model
PbB>25//g/dL(%»
PbB>10//g/dL(%)
IQ<70(%)
ia decrement >1 (%)
IQ decrement > 2 (%)
IQdecrement>3(%)
Avg. IQ decrement
SD of IQ decrement
0.27
7.8
0.53
49
15
4.9
1.24
0.94
0.25
7.5
0.53
49
14
4.7
1.23
0.92
0.24
7.3
0.52
49
14
4.6
1.22
0.91
0.22
7.1
0.52
48
14
4.4
1.21
0.90
0.21
7.0
0.52
48
14
4.3
1.20
0.89
0.21
6.9
0.52
48
14
4.3
1.20
0.89
Hearth Effects Projected by IEUBK Model
PbB>25/ig/dL(%)
PbB>10//g/dL(%)
IQ<70(%)
IQ decrement >1 (%>
IQ decrement > 2 (%»
IQ decrement > 3 {%>
Avg. IQ decrement
SD of IQ decrement
0.071
4.8
0.49
46
11
2.7
1.13
0.75
0.031
3.6
0.48
45
8.8
1.9
1.09
0.67
0.026
3.4
0.48
44
8.4
1.8
1.08
0.65
0.023
3.2
0.48
44
8.1
1.7
1.07
0.64
0.015
2.7
0.47
43
7.2
1.4
1.05
0.61
O.011
2.5
0.47
42
6.7
1.2
1.04
0.60
Draft - Do Not Cite or Quote
154
September 27, 1996
-------
However, this is not always the case because there is often overlap between these two sets of
homes. That is, often a house exceeding the sill dust-lead standard will also exceed the floor
dust-lead standard. For example, for the third set of dust-lead standards listed in Table 5-7, the
floor and window sill dust-lead standards are 100 and 500 fig/ft2, respectively. 8.94 percent of
homes are expected to exceed the floor dust-lead standard, and 14.1 percent are expected to
exceed the window sill dust-lead standard. Eighteen percent are expected to exceed either
standard. This means that 23.04 (=8.94 + 14.1) - 18.0 = 5.04 percent are expected to exceed
both floor and window sill standards.
Continuing down the rows of Table 5-7, we see that 24.3 percent of the nation's homes
would be expected to exceed any of the standards considered in this column. These are:
floor dust-lead loading: 400 ug/ft2
window sill dust-lead loading: 800 ug/ft2
soil (cover): 400 ug/g
soil (removal): 3000 ug/g
paint (maintenance): Greater than 5 ft2 of damaged LBP
paint (abatement): Greater than 20 ft2 of damaged LBP
As mentioned above, only 10.9 percent of the homes were projected to exceed either of the
dust-lead standards. This means that 24.3 -10.9 = 13.4 percent of the nation's homes are
projected to exceed 400 ug/ft in soil-lead concentration, or exceed 5 ft2 of damaged LBP but
not exceed either of the dust-lead standards.
Estimates of the selected health effects projected after implementation of these standards,
based on the EPI model are presented in the middle section of Table 5-7. For example, if §403
were formulated with the standards associated with the first column of Table 5-7 (described
above), 0.27 percent of the nation's children would be projected to have PbB above 25 ug/dL
using the EPI model. Approximately 8 percent of children would be projected to have PbB
exceeding 10 ug/dL. Only 0.53 percent of kids would be expected to have IQ scores below 70
due to elevated blood-lead concentration. The next three lines describe the proportion of kids
expected to have IQ decrements of 1 or more, 2 or more, and 3 or more, based on the EPI
model, after intervention on the basis of these standards. The predictions are 49 percent, 15
percent, and 4.9 percent, respectively. The next two lines describe the distribution of IQ
Draft - Do Not Cite or Quote 155 September 27, 1996
-------
decrements associated with elevated PbB. Interventions triggered by the first set of standards
would be projected to result in (arithmetic) average IQ decrement of 1.24 with standard
deviation of 0.94. Note that the distribution of IQ decrements is not symmetric. This was
illustrated in Figure 5-2 for NHANES III. In fact, it is well described by a lognormal
distribution. Therefore, for estimating the proportion of children with IQ decrement of
specified levels, the three previous lines in the table should be used.
The bottom part of Table 5-7 presents the same information but with the projected health
effects determined using the IEUBK model to predict blood-lead concentrations instead of the
EPI model.
5.3.1.1 Varying Dust Standard Options
Table 5-7 examines the impact of various options for the dust-lead standard on
childhood health effects and blood-lead concentration. Options for floor and window sill dust-
lead loadings were varied simultaneously from 400 and 800 ug/ft2 for floor dust-lead loading
and window sill dust-lead loading, respectively, to 25 and 25 ug/ft2, respectively. A total of
six combinations were assessed. The first few rows of Table 5-7 predict that the number of
houses that would be affected by any of the selected standards represented in these tables
ranges from 24 percent to 65 percent. Examining the associated health effect and blood-lead
concentration endpoints reveals that the most dramatic improvement is achieved between the
least stringent sets of standards considered, with diminishing returns evident between
successive reductions in the standards. This is most clearly evident for the percentage of
children with blood-lead concentration exceeding 25 ug/dL or 10 ug/dL, the percentage of
children that will have an IQ decrement of at least 2 or 3, and the average IQ decrement. For
example, based on the EPI model, 7.8 percent of the nation's children would be anticipated to
have PbB exceeding 10 ug/dL if the least stringent set of standards (floor: 400 ug/ft2; window
sill: 800 ug/ft2). The IEUBK model predicts 4.8 percent. These percentages can be compared
with the baseline (current) estimate of 10.5 percent. Under the second option for dust standards
(floor 200 ug/ft2; window sill 500 ug/ft2) without changing the options for the soil and paint
standards, the projections come down to 7.5 percent (EPI) and 3.6 percent (IEUBK). By
reducing both of the dust standards to the lowest considered in this analysis (floor 25 ug/ft2,
Draft - Do Not Cite or Quote 156 September 27, 1996
-------
window sill 25 ug/ft2), the estimates of the proportion of children with blood-lead
concentration exceeding 10 jig/dL are reduced to 6.9 percent (EPI) and 2.5 percent (IEUBK).
Figures 5-3a and 5-3b display the eight health endpoints presented in Table 5-7 in
graphical form. The top left graph in Figure 5-3a presents the projected percentage of children
that would have PbB greater than 25 ug/dL as a result of implementing §403 for each set of
standards considered for floor and window sill dust versus the percentage of homes exceeding
any of these standards. There are two curves in the graph, the top curve reflects predictions
based on the EPI model, and the bottom curve is based on the IEUBK model. Note that on
each curve there are six dots, corresponding to the six sets of standards considered for dust-lead
loadings. A reference line is drawn at the top of the graph to indicate the baseline levels for
this response determined from NHANES III, (0.58 percent of children are currently estimated
to have PbB above 25 ug/dL).
The top right graph in Figure 5-3a presents the analogous information for the projected
percentage of children that would have PbB greater than 10 ng/dL as a result of implementing
§403 with these standards. The lower left graph presents the average IQ decrement resulting
from elevated blood-lead concentration, and the lower right graph presents the standard
deviation of IQ decrements resulting from elevated blood-lead concentration. All of these are
plotted versus the percentage of homes anticipated to exceed any of the standards over the
range of standards considered. Figure 5-3b presents, in the same format, the percentage of
children with IQ below 70, and the percentage of children expected to have IQ decrements of at
least 1,2, and 3 points as a result of elevated blood-lead concentration.
These graphs show the impact of various options for the dust standards on health effects,
children's blood-lead concentrations, and the number of homes impacted by the standards.
Note the generally consistent shape of each of the curves in these figures. A sharp decline in
the curve indicates a large change in the health effects or blood-lead concentrations relative to a
small difference in the number of homes requiring an intervention. A less steep decline
indicates either a large increment in the number of homes requiring an intervention, or
Draft - Do Not Cite or Quote 157 September 27. 1996
-------
I
1
5:
g
i
CD
00
1
M
1
0.60
0.55
0.50
0.45
0.40
S 0.35
in
7 0.30
0.20
0.15
0.10
0.05
0.00
1.5
1.4
1.3
1.2
"c
J ..o
g
o*> 0.9
0.8
0.7
0.6
0.5
**^^_^
Model: 0
e-a o IEUBK ° -o - - - ^ o
i i i '• i1 • p- r 1 1 •
10 20 30 40 SO 60 7Q 80
Percentage of Homes Exceeding Any Standard
a
Modd:
0-3 0 IEUBK
10 20 30 40 SO 60 70 80
Percentage of Homes Exceeding Any Standard
12
10
B
o
6
/
ffi
£
4
2
0
10 decrement
o p — —
b> bo o to
O
<" 0.4
0.2
0.0
•L
6
° o -~ ..
o
Uod«l:
0-3 0 IEUBK
10 20 30 40 50 60 70 80
Percentage of Homes Exceeding Any Standard
.^
o
Model:
•-•-* EPI
O-O O IEUBK
10 20 30 40 50 60 70 80
Percentage of Homes Exceeding Any Standard
Figure 5-3a. Projected Health Endpoints Based on Various Options for Dust Standards, Part 1 ; Soil Cover 400 //g/g,
Removal 3000 //g/g. Paint Maintenance 5 ft2 Paint Abatement 20 ft2. (Dashed reference line represen
hacolino rielr \
-------
s>
01 0.6
1
g> 0.5
§
•5
*0 i?
5 g 0.3
rN v
O g
§
5 0.2
0.1
0.0
20
C71
CD
16
-^ 14
JS
N 12
A
1 10
I
» 8
•o
O
6
4
2
0
Co
*•» •
»-• • •••• • ' e
G
Model-
O-OO JCU8K
1 , , , , , , 1 1
10 20 30 40 50 60 70 80
Percentage of Homes Exceeding Any Standard
•^*— -__
* • ••
0
a
& e* . ,
" * ~ -o- -
o
M0d.,:
•-** EPI
e-o o ICUBK
10 20 30 40 50 60 70 SO
55
50
45
40
^ 35
- 30
c
V
L.
5 20
o
15
10
5
0
8
7
6
^
t*T ^
1 4
i
o *
•o 3
5
2
1
0
•^*—
O.
" ~° - -- -o
Model:
e-o o ICUBK
10 20 30 40 50 60 70 80
Percentage of Homes Exceeding Any Standard
^^-«—
— •
• •
0
'$
*- o..,_
Uodrt:
• • • tPI
0-0 C IEU8K
10 20 30 40 50 60 70 80
Percentage of Homes Exceeding Any Standard
Percentage of Homes Exceeding Any Standard
<0
(0
Figure 5-3b. Projected Health Endpoints Based on Various Options for Dust Standards, Part 2; Soil Cover 400 //g/g. Soil
Removal 3000 //g/g, Paint Maintenance 5 ft2. Paint Abatement 20 ft2. (Dashed reference line represents
baseline risk.)
-------
a small net health benefit. In each case, the steepest drop occurs between the two least-
stringent sets of standards (floor: 400 ng/ft2; window sill: 800 ug/ft2) and (floor: 200 ug/ft2;
window sill: 500 ug/ft2), and then gradually levels off as the standards affect greater and greater
numbers of homes. This pattern is consistent between the EPI and IEUBK models, and across
health effects, with some endpoints reflecting the pattern more drastically than others.
This suggests that with regard to changes in dust-lead standards, there does not appear to
be much benefit to enforcing dust-lead loading standards more stringent than perhaps the
second or third most stringent standards considered in this analysis. Although the number of
additional houses affected by the more stringent standards is very large, the incremental gains
in health benefits are small.
The projected health effects as a result of implementing §403 with the various standards
can be compared to the baseline (current estimated) health effects and blood-lead
concentrations using the reference line in each graph. For example, each of the sets of
standards considered for dust would result in a substantial improvement relative to the baseline
for the percentage of children exceeding 25 ug/dL and 10 ug/dL, and the percentage of children
anticipated to have an IQ decrement of at least 2 or 3 resulting from elevated blood-lead
concentration. The improvement from baseline is much less for IQ < 70 and IQ decrements of
at least 1. Note that the graph of average IQ decrement has a lower axis bound of 0.5.
There is little reduction in the percentage of children predicted to have IQ below 70 or in
the percentage of children expected to have IQ decrements greater than 1 over the range of
standards considered.
5.3.1.2 Varying Soil Standard Potions
Table 5-8 presents results for a range of options for the §403 soil cover and soil removal
standards with the floor dust-lead loading standard set at 100 ug/ft2, the window sill dust-lead
loading set at 500 ug/ft2, the paint maintenance standard set at 5 ft2 of damaged LBP, and the
paint abatement standard set at 100 ft2 of damaged LBP. The options for requiring soil
covering range from 50 to 1500 ug/g and the options for requiring soil abatement range from
1000 to 5000 ug/g. For each of these options, the top portion of Table 5-8 indicates the
percentage of homes that would be affected specifically by the soil covering standard, the
Draft - Do Not Cite or Quote 160 September 27, 1996
-------
percentage that would be affected by the soil abatement standard, and the percentage of homes
that would be affected by any one of the standards for dust, soil, or paint specified in the table.
The remaining rows of Table 5-8 are analogous to those displayed in Table 5-7. Projected
health effects are predicted first based on the EPI model, then based on the IEUBK model.
Table 5-8 predicts that the number of houses that would be affected by any of the selected
standards represented in these tables ranges from 25 percent to 52 percent. The least stringent
standards considered were 1500 fig/g for soil cover and 5000 (ig/g for soil removal. The most
stringent standards considered were 50 ug/g and 1000 ug/g for soil cover and soil removal,
respectively. Over this range of standards, the projected post-§403 proportion of children with
blood-lead concentration exceeding 25 ug/dL ranges from 0.27 percent to 0.16 percent based
on the EPI model and from 0.074 percent to 0.005 based on the IEUBK model. The
corresponding projected proportions for blood-lead concentration exceeding 10 ug/dL range
from 7.8 to 6.2 percent based on the EPI and 4.9 to 1.8 percent based on the IEUBK. Thus,
although for each set of standards the IEUBK projects lower incidence of elevated blood-lead
concentrations, both models project substantial reductions in this incidence over the range of
standards.
The proportion of children projected to have IQ scores below 70 due to elevated blood-
lead concentration only ranges from 0.53 percent to 0.51 percent based on the EPI model, and
from 0.50 percent to 0.46 percent based on the IEUBK model. Thus, little benefit is
anticipated from implementation of these standards for this health effect. However, of the three
sets of standards considered (dust, soil, and paint), varying soil standards has the greatest
potential impact on this endpoint.
For IQ decrements, the EPI model projects reductions in the proportions of children with
the greatest relative reductions seen for the larger decrements (IQ decrement >2,3). The
IEUBK predicts greater relative reductions for each of the three thresholds (>1, >2, >3).
Draft - Do Not Cite or Quote 161 September 27, 1996
-------
Table 5-8. Characterization of Impact of Various Options for Soil Standards: Dust and
Paint Standards fixed (100 jig/ft2 for Dust Lead Loading, 500 //g/ft2 for
Window Sill Dust Lead Loading, 5 ft2 damaged LBP for Paint
Maintenance, 20 ft2 damaged LBP for Paint Abatement).
Soil Cover
. Soil Removal
Percentage of Homes Exceeding
Soil Cover Standard
Percentage of Homes Exceeding
Soil Removal Standard
Percentage of Homes Exceeding
Any Standard
Options for Soil Lead Concentration Standard 0/g/g)
1500
5000
3.27
0.215
25.2
800
4000
8.11
0.746
25.9
400
3000
12.8
0.746
28.2
200
3000
20.1
0.746
30.9
. 100
2000
27.1
2.71
36.6
50
1000
42.8
6.14
52.1
: : Health Effects Projected by EPi model ;; . :
PbB>25/ig/dL(%>
PbB>.10/fg/dL(%)
IQ<70(%)
IQ decrement >1 (%>
IQ decrement>2 (%) .
IQ decrement > 3 (%> .
• Avg. IQ decrement ... -
SD of IQ decrement
0.27
7.8
0.53
49
15
4.9
1.24
0.94
0.25
7.5
0.53
49
14
4.7
1.23
0.92
0.24
7.3
O.52
49
14
4.6
1.22
0.91
0.22
7.1
0.52
48
14
4.5
1.21
0.9
0.19
6.7
0.52
47
13
4.1
1.19
0.88
0.16
6.2
0.51
46
12
3.8
1.16
0.85
Health Effects Projected by IEUBK model • .
.PbB>2Bjfg/dL(%>
PbB>10//g/dL(%)
IQ<70<%)
IQ decrement > 1 (%)
IQ decrement > 2 (%)
IQ decrement>3 (%)
Avg. IQ decrement!- •--• *
SD of IQ decrement - . :
0.074
4.9
0.50
47
11
2.8
1.14
0.75
0.034
3.7
0.48
45
9.0
2.0
1.10
0.68
0.026
3.4
0.48
44
8.4
1.8
1.08
0.65
0.020
3.1
0.47
44
7.8
1.6
1.06
0.63
0.010
2.3
0.47
41
6.5
1.1
1.03
0.59
0.0050
1.8
0.46
39
5.2
0.81
0.99
0.55
Draft - Do Not Cite or Quote
162
September 27. 1996
-------
Figures 5-4a and 5-4b display the eight health endpoints presented in Table 5-8 in a
graphical form similar in format to Figures 5-3a and 5-3b. There are two curves in each graph
representing the predictions from the EPI and IEUBK models, and the baseline level of each
endpoint (as determined from NHANES) is drawn as a reference line for comparison on each
graph.
As with the predicted health effects associated with changes in dust-lead standards, the
health effects and blood-lead concentrations predicted as a result of various options for soil
standards indicate that the greatest improvement in health effects is achieved between the two
least stringent sets of standards considered. The first and least stringent set of soil standards
considered is (soil cover: 1500 |ig/g; soil removal: 5000 ng/g) and the second set was (soil
cover: 800 ng/g; soil removal: 4000 ng/g). There are reduced incremental benefits achieved
for more stringent soil standards, but there are still gains to be made between successive
reductions in the standards up to about the fifth set of standards (soil cover: 100 ng/g; soil
removal: 2000 |ig/g).
Again, note the generally consistent shape of each of the curves in these figures. In each
case, the steepest drop occurs between the two least-stringent sets of standards, the next four
points follow almost a straight line, and then the most stringent set of standards results in only
a small incremental health benefit. This pattern is consistent between the EPI and IEUBK
models and across responses, with some endpoints reflecting the pattern more drastically than
others. The endpoints most sensitive to changes in soil standards based on the EPI and IEUBK
models are the proportion of children projected to have blood-lead concentrations exceeding 10
and 25 jag/dL and the proportion of children expected to have IQ decrements at least 2 or 3.
The health effect least sensitive to changes in soil standards is the projected proportion of
children with IQ less than 70 due to elevated blood-lead concentration.
Draft - Do Not Cite or Quote 163 September 27, 1996
-------
a> o.6o
1* 0.55
CD 0.50
o
> 0.45
5"
** 0.40
O „
5' E °35
_» tn
^ ™ 0.30
r? 2 0.25
C a.
O
5 0.20
0.15
0.10
0.05
0.00
1.5
2
•i"
1.3
1.2
c
I '•'
u
« 1.0
g
a, 0.9
>
<
O.B
0.7
0.6
0.5
V
"••^
~».^
^•____^^
Model: 0.4
0.2
0.0
**-»-
~*~\^^^
*
o
0
"o
" tl
3 """"•----.
Hodtl:
•-•-• EPI
e-3 0 IEU8K
10 20 30 40 50 60 70 80
Percentage of Homes Exceeding Any Standard
v^_^^
* .
0
" O . _
~~ " - ~o
Model:
0-0 0 IEUBK
* I
*
20 30 40 50 60 70
Percentage of Homes Exceeding Any Standard
80
10 20 30 40 50 60 70
Percentage of Homes Exceeding Any Standard
M
-^ Figure 5-4a. Projected Health Endpoints Based on Various Options for Soil Standards, Part 1; Floor Dust 100 //g/ft2,
2 Window Sill Dust 500 //g/ft2. Paint Maintenance 5 ft2. Paint Abatement 20 ft2. (Dashed reference line
01 represents baseline risk.)
-------
Cu 0.6
1
g> 0.5
^
s
*0 fc?
5 g 0.3
0
3 "
0.1
0.0
*-•-• — . «
^
Nod«i'
•-•-• CPI
0-0 0 ICUBK
55
50
45
40
2 35
- 30
c
! »
u
5 20
g
15
10
5
0
*^.
V"0--J "
' 0- . . _
~ ~ ~ - o
Uodil:
•-•-• CPI
0-3 0 ICU8K
10 20 30 40 SO 60 70 80 10 20 30 40 SO 60 70 80
Percentage of Homes Exceeding Any Standard Percentage of Homes Exceeding Any Standard
20
o> 18
Ol
16
s? '*
*"^
N 12
A
C
£
• B
o
~ 6
4
2
0
v^
^*^*"^*-—-^
•
o
6.
'"a.
"vo-
" " * -o
Uodtl:
•-.»-• £p|
O-O O ICUBK
8
7
6
t?
5
10
A
I *
•a 3
2
2
1
0
*-•
-_^
°
a.
°"tt-..
UwJ«l: ^"""""••----3
•-*>-»» CPI
€>•& C ICUBk
C/) 10 20 30 40 SO 60 70 80 10 20 30 40 50 60 70 80
^ Percentage of Homes Exceeding Any Standard Percentage of Homes Exceeding Any Standard
|
^ Figure 5-4b. Projected Health Endpoints Based on Various Options for Soil Standards, Part 2; Floor Dust 100//g/ft2
jo Window Sill Dust 500 //g/ft2, Paint Maintenance 5 ft2. Paint Abatement 20 ft2. (Dashed reference line
°> renresents baseline risk.)
-------
The projected health effect and blood-lead concentration endpoints as a result of
implementing §403 with the various standards can be compared to the baseline (current
estimated) effects using the reference line in each graph. Each of the sets of standards
considered for soil would result in a substantial improvement relative to the baseline for the
percentage of children exceeding 25 ug/dL and 10 ug/dL, and the percentage of children
anticipated to have an IQ decrement of at least 2 or 3 resulting from elevated blood-lead
concentration.
There is little reduction in the percentage of children predicted to have IQ below 70 or hi
the percentage of children expected to have IQ decrement greater than 1 over the range of
standards considered.
There is a clear benefit projected for even the least stringent soil-lead standards, and there
is some additional benefit predicted for the more stringent standards. There are gains to be
made in health benefits for more stringent standards as low as (soil cover: 100 ug/g; soil
removal: 2000 ug/g). The most stringent set of standards (soil cover: SO ug/g; soil removal:
1000 ^g/g) affects a large number of houses with little incremental gains hi health effects.
5.3.1.3 Varying Paint Standard Options
Table 5-9 presents results for a range of options for paint intervention standards with the
floor dust-lead loading standard set at 100 ug/ft2, the window sill dust-lead loading standard set
at 500 fig/ft2, the soil covering standard set at 400 ug/g, and the soil abatement standard set at
3000 |ig/g. The options for requiring a paint maintenance range from O.to 10 ft2 of damaged
LBP, and options for requiring a paint abatement range from 5 to 100 ft2 of damaged LBP. For
each of these options, the top portion of Table 5-9 indicates the percentage of homes that would
be affected specifically by the standard for either interior or exterior paint maintenance
standard, the percentage that would be affected specifically by either interior or exterior paint
abatement standard, and the percentage of homes that would be affected by any one of the
standards for dust, soil, or paint specified in the table. The remaining rows of Table 5-9 are
analogous to those displayed in Tables 5-7 and 5-8. Results are first presented based on the
EPI model, then based on the ffiUBK model.
Draft - Do Not Cite or Quote 166 September 27, 1996
-------
Table 5-9 predicts that the number of houses that would be affected by any of the selected
standards represented in these tables ranges from 27 percent to 29 percent. The standards for
paint intervention are defined in terms of a specified amount of damaged LBP. The least
stringent standards considered were 10 square feet and 100 square feet, respectively. The most
stringent standards considered for paint maintenance and paint abatement were zero square feet
of damaged LBP and five square feet of damaged LBP, respectively.
Figures 5-Sa and S-Sb display the eight health endpoints presented in Table 5-9 in
graphical form in the same format as Figures 5-3a and 5-3b. There are two curves in each
graph representing the predictions from the EPI and IEUBK models. The baseline level of each
endpoint (as determined from NHANES HI) is drawn as a reference line for comparison on
each graph.
These figures highlight an important fact: The amount of damaged LBP is not a very
useful discriminant for determining whether an intervention should be performed. The range of
the percentage of houses affected by different thresholds of damaged LBP is not very wide. In
fact, 24 percent of the nation's houses have dust- or soil-lead levels exceeding the standards for
these media regardless of the amount of damaged LBP present in the houses. Imposing the
standards considered for paint increases this number by only 3 to 5 percent. However, the tools
available for assessing the impact of damaged lead-based paint are limited. Both the EPI and
IEUBK models for predicting blood-lead concentrations based on environmental-lead levels
are limited in their usage of paint-lead measurements. Paint-lead is incorporated into the
IEUBK model by considering paint ingested due to pica as discussed in Section 4.3. The EPI
model is based on the Rochester study data which does not have damaged LBP variables
similar to the HUD National Survey damaged LBP variables. Pica for paint also plays a role in
this model. Our estimate of the prevalence of pica for paint may not be accurate.
Draft - Do Not Cite or Quote 167 September 27, 1996
-------
o
S> O.BO
1 0.55
O 0.50
§ 0.45
0 0.40
*** "^
Cft i? 0.35
O *>
*« " 0.30
| z -
C? 0.20
0.15
0.10
0.05
0.00
1.5
^
O) '.4
09
1.3
1.2
"c
I ' '
O
"D
—
a. 0.9
0.8
0.7
0.6
0.5
1
3
3
1
Nl
S|
.,
Model:
•>*-* CPi
0-0 0 IEUBK 0
-------
Q> 0.6
1
g> 0.5
§
"* 0.4
g
* (7
§ £ 0.3
V
O 0
§
5? 0-2
0.1
0.0
20
o> '8
(0
16
,-, M
A
1 10
E
1 8
S 6
4
2
0
OOT
1 Modal1
Mod...
+ + + EP1
6-3 0 IEUW
10 20 30 40 50 60 70 80
Percentage of Homes Exceeding An/ Standard
-
am
Modal:
• e • CPI
e-o o ICUBK
55
50
45
40
C7"
A
- 30
c
V
f
I 25
o
T> 20
g
15
10
5
0
8
7
6
A
1 4
S
1 3
o
2
1
0
^
..
Mod.l:
«-e-e CPI
e-3 0 ICUBK
10 20 30 40 50 60 70 BO
Percentage of Homes Exceeding Any Standard
•«.
o an>
Modal:
•-»-• CPI
0-9 0 ICUBK
&
10 20 30 40 SO 60 70
Percentoge af Homes Exceeding Any Standard
80
10 20 30 40 50 60 70
Percentoge of Homes Exceeding Any Standard
N)
£
O)
Figure 5-5b. Projected Health Endpoints Based on Various Options for Paint Standards, Part 2; Floor Dust 100 //g/ft2.
Window Sill Dust 500 //g/ft2. Soil Cover 400 //g/g. Soil Removal 3000 //g/g. (Dashed reference line represents
baseline risk.)
-------
Table 5-9. Characterization of Impact of Various Options for Paint Standards: Dust and
Soil Standards fixed (100 //g/ft2 for Dust Lead Loading, 500 //g/ft2 for
Window Sill Dust Lead Loading, 400 //g/g for Soil Covering, 3000 //g/g for
Soil Removal).
Options for Paint Standard (ft1 damaged LBP)
Paint Maintenance '
Paint Abatement
Percentage of Homes
Exceeding Interior Paint
Maintenance Standard
Exceeding Exterior Paint
Maintenance Standard
Percentage of Homes
Exceeding Interior Paint
Abatement Standard
Percentage of Homes
Exceeding Exterior Paint
Abatement Standard
Percentage of Homes
Exceeding Any Standard
10
100
2.80
3.84
0.453
3.03
26.8
5
40
4.37
4.80
0.980
4.46
28.2
2
20
3.03
4.20
2.43
5.77
28.6
1
10
2.75
3.22
3.25
6.87
28.9
0
5
1.08
1.15
5.35
9.26
29.2
Health Effects Projected by En model
PbB>25(%)
PbB>10(%)
IQ<70(%)
IQ decrement >1 (%>
IQ decrement > 2 (%>
IQ decrement > 3 (%)
Avg. IQ decrement
SD of IQ decrement
0.25
7.4
0.52
49
14
4.7
1.22
0.92
0.24
7.3
0.52
49
14
4.6
1.22
0.91
0.24
7.3
0.52
49
14
4.6
1.22
0.91
0.24
7.3
0.52
48
14
4.6
1.22
0.91
0.24
7.3
0.52
48
14
4.6
1.22
0.91
Health Effects Projected by IEUBK model
PbB>25(%)
PbB>10(%)
IQ<70(%)
IQ decrement >1 (%)
IQ decrement > 2 {%)
IQ decrement > 3 (%)
Avg. IQ decrement
SD of IQ decrement
0.027
3.4
0.48
45
8.5
1.8
1.08
0.66
0.026
3.4
0.48
44
8.4
1.8
1.08
0.65
0.026
3.4
0.48
44
8.3
1.7
1.08
0.65
0.026
3.3
0.48
44
8.3
1.7
1.08
0.65
0.025
3.3
0.48
44
8.3
1.7
1.08
0.65
Draft - Do Not Cite or Quote
170
September 27, 1996
-------
5.3.1.4 Varying AH Standard Options
Analyses summarized in Tables 5-7, 5-8, and 5-9 permit an assessment of the impact on
the nation's housing and health effects of children for various standard options for each
individual environmental medium. However, those results do not show the effect of varying
the levels simultaneously for dust, soil, and paint. Table 5-10 presents the results of analyses
when the standards for all media are varied. The table is structured similarly to Tables 5-7, 5-8,
and 5-9. Each column represents a unique combination of standards displayed at the very top
in the shaded rows. For instance, the second column represents an option for the standards of
400 fig/ft2 for floor dust-lead loading, 500 ug/ft2 for window sill dust-lead loading, 800 |ig/g
for soil cover, 4000 ug/g for soil removal, 10 ft2 of damaged lead-based paint for paint
maintenance, and 40 ft2 for paint abatement. Below these rows, the table displays the estimated
number of homes affected by each standard. The first row in the middle portion provides the
estimated number of homes above the option for the floor dust standard. Analogous
information are provided in the next seven rows for window sill dust, soil cover, soil removal,
interior and exterior paint maintenance, and interior and exterior paint abatement. Finally, the
estimated number of housing units that would be affected by any one of the standards is
presented.
The bottom portion of Table 5-10 provides the estimated health and blood-lead
concentration effects, first based on the EPI model, then based on the IEUBK model, for each
selected health endpoint in the post-§403 environment. For instance, health endpoints are
shown in the second column for the combination of standards defined in the second column in
the top half of the table. Presentation of the health effect results is in the same format as that
employed in Tables 5-7 through 5-9.
A total of seven complete options for the standards were assessed. The least stringent
option, given in the first column, is (floor: 400 fig/ft2; window sill: 800 ug/ft2; soil cover:
1500 |ig/g; soil removal: 5000 ug/g; paint maintenance: 10 ft2 damaged LBP; paint abatement:
100 ft2 damaged LBP). The most stringent option, given in column 7, is (floor: 25 ug/ft2;
window sill: 25 ug/ft2; soil cover: 50 ug/g; soil removal: 1000 ug/g; paint maintenance: 0 ft2
damaged LBP; paint abatement: 5 ft2 damaged LBP). hi addition, an option corresponding to
the interim standards presented in the interim rule (floor: 100 ug/ft2; window sill: 500 ug/ft2;
Draft - Do Not Cite or Quote 171 September 27, 1996
-------
soil cover: 400 ng/g; soil removal: 5000 ug/g; paint maintenance: 2 ft2 damaged LBP; paint
abatement: 10 ft2 damaged LBP), is assessed in the last column. Table 5-10 illustrates, in a
rough sense, the costs and benefits that would be realized as a result of implementing §403
with various sets of standards. The cost is measured as the number of housing units affected
and the benefits are expressed as probabilities of observing various health effects in children
residing in these housing units. The number of units affected and the proportion of children
with health effects were estimated using the methods presented in Chapter 3.
Table 5-10 predicts that the number of houses that would be affected by the sets of
standards selected ranges from 19 percent to 74 percent. This is a wider range than was
observed for any of the individual media. This is because the options considered in these tables
represent the broadest range of standards considered in this risk assessment.
Over this range of standards, the proportion of children expected to have PbB exceeding
25 ng/dL ranged from 0.32 to 0.14 percent based on the EPI model and 0.16 to 0.002 percent
based on the IEUBK model. The proportion of children expected to exceed 10 ug/dL ranged
from 8.3 to 5.8 percent for the EPI model and from 6.5 to 3.7 percent based on the IEUBK
model. The proportion of kids expected to have IQ below70 only ranged from 0.54 to 0.50
percent based on the EPI model and 6.5 to 11 percent based on the IEUBK model.
Figures 5-5a and 5-5b display the eight health endpoints presented in Table 5-10 in
graphical form in the same format as Figures 5-3a and 5-3b. There are two curves in each
graph representing the predictions from the EPI and IEUBK models, and the baseline level of
each endpoint (as determined from NHANES III) is drawn as a reference line for comparison
on each graph. On each curve, a diamond is overlaid to represent the health effects that would
be projected if the interim standards were used for §403.
The incremental improvement in aggregate health effects per house affected can be
judged by the slope of the curve between two points on each of the graphs. The slope is
steepest on the left side of each of these graphs, between the first, second, and third sets of
standards. This property was generally present in the graphs illustrating the effects of changes
in standards for the individual media. However, allowing each of the standards to vary from
the greatest option considered to the lowest, these graphs illustrate that greater benefits are
achievable than those reflected in the graphs for the individual media.
Draft - Do Not Cite or Quote 172 September 27, 1996
-------
Table 5-10. Characterization of Impact of Various Sets of Dust, Soil, and Paint
Standards.
STANDARDS
Floor Dust Lead Loading
O/g/ffl
Window Sill Dust Lead
Loading (fig/ft1)
Soil Cover (figlg)
Soil Removal (pg/g)
Paint Maintenance
(ft2 damaged LBP)
Paint Abatement
(ft2 damaged LBP)
400
800
1500
5000
10
100
400
500
800
4000
10
40
200
500
400
3000
6
20
100
200
200
3000
2
10
50
100
100
2000
1
10
25
25
50
1000
0
5
Current
Interim
Guidance
100
500
400
5000
2
10
PERCENTAGE OF HOMES EXCEEDING STANDARDS
Floor Dust
Window SHI Dust
Soil Cover
Soil Removal
Interior Paint Maintenance
.
Exterior Paint Maintenance
Exterior Paint Abatement
Percentage of Homes
Exceeding Any Standard
0.297
10.6
3.27
0.215
2.80
3.84
0.453
3.03
18.6
0.297
14.1
8.11
0.746
2.27
2.41
0.980
4.46
22.0
1.98
14.1
12.8
0.746
2.92
3.49
2.43
5.77
26.1
8.94
26.9
20.1
0.746
2.22
3.09
3.25
6.87
38.3
16.7
37.0
27.1
2.71
2.75
3.22
3.25
6.87
52.9
33.4
54.6
42.8
6.14
1.08
1.15
5.35
9.26
74.2
8.94
14.1
13.4
0.215
2.22
3.09
3.25
6.87
28.6
HEALTH EFFECTS PROJECTED BY EPI MODEL
PbB>25 (%)
PbB>10(%)
IQ<70(%)
IQ decrement >1 (%)
IQ decrement > 2 (%)
IQ decrement > 3 <%)
Avg. IQ decrement
SD of IQ decrement
0.32
8.3
0.54
50
15
5.3
1.26
0.97
0.27
7.8
0.53
49
15
4.9
1.24
0.94
0.25
7.5
0.53
49
14
4.7
1.23
0.92
0.20
6.9
O.52
48
13
4.3
1.20
0.89
0.17
6.3
0.51
47
13
3.9
1.17
0.85
0.14
5.8
0.50
45
12
3.5
1.15
0.83
0.24
7.4
0.52
49
14
4.7
1.22
0.92
HEALTH EFFECTS PROJECTED BY IEUBK MODEL
PbB>25(%)
PbB>10(%)
IQ<70(%>
IQ decrament> 1 (%)
IQ decrement 2 (%)
IQdecrement>3(%)
Avg. IQ decrement
SD of IQ decrement
0.16
6.5
0.51
49
13
4.0
1.20
0.85
0.081
5.1
0.50
47
11
2.9
1.15
0.76
0.031
3.6
0.48
45
8.8
1.9
1.09
0.67
0.018
2.9
0.47
43
7.5
1.5
1.06
0.63
0.0052
1.8
O.46
40
5.4
0.83
1.00
0.55
0.0017
1.1
0.45
36
3.9
0.48
0.95
0.50
0.035
3.7
0.48
45
9.0
2.0
1.10
0.68
Draft - Do Not Cite or Quote
173
September 27, 1996
-------
5
0
-O O IEUBK
0 "- c INTERIM ICUBK
e
-o
10 20 30 40 50 60 70
Percentage of Homes Exceeding Any Standard
80
12
Modtl:
•-•-• EPI
0-acicuBK
• ••INTERIM EPI
000 INTERIM ICUBK
tO 20 30 40 50 60 70
Percentage of Homes Exceeding Any Standard
80
r
Ki
1.5
1.4
1.3
1.2
g
a, 0.9
0.8
0.7
0.6
0.5
Model:
•-•-• tPI
e-0 o ICUBK
• •• INTERIM CPI
000 INTERIM ICUBK
20 30 40 50 60 70
Percentage of Homes Exceeding Any Standard
80
1.2
1.0
decrement
P P
at ba
0.4
0.2
0.0
'"-*-..
o-o « IEUBK
• ••iNICRIMEPI
000 INTERIM IEUBK
1 - 1
20 30 40 50 60 70
Percentoge at Homes Exceeding Any Standard
80
-------
£
Qt 0.6
1
Q 0.5
^
§
K
o £
C 2
» 0.2
0.1
0.0
20
«J
(1) 18
16
-* 14
W « 2
A
• 10
1 8
O
E
2
00 0
3
0-*—^___
Hod«l:
•-*-* £PI
0-30 ICUBK
• t+INTCRIUCPI
000 iNfCRIU IEU8K
10 20 30 40 50 60 70 80
Percentage of Homes Exceeding Any Standard
^"^^-^^_
" * — ~~— -•
3
" ' - ' • c.
0-9 c IEUBK
• •• INTERIM EPI
> « « INTERIM IEUBK
10 20 30 40 50 60 70 80
55
50
45
40
t 30
f 25
•S 20
0
15
10
5
0
8
t?
^ 5
"c
« 4
•
u
« T
•o 3
g
2
'
?rr^>— — — __ ^__
' '3 -•:-.. •
" "° - - .
* ~ '- - c- - . .
' " • -o
Model:
•-•-• EPI
• ••INTERIM EPI
0 0 « INTERIM ICUBK
10 20 30 40 50 60 70 80
Percentage of Homes Exceeding Any Standard
.
•^^^
' ^ ^_
'=!
'o. .« _
Modal: " " "0 - .^
e-e o IEUBK " "°
• ••INTERIM EPI "• c
o 0 ° INTERIM IEUBK
10 20 30 40 50 60 70 80
Percentage of Homes Exceeding Any Standard
Percentage of Homes Exceeding Any Standard
(o Figure 5-6b. Projected Health Endpoints Based on Various Sets of Options for Dust, Soil, and Paint, Part 2. (Dashed
°> reference line represents baseline risk.)
-------
There is, again, a generally consistent shape of each of the curves in these figures. In
each case, the steepest drop occurs between the three least-stringent sets of standards, the next
four points follow almost a straight line, and then a gradual reduction in the slope of each
curve. This pattern is consistent between the EPI and IEUBK models. In each case, the
projected health endpoints associated with the interim standards (black and white diamonds in
the graphs) appear just to the right of the point representing the third set of standards, affecting
about 29 percent of the homes. The estimated health and blood-lead concentration effects for
the interim standards are always slightly above the line connecting the third and fourth set of
options considered. This implies that slightly larger benefits might be achieved without
affecting more houses if standards slightly different from the interim guidelines were
employed.
Similarly to the results for the individual media standards, each of the sets of standards
considered would result in a substantial improvement relative to the baseline for the percentage
of children exceeding 25 ug/dL and 10 ug/dL, and the percentage of children anticipated to
have an IQ decrement of at least 2 or 3 resulting from elevated blood-lead concentration. Even
by varying all standards, there is little reduction in the percentage of children predicted to have
IQ below 70 due to elevated blood-lead concentration or in the percentage of children expected
to have IQ decrement greater than 1 due to elevated blood-lead concentration over the range of
standards considered.
In general, the IEUBK model predicts a larger impact of Section 403 on health effects and
children's blood-lead concentrations than the EPI model. This is evident, first, in that it
predicts a greater reduction from the baseline. For instance, for incidence of PbB>25 ug/dL,
the EPI model predicts 0.32 percent for the first set of standards, whereas the IEUBK model
predicts 0.16. Both of these are compared to a baseline prediction of 0.58 percent. Second, the
IEUBK generally predicts a greater reduction in health effects over the range of standards. For
example, for the percentage of children exceeding 25 ug/dL, the drop was
56 percent = —• :— over the range of standards for the EPI model, and the drop was
( 0.32 )
Draft - Do Not Cite or Quote 176 September 27, 1996
-------
99 percent = — ' over the range of standards for the IEUBK model. This pattern
F ( 0.16 J 5 V
persists across all ranges of standards considered in this risk assessment.
5.3.2 Detailed Characterization for a Particular Set of Standards
This section provides a more detailed characterization of projected blood-lead
concentrations and health effects associated with a particular option for the §403 standards for
dust, soil, and paint. These particular levels of the standards were chosen as "central" values,
not to promote a single set of standards as optimal. Those standards are 200 ug/ft2 for dust-
lead loading, 500 ug/ft2 for window sill dust lead loading, 400 ug/g for soil cover, 3000 ug/g
for soil removal, 5 ft2 damaged LBP for paint repair, and 20 ft2 damaged LBP for paint
abatement.
Figure 5-7 displays the projected post-§403 distribution of blood-lead levels based on the
EPI model and the IEUBK model in both histogram and cumulative distribution function (cdf)
format. The former allows the reader to clearly understand the general shape of the
distribution. The cdf displays the probability that a child has blood-lead level below a specified
value. This also enables the reader to infer the proportion of children projected to have blood-
lead concentrations within a particular interval.
Qualitatively, the curve associated with the lEUBK-predicted, post-intervention blood-
lead levels appears slightly to the left of the corresponding EPI curve, which is slightly to the
left of the baseline curve determined from NHANES III. This means that the decrease in
blood-lead concentrations predicted by the IEUBK model is greater than the decrease predicted
by the EPI model.
Table 5-11 compares blood-lead concentrations and health effects as reported in
NHANES in and those projected post-§403 based on both the EPI model and the IEUBK
model for the third set of standards assessed in Section 5.3.2: 200 ug/ft2 for dust-lead loading,
500 ug/ft2 for window sill dust-lead loading, 400 ug/g for soil cover, 3000 ug/g for soil
removal, 5 ft2 damaged LBP for paint repair, and 20 ft2 damaged LBP for paint abatement. The
top half of the table characterizes the distribution of children's blood-lead concentrations.
Draft - Do Not Cite or Quote 177 September 27, 1996
-------
10
Blood-Uad Dlttribullon
Nhdn.t III P«.-lnl.iv.nl,un
• HUD/tEUBK Poil-lnl.rv.nliot
HUO/EW Poil-lnl.rv«nl.oo
Blood-Lead Concentration
5 60
0-
1
E 40
9 JO 31 32
Blood-Lead Concentration
Figure 5-7. Projected Post-5403 Blood-Lead Concentration Distributions Based on EPI and
IEUBK Models at Standards of Floor Dust-Lead -200 //g/ft2; Window Sill Dust-
Lead - 500 //g/ft2; Soil Cover - 400 //g/g; Soil Removal - 3000 /t/g/g; Paint
Maintenance - 5 ft2; Damaged LBP, and Paint Abatement - 20 ft2 Damaged
LBP
Draft - Do Not Cite or Quote
178
September 27, 1996
-------
Contained are the estimated numbers and proportions of children with blood-lead concentration
in various intervals. The bottom half of the table displays the various health endpoints
discussed thus far in this risk assessment for the baseline and post-§403 projections based on
this set of standards.
Table 5-11. Comparison of Blood-Lead Concentrations Before and After §403.
PbB (pg/dL)
Total
10,1)
11.3)
I3.5)
[5,10)
[10.15)
(15,20)
120,25)
*25
NHANES III
# Children*
7.961,000
209,000
2,490,000
2.200,000
2,227,000
559,000
169,000
60,000
46.000
Inferred Hoalth Effects
IQ < 70
Id decrement > 1
IQ decrement > 2
IQ decrement > 3
Average IQ decrement
Houses Affected
45,000
4,153,000
1,451,000
564,000
P6fC6nt
100
2.6
31
28
28
7.0
2.1
0.8
0.6
EPI Model1
9 Children
7,961,000
183,000
2,689,000
2,363,000
2,129,000
434.000
109.000
33,000
20,000
0.6
52
18
7.1
1.35
1.11
# Houses
0
Danwavw*
rBrcenr
0
42,000
3,879,000
1,140,000
377,000
tYna-f. *»«-»•
r*8iC8nT
100
2.3
34
30
27
5.5
1.4
0.4
0.3
0.5
49
14
4.7
1.23
0.92
# Houses
26,210,000
•%—_———*
I*6rC6flT
26
IEUBK Model1
# Children
7.961,000
90,000
2,850,000
2,760,000
1,974,000
240,000
37,000
7,000
2.000
Porcont
100
1.1
36
35
25
3.0
0.5
0.1
0.03
38.313
3,577,914
698,810
1 52,099
0.5
45
8.8
1.9
1.09
0.67
# Houses
26.210.000
DBHVBM*
rerconr
26
'Predicted distribution of blood-lead concentration following the rule-making for standards of 200 ug/ft2
for dust lead loading, 500 ug/ft2 for window sill dust-lead loading, 400 ug/g for soil cover, 3000 ug/g for soil
removal, 5 ft2 damaged LBP for paint repair, and 20 ft2 damaged LBP for paint abatement.
Predicted distribution of blood-lead concentration and health effects following the rule-making for
standards of 200 ug/ft2 for dust lead loading, 500 ug/ft2 for window sill dust-lead loading, 400 ug/g for soil cover,
3000 ug/g for soil removal, 5 ft2 damaged LBP for paint repair, and 20 ft2 damaged LBP for paint abatement.
3Numbers of children in thousands
Draft - Do Not Cite or Quote
179
September 27. 1996
-------
5.4 SENSITIVITY AND UNCERTAINTY ANALYSES
The results presented in this risk assessment are dependent on a number of factors,
including the various assumptions and data analysis approaches taken, the outcomes of
supporting data analyses, and the availability of sufficient data. Sensitivity analyses address the
extent to which variations in key assumptions and approaches affect the outcome of the risk
assessment. These variations are associated with overall uncertainty. Thus, sensitivity analysis
evaluates how sensitive the results and conclusions of the risk assessment are to the uncertainty
present in the analysis.
There are numerous procedures and assumptions discussed and presented in Chapters 3
through 5 that contribute to the final results of the risk assessment. As it was not feasible to
consider variations in all aspects of the risk assessment data analysis, the sensitivity analysis
considered approaches and assumptions which had the potential for producing the largest
expected deviation from the final results. The alternative approaches considered in the
sensitivity analysis and the comparison of their findings with the final results had to be
manageable within the context of the sensitivity analysis. Table 5-12 summarizes the six
factors addressed by the sensitivity analysis and the alternative approaches) considered for
each factor.
5.4.1 Components of the Sensitivity Analysis
This subsection discusses each component of the sensitivity analysis portrayed in Table
5-12. Justification for each alternative approach considered in the analysis is provided, and
reasons for not including certain factors of the risk assessment in the sensitivity analysis are
discussed.
One aspect of sensitivity analysis on predicting post-intervention blood-lead
concentrations from environmental-lead levels does not appear within Table 5-12, as it has
been incorporated directly within the risk assessment. Two models were used to predict post-
intervention blood-lead concentration: the IEUBK model and the EPI model. As these two
models represent different approaches for predicting blood-lead concentration from
environmental-lead levels, their results can be compared as part of a sensitivity analysis.. In
Draft - Do Not Cite or Quote 180 September 27, 1996
-------
addition, the use of two models led to the decision to not consider modifying parameter
estimates or model forms within either the IEUBK. or EPI models in the sensitivity analysis.
Table 5-12. Procedures and Their Alternatives That Were Included in the Sensitivity
Analysis
Procedure
Approach Taken in the
Risk Assessment
Altemative(s) Considered in the
Sensitivity Analysis
Determine an appropriate age
group of children to consider for
risk assessment
Age group = 1 to 2 years
(i.e., 12 to 35 months)
Age group = 1 to 5 years
(i.e., 12 to 71 months)
Determine an average IQ point
loss associated with every 1
microgram of lead per deciliter
of blood in children
Average IQ point loss = 0.257.
Alt. #1: Average IQ point loss = 0.185
Alt. tt2: Average IQ point loss = 0.323
Determine a baseline (pre-
intervention) distribution of
blood-lead concentrations from
NHANES III data
Assume lognormal distribution
(See Section 5.1)
Use the empirical distribution reported in the
NHANES III without any type of modeling.
Convert Blue Nozzle vacuum
dust-lead loadings reported in
the National Survey to wipe
dust-lead loadings, so their area-
weighted geometric mean can
be compared to S403
environmental-lead standards
Convert each sample result using
the following formulas:
Floors:
Wipe = 11.4'(Vac)0890
Window Sills:
Wipe = 5.79'(Vac)1078
where 'Wipe* is the estimated
wipe dust-lead loading and "Vac"
is the measured vacuum dust-
lead loading (see Section 4.2)
Alt. #1 (low estimatel: Assign the lower 90%
confidence bound on the estimated wipe dust-
lead loading obtained from the adjacent formula
to each sample result.
Alt. »2 (high estimatel: Assign the upper 90%
confidence bound on the estimated wipe dust-
lead loading obtained from the adjacent formula
to each sample result.
Determine a post-S 403 blood-
lead concentration distribution
as a function of post-
intervention environmental-lead
levels (Section 5.2)
Consider post-intervention
environmental-lead levels
summarized in Table 5-3 of
Section 5.2.
Consider the following alternative post-
intervention environmental-lead levels:
- 20 //g/ft* for floor dust-lead loading and
50 itgl\t2 for window sill dust-lead
loading
- 100 //g/ft2 for floor dust-lead loading and
250 //g/ft2 for window sill dust-lead
loading
- 20% and 80% decline in pre-intervention
soil-lead concentration when soil cover is
performed.
Determine a method for
characterizing the post-$403
distribution of blood-lead
concentration, and comparing
health effects between pre- and
post-5403.
Apply the methods in Section
5.3 to obtain pre- and post-
intervention distributions.
Alt. 01: Apply the alternative method detailed
in Approach #1 in Section 5.4.1.6.
Alt, if2: Rather than predicting post-5403
blood-lead concentration as a function of
environmental-lead levels, conduct the
prediction based on efficacy seen in abatement
studies with an adjustment for bone-lead
stores.
Draft - Do Not Cite or Quote
181
September 27, 1996
-------
The sensitivity analysis does not consider other options for obtaining estimated
numbers of housing units in the 1997 housing stock or numbers of children residing in the
housing stock (presented hi Chapter 3). In preliminary analyses, it was observed that regardless
of the method used to obtain an estimated number of units (or children) within the four
categories determined by housing age, the percentage of the total housing stock (or the total
population of children) within each group remained relatively constant. Therefore, it was not
deemed necessary to consider alternative methods for determining numbers of housing units or
children.
Note that the sensitivity analysis does not address various options for the §403
standards. Results to be used in evaluating these options are included within the risk
assessment presented in Section 5.3.
5.4.1.1 Alternative Aoe Ranae of Children
For reasons discussed in Section 2.4, the §403 risk assessment characterized lead
exposures and health effects for children aged 12-35 months (i.e., 1-2 years). However, as the
interventions that result from §403 are expected to benefit young children of other ages as well,
the sensitivity analysis calculated health effects associated with children in a broader age range:
children aged 12-71 months (i.e., 1-5 years). This alternative age range is considered as Title X
has defined target housing as housing built prior to 1978 hi which children less than six years
of age may reside. Broadening the age range to include older children will likely result in an
emphasis on lower blood-lead concentrations hi the overall distribution.
5.4.1.2 Alternative Assumptions on Average IQ Score Decline Per Unit Increase in
Blood-Lead Concentration
As discussed in Chapter 4, results of the meta-analysis documented in Schwartz (1994)
indicate that an average IQ point loss of 0.257 is predicted for every 1.0 ug/dL increase in
blood-lead concentration. This relationship was used hi Sections 5.1 and 5.3 to characterize
health effects associated with elevated blood-lead concentration. In the sensitivity analysis,
two alternative average IQ point loss estimates were considered when calculating pre-
intervention reduction in IQ points associated with blood-lead concentration: 0.185 and 0.323.
Draft - Do Not Cite or Quote 182 September 27. 1996
-------
The lower value of 0.185 was selected based on findings of a prospective study of blood-lead
concentration for children approximately two years of age, as reported in Pocock et al. (1994).
The higher value of 0.323 corresponds to an examination in Schwartz (1994) on the existence
of a threshold in the relationship between IQ score and blood-lead concentration. In studies
that involve primarily children with blood-lead concentrations of 15 ug/dL or lower, the
estimated average IQ point loss was reported to be 0.323. The estimates of 0.185 and 0.323
result in a lower and higher estimate, respectively, of the benefits associated with §403. The
sensitivity analysis did not consider alternative methods for estimating other health effect
endpoints, such as the probability of observing IQ scores less than 70 or the probability of
observing elevated blood-lead concentrations.
5.4.1.3 Alternative Approach to Characterizing a Baseline Blood-Lead Distribution from
NHANES III Data
As discussed in Section 5.1, the baseline, pre-§403 distribution of blood-lead
concentrations in children aged 12 to 35 months was assumed to be lognormal, with geometric
mean and standard deviation calculated from NHANES ID for this age group. Health
endpoints were then calculated from this distribution. An alternative approach to
characterizing the baseline distribution using the NHANES ffl data considered an empirical
distribution. This alternative approach was applied in the sensitivity analysis.
The NHANES HI database contained blood-lead concentrations for 924 children aged
12-35 months at the time of their survey interview. Each child in the survey was assigned a
sampling weight corresponding to the number of children in the country being represented by
the child. This combination of blood-lead concentration and sample weight for each surveyed
child provided an empirical distribution of blood-lead concentration for children aged 12-35
months. Percentiles, such as the probability of observing a blood-lead concentration less than
10 ug/dL, were calculated from this distribution by summing the sample weights for children
with blood-lead concentrations less than 10 ug/dL, then dividing by the total of all sampling
weights. Percentiles were used to obtain the probability of elevated blood-lead concentration
and the probability of observing a specific decrement in IQ score. The probability of observing
an IQ score less than 70 was calculated for each surveyed child based on his/her blood-lead
Draft - Do Not Cite or Quote 183 September 27, 1996
-------
concentration (see Section 4.4), then was multiplied by the child's sample weight and summed
across children to obtain an expected number (and percentage) of children in the nation with IQ
less than 70.
Providing an alternative approach to characterizing the pre-§403 blood-lead
concentration distribution and the resulting health effects provided a means of evaluating the
lognormality assumption placed on this distribution.
5.4.1.4 Uncertainty in Converting Dust-Lead Loadings for Comparison to Standards
Because the §403 dust-lead standards will be defined in terms of a lead loading of a
wipe sample, and because dust samples in the HUD National Survey were collected using a
Blue Nozzle vacuum, methods in Section 4.2 were used to convert the National Survey dust-
lead loadings (for both floors and window sills) to wipe dust-lead loadings. In the risk
assessment, two formulas were used (Table 5-12) to predict a wipe dust-lead loading from a
Blue Nozzle vacuum dust-lead loading, depending on whether a floor or window sill was
sampled. These formulas indicate that the expected value of the log-transformed wipe dust-
lead loading (log(Wipe)) takes the form
a + p*log(Vac)
where a and P are parameter estimates. Therefore, assuming lognormality, upper and lower
one-sided 90% confidence bounds on the expected value of log(Wipe) are
log(Wipe) ±1.3* SE(a + p * log(Vac))
where SE(a + P * log (Vac)) is the standard error of the predicted average wipe dust-lead
loading of a vacuum sample with a dust-lead loading of Vac. Upper and lower 90%
confidence bounds on the untransformed wipe dust-lead loadings are obtained by
exponentiating the bounds for the log-transformed loading.
The confidence bounds were used to define two alternative sets of converted dust-lead
loadings in the sensitivity analysis:
Draft - Do Not Cite or Quote 184 September 27, 1996
-------
Alternative set #1: Wipe dust-lead loading equals the lower 90% confidence bound on the
expected value of wipe dust-lead loading
/
Alternative set #2: Wipe dust-lead loading equals the upper 90% confidence bound on the
expected value of wipe dust-lead loading
Note that alternative set #1 is a low estimate of the converted loading value, while alternative
set #2 is a high estimate. Under both sets, area-weighted arithmetic mean dust-lead loadings
for both floors and window sills were calculated for each National Survey unit. The means
were used to determine whether dust-lead loading standards were exceeded for a given unit, hi
the sensitivity analysis, numbers and percentages of units exceeding various combinations of
environmental-lead standards were calculated under each set of converted dust-lead loadings.
5.4.1.5 Alternative Assumptions on Post-Intervention Environmental-Lead Levels
Estimates of expected post-intervention environmental-lead levels were provided in
Table 5-3. The sensitivity analysis considered alternatives to the post-intervention dust-lead
loading following dust cleaning, interior LBP intervention, or soil removal; as well as in the
soil-lead concentration following soil cover, in order to observe how the health effect estimates
were affected by assumptions on post-intervention environmental-lead levels. Two sets of
alternative post-intervention dust-lead loadings for floors and window sills were considered as
a result of dust cleaning:
• 20 ug/ft2 for floors and SO ug/ft2 for window sills, and
• 100 jag/ft2 for floors and 250 ug/ft2 for window sills.
(The loadings used in the risk assessment were 40 ug/ft2 for floors and 100 ug/ft2 for window
sills.) As post-intervention soil-lead concentration following soil cover was assumed to be
50% of the pre-intervention concentration in the risk assessment, the sensitivity analysis
considered a low alternative value of 20% and a high alternative value of 80%. The sensitivity
analysis did not address alternative soil-lead concentration values following soil removal (150
Hg/g), or amounts of deteriorated lead-based paint following paint interventions (0 ft2). The
Draft - Do Not Cite or Quote 185 September 27, 1996
-------
same approaches to determining post-§403 blood-lead distributions using either the IEUBK
model or the EPI model were repeated in this analysis.
5.4.1.6 Alternative Methods to Observing Differences in Health Effects Between Pre-
and Post-Intervention
As presented in Section 5.3 and Appendix E2, the method for characterizing a post-
§403 distribution of blood-lead concentration used in the risk assessment effort involved the
;
following: 1) obtain a predicted distribution for both pre- and post-§403 by applying either the
IEUBK or the EPI model to environmental-lead levels; 2) calculate the ratio of the post-§403
geometric mean to the pre-§403 geometric mean; 3) multiply the ratio by the geometric mean
observed for the pre-§403 distribution obtained from NHANES TO. data, resulting in an
estimated geometric mean for the post-§403 distribution; and 4) assume that the distribution is
lognormal. This method yielded a post-§403 distribution that was directly comparable with the
pre-§403 distribution. Two alternative approaches to obtaining a post-§403 blood-lead
concentration distribution were considered in the sensitivity analysis.
Approach #1: Alternative approach to obtaining comparable pre- and post-§403 blood-lead
distributions
In the first alternative approach, two predicted distributions were obtained (one for pre-
§403 one for post-§403) simply by applying the same model (IEUBK or the EPI model) to
either pre- or post-§403 environmental-lead levels. This approach is step #1 in the preceding
paragraph. Note, that the resulting distributions are purely model-based; the NHANES HI data
are not used as a basis for characterizing the distributions. In the sensitivity analysis, the health
endpoints were calculated using the risk assessment approach directly on the model-based
blood-lead distributions.
Approach #2: Alternative approach to determining a post-intervention blood-lead
distribution using directly-measured blood lead changes
A second approach to characterizing the post-§403 blood-lead concentration
distribution was performed utilizing published results on the effectiveness of lead hazard
Draft - Do Not Cite or Quote 186 September 27, 1996
-------
intervention strategies among children exposed to residential lead hazards. This approach is
desirable since blood-lead concentrations are a more direct measure of intervention
effectiveness than environmental lead levels. The scientific literature reports the results of a
range of non-medical intervention strategies conducted to reduce the lead exposure of children
residing at the targeted residences (EPA, 199Sb). The strategies studied included lead-based
paint abatement, interior dust abatement via routine cleaning procedures, elevated soil lead
abatement, and intensive educational efforts (EPA, 199Sb). The effectiveness of these
strategies as measured by declines in children's blood-lead concentrations may be used to
estimate the post-§403 blood-lead concentration distribution. As such, this approach represents
a somewhat independent (of many of the procedures and data used for the risk assessment)
estimation of the post-§403 distribution.
As summarized in the EPA technical report, "Review of Studies Addressing Lead
Abatement Effectiveness," the intervention strategies reported 18-34% declines in the blood-
lead concentrations of exposed children six to twelve months following the conduct of the
intervention (EPA, 1995b). Lead-based paint abatement (of all deteriorated LBP), biweekly
dust abatement (of areas with elevated dust lead), soil abatement (removal and replacement of
top 6"), and intensive education (visit by semi-professional outreach worker) reported
comparable declines of approximately 25% one year following conduct of the intervention
(EPA, 199Sb). Each of these four intervention studies reported significantly greater declines
among the study population than among a suitable control population—no control population
was studied for the educational intervention associated with the 34% decline—providing
reassurance that the interventions themselves were responsible for much of the reported
declines. For the purpose of this sensitivity analysis, therefore, the average decline in
children's blood-lead concentration resulting from an intervention was taken to be 25%4.
This degree of effectiveness may not be suitable for estimating the post-§403 blood-
lead distribution since the reported declines were for children already exposed (i.e., already
exhibiting elevated blood-lead concentrations due to exposure to the targeted lead source). By
4 In all four studies, the control population did exhibit some decline which may be attributed to increased
awareness of environmental lead and its hazards. As similar awareness may be expected to accompany §403
prompted interventions, it was not deemed necessary to adjust the reported study population declines by the declines
associated with the control populations.
Draft - Do Not Cite or Quote 187 September 27, 1996
-------
contrast, the promulgation of §403 will prompt preventive interventions (primary prevention)
conducted prior to any lead exposure to resident children. Measures of secondary prevention
effectiveness may not be representative of primary intervention effectiveness because lead
present in blood is a combination of current environmental exposure and internal reservoirs of
lead stored in bone and soft tissue (Gulson et al., 1995; Smith et al., 1996; Rabinowitz et al.,
1976; Manton, 1985). The reported declines in exposed children's blood-lead concentrations,
therefore, may underestimate the primary prevention effectiveness of a intervention (Gulson
etal., 1995).
A methodology was developed to estimate the impact of body lead burdens on measures
of secondary intervention effectiveness in order to adjust the reported secondary prevention
effectiveness (see Appendix E3). For a two-year-old child (the target population of interest), it
is estimated that an intervention prompting 25% declines one year following intervention
among exposed children would actually prompt 33% declines were the intervention primary
prevention in character. Based on this result a 33% efficacy will be utilized for the purposes of
this portion of the sensitivity analysis.
As a comparison, the IEUBK model indicates a 54% primary prevention efficacy were
lead-based paint hazards eliminated and dust- and soil-lead levels lowered to background
levels. Specifically, the geometric mean blood-lead concentration reported by NHANES III for
children 1-2 years of age (4.1 ug/dL) (Brody et al., 1994) was contrasted with the geometric
mean predicted by the IEUBK model with inputted environmental lead levels at national
background levels (1.9 ug/dL). Shacklette et al. reported a background national geometric
mean soil-lead concentration of 20 ppm. The background dust-lead concentration
corresponded to the default dust lead assumed by the IEUBK Multiple Source Analysis (dust-
lead concentration = 0.70 x soil-lead concentration +100 ug/g / ug/m3 x air-lead concentration
[air-lead concentration = 0.01 ug/m3]). All other default values defined by the IEUBK model
were used in these analyses.
It is worth noting that the scientific literature also includes two recent journal articles
regarding the percentage of lead in blood that may be attributed to body lead stores (Gulson et
al., 1995; Smith et al., 1996). Such results, of course, have relevance to this aspect of the
sensitivity analysis. Both articles indicate that between 40-70% of an adult woman's blood
Draft - Do Not Cite or Quote 188 September 27, 1996
-------
lead may be attributed to mobilized bone-lead stores. The fact that these studies examined
adult women is critical because the percentage of blood lead attributable to bone-lead stores
varies considerably with age (Rabinowitz, 1991). Higher percentages are associated with older
individuals (Rabinowitz, 1991). Thus, the target population considered by §403 may have
lower percentages of their blood lead attributable to mobilized bone lead. Greater primary
prevention efficacy is reported for, say, 7 year old children than for 2 year old children (see
Table 1 in Appendix E3). If the methodology used in this alternative approach were extended
to adults, it would also suggest that 40-70% of blood lead is attributable to mobilized bone-lead
stores.
This alternate approach to estimating a post-§403 national distribution of blood-lead
concentrations for 1997 children aged 12 to 35 months (1 to 2 years) was implemented based
on the estimated 33% decline in blood-lead concentration following an intervention. This
alternative estimate of primary prevention effectiveness, which adjusts the blood-lead changes
for body-lead stores and hereafter is denoted the 'adjusted blood lead effects model', was then
compared to post-§403 distribution using the DEUBK. model and the HUD National Survey
data.
The methodology for this comparison is summarized as follows:
1. Environmental lead levels for each HUD National Survey unit were used as input to
the IEUBK. model to predict the geometric mean blood-lead concentration for
children aged 1-2 years old exposed to environmental lead levels similar to that in
the National Survey unit. The contribution of pica was estimated using the
methodology documented in Section 4.3.
2. For each unit in the National Survey, lead levels in paint, dust, and soil were
compared to the following options for standards:
• 100 ug/ft2 as an interior floor dust-lead loading and 500 (ig/ft2 as an interior
window sill dust-lead loading; and,
• 400 ug/g as a soil-lead concentration for soil cover and 3000 ug/g for soil
abatement.
• Oft2 of deteriorated lead-based paint;
If environmental lead levels exceeded the standard for at least one media, then an
intervention will be conducted in the unit.
Draft - Do Not Cite or Quote 189 September 27, 1996
-------
3. For each National Survey unit, if an intervention was triggered then the geometric
mean blood-lead concentration for the post-§403 result was set equal to 67% of the
geometric mean computed in (1). If an intervention was not triggered then the
geometric mean blood-lead concentration for the post-§403 result equaled the
geometric mean calculated in (1).
4. The geometric mean blood-lead concentration and the geometric standard deviation
of 1.6 ng/dL were used to generate a frequency distribution of blood-lead
concentrations for each unit in the National Survey. The frequency distributions
were then combined over all of the National Survey units to yield the HUD/IEUBK
post-§403 blood-lead distribution. The details for generating the frequency
distribution of blood-lead concentration at each unit and over all units in the
National Survey are presented in Appendix El.
5.4.2 Results of the Sensitivity Analysis
The six factors considered in the sensitivity analysis and presented in Table 5-12
address various segments of the risk assessment effort. The results of the sensitivity analysis
are presented in this subsection and are grouped according to these segments.
5.4.2.1 Alternative Age Range of Children
Table 5-1 of Section 5.1 presented baseline estimates (pre-§403) of numbers and
percentages of children in the U.S. aged 12-35 months in 1997 who exhibited the health effects
of interest. Tables 5-13a and 5-13b present these estimates, along with estimates for the age
group 12-71 months (i.e., 1-5 years). Table 5-13a presents estimated percentages of children
with elevated blood-lead levels, IQ less than 70, and specified decrements in IQ score. The
percentages for the 12-71 month age group are approximately 15%-25% lower than those for
the 12-35 month age group. For example, Table 5-13a indicates that the expected percentage
of children aged 12-35 months having blood-lead concentration of at least 10 jig/dL is 10.5%,
compared to approximately 8% for children aged 12-71 months. Table 5-13b presented
estimated average IQ score loss; similar declines are observed here. The declines are the result
of lower blood-lead concentrations introduced to the distribution by including older children.
Draft - Do Not Cite or Quote 190 September 27, 1996
-------
Table 5-13a. Estimated Baseline1 Number and Percentage of Children Having Specific
Health Effects, for Two Age Groups of Children and Under Three
Assumptions on Average Decline in IQ Score per Unit Increase in Blood-
Lead Concentration
Health Effect2
Blood-lead concentration of
at least 10/;g/dL
Blood-lead concentration of
at least 25 fjg/dL
IQ score less than 70
IQ score decrement
of greater than 1
IQ score decrement
of greater than 2
IQ score decrement
of greater than 3
0.1 85 decline/
Ijug/dL increase
0.257 decline/
1|/g/dL increase
0.323 decline/
1/yg/dL increase
0.1 85 decline/
1//g/dL increase
0.257 decline/
1yug/dL increase
0.323 decline/
1/sg/dL increase
0.1 85 decline/
Ifjg/dL increase
0.257 decline/
1/yg/dL increase
0.323 decline/
1/yg/dL increase
Children Aged 12-35
Months Having the Given
Health Effect
Number
(millions)
0.83
0.046
0.045
2.74
4.15
5.13
0.69
1.45
2.21
0.22
0.56
0.99
Percentage
10.5
0.578
0.566
34.4
52.2
64.5
8.65
18.2
27.8
2.71
7.09
12.5
Children Aged 12-71 Months
Having the Given Health
Effect
Number
(millions)
1.64
0.082
0.11
5.84
9.25
11.77
1.34
2.95
4.63
0.40
1.09
1.97
Percentage
8.04
0.403
0.531
28.6
45.3
57.7
6.57
14.4
22.7
1.97
5.34
9.65
1 'Baseline* refers to projected 1997 conditions prior to implementing any interventions under Section 403 rules. For a
given age group, the baseline blood-lead distribution used to determine health effects was characterized using methods in
Section 5.1.
2 For IQ score decrement, this column also includes the assumption on average IQ score decline per 1 //g/dL increase in
blood-lead concentration.
Shaded cells correspond to results that were presented in Table 5-1.
Draft - Do Not Cite or Quote
191
September 27, 1996
-------
Table 5-13b. Estimated Baseline1 Average (and Standard Deviation) IQ Loss, for Two
Age Groups of Children and Under Three Assumptions on Average Decline
in IQ Score per Unit Increase in Blood-Lead Concentration
Assumption on Average
IQ Score Decline per 1 .0
/ig/dL Increase in Blood-
Lead Concentration
0.185
0.257
0.323
Average IQ Loss
(Standard Deviation)
Children Aged
12-35 Months
0.97 (0.80)
1.35(1.11)
1.70(1.40)
Children Aged
12-71 Months
0.87 (0.73)
1.20(1.02)
1.51 (1.28)
5.4.2.2
1 'Baseline* refers to projected 1997 conditions prior to implementing any
interventions under Section 403 rules. For a given age group, the baseline blood-
lead distribution used to determine health effects was characterized using methods
in Section 5.1.
Shaded cell corresponds to results that were presented in Table 5-1.
Alternative Assumptions on Average IQ Score Decline Per Unit Increase in
Blood-Lead Concentration
In Tables 5-13a and 5-13b, the percentage of children with IQ score decrements of a
certain magnitude, and average IQ score loss across all children, were calculated under the
assumption of a decline of 0.257 in IQ score for each 1 ug/dL increase in blood-lead
concentration. The tables also include these percentages as calculated under the alternative
assumptions of an average decline of 0.185 and 0.323 in IQ score (Section 5.4.1.2).
The low and high estimates for the decline in IQ score associated with a 1 ug/dL
increase in blood-lead concentration has a considerable impact on the likelihood that a child
will experience a specific decrement in IQ score, with this effect increasing as the decrement of
interest increases. As seen in Table 5-13a, the estimated percentage of children with an IQ
score decrement of greater than one, as calculated using the low estimate of IQ point loss
(0.185), nearly doubles when the high estimate (0.323) is used instead (from 34% to 65%).
When the decrement is greater than three, this difference is over four times the result associated
with the low estimate (from 2.7% to 12.5%). In Table 5-13b, average IQ point loss increases
from 0.97 to 1.70 for children aged 12-35 months, with a similar increase for children aged 12-
71 months.
Draft - Do Not Cite or Quote
192
September 27. 1996
-------
5.4.2.3 Alternative Approach to Characterizing a Baseline Blood-Lead Distribution from
NHANES III Data
Using the empirical distribution method in Section 5.4.1.3, an alternative baseline blood-
lead distribution based on NHANES III data was derived. The values of the health endpoints
under this alternative distribution, as well as for the baseline distribution used in the risk
assessment, are provided in Table 5-14. Between the two distributions, the values differ by
small amounts (e.g., from one to ten percent). While the alternative method estimates a higher
percentage of children with blood-lead concentrations exceeding 10 ug/dL (11.1% versus
10.5% for the risk assessment baseline distribution), it estimates a slightly smaller percentage
of children with blood-lead concentrations exceeding 25 ug/dL (0.52% versus 0.58%).
Estimated average IQ score loss is nearly identical between the two distributions.
Table 5-14. Estimated Baseline Health Effects, As Calculated Under Two Approaches to
Calculating the Baseline Distribution of Blood-Lead Concentration Using
NHANES III Data
Health Effect
Percent with blood-lead concentration of
at least 10/ig/dL
Percent with blood-lead concentration of
at least 25 //g/dL
Percent with an IQ score less than 70
Percent with an IQ score decrement of
greater than 1
Percent with an IQ score decrement of
greater than 2
Percent with an IQ score decrement of
greater than 3
Average IQ Loss
(Standard Deviation)
Approach
Used in the
Risk
Assessment
10.5
0.58
0.566
52.2
18.2
7.09
1.35
(1.11)
Alternative
Approach
11.1
0.52
0.56
51.1
17.9
7.96
1.35
(1.12)
Draft - Do Not Cite or Quote
193
September 27. 1996
-------
5.4.2.4 Uncertainty in Converting Dust-Lead Loadings for Comparison to Standards
As discussed in Section 5.4.1.4, the dust-lead loadings (both floors and window sills)
measured in the HUD National Survey were converted to wipe equivalent dust-lead loadings.
In the sensitivity analysis, two alternative sets of converted dust-lead loadings were considered
for the dust samples:
Alternative set #1: Wipe dust-lead loading equals the lower 90% confidence bound on the
predicted average value of wipe dust-lead loading
Alternative set #2: Wipe dust-lead loading equals the upper 90% confidence bound on the
predicted average value of wipe dust-lead loading
Table 5-15 contains estimates of the total number of 1997 housing units which exceed various
environmental-lead standards, given that the wipe-converted dust-lead loadings for the units
were determined based on the risk assessment method, alternative set #1, or alternative set #2.
Table 5-15 considered numbers of units exceeding the floor-dust standard of 200 ug/ft2,
exceeding the window sill dust standard of 500 (ig/ft2, any of these two standards, or any of the
standards for dust, soil, or paint.
The largest variation between the two alternative sets of dust-lead loadings occurred
when considering only the floor-dust standard. When a high conversion value is used for each
dust-lead loading, over four million units fail the floor-dust standard, compared to nearly two
million units under the risk assessment conversion, and 1.5 million units under the low
conversion values. This finding implies that the risk assessment may be underestimating the
numbers of homes affected by the §403 floor dust standard by a factor of two if the average
predicted wipe equivalent dust-lead loadings are being underestimated. However, a dust-
cleaning intervention is triggered if either the floor or window sill dust-lead loading standard is
exceeded. The impact of the uncertainty in the dust-lead loading conversion equation was
much less for the number of homes affected by either the §403 floor dust or window sill dust
standard. The number of units triggering an intervention because of either dust standard ranged
from a low estimate of 12.7 million to a high estimate of 16.3 million, which is a range of
about 30%.
Draft - Do Not Cite or Quote 194 September 27, 1996
-------
Table 5-15. Number (and Percentage) of Units in the 1997 National Housing Stock
Projected to Exceed Various Combinations of Environmental-Lead Standards
Under Section 403 Rules, As Determined from Three Different Sets of
Converted Dust-Lead Loadings
Environmental-Lead
Standards Exceeded
Floor-dust standard of 200
//g/ft2
Window sill-dust standard
of 500 A/g/ft2
Floor- or window sill- dust
standard
At least one dust, soil, or
paint standard3
Number {%) of Units
Using Risk
Assessment
Estimates for
Converted Dust-
Lead Loading1
1,968,000
(1.98%)
13,979,000
(14.1%)
1 5,269,000
(15.4%)
25,957,000
(26.1%)
Using Low
Alternative
Estimates for
Converted Dust-
Lead Loading2
1,509,000
(1.52%)
11,803,000
(11.9%)
12,743,000
(12.8%)
24,918.000
(25.1%)
Using Hioh Alternative
Estimates for
Converted Dust-Lead
Loading1
4,270,000
(4.30%)
14.597,000
(14.7%)
16.275,000
(16.4%)
26,780,000
(27.0%)
1 See Section 4.2 on the methods for performing conversions from Blue Nozzle vacuum to wipe dust-lead loadings.
2 Low and high estimates correspond to the lower 90% confidence bound and upper 90% confidence bound, respectively,
of the risk assessment estimates considered in the second column of this table.
3 Soil and paint standards are as follows: soil-lead concentration of 400 j/g/g for soil cover, soil-lead concentrations of
3000 i/g/g for soil removal, 5 ft2 of deteriorated lead-based paint for paint repair, and 20 ft2 of deteriorated lead-based
paint for paint removal.
5.4.2.5 Alternative Assumptions on Post-Intervention Environmental-Lead Levels
Tables 5-16a and 5-16b summarize the childhood health and blood lead effects post-§403
based on the IEUBK and EPI models, respectively, for alternative post-intervention
environmental-lead levels. These tables show the impact of alternative assumptions on the
efficacy of the interventions on the risk assessment. Results in these two tables were calculated
assuming the following §403 standards:
• Dust-lead loading (under wipe sampling techniques) of 200 ug/ft2 for floors and 500
ug/ft2 for window sills
• Soil-lead concentration of 400 ug/g for soil cover and 3000 ug/g for soil removal
• 5 ft2 of deteriorated lead-based paint for paint repair, and 20 ft2 for paint removal
Draft - Do Not Cite or Quote
195
September 27, 1996
-------
Table 5-16a. Estimated Percentages of Children Aged 12-35 Months Having Specific Health and Blood-Lead Effects,
Based on the IEUBK Model, for Various Options for Post-Intervention Environmental-Lead Levels
Health Effect
Percent with blood-lead concentration of
at least 10/sg/dL
Percent with blood-lead concentration of
at least 25^g/dL
Percent with an IQ score less than 70
Percent with an IQ score decrement of
greater than 1
Percent with an IQ score decrement of
greater than 2
Percent with an IQ score decrement of
greater than 3
Average IQ Loss
(Standard Deviation)
0 ft1 Deteriorated Lead-Based Paint after all Paint Interventions
Soil-Lead Concentration after Soil Removal Intervention - 160 //g/g
Dust-Lead Loading after Dust
Cleaning Intervention:
Floors = 20 /i g/ft2
Window Sills = 50 jig/ft*
Soil-Lead
Interventic
20%
2.01
0.007
0.462
40.5
5.81
0.940
1.01
(0.56)
1 Cone, after
m (% of Pre
Cone.)
50%
3.10
0.022
0.475
43.6
7.88
1.60
1.07
(0.64)
Soil Cover
•Intervention
80%
4.22
0.049
0.488
45.6
9.73
2.32
1.11
(0.71)
Dust-Lead Loading after Dust
Cleaning Intervention:
Floors = 40/ig/ft*
Window Sills- lOOpg/ft*
Soil-Lew
Interventk
20%
2.44
0.011
0.468
42.2
6.70
1.19
1.04
(0.59)
1 Cone, aftei
m (% of Pre
Cone.)
50%
3.60
0.031
0.481
44.9
8.78
1.91
1.09
(0.67)
Soil Cover
-Intervention
80%
4.74
0.066
0.494
46.8
10.6
2.68
1.14
(0.74)
Dust-Lead Loading after Dust
Cleaning Intervention:
Floors = 100 //g/ft2
Window Sills = 250 //g/ft2
Soil-Lead
Interventic
20%
3.29
0.024
0.478
44.4
8.26
1.71
1.08
(0.65)
1 Cone, aftei
in (% of Pre
Cone.)
50%
4.51
0.056
0.492
46.7
10.3
2.51
1.13
(0.72)
Soil Cover
-Intervention
80%
5.66
0.104
0.505
48.2
12.0
3.31
1.17
(0.79)
i
CO
O)
I
IS)
M
This analysis assumes the following environmental-lead standards determine whether or not a particular intervention is performed in a
housing unit:
• Dust-lead loading (under wipe techniques) of 200 //g/ft2 for floors and 500 //g/ft2 for window sills for dust cleaning
• Soil-lead concentration of 400 //g/g for soil cover and 3000 //g/g for soil removal
• 5 ft2 of deteriorated lead-based paint for paint repair and 20 ft2 for paint removal
Shaded cells correspond to results that were presented in Table 5-7.
-------
Table 5-16b. Estimated Percentages ol Children Aged 12-35 Months Having Specific Health and Blood-Lead Effects,
Based on the EPI Model, for Various Options for Post-Intervention Environmental-Lead Levels
Health Effect
Percent with blood-lead concentration of
at least 10//g/dL
Percent with blood-lead concentration of
at least 25 /ig/dL
Percent with an IQ score less than 70
Percent with an IQ score decrement of
greater than 1
Percent with an IQ score decrement of
greater than 2
Percent with an IQ score decrement of
greater than 3
Average IQ Loss
(Standard Deviation)
0 ft* Deteriorated Lead-Based Paint after all Paint Interventions
Soil-Lead Concentration after Soil Removal Intervention = 150/ig/g
Dust-Lead Loading after Dust
Cleaning Intervention:
Floors = 20/ig/ft1
Window Sills = 50 //a/ft7
Soil-Lead Cone, after Son Cover
Interventir
20%
6.46
0.177
0.512
47.0
12.8
3.97
1.18
(0.86)
m i TO ai rre
Cone.)
60%
6.94
0.210
0.519
47.8
13.5
4.33
1.20
(0.89)
-imervBnuun
80%
7.22
0.231
0.522
48.2
13.9
4.54
1.21
(0.91)
Dust-Lead Loading after Dust
Cleaning Intervention:
Floors = 40 jig/ft*
Window Sills - 100 //g/ft1
Soil-Lea*
Interventk
20%
6.96
0.209
0.519
47.9
13.6
4.34
1.20
(0.89)
1 Cone, after
in (% of Pre
Cone.)
60%
7.49
0.249
0.526
48.7
14.3
4,73
1.23
(0.92)
Soil Cover
•Intervention
80%
7.79
0.275
0.529
49.1
14.7
4.96
1.24
(0.94)
Dust-Lead Loading
Cleaning Interve
Floors = 100/
Window Sills = 2
Soil-Leat
Interventk
20%
7.63
0.260
0.527
49.0
14.5
4.84
1.23
(0.93)
1 Cone, after
m (% of Pre
Cone.)
50%
8.20
0.310
0.535
49.8
15.3
5.28
1.26
(0.97)
after Dust
ntion:
vg/ft1
50/ig/ft*
Soil Cover
•Intervention
80%
8.53
0.342
0.539
50.2
15.7
5.53
1.27
(0.99)
CO
vl
Kj
This analysis assumes the following environmental-lead standards determine whether or not a particular intervention is performed in a
housing unit:
• Dust-lead loading (under wipe techniques) of 200 ^g/ft2 for floors and 500 ^g/ft2 for window sills for dust cleaning
• Soil-lead concentration of 400 fjg/g for soil cover and 3000 //g/g for soil removal
• 5 ft2 of deteriorated lead-based paint for paint repair and 20 ft2 for paint removal
Shaded cells correspond to results that were presented in Table 5-7.
-------
The following alternative of post-intervention environmental-lead levels, in addition to the
levels assumed in the risk assessment effort, were evaluated:
• two combinations of post-intervention dust-lead loadings for floors and window sills
(20 ug/ft2 for floors and 50 ug/ft2 for window sills; and 100 ug/ft2 for floors and 250
ug/ft2 for window sills)
• two settings of soil-lead concentrations following soil cover (20% and 80% of pre-
intervention levels).
Table 5-16a indicates that the health effects and blood-lead most affected by uncertainty
in the post-intervention environmental-lead levels are those indicating the most extreme effects
(e.g., IQ decrement greater than 3, blood-lead concentrations of at least 25 ug/dL). The
uncertainty in the assumed efficacy of the soil cover intervention has a larger impact on the
predicted health endpoints than the uncertainty in the efficacy of the dust cleaning. For
example, the percentage of children predicted with an IQ decrement greater than three varies
from 1.19 to 2.68% (columns 5,6,7) when the assumed efficacy of soil cover is varied from 20
to 80% of pre-intervention levels. On the other hand, the percentage of children predicted with
an IQ decrement greater than three varies from 1.6 to 2.5% (columns 3,6, and 9) when the
assumed efficacy of the dust cleaning is varied from 20 to 100 ug/ft2 for floor dust-lead
loadings and from 50 to 250 ug/ft2 for window sill dust-lead loadings.
A slightly different conclusion is made when considering the EPI model (Table 5-16b).
The uncertainty in the assumed efficacies of the interventions has little impact on the predicted
childhood health and blood-lead effects. For instance, the percentage of children predicted to
have an IQ decrement greater than three varies from 4.34 to 4.96% (columns 5,6, and 7) when
the assumed efficacy of soil cover is varied from 20 to 80% of pre-intervention levels.
5.4.2.6 Alternative Methods to Observing Differences in Health Effects Between Pre-
and Post-Intervention
Approach #1: Alternative approach to obtaining comparable pre- and post-intervention
blood-lead concentration distributions
Draft - Do Not Cite or Quote 198 September 27. 1996
-------
Section 5.4.1.6 described an alternative approach to obtaining comparable pre- and post-
§403 blood-lead distributions. This approach used the same modeling techniques (using either
the IEUBK or the EPI model) to obtain a geometric mean blood-lead distribution and
associated geometric standard deviation for both the pre- and post-§403 distributions, and then
assumed lognormality to characterize the distribution.
Table 5-17a presents estimated health and blood-lead endpoints at both pre- and post-
§403 for the risk assessment approach and for this alternative approach, based on the IEUBK
model. The results for the EPI model are presented in Table 5-17b. The following standards
were employed in the analyses:
• Dust-lead loading of 200 ug/ft2 for floors and 500 fig/ft2 for window sills
• Soil-lead concentration of 400 ug/g for soil cover and 3000 ug/g for soil removal
• 5 ft2 of deteriorated lead-based paint for paint repair and 20 ft2 for paint removal
Table 5-17a shows that for the IEUBK model, the estimated reduction in health and
blood-lead concentration risks for the two approaches are very similar. For example, the risk
assessment approach calculated an approximate 95% drop in the incidence of blood-lead
concentration exceeding 25 ug/dL, compared to an 88.5% drop under the alternative approach.
In contrast, more substantial differences between the two approaches were observed when
the EPI model was used, as indicated in Table 5-17b. The estimated risk reduction is greater
for the alternative approach employed in the risk assessment. However, the differences in the
estimated risk reduction for the two approaches are not substantial. For instance, the reduced
risk of a blood-lead concentration greater than 25 ng/dL are 57% and 75% based on the risk
assessment and alternative approaches, respectively.
Draft - Do Not Cite or Quote 199 September 27, 1996
-------
Table 5-17a. Estimated Percentages of Children Aged 12-35 Months Having Specific
Health and Blood-Lead Effects, Based on the IEUBK Model Under the
Approach Used in the Risk Assessment and an Alternative Approach
Health Effect
Percent with blood-lead
concentration of at least 10//g/dL
Percent with blood-lead
concentration of at least 25 fjg/dL
Percent with an IQ score less than
70
Percent with an IQ score decrement
of greater than 1
Percent with an IQ score decrement
of greater than 2
Percent with an IQ score decrement
of greater than 3
Average IQ Loss
(Standard Deviation)
Geometric Mean Blood-Lead
Concentration (Geometric Standard
Deviation)
Approach Used in the Risk
Assessment
Pre-lnt.
10.5
0.578
0.566
52.2
18.2
7.09
1.35
(1.11)
4.05
(2.06)
Post-
Int.
3.60
0.031
0.481
44.9
8.78
1.91
1.09
(0.67)
3.62
(1.76)
%
Change
-65.6
-94.6
-15.0
-13.8
-51.8
-73.0
-19.1
-10.5
Alternative Approach
Pre-lnt.
12.3
1.08
0.598
50.6
19.9
8.83
1.40
(1.33)
3.94
(2.23)
Post-hit.
5.37
0.124
0.498
44.0
11.1
3.22
1.12
(0.81)
3.53
(1.91)
%
Change
-56.4
-88.5
-16.8
-13.1
-44.2
-63.5
-20.1
-10.5
This analysis assumes the following environmental-lead standards determine whether or not a particular intervention is
performed in a housing unit:
• Dust-lead loading (under wipe techniques) of 200 //g/ft2 for floors and 500 //g/ft2 for window sills for dust
cleaning
• Soil-lead concentration of 400 //g/g for soil cover and 3000 //g/g for soil removal
• 5 ft2 of deteriorated lead-based paint for paint repair and 20 ft2 for paint removal
Draft - Do Not Cite or Quote
200
September 27. 1996
-------
Table 5-17b. Estimated Percentages of Children Aged 12-35 Months Having Specific
Health Effects, Based on Blood-Lead Concentrations at Pre- and Post-
Intervention as Determined from the EPI Model Under the Approach Used in
the Risk Assessment and an Alternative Approach
Health Effect
Percent with blood-lead
concentration of at least 10//g/dL
Percent with blood-lead
concentration of at least 25 //g/dL
Percent with an IQ score less than
70
Percent with an IQ score decrement
of greater than 1
Percent with an IQ score decrement
of greater than 2
Percent with an IQ score decrement
of greater than 3
Average IQ Loss
(Standard Deviation)
Geometric Mean Blood-Lead
Concentration (Geometric Standard
Deviation)
Approach Used in the Risk
Assessment
Pre-lnt.
10.5
0.578
0.566
52.2
18.2
7.09
1.35
(1.11)
4.05
(2.06)
Post-
Int.
7.49
0.249
0.526
48.7
14.3
4.73
1.23
(0.92)
3.81
(1.95)
%
Change
-28.6
-56.9
-7.19
-6.59
-21.4
-33.2
-9.15
-5.87
Alternative Approach
Pre-lnt.
6.65
0.117
0.520
53.5
14.0
3.90
1.26
(0.82)
4.10
(1.81)
Post-lnt.
3.97
0.029
0.489
49.3
9.80
2.07
1.15
(0.67)
3.86
(1.72)
%
Change
-40.4
-75.2
-5.84
-7.70
-30.0
-46.9
-8.56
-5.87
This analysis assumes the following environmental-lead standards determine whether or not a particular intervention is
performed in a housing unit:
• Dust-lead loading (under wipe techniques) of 200 pg/ft2 for floors and 500 //g/ft2 for window sills for dust
cleaning
• Soil-lead concentration of 400 //g/g for soil cover and 3000 i/glg for soil removal
• 5 ft2 of deteriorated lead-based paint for paint repair and 20 ft2 for paint removal
Approach #2: Alternative approach to determining a post-intervention blood-lead distribution
using directly-measured blood lead changes
An alternative approach to estimating a post-§403 national distribution of blood-lead
concentrations for 1997 children aged 12 to 35 months (1 to 2 years) assumed an estimated
33% decline in blood-lead concentration following an intervention, based on results of
published intervention efficacy studies and information on mobilized bone-lead stores (see
Section 5.4.1.6). This alternative estimate of primary prevention effectiveness is denoted the
'adjusted blood-lead effects model', and was compared to post-§403 distribution using the
IEUBK model and the HUD National Survey data.
Draft - Do Not Cite or Quote
201
September 27, 1996
-------
In this approach, the summarized environmental-lead levels within each National
Survey unit (Table C-7) were compared to the following to determine necessary intervention
strategies:
• 100 (ig/ft2 as an interior floor dust-lead loading and 500 fig/ft2 as an interior
window sill dust-lead loading;
• 400 ug/g as a soil-lead concentration for soil cover and 3000 ug/g for soil
removal; and
• 0 ft2 of deteriorated lead-based paint.
(Note that other analyses within the sensitivity analysis considered a floor dust-lead loading
standard of 200 ug/ft2).
Table 5-18 summarizes the pre-§403 blood-lead distribution and the post-§403
distribution generated under the adjusted blood-lead effects model. This table also presents the
post-§403 blood-lead distribution generated in Section 5.3 for candidate standards of 100 (ig/ft2
for floor dust, 500 ug/ft2 for window sill dust, 400 ppm for soil cover, 3000 ppm for soil
removal, 0 ft2 for paint maintenance, and 20 ft2 for paint abatement. The table presents the
geometric mean and geometric standard deviation of the distributions, and the probabilities of
exceeding certain concentration threshold values.
According to Table 5-18, the post-intervention geometric mean blood-lead
concentration under both the adjusted blood-lead effects model and the post-§403 risk
assessment method was 11% lower than the pre-intervention geometric mean. Also, as
compared to the adjusted blood-lead effects model, a smaller percentage of children exceeded
the various blood-lead levels under the approach used in the risk assessment.
Figure 5-8 contains a plot of the three blood-lead distributions documented in Table
5-18. As noted in the accompanying legend, distinct line types are utilized for each of the three
distributions (e.g., the solid line denotes the pre-intervention distribution). The difference
between the geometric standard deviation estimated for both post-§403 approaches is evident
in this figure. The adjusted blood-lead effects model suggests a wider post-§403 distribution
than does the post-§403 risk assessment method.
Draft - Do Not Cite or Quote 202 September 27, 1996
-------
o
s
i
Table 5-18. Estimated Distribution of Post-§403 Blood-Lead Concentrations for Children 1-2 Years Old Based on the
IEUBK Model for Both the Risk Assessment and the Adjusted Blood-Lead Effects Model Approach
Distribution
Estimation
Procedure
Baseline
(Pre-§403)
Post-§403
Under the
Adjusted Blood
Lead Effects
Model1
Post-§403
Under the Risk
Assessment
Method'
Geometric
Mean
PbB
4.05
3.60
3.60
Geometric
Std. Dev.
PbB
2.06
1.89
1.75
%with
PbB > 5
W/dL
38.5
30.2
27.8
% with PbB
> 10//g/dL
10.5
5.4
3.4
%with
PbB > 15
//g/dL
3.5
1.2
0.5
% with PbB
> 20//g/dL
1.3
0.3
0.1
% with PbB
> 25//g/dL
0.6
0.1
0.03
IS)
O
u
1 Based on the IEUBK model with the following options for standards: 100/yg/ft2 for floor dust-lead loading; 500/yg/ft2 for window sill
dust-lead loading; 400 //g/g for soil cover; 3000 pg/g for soil removal; and 0 ft2 of deteriorated lead-based paint for paint
maintenance.
-------
!
i
NO
>J
0.290-
0.258-
0.226-
0.193-
|- 0.161-
o
I 0.129-
0.097^
0.064
0.032-
0.000-
0 2
NHANES III
— — Adjusted Blood Lead Effects Model
—— — Post-403 Risk Assessment Method
8 10 12 14 16 18 20 22 24 26 28 30 32
Blood-Lead Concentration (/.ig/dL)
Figure 5-8. Comparison of NHANES III Blood-Lead Concentration Distribution to Distributions Estimated Using the
Adjusted Blood Lead Effects Model and the Post-§403 Risk Assessment Method
-------
6.0 CONFIRMATION STUDIES FOR IEUBK MODEL
CHAPTER 6 SUMMARY
Chapter 6 evaluates the IEUBK model for its implementation in the Risk
Assessment by using relevant data from 3 studies: NHANES III / HUD National
Survey, Rochester Lead-in-Dust study, and Baltimore Repair and Maintenance
study. Model performance is evaluated by comparing the geometric mean of the
model-predicted blood-lead levels across all the children to the observed geometric
mean blood-lead level. The proportions of children exceeding threshold
concentrations, such as 10 ug/dL, 15 ug/dL, 20 ug/dL, and 25 ug/dL are also
examined. The differences between observed and model-predicted percentages of
children having high blood-lead levels ranged from 0 to 1% when using national
data from the NHANES HI / HUD National Survey study. Results confirm IEUBK
model's ability to predict a national distribution of blood-lead levels based on a
national distribution of environmental lead levels.
As described in Chapter 4, the IEUBK model is a biological simulation model designed
to predict the probability of elevated blood-lead levels in children. For the §403 Risk
Assessment, the IEUBK model (version 0.99D) was applied to the environmental-lead
measurements collected in the HUD National Survey to estimate a national distribution of
blood-lead levels in children, assuming promulgation of the §403 health-based standards. The
model predicts children's blood-lead levels using information on their multimedia exposure to
environmental lead. The model has been developed using data from many different scientific
studies of lead biokinetics, contact rates of children exposed to contaminated media, and data
on the presence of environmental lead in residences. As part of applying this model to the
§403 Risk Assessment, it is important to understand the extent to which model predictions are
supported by comparisons with real-world data.
This chapter presents a comparison of IEUBK model predictions against data from
relevant epidemiological studies:
Phase I of the Third National Health and Nutrition Examination Survey
(NHANES III) in combination with the HUD National Survey (Brody, et al., 1994;
EPA, I995a);
Draft - Do Not Cite or Quote 205 September 27, 1996
-------
• Rochester Study (HUD, 1995a);
• Baltimore R&M Study (EPA, 1994a).
As stated in Chapters 3 and 5, the NHANES III Survey and the HUD National Survey are
the fundamental sources of baseline data for the §403 Risk Assessment. The NHANES III
Survey was used to estimate a baseline national distribution of blood-lead concentrations in
children. The HUD National Survey (the only national survey of environmental lead levels)
was employed to estimate a distribution of environmental lead levels, which were then inputted
to the IEUBK model to predict a national distribution of children's blood-lead levels.
Comparison of model-predicted blood-lead levels based on the HUD National Survey data
versus the baseline levels from the NHANES III Survey data provides a basis for a general
assessment of model predictions.
The Rochester and Baltimore R&M studies are blood-lead studies that provide useful
information from urban residential communities over a range of environmental conditions. A
comparison of IEUBK model predictions against data in these studies is relevant to the
examination of model performance for urban lead sources. However, there are some
difficulties in using Rochester and Baltimore R&M studies in evaluating IEUBK model:
• Environmental measurements in Rochester and Baltimore may not have been
representative of a child's cumulative lead exposure. Given representative
measurements of all relevant lead sources at a residence, the model estimates a
plausible distribution of blood-lead levels associated with full-time exposure to those
measured sources. Environmental measurements collected in childhood lead
exposure studies are usually selected in a systematic fashion across residences. This
may involve looking for areas containing high lead levels (dripline soil or window
wells), or surfaces with ample medium to sample (sides of rooms where more dust
may collect). While such samples may indeed represent real hazards to some
children, they can be over-represented in exposure assessments.
• Model predictions were based on a generalized activity assessment. Residence-
specific soil- and dust-lead concentrations should come from areas where children are
most likely to play or spend their time and water-lead concentrations should reflect
the water consumed by the children. Because environmental lead levels can vary
considerably within a child's exposure unit and children vary in their activities, it
would be preferable to weight individual dust or soil measurements by information on
Draft - Do Not Cite or Quote 206 September 27, 1996
-------
childhood activity patterns. However, this information is not available. The
particular environmental measures (average of play area and bedroom dust, and
bare/play area fine fraction soil from the Rochester study and composite floor dust
and dripline soil from the Baltimore R&M study) were evaluated because they may
represent typical locations where most children are exposed to lead. This generalized
approach to activity assessment will overestimate exposure for some children and
underestimate it for others, and therefore, predictions of blood-lead concentrations are
best assessed on an overall group basis.
• Variability in environmental lead samples. House dust, in particular, has substantial
spatial variability and temporal variability. These sources of measurement error are
difficult to assess quantitatively without more extensive data. In the Rochester data
set, bedroom and play area dust-lead concentrations differed by at least 200 ppm for
50% of the residences. Although the selected model inputs of environmental
exposure may represent the average lead exposure for an average child, there is
considerable variability in their application to individual children.
• Relation between blood and environmental lead in Rochester and Baltimore may be
biased due to study selection criteria. In Rochester, children in lower income
families living in older homes were purposely over-sampled. Children were excluded
if 1) medical treatment for an elevated blood-lead level or an environmental
intervention was conducted 2) child had taken an iron supplement in the past 2
months, 3) any major renovation of residence occurred during past 12 months, or 4)
an adult employed in an industry or involved in a hobby that would expose to lead
lived in the household. Of 1536 families who were interviewed, about 75% were
considered ineligible by these criteria. On one hand, the Rochester study may have
targeted older homes with the potential for containing higher concentrations of lead in
dust and soil. On the other hand, the selection criteria for children may have
eliminated children with the potential for higher blood lead. Therefore, the study
design may have undersampled children with a potential for both higher lead
exposures and blood-lead concentrations. Therefore, it is not unexpected that the
IEUBK model predicted a larger number of children with higher blood-lead
concentrations compared to that observed in the study. In Baltimore, children were
included because their homes were chosen for intervention or for control/baseline
measurements. This design comes closer to a control versus treatment framework
than a random sampling approach.
The validation approach used is the same as that recommended in an EPA guidance
document, Validation Strategy for the Integrated Exposure Uptake Biokinetic Model for Lead
in Children (EPA, 1994b). EPA developed this document to outline a working strategy for
conducting IEUBK model empirical comparisons which includes the data requirements for
generating and interpreting model predictions. Specifically, three steps are involved:
Draft - Do Not Cite or Quote 207 September 27, 1996
-------
1. Identify the children that have adequate data to characterize exposure or dose
(approximately). This requires a review and the assessment of available data on each
child relative to observed measurements on lead in dust, soil, and water.
2. Calculate the inputs to the IEUBK model. Because the model requires single
estimates of lead concentrations in dust, soil, and water for each child, multiple
measurements available at the target locations for a medium need to be combined into
a composite measurement.
3. Compare the model predictions of central values (e.g., geometric mean) and
population percentiles (e.g., percent above 10 ug/dL) to those observed in the studies.
These comparisons are made across all the children and for subsets of children
according to variables (e.g., behavioral) that may influence or qualify their exposure
characterization.
Descriptive measures used to evaluate model performance and the criteria for selecting data
from each study for IEUBK model input are described in Section 6.1. Section 6.2 presents the
results of comparing observed blood-lead levels to IEUBK model predictions. Conclusions are
drawn in Section 6.3.
6.1 METHODS
6.1.1 Descriptive Measures
For each of the studies, model performance is evaluated by comparing the geometric
mean of the model-predicted blood-lead levels across all the children to the observed geometric
mean blood-lead level. In addition, the proportions of children exceeding threshold
concentrations, such as, 10 ug/dL, 15 ug/dL, 20 ug/dL, and 25 ug/dL are examined.
While the observed proportion of children exceeding a threshold is estimated directly
from the actual blood-lead levels for all studies, the procedure used to calculate the predicted
proportion for the HUD National Survey is different from that used for the other two studies.
For the HUD National Survey, the procedure is the same as that described in Chapter 5. This
procedure takes into account the HUD National Survey weights which represent the number of
children in the nation associated with a specific set of environmental lead measurements. For
each study, the predicted proportion is estimated by calculating the exceedance probability for
Draft - Do Not Cite or Quote 208 September 27, 1996
-------
each hypothetical child, and averaging them over the entire group. The exceedance probability
for each hypothetical child is calculated using
f [ln(Tc)-ln(GM)])
Prob Z >
( ln(GSD) J
where Z is a standard normal variate, Tc is a threshold concentration, GM is the predicted
blood-lead level, GSD is the geometric standard deviation assumed equal to the IEUBK default
value of 1.6, and In is the natural logarithm. The GSD pertains to the inter-individual and
biological variability in the blood-lead levels of children exposed to similar environmental lead
levels. The IEUBK default value of 1.6 was estimated by EPA using several epidemiological
studies, as described in the Guidance Manual (EPA, 1994a).
Predicted blood-lead levels are also compared to the observed levels graphically. For the
NHANES III Survey, histograms and probability density functions are used to compare the
observed and predicted distributions of blood-lead levels. For the Rochester and Baltimore
R&M studies, scatter plots are used to display the observed and predicted individual blood-lead
levels in combination with the line of perfect agreement and the 95% prediction intervals based
on the model. The prediction intervals represent ranges which are intended to encompass with
95% confidence the true blood-lead levels of children exposed to similar environmental levels.
6.1.2 Input Data Selection
The IEUBK model is a dose-response model and therefore requires environmental lead
levels that represent children's typical lead exposure or dose as inputs. As none of the
studies was designed to measure children's exposure, the exposure information available varied
across children. Therefore, subsets of children that had data judged to reasonably characterize
exposure were selected from the Rochester and R&M studies. As recommended in both the
Guidance Manual and Validation Strategy, only those children whose exposure to lead is
relatively well characterized were included hi the analysis. Children without dust-, soil-, or
water-lead measurements and those who did not live for at least three months in their residence
prior to blood collection were excluded from the analysis. These children were not expected to
have enough information to predict blood-lead levels reliably. Environmental lead levels
Draft - Do Not Cite or Quote 209 September 27, 1996
-------
corresponding to each child having sufficient exposure data were provided as input to the
model. Note that the small sample size in the input data sets for the Rochester study is because
play yard soil-lead concentration was reported for only approximately 40% of the 205 study
children. Similarly, for the Baltimore R&M study, only one-third of the 163 study children
from the pre-intervention round had dripline soil-lead concentration measurements. Due to the
availability of only dripline soil and BRM (HVS-3 vacuum sampler) dust measurements in the
Baltimore R&M study, IEUBK model predictions may be biased in estimating the lead
exposures. For the HUD National Survey, we used environmental lead levels for each housing
unit in the survey as input to the model. The model was then used to predict the blood-lead
concentration for a hypothetical child of age 24 months in each housing unit.
Rochester
Two different analysis data sets were constructed to analyze the data from Rochester
study based on different groups of children having adequate exposure to lead. These two data
sets were labeled as Input Data Sets A and B.
(1) Input Data Set A
From the Rochester study, only 87 children had play yard fine-sieved soil concentration
measurement. Among those children, 84 children had at least one floor dust-lead concentration
measurement from the bedroom or principal play area collected by the DVM (Dust Vacuum
Method). Of the three dust collection methods used in this study, wipe sampling, DVM, and
BRM, the DVM method was the closest to that used in calibrating the IEUBK model.
Therefore, only 84 of the 205 children in the Rochester study were considered to have
sufficient exposure data and were included in the IEUBK model empirical comparisons. These
children's blood-lead concentration measurements and environmental lead concentration
measurements (dust-lead, soil-lead, water-lead) constitute Input Data Set A.
Dust-lead, soil-lead, and water-lead concentrations were computed or selected as
parameters to be input into the IEUBK model. The model used a composite dust-lead
concentration which was computed as a dust mass-weighted arithmetic mean of the floor
Draft - Do Not Cite or Quote 210 September 27, 1996
-------
measurements from the bedroom floor and principal play area. A weighted average is
meaningful if the precision with which lead is measured increased with large sample masses.
The play yard fine-sieved soil-lead concentration and one-minute flushed water-lead
concentration were also provided as inputs to the IEUBK model. Fine-sieved soil fraction was
selected rather than coarse soil because fine soil is more likely to be ingested by children.
(2) Input Data Set B
Among 87 children who had play yard fine soil-lead concentration measurements, 82
children also had foundation fine soil-lead concentration measurements in the Rochester data
set. Among those, 80 children had at least one floor dust-lead concentration measurement from
all sampled indoor locations (bedroom, principal play area, living room, kitchen, and
entryway). These 80 children make up Input Data Set B.
The composite dust-lead concentration was computed as a dust mass-weighted arithmetic
mean of the floor measurements from all sampled indoor locations. An arithmetic average of
play yard find soil-lead concentration and foundation fine soil-lead concentration was
calculated and used in the IEUBK model. One-minute flushed water lead concentration also
provided input to the lead model computation.
Baltimore R&M
Similar to the Rochester study, two data sets were constructed to analyze the data from
Baltimore R&M study. These two data sets were labeled as Input Data Sets C and D.
(1) Input Data Set C
Only 54 children from the Baltimore R&M study had dripline soil-lead concentration
measurements. Among those, 7 children were excluded from the IEUBK model empirical
comparisons due to children living in the current residence less than three months or because
the surveyed houses were vacant at the time of measurement collection. Therefore, only 47 of
the 163 children from the pre-intervention round in the Baltimore R&M study who had blood-
lead concentration measurements were considered to have sufficient exposure data and were
used in the IEUBK model comparison. These 47 children constitute Data Set C in Table 6-1.
Draft - Do Not Cite or Quote 211 September 27, 1996
-------
Dripline soil-lead concentration measurements were used in the IEUBK model validation
due to lack of other data representing outdoor lead exposure. Almost all soil samples were
collected from the driplines in the pre-intervention round for this study. Floor dust samples
collected by the BRM method were composited over multiple rooms and a single composite
floor dust-lead concentration based on all samples was computed using a dust mass-weighted
arithmetic mean for the input to the model. The two-hour stagnation water-lead concentration
was also used in the IEUBK model.
(2) Input Data Set D
Input Data Set D was constructed using the 47 children from Input Data Set C. The
composited dust-lead concentration employed hi Input Data Set D included entryway dust-lead
concentration together with floor dust-lead concentration. The composite dust-lead
concentration was calculated using a dust mass-weighted arithmetic mean. The dripline soil-
lead concentration and the two-hour stagnation water-lead concentration were also included in
the Input Data Set D.
NHANES III/HUD
Soil-lead concentrations and dust-lead concentrations were generated from the HUD
National Survey database for use in calculating blood-lead levels by the IEUBK model.
Because water-lead measurements were not reported in the survey, default values assigned by
the IEUBK model were used (Table 4-1). Soil samples collected at the dripline, entryway, and
a remote location were used to calculate a weighted soil-lead concentration average for each
house. The soil-lead concentration average was calculated by giving the remote location
sample twice the weight as the weight given to each of the samples collected near the house.
Floor dust samples were used to calculate a dust mass-weighted average of dust-lead
concentrations from each house. The weighting of the samples in the average was based on the
reported tap weights of the samples.
Draft - Do Not Cite or Quote 212 September 27, 1996
-------
6.2 RESULTS
The IEUBK model performance was analyzed using the input data sets containing soil,
dust, and water lead concentrations for children with adequate exposure data as described in
Section 6.1. Default values defined by the model were used for all other required data in these
runs. Table 6-1 presents the observed and predicted geometric mean blood-lead levels and
proportions of children exceeding 10 |ig/dL, 15 ug/dL, 20 ug/dL, and 25 ug/dL for each of the
studies. The geometric mean blood-lead levels and exceedance proportions for each input data
set were calculated by combining the data across all the children in that data set. Results of the
comparison are separately discussed below for each study.
Rochester
As indicated in Table 6-1, the observed geometric mean blood-lead level for Input Data
Set A is 6.3 (ig/dL, while the model predicted geometric mean blood-lead level is 6.4 |ig/dL.
Thus, utilizing soil-lead levels from the play yard and dust-lead levels from the bedroom and
principal play area as inputs to the IEUBK model results in predicted blood-lead levels that on
the average agree with the observed blood-lead levels. The observed and predicted proportions
of children exceeding 10 ug/dL in blood-lead, when using Input Data Set A, are 17% and 26%,
respectively; the observed and predicted proportions of children exceeding 25 |ig/dL are 1%
and 6%, respectively.
The results of using Input Data Set B show that the observed geometric mean blood-lead
level is 6.3 ng/dL and the predicted geometric mean blood-lead level is 9.2 ug/dL. The
observed and predicted proportions of children exceeding 10 |ig/dL in blood-lead are 18% and
44%, respectively; the observed and predicted proportions of children exceeding 25 ng/dL are
1% and 11%, respectively.
For both input data sets, the predicted proportions of children exceeding 10 ug/dL, 15
ug/dL, 20 ug/dL, and 25 ug/dL in blood-lead are all higher than the corresponding observed
proportions. Figure 6-1 presents a scatter plot of the observed blood-lead levels against model
predictions for the same set of children from Input Data Set A. Each symbol in the figure
represents an individual child. Note from the figure that 83% of the observed blood-lead levels
Draft - Do Not Cite or Quote 213 September 27, 1996
-------
Table 6-1. Comparison of Observed and IEUBK Model Predicted Blood-Lead Levels
Study
NHANES
HI/HUD
Input
Data
Set
A
B
C
0
No. of
Children
84
80
47
47
5,272.068
(NHANES)
7.960.614
(HUD)
Geometric Mean Blood-lead (pg/dU
Observed
6.3
6.3
6.6
6.6
4.1
(GSD-2.1)
Predicted
6.4
9.2
9.1
9.7
3.9
(GSD-2.2)
Relative
Difference
(Prad-
Obs)/0bs
0%
46%
38%
46%
6%
Percent of
Blood-Lead
> 10jig/dl
Observed
17%
18%
34%
34%
11%
Predicted
26%
44%
53%
54%
12%
Percent of
Blood-lead
> 15/ig/dl
Observed
2%
3%
17%
17%
4%
Predicted
14%
26%
35%
38%
5%
Percent of
Blood-teed
>20«ig/dl
Observed
2%
3%
9%
9%
1%
Predicted
8%
17%
21%
25%
2%
Percent of
Blood-lesd
> 26/ig/dl
Observed
1%
1%
6%
6%
1%
Predicted
6%
11%
12%
17%
1%
10
Nl
N
(0
-------
lie within the 95% prediction intervals, and there is a pattern for those outside of the 95%
prediction intervals. In six of the seven pairs where the observed blood-lead level is greater
than the upper prediction bound, predicted blood-lead levels were less than 10 u,g/dL.
Similarly, for four of the seven pairs where the observed blood-lead level is less than the lower
prediction bound, predicted blood-lead levels were greater than 10 ug/dL.
100
o>
3
c
o
o
o
o
T>
O
o>
TJ
O
m
a>
(A
-a
O
10
10
IEUBK Predicted Blood - Lead Concentration (/zg/dL)
o o o Hours Away From Home Not Reported
o o o Never Away From Home
+ + + Less Than 10 Hours/Week Away From Home
• • • Greater Than 10 Hours/Week Away From Home
95% Prediction Intervals for PbB Based on Assumed GSD of 1.6
Line of Observed and Predicted Concentrations in Agreement
100
Dust = DVM Bedroom and Play Area floor (Weighted Average)
Soil = Play Yard Fine
Number of Children = 84
Figure 6-1. Comparison of Predicted and Observed Blood-Lead Levels for the Rochester
Study.
(Symbols are Coded for Time Away from Home in Hours/Week)
Draft - Do Not Cite or Quote
215
September 27, 1996
-------
Baltimore R&M
When using Input Data Set C, the observed geometric mean blood-lead level is 6.6
ug/dL, compared to the predicted geometric mean blood-lead level of 9.1 ug/dL. The
observed and predicted proportions of children exceeding 10 ug/dL in blood-lead are 34% and
53%, respectively; the observed and predicted proportions of children exceeding 25 ug/dL are
6% and 12%, respectively.
For Input Data Set D, the observed and predicted geometric mean blood-lead levels are
6.6 ug/dL and 9.7 ug/dL respectively. The observed and predicted proportions of children
exceeding 10 ug/dL in blood-lead are 34% and 54%, respectively; the observed and predicted
proportions of children exceeding 25 ug/dL are 6% and 17%, respectively.
As before, the predicted proportions are higher than the observed proportions. Figure 6-2
displays the observed blood-lead levels plotted against the predicted blood-lead levels from
Input Data Set C with 95% prediction intervals. More than 80% of the observed blood-lead
levels lie within the 95% prediction intervals. Note that in one of the three pairs where the
observed blood-lead level is greater than the upper prediction bound, predicted blood-lead
levels were less than 10 ug/dL. Similarly, for five of the six pairs where the observed blood-
lead level is less than the lower prediction bound, predicted blood-lead levels were greater than
10 ug/dL. Also note that a higher percentage of children in the Control Modem Urban homes
are bracketed by the prediction limits. These children's blood-lead levels span the range of
blood-lead levels which §403 aims to achieve.
NHANES III/HUD
Table 6-1 shows that the predicted geometric mean blood-lead level is 3.9 ug/dL using
IEUBK/HUD data, and the observed value is 4.1 ug/dL for the NHANES ffl data. The
predicted and observed geometric standard deviations are 2.2 and 2.1, respectively. The
observed and predicted percentages of children having blood-lead levels exceeding 10 ug/dL
are 11% and 12%, respectively; and the observed and predicted proportions of children
exceeding 25 iig/dL, are both 1%. Figure 6-3 displays the distribution of blood-lead levels for
Draft - Do Not Cite or Quote 216 September 27, 1996
-------
children aged 1-2 years as reported in NHANES III. A smooth probability density function for
blood-lead levels is overlaid on the histogram. Similarly, Figure 6-4 displays the
corresponding projected distribution of blood-lead concentrations for children aged 1-2 years
based on the IEUBK model. In Figure 6-5, the probability density functions based on
NHANES III and the IEUBK model are presented simultaneously for purposes of easier
comparison. These two curves are comparable.
100
0)
3
o
o
§
u
•o
o
•o
o
CD
0>
in
-Q
O
10
10 100
IEUBK Predicted Blood - Lead Concentration
+ + + Control Modern Urban Home
ooo Control Previously Abated Home
o a a R&M Home
95% Prediction Intervals for PbB Based on Assumed GSD of 1.6
Line of Observed and Predicted Concentrations in Agreement
Dust = Floor (Weighted Average)
Soil = Dripline
Number of Children = 47
Figure 6-2. Comparison of Predicted and Observed Blood-Lead Levels for the Baltimore
R&M Study.
(Symbols are Coded for Study Group)
Draft - Do Not Cite or Quote
217
September 27, 1996
-------
25-r
20-
15 -
o
5 -
T
0 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
Blood-Lead Concentration (/.tg/dL)
Figure 6-3. Distribution of Blood-Lead Concentrations (//g/dL) for NHANES III (1-2 Year
Old Children).
Draft - Do Not Cite or Quote
218
September 27, 1996
-------
25 i
20
15
c
0>
0)
Q,
10
5 6
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
Blood-Lead Concentration (/zg/dL)
Figure 6-4. Distribution of Blood-Lead Concentrations (//g/dL) for HUD National Survey
Data (1-2 Year Old Children) Predicted from IEUBK Model.
Draft - Do NOT Cite or Quote
219
September 27, 1996
-------
20
15-
c
o
o
c
c
0)
o
10 -
5-
0-
Blood-Leod Distribution
Nhanes III Pre-lntervention
IEUBK/HUD Pre-lntervention
l l 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 [—
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
Blood-Lead Concentration
Figure 6-5. Distribution of Blood-Lead Concentrations (//g/dL) Based on NHANES III and
IEUBK Model (1-2 Year Old Children).
Draft - Do Not Cite or Quote
220
September 27, 1996
-------
6.3 CONCLUSIONS
The IEUBK model-predicted percentages of children exceeding 10 ug/dL, 15 ug/dL, 20
ug/dL, and 25 ug/dL are consistently higher than the observed percentages. Observed and
Predicted percentages of children having blood-lead levels exceeding 25 ug/dL are 1% vs 6%
for Rochester study Input Data Set A; 1% vs 11% for Rochester Input Data Set B; 6% vs 12%
for Baltimore R&M study Input Data Set C; 6% vs 17% for Baltimore R&M study Input Data
Set D; and 1% vs 1% for NHANESIII/HUD.
Differences between observed and model-predicted geometric means ranged from 1.6 to
47%. Observed and model predicted geometric means are 6.3 and 6.4 ug/dL for Rochester
study Input Data Set A; 6.3 and 9.2 ug/dL for Rochester study Input Data Set B; 6.6 and 9.1
ug/dL for Baltimore R&M study Input Data Set C; 6.6 and 9.7 ug/dL for Baltimore R&M
study Input Data Set D; and 4.1 and 3.9 ug/dL for NHANES IE/HUD.
As shown in Table 6-1, differences between observed and model-predicted percentages of
children having high blood-lead levels ranged from 0 to 1% when using national data from the
HUD survey and NHANES III. The statistical design employed in the HUD National Survey
produced a sample of environmental lead levels which was representative of actual levels in
U.S. housing. The IEUBK model is designed to predict distributions of blood-lead levels from
environmental lead levels. Results here confirm its ability to predict a national distribution of
blood-lead levels based on a national distribution of environmental lead levels, hi conclusion,
the analyses conducted to compare IEUBK model predicted blood-lead concentrations to those
of observed blood-lead concentrations did not provide any evidence that use of the IEUBK
model in the analyses conducted for §403 is inappropriate.
Draft - Do Not Cite or Quote 221 September 27. 1996
-------
7.0 RISK SUMMARY
This risk assessment documents the scientific basis for setting regulatory standards for
lead-based paint hazards, lead-contaminated dust, and lead-contaminated soil under §403 and
provides estimates of health risk reductions (for children under six) and numbers of children
and housing units affected for various §403 regulatory standards options. This risk assessment
does not include selection of the §403 standards.
Figures included in Chapter 5, plotting quantitative health risks against percentages of
housing units in which interventions are required, illustrate the main conclusions of this report.
Varying options for standards for lead-contaminated paint, dust, and soil from intermediate
values to more stringent values provides diminishing gains in health risk reduction while
requiring interventions at many more housing units. Although health risks continue to decline
as the standards for lead-based paint hazards, lead-contaminated dust, and lead-contaminated
soil are made more stringent, the rate of decline diminished relative to the number of
interventions required. Conversely, standards less stringent than the intermediate options
provide only marginal decreases in the numbers of housing units requiring intervention, but
significantly decrease gains in health risk reduction.
To summarize, the greatest quantitative reduction in health risks is always achieved by
the most stringent standards. However, these reductions require intervention in the largest
number of housing units. At intermediate values for the standards, many of the health risk
endpoints are nearly at their lowest levels, but many fewer housing units are estimated to be
affected by these standards in comparison to the number of units affected by the more stringent
standards.
Sensitivity analyses are performed in this risk assessment to gauge the robustness of the
risk analysis methodology employed. These results suggest relative insensitivity to some
changes in some assumptions, but strong dependence on other assumptions. The applicability
of this risk assessment to risk managers making policy decisions to set appropriate §403
standards hinges on the robustness of the risk assessment methodology.
Section 7.1 includes a summary of the scientific evidence, presented in Chapters 2, 3,
and 4, identifying the need for §403 standards. Tools developed for implementation of the risk
Draft - Do Not Cite or Quote 222 September 27, 1996
-------
assessment methodology are also summarized in this section. Section 7.2 provides conclusions
on the potential health risk reductions predicted by this risk assessment and numbers of
children and housing units affected for various standards options. Comments on the robustness
of the methodology employed are made in Section 7.3.
7.1 SCIENTIFIC BASIS FOR §403 AND TOOLS FOR THE RISK ASSESSMENT
The toxicologic effects of lead exposure on young children were documented based on
numerous studies reported in the scientific literature. Typically, in studies assessing adverse
health effects associated with lead exposure, relationships between health effects and exposure
are established using a measure of internal rather than external exposure. Blood-lead
concentration is the most common measure of internal exposure.
Though lead causes a wide array of adverse health effects, particularly at high dose
levels, lead is best known for its adverse effects on the central nervous system. IQ score
decrements and incidence of IQ scores less than 70 due to lead exposure were examined as
health endpoints in this risk assessment. As surrogates for the wide array of other, non-IQ
related health risks to both the central nervous system and other organs, incidence of elevated
blood-lead concentrations were estimated for specified thresholds.
Children aged 1-2 were targeted for estimation of health risks in this risk assessment for
two reasons. The first is related to the increased vulnerability of this age group due to their
rapidly developing central nervous system. One to two year old children may be the most
appropriate age group for relating blood-lead concentrations to adverse IQ effects. This is
documented in Chapter 2. Second, both the normal hand-to-mouth activities of this age group
and pica tendencies observed in some children may put children aged 1-2 most at risk to lead
exposure.
It was shown, based on a number of lead studies, that elevated lead levels continue to
exist in residential environments (Chapter 3) providing an on-going threat of childhood lead
exposure. Lead is particularly a threat in older homes. The HUD National Survey, discussed
in detail in Section 3.3, provides nationally representative estimates of numbers of homes
exceeding certain exposure levels. For instance, 10.4% and 6.4% of all homes are predicted to
Draft - Do Not Cite or Quote 223 September 27, 1996
-------
have some deteriorated lead-based paint in the exterior and interior, respectively. At least 10%
of homes are estimated to have lead loadings above the Interim Guidance standards.
NHANES III data indicate that concentrations of lead in children's blood are still high.
Section 3.4 documents these elevated levels. There is also evidence in the existing scientific
literature of a dose-response relationship between residential environmental lead (external
exposure) and elevated blood-lead concentrations of resident children (internal exposure).
Section 3.2 documents the overwhelming evidence from epidemiologic studies on the existence
of a positive relationship between environmental-lead levels and blood-lead concentrations.
In this risk assessment, the connection between environmental-lead levels and adverse
health effects (dose-response) was estimated in two steps because there is little scientific data
for estimating this relationship directly. First blood-lead concentrations were estimated based
on environmental-lead levels, and then health risks were estimated from those blood-lead
concentrations. This was necessary because the majority of the existing scientific evidence for
the relationship between lead exposure and adverse health effects was available in this form.
Two approaches were taken to mapping environmental-lead levels to blood-lead
concentrations, an EPI model and the IEUBK model. The EPI model was constructed based
on data from the Rochester study. This study is presented in Section 3.2, and the EPI model is
discussed in Section 4.1. The EPI and IEUBK models reflect qualitatively different
relationships between environmental-lead levels and blood-lead concentrations. Estimating
blood-lead concentrations (and thus adverse health effects) using these two different models
strengthens the analysis by providing two separate estimates of health risk reductions
associated with §4.03.
Analyses were conducted to select appropriate ranges of options for the §403 standards.
The selected ranges are presented in Table 4.6.
Decisions regarding the appropriate interventions for §403 and the efficacy of these
interventions are documented in Section 5.2. In this risk assessment, in which potential health
risk reductions achievable under various standards options are evaluated, interventions are
assumed to occur whenever a housing unit exceeds a proposed standard. One dust intervention
Draft - Do Not Cite or Quote 224 September 27, 1996
-------
was proposed; two levels of soil intervention; and two levels of exterior and interior paint
intervention. Abatement efficacy assumptions are detailed in Table 5.3.
7.2 HEALTH RISK REDUCTIONS AND NUMBERS OF CHILDREN AND HOUSING UNITS
EFFECTED
Risk comparisons between baseline and predicted post-§403 values of seven health
endpoints (IQ related health effects and elevated blood-lead concentration incidence) for 1997
were conducted for each set of standards evaluated. Baseline health risks were estimated using
NHANES III. Post-§403 health endpoints were estimated by modeling blood-lead
concentrations, both before and after interventions triggered by §403, and applying that
difference to the estimated baseline blood-lead concentration distribution.
This summary focuses on four of the health endpoints: blood-lead concentrations greater
than 25 |ig/dL or 10 ng/dL and IQ score decrements of greater than 2 or 3 points. The pre-
§403 (baseline) health risks, as quantified by these four measures, were presented in Table 5-1.
Over 10% of children aged 1-2 years are predicted to have blood-lead concentrations over 10
ug/dL. 10 ug/dL is the CDC level of community concern. These estimates represent current
predicted health risks for the United States in 1997.
Chapter 5 presents post-§403 adverse health effects estimates for a wide array of
standards options. First, results for varying dust-lead standard options are presented with soil
and paint standards held fixed at central values. Likewise results for different soil and paint
standard options are presented with the other media held fixed at central values. Lastly, post-
§403 adverse health effects estimates are presented for different combinations of options. In
this case, the standards for all media are simultaneously varied for lead-contaminated paint,
dust, and soil from options at the upper end of each media's range down to the more stringent
end. This summary focuses on this last set of results (Table 5-10 and Figures 5-6a and 5-6b).
The IEUBK model predicts larger reductions in adverse health effects due to
promulgation of §403 than the EPI model in all cases. The IEUBK model also predicts greater
incremental reduction in health risks across the range of standards.
The characterization of post-§403 health risks (particularly those based on the IEUBK
model) and the numbers of housing units requiring interventions, as plotted in Figures 5-6a and
Draft - Do Not Cite or Quote 225 September 27, 1996
-------
5-6b, are well described by a piecewise linear function with two pieces; the first piece is
steeper than the second. In summarizing the results of the integrated risk assessment, these
conclusions focus on the intermediate set of standards, where the two linear pieces meet.
Performance of the intermediate set of standards is compared to performance of sets of
standards at the most stringent and least stringent end of the range of options.
The incremental reductions in health risks per number of housing units requiring
interventions between the least stringent set of standards and the intermediate set is greater than
the incremental reductions between the intermediate and most stringent set of standards.
Table 7-1 summarizes results for these three selected sets of standards, both in terms of
the numbers of interventions that would be required by each set of standards, and the resulting
risk reductions. At the intermediate set of standards, 26% of all housing units require at least
one of the interventions considered. Less than 1% of all housing units need costly Soil
Removal interventions with an additional 13% requiring Soil Cover interventions; 6% need
costly Exterior Paint Abatement interventions with an additional 2% requiring Exterior Paint
Maintenance and; 2% require the more intensive treatment of Interior Paint Abatement while
less than 1% require Interior Paint Maintenance.
The bottom portion of Table 7-1 describes the estimated health risk reductions anticipated
after promulgation of §403 based on the selected standards. Results are presented as the
percentage decline in numbers of children affected by each health endpoint. According to the
EPI model only half as many children (20,000 as compared to 46,000) would have blood-lead
concentrations greater than 25 ug/dL if the standards are set at the intermediate levels identified
earlier. The reduction is even greater for the IEUBK model, 95% (2,500 as compared to
46,000). Similarly, for the EPI model, elevated blood-lead concentrations above 10 |ig/dL
would be reduced from 834,000 to 600,000 or approximately 28%. The reduction is predicted
to be even greater using the IEUBK model, with approximately 548,000 (=66%) fewer children
above 10 ^ig/dL.
To summarize, the EPI model predicts 57% declines in blood-lead concentrations above
25 ng/dL and 28% declines in blood-lead concentrations above 10 ug/dL. Based on the
IEUBK model, declines of 95% and 67% are predicted. According to the EPI model, the
number (1.45 million) of IQ decrements greater than 2 would be reduced by 23% with the
Draft - Do Not Cite or Quote 226 September 27. 1996
-------
Table 7-1. Percent of Housing Units Requiring §403 Interventions and Estimated
Reduction from Baseline Health Risks Due to §403 for Three Sets of
Standards Options
Intervention
Total
Soil
Interior
Paint
Exterior
Paint
Dust
Cover
Removal
Maintenance
Abatement
Maintenance
Abatement
Cleaning
Health Effects
Blood-Lead Concentration
Exceeding 10pg/dL <%)
Blood-Lead Concentration
Exceeding 25 /ig/dL (%)
IQ Decrement Exceeding 2
(%)
IQ Decrement Exceeding 3
(%)
Percent of Housing Requiring Intervention
for Three Sets of §403 Standards
Least Stringent1
19
3
<0.5
3
<0.5
4
3
13
Intermediate2
26
13
1
3
2
3
6
18
Most Stringent 3
74
43
6
1
5
1
9
64
Estimated Percent Reduction in Risk from Baseline
EPI
45
21
18
25
IEUBK
72
38
29
44
EPI
57
28
23
34
IEUBK
95
66
52
73
EPI
76
45
34
51
IEUBK
100
90
79
93
1 These standards are floor dust-lead loading 400 //g/ft2, window sill dust-lead loading 800 /ig/ft2, soil cover soil-lead
concentration 1500 //g/g, soil removal soil-lead concentration 5000 //g/g, paint maintenance square footage of
deteriorated lead-based paint 10 ft2, and paint abatement square footage of deteriorated lead-based paint 100 ft2.
2 These standards are floor dust-lead loading 200 //g/ft2, window sill dust-lead loading 500 //g/ft2, soil cover soil-lead
concentration 400 //g/g, soil removal soil-lead concentration 3000 //g/g, paint maintenance square footage of deteriorated
lead-based paint 5 ft2, and paint abatement square footage of deteriorated lead-based paint 20 ft2.
3 These standards are floor dust-lead loading 25 //g/ft2, window sill dust-lead loading 25 //g/ft2, soil cover soil-lead
concentration 50 //g/g, soil removal soil-lead concentration 1000 //g/g, paint maintenance square footage of deteriorated
lead-based paint 0 ft2, and paint abatement square footage of deteriorated lead-based paint 5 ft2.
Draft - Do Not Cite or Quote
227
September 27, 1996
-------
intermediate set of standards; the number of cases of IQ decrements of greater than 3 would be
reduced by 34%. Predicted reductions based on the IEUBK model are 52% and 73% for IQ
decrements of at least 2 and 3, respectively.
Moving to the most stringent standards, slight reductions in health risks can be achieved
compared to the intermediate standards but 74% of housing units would require intervention.
This is a threefold increase in total number of interventions required from the intermediate set
of standards. In comparison, the increased reductions in health risks gained by moving to the
most stringent set of standards are only marginal.
Decreasing to the least stringent standards yields smaller reductions in health risk from
the baseline without greatly decreasing the number of housing units which would be affected.
For example, at the least stringent set of standards, 18.6% of housing units require
interventions. This is nearly three quarters as many interventions as are estimated to be
required at the intermediate standards.
Based on our analysis of health risks reductions, it may be possible to discover some
slight changes to the standards, i.e., shifting the standard for a single media, which further
reduces health risks without increasing numbers of housing units requiring interventions.
However, given the uncertainties associated with estimates of health risk reductions, discussed
in the context of sensitivity analyses in the subsequent section, it is doubtful that this will
generate additional information to aid risk managers in selection of standards.
7.3 ROBUSTNESS OF RISK ASSESSMENT DATA SOURCES AND METHODOLOGY
Several analyses were conducted to assess the sensitivity of the estimated reductions in
risks to the uncertainty in the underlying assumptions and methods utilized in the risk
assessment.
1. Using both the IEUBK and EPI models to predict blood-lead concentrations from
environmental-lead levels.
2. Estimating baseline numbers and percentages of children having specific health
effects for two age groups, 1-2 year olds and 1-5 year olds, using three different
coefficients to quantify the relationship between decline in IQ score and increases in
blood-lead concentration.
Draft - Do Not Cite or Quote 228 September 27, 1996
-------
3. Estimating baseline health effects using both an empirical and a model-based
approach.
4. Estimating numbers and percentage of housing units expected to exceed a 200 jig/ft2
floor dust-lead standard and/or a 500 fig/ft2 window sill dust-lead standard using three
different approaches to converting dust-lead loadings to wipe equivalent dust-lead
loadings.
5. Estimating post-§403 health effects using three different assumptions for the
effectiveness of the Dust Cleaning Intervention and three different assumptions for
the effectiveness of the Soil Cover Intervention.
6. Comparing pre- and post-§403 estimates of health effects using the methodology
employed in the risk assessment and an alternative approach that makes direct
comparison of model-predicted pre- and post-§403 distributions of blood-lead
concentrations (approach #1 in Section 5.4.2.6).
7. Comparing pre- and post-§403 estimates of health and blood-lead effects using the
methodology employed in risk assessment approach which relies on assumptions of
the effectiveness of interventions on environmental-lead levels to an alternative
approach which relies on assumptions of the effectiveness of interventions on
childhood blood-lead concentrations (adjusted blood-lead effects model in Section
5.4.2.6).
It may be concluded, based on the results of the sensitivity analysis presented in Section
5.4.2, that the reductions in health risks predicted in this risk assessment are sensitive to at least
three sources of uncertainty:
1. Uncertainty in the relationship between declines in IQ score and increases in blood-
lead concentration (Section5.4.2.2).
2. Uncertainty in converting vacuum dust-lead loadings to wipe equivalent dust-lead
loadings (Section 5.4.2.4).
3. Uncertainty in the assumed efficacy of the environmental interventions on the
environmental-lead levels (Section 5.4.2.5).
The assumption of an 0.257 decrease in IQ score for an increase of one ng/dL in blood-
lead concentration has considerable impact on the estimates of numbers of children with
specified IQ decrements and average decline in IQ due to lead exposures. However, even if the
Draft - Do Not Cite or Quote 229 September 27, 1996
-------
decline is less severe (0.185 vs. 0.257), approximately 700,000 children 1-2 years old, and 1.3
million children 1-5 years old suffer IQ decrements greater than 2 points due to exposures to
lead-based paint hazards, lead-contaminated dust, and lead-contaminated soil.
The number and percentage of housing units expected to exceed a 200 ug/ft2 floor dust-
lead standard varies considerably among the three conversion approaches taken to estimate
wipe dust-lead loading. Over twice as many housing units' floor dust-lead loadings exceed 200
Ug/ft2 using the "high" alternative as do using the risk assessment approach. Exceedance
proportions for window sill wipe dust-lead loadings are considerably less sensitive. The
impact of uncertainty in the dust-lead loadings conversion on the number of homes requiring
Dust Cleaning interventions based on either the floor or window sill dust standard is much less.
This number ranges from 12.7 million to 16.3 million when the standards are set at 200 (ig/ft2
for floor dust-lead loadings and 500 for window sill dust-lead loadings.
Estimated reductions in health and blood-lead concentration risks are dependent on the
assumed efficacy of interventions performed under §403. Under the alternative efficacies
considered, which varied post-intervention soil-lead concentrations following soil cover
intervention and dust-lead loading following Dust Cleanings, estimates of reductions in health
risks varied considerably. The range of health risk reductions observed in the intervention
effectiveness sensitivity analysis, calculated as percent reduction in numbers of children
affected for each health endpoint, are presented in Table 7-2. These reductions are all
calculated with standards set at the intermediate option for the standards.
Although the results in Table 7-2 indicate that there is considerable variability in the
estimated reduction in health risks due to the uncertainty in the assumed efficacy of the
interventions, the estimated risk reductions are nevertheless considerable and would impact the
health of hundreds of thousands of children, even if the assumed efficacy of the intervention
were decreased by a factor of 2.5.
Two analyses were performed to evaluate the risk assessment methodology. The first
method provides an alternative characterization of the pre-§403 health risks based on mapping
the HUD National Survey environmental-lead levels to a blood-lead concentration distribution
«
rather than using NHANES in to characterize pre-§403 health risks. The second method
Draft - Do Not Cite or Quote 230 September 27, 1996
-------
Table 7-2. Reductions From Baseline Health Risks (in percent) Obtained by Varying
Intervention Effectiveness Due to §403 Intermediate Standards For Selected
Health Endpoints and Risk Assessment Methodology
Health Effects
Blood-lead
concentration
exceeding 25
Blood-lead
concentration
exceeding 10
IQ decrement
exceeding 2
IQ decrement
exceeding 3
Oust and Soil Intervention Effectiveness
Reduced1
EM
41
19
14
22
IEUBK
82
46
34
53
RAZ
EPI
57
28
23
34
IEUBK
95
66
52
73
Enhanced*
EPt
71
38
30
44
IEUBK
99
80
68
87
Direct
Comparison of
Model-Predicted
Distribution4
EPI
75
40
30
47
IEUBK
89
56
44
Adjusted
Blood-Lead
Effects
Model5
IEUBK
83
49
1 Soil Cover intervention is assumed to reduce soil-lead concentration to 80% of pro-intervention concentration. Dust
Cleaning intervention is assumed to reduce floor dust-lead loading to 100//g/ft2 and window sill dust-lead loading to 250
mat.
2 Soil Cover intervention is assumed to reduce soil-lead concentration to 50% of pro-intervention concentration. Dust
Cleaning intervention is assumed to reduce floor dust-lead loading to 40 //g/ft2 and window sill dust-lead loading to 100
//g/ft2.
3 Soil Cover intervention is assumed to reduce soil-lead concentration to 20% of pro-intervention concentration. Dust
Cleaning intervention is assumed to reduce floor dust-lead loading to 20 //g/ft2 and window sill dust-lead loading to 50
//g/ft2.
* This risk assessment methodology involves comparing environmentally predicted pro- and post- 5 403 blood-lead
concentration distributions without considering NHANES III data.
5 An alternative approach to determining intervention effectiveness based on observed changes in blood-lead concentration
following an intervention and adjusted to be reflective of primary prevention interventions. Note that the floor dust-lead
loading standard used in this analysis was 100 //g/ft2 rather than 200 //g/ft2 which was used in all other sensitivity
analyses. The paint standards were also different. Paint maintenance interventions were conducted at hourly units
exceeding 0 ft* (rather than 5 ft1) of deteriorated LBP.
provides an alternative characterization of the post-§403 blood-lead concentration distribution
based on the adjusted blood-lead effects model (Section 5.4.1.6). Predicted reductions in
health risks based on these methodologies are also presented in Table 7-2.
Results of directly comparing model-predicted post-§403 distribution of blood-lead
concentration to model-predicted pre-§403 distributions are shown in columns 8 and 9 of the
table. Estimated risk reduction are comparable to those based on the risk assessment
methodology. Estimates are slightly larger for the EPI model and slightly less for the IEUBK
model than those generated by the risk assessment methodology.
Estimated reductions in blood-lead concentration based on the adjusted blood-lead effects
model are shown in the last column of Table 7-2. They are remarkably similar to those
Draft - Do Not Cite or Quote
231
September 27, 1996
-------
estimated based on the main risk assessment methodology, suggesting that estimates of post-
intervention environmental-lead levels are reasonable.
Obviously, estimates of health risk reductions vary when the methodology used to
estimate them changes. The results presented in Table 7-2 reflect this variability and provide a
range of reasonable estimates for potential reductions in health risks due to §403 at the
intermediate standards.
7.4 CONCLUSIONS OF RISK ASSESSMENT
On an overall basis after performing, integrating, and assessing the uncertainty in the
hazard identification, exposure assessment, and dose-response assessment, the following
conclusions were made:
1. The health risks of young children from exposure to lead-based paint hazards, lead
contaminated dust, and lead-contaminated soil are severe. Forty-six thousand
children aged 1-2 years, and 82,000 children aged 1-5 years have a blood-lead
concentration exceeding 25 ug/dL.
2. The health risks of children can be reduced.
3. The standards defined by §403 will help reduce the health risks to our nation's
children. Depending on the methodology implemented (assumptions on intervention
efficacy, predictive model used, and methodology for computing risk reductions) the
reduction in the number of children with a blood-lead concentration exceeding 25
ug/dL for the intermediate option for the §403 standards ranged from 41% to 99%.
Draft - Do Not Cite or Quote 232 September 27, 1996
-------
8.0 REFERENCES
Adebonojo, P.O. (1974) "Hematologic Status of Urban Black Children in Philadelphia:
Emphasis on the Frequency of Anemia and Elevated Blood Lead Levels." Cli Pediatr. 13:874-
Alessio, L., Bertazzi, P.A., Monelli, O., et al. (1976) "Free Erythrocyte Protoporphyrin as an
Indicator of the Biological Effect of Lead in Adult Males: II. Comparison Between Free
Erythrocyte Protoporphyrin and Other Indicators of Effect." International Archives of
Occupational and Environmental Health. 37:89-105.
Alomran, A.H., Shleamoon, M.N. (1988) "The Influence of Chronic Lead Exposure on
Lymphocyte Proliferative Response and Immunoglobulin Levels in Storage Battery Workers."
Journal of Biological Science Research. 19:575-585.
Alvares, A.P., Kapelner, S., Sassa, S., et al. (1975) "Drug Metabolism in Normal Children,
Lead-Poisoned Children, and Normal Adults." Clin Pharmacol Ther. 17:179-183.
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(5):893-897.
Angle, C. R., Mclntire, M. S., Colucci, A. V. (1974) "Lead in Air, Dustfall, Soil, House Dust,
Milk and Water: Correlation with Blood Lead of Urban and Suburban School Children." In:
Trace Substances in Environmental Health - VIII, D. D. Hemphill, Ed., 23-29.
Angle, C.R., Mclntire, M.S. (1978) "Low Level Lead and Inhibition of Erythrocyte Pyrimidine
Nucleotidase." Environmental Research. 17:296-302.
Angle, C.R., Mclntire, M.S. (1979) "Environmental Lead and Children: The Omaha Study."
Journal of Toxicological and Environmental Health. 5:855-870.
Angle, C.R., Mclntire, M.S., Swanson, M.S., et al. (1982) "Erythrocyte Nucleotides in
Children—Increased Blood Lead and Cytidine Triphosphate." Pediatr Res. 16:331-334.
Angle, C.R., Marcus, A., Cheng, I-H., et al. (1984) "Omaha Childhood Blood Lead and
Environmental Lead: A Linear Total Exposure Model." Environmental Research. 35:160-170.
Angle, C.R., Kuntzelman, D.R. (1989) "Increased Erythrocyte Protoporphyrins and Blood Lead
— A Pilot Study of Childhood Growth Patterns." J Toxicol Environ Health. 26:149-156.
Annest, J.L. (1983) "Trends in the Blood-Lead Levels of the U.S. Population: The Second
National Health and Nutrition Examination Survey (NHANES II) 1976-1980." In: Lead Versus
Health: Sources and Effects of Low Level Lead Exposure, Rutter, M., Russell Jones, R., eds.
New York: John Wiley and Sons, 33-58.
Draft - Do Not Cite or Quote 233 September 27. 1996
-------
Araki, S., Honma, T., Yanagihara, S., et al. (1980) "Recovery of Slowed Nerve Conduction
Velocity in Lead-Exposed Workers." 46:151-157.
Arnvig, E., Grandjean, P., Beckmann, J. (1980) "Neurotoxic Effects of Heavy Lead Exposure
Determined with Psychological Tests." ToxicolLett. 5:399-404.
Aschengrau, A., Beiser, A., Bellinger, D., Copenhafer, D., Weitzman, M. (1994) "The Impact
of Soil Lead Abatement on Urban Children's Blood Lead Levels: Phase II Results from the
Boston Lead-in-Soil Demonstration Project." Environmental Research. 67:125-148.
Assennato, G., Baser, M., Molinini, R., et al. (1987) "Sperm Count Suppression Without
Endocrine Dysfunction in Lead-Exposed Men." Archives of Environmental Health. 42:124-
127.
ATSDR (1988a) "The Nature and Extent of Lead Poisoning in Children in the United States: A
Report to Congress." U.S. Department of Health and Human Services, Public Health Service,
Agency for Toxic Substances and Disease Registry.
ATSDR (1988b) "The Silver Creek Mine Tailings Exposure Study, Park City, Utah." Final
Report of the Agency for Toxic Substances and Disease Registry, Public Health Service, U.S.
Department of Health and Human Services, June 1988.
ATSDR (1991a) "Child Lead Exposure Study, Leeds, Alabama." Final Report of the Agency
for Toxic Substances and Disease Registry, Public Health Service, U.S. Department of Health
and Human Services, September 1991.
ATSDR (1991b) "Philadelphia Neighborhood Lead Study." Final Report of the Agency for
Toxic Substances and Disease Registry, Public Health Service, U.S. Department of Health and
Human Services, September 1991.
ATSDR (1992) "ClearCreek/Central City Mine Waste Exposure Study, Part I: Smuggler
Mountain Site." Final Report of the Agency for Toxic Substances and Disease Registry, Public
Health Service, U.S. Department of Health and Human Services, September 1992.
ATSDR (1993) "Toxicological Profile for Lead." Final Report of the Agency for Toxic
Substances and Disease Registry, Public Health Service, U.S. Department of Health and
Human Services, April, 1993.
Awad El Karim, M.A., Hamed, A.S., Elhaimi, Y.A.A., et al. (1986) "Effects of Exposure to
Lead among Lead-Acid Battery Factory Workers in Sudan." Archives of Environmental Health.
41:261-265.
Baghurst, P.A., Robertson, E.F., McMichael, A.J., et al. (1987) "The Port Pirie Cohort Study:
Lead Effects on Pregnancy Outcome and Early Childhood Development." Neurotocicology.
8:395-401.
Draft - Do Not Cite or Quote 234 September 27, 1996
-------
Baghurst, P.A., McMichael, A.J., Wigg, N.R., Vimpani, G., Robertson, E.F., Roberts, R.J.,
Tong, S. (1992) "Life-long Exposure to Environmental Lead and Children's Intelligence at Age
Seven: The Port Pirie Cohort Study." New England Journal of Medicine. 327:1279-1284.
Baker, E.L. Jr., Landrigan, P.J., Harbour, A.G., et al. (1979) "Occupational Lead Poisoning in
the United States: Clinical and Biochemical Findings Related to Blood Lead Levels." British
Journal of Industrial Medicine. 36:314-322.
Baker, E.L., Feldman, R.G., White, R.F., et al. (1983) "The Role of Occupational Lead
Exposure in the Genesis of Psychiatric and Behavioral Disturbances." Acta Phsychiatr Scand
Suppl. 67:38-48.
Baloh, R.W., Spivey, G.H., Brown, C.P., et al. (1979) "Subclinical Effects of Chronic
Increased Lead Absorption — A Prospective Study: H Results of Baseline Neurologic
Testing." Journal of Occupational Medicine. 21:490-496.
Barry, P.S.I., Mossman, D.B. (1970) "Lead Concentration in Human Tissues." British Journal
of Industrial Medicine. 27:339-351.
Barry, P.S.I. (1975) "A Comparison of Concentrations of Lead in Human Tissues." British
Journal of Industrial Medicine. 32:119-139.
Barry, P.S.I. (1981) "Concentrations of Lead in the Tissues of Children." British Journal of
Industrial Medicine. 38:61-71.
Batschelet, E., Brand, L., Steiner, A. (1979) "On the Kinetics of Lead in the Human Body."
J. Math Biology. S:\S-23.
Battelle (1995) "Studies of the Lead Problem in Paint, Dust, and Soil - Volume II:
Appendices." Final Report, September 1995.
Battelle (1995) "A Summary of the Relationships Between Blood Lead and Lead-
Contaminated Soil and Lead-Contaminated Dust, as Reported in the Scientific Literature."
November 1995.
Batuman, V., Maesaka, J.K., Haddad, B., et al. (1981) "The Role of Lead in Gout
Nephropathy." New England Journal of Medicine. 304:520-523.
Batuman, V., Landy, E., Maesaka, J.K., et al. (1983) "Contribution of Lead to Hypertension
with Renal Impairment." New England Journal of Medicine. 309:17-21.
Bauchinger, M., Dresp, J., Schmid, E., et al. (1977) "Chromosome Analyses of Children After
Ecological Lead Exposure." Mut. Res. 56:75-79.
Draft - Do Not Cite or Quote 235 September 27, 1996
-------
Bellinger, D.C., Needleman, H.L. (1983) "Lead and the Relationship Between Maternal and
Child Intelligence." Journal of Pediatrics. 102:523-527.
Bellinger, D.C., Needleman, H.L., Leviton, A., et al. (1984) "Early Sensory-Motor
Development and Prenatal Exposure to Lead." Neurobehav Toxicol Teratol. 6:387-402.
Bellinger, D.C., Leviton, A., Waternaux, C., et al. (1985a) "A Longitudinal Study of the
Developmental Toxicity of Low-Level Lead Exposure in the Prenatal and Early Postnatal
Periods." In: Lekkas TD, ed. International Conference on Heavy Metals in the Environment,
Athens, Greece, September, Vol 1. Edinburgh, United Kingdom: CEP Consultants, Ltd, 32-34.
Bellinger, D.C., Leviton, A., Waternaux, C., et al. (1985b) "Methodological Issues in Modeling
the Relationship Between Low-Level Lead Exposure and Infant Development: Examples from
the Boston Lead Study." Environmental Research. 38:119-129.
Bellinger, D.C., Leviton, A., Needleman, H.L., et al. (1986a) "Low-Level Lead Exposure and
Infant Development in the First Year." Neurobehav Toxicol Teratol. 8:151-161.
Bellinger, D.C., Leviton, A., Rabinowitz, M., et al. (1986b) "Correlates of Low-Level Lead
Exposure in Urban Children at Two Years of Age." Pediatrics. 77:826-833.
Bellinger, D., Needleman, H.L., Bromfield, R., Mikntz, M. (1986c) "A Follow-up Study of the
Academic Attainment and Classroom Behavior of Children with Elevated Dentine Lead
Levels." Bio Trace Elem Res. 6:207-23.
Bellinger, D.C., Leviton, A., Waternaux, C., et al. (1987a) "Longitudinal Analyses of Prenatal
and Postnatal Lead Exposure and Early Cognitive Development." New England Journal of
Medicine. 316:1037-1043.
Bellinger, D., Sloman, J., Leviton, A., et al. (1987b) "Low Level Lead Exposure and Child
Development: Assessment at Age 5 of a Cohort Followed from Birth." In: Lindberg, S.E.,
Hutchinson, T.C., eds. International Conference on Heavy Metals in the Environment. New
Orleans, LA, September, Vol. 1. Edinburgh, UK: CEP Consultants, Ltd., 49-53.
Bellinger, D., Leviton, A., Waternaux, C., et al. (1989) "Low-Level Lead Exposure and Early
Development in Socioeconomically Advantaged Urban Infants." In: Smith M., Grant L.D.
Sors, A., eds. Lead Exposure and Child Development: An International Assessment. Lancaster,
UK: Kluwer Academic Publishers.
Bellinger, D.C., Leviton, A., Waternaux, C., et al. (1989b) "Low-level Lead Exposure, Social
Class, and Infant Development." Neurotoxicol Teratol. 10:497-504.
Bellinger, D., Sloman, J., Leviton, A., Rabinowitz, M., Needleman H.L., Watemaux, C. (1991)
"Low-level Lead Exposure and Children's Cognitive Function in the Preschool Years."
Pediatrics. 87(2):219-227.
Draft - Do Not Cite or Quote 236 September 27, 1996
-------
Bellinger, D., Stiles, K.M., Needleman, H. (1992) "Low-level Lead Exposure, Intelligence, and
Academic Achievement: a Long-term Follow-up Study." Pediatrics. 90(6):855-861.
Bellinger, D.C. (1995) "Interpreting the Literature on Lead and Child Development: The
Neglected Role of the 'Experimental System'." Neuratoxicology and Teratology. 17(3):201-
212.
Bergomi, M., Borella, P., Fantuzzi, G., et al. (1989) "Relationship Between Lead Exposure
Indicators and Neuropsychological Performance in Children." Dev Med Child Neurol. 31:181-
190.
Bernard, B.P., Becker, C.E. (1988) "Environmental Lead Exposure and the Kidney." Clin
Toxicol. 26:1-34.
Bert, J.L., van Dusen, L.J., Grace, J.R. (1989) "A Generalized Model for the Prediction of Lead
Body Burdens." Environmental Research. 48:17-127.
Belts, P.R., Astley, R., Raine, D.N. (1973) "Lead Intoxication in Children in Birmingham."
British MedicalJournal. 1:402-406.
Biagini, G., Caudarella, R., Vangelista, A. (1977) "Renal Morphological and Functional
Modification in Chronic Lead Poisoning." In: Brown, S.S., ed. Clinical Chemistry and
Chemical Toxicology of Metals. Elsevier/North-Holland Biomedical Press, 123-126.
Bomschein, R.L., Hammond, P.B., Dietrich, K.N., Succop, P.A., Krafft, K.M., Clark, C.S.,
Pearson, D., Que Hee, S. (1985a) "The Cincinnati Prospective Study of Low-Level Lead
Exposure and Its Effects on Child Development: Protocol and Status Report." Environmental
Research 38:4-18.
Bomschein, R.L., Succop, P.A., Dietrich, R.N., Clark, C.S., Que Hee, S., Hammond, P.B.
(1985b) "The Influence of Social and Environmental Factors on Dust Lead, Hand Lead, and
Blood Lead Levels in Young Children." Environmental Research 38:108-118.
Bomschein, R.L., Succop, P.A., Krafft, K.M., Clark, C.S., Peace, B., Hammond, P.B. (1986).
"Exterior Surface Dust Lead, Interior House Dust Lead and Childhood Lead Exposure in an
Urban Environment." In: Trace Substances in Environmental Health n, 1986, A Symposium,
D.D. Hemphill, ed. Columbia, MO: University of Missouri.
Bomschein, R.L., Clark, S., Grote, J., Peace, B., Roda, S., Succop, P. (1988) "Soil Lead-Blood
Lead Relationship in a Former Lead Mining Town." In: Lead in Soil: Issues and Guidelines,
Supplement to Volume 9 of Environmental Geochemistry and Health, Devis, B.E., and
Wixson, E.G., Eds., 149-160.
Draft - Do Not Cite or Quote 237 September 27, 1996
-------
Bornschein, R.L., Grote, J., Mitchell, T., et al. (1989) "Effects of Prenatal Lead Exposure on
Infant Size at Birth." In: Smith, M., Grand, L.D., Sors, A., eds. Lead Exposure and Child
Development: An International Assessment. Lancaster, UK: Kluwer Academic Publishers.
Bornschein, R., Clark, S., Pan, W., Succop, P. (1990) "Midvale Community Lead Study." Final
Report of the University of Cincinnati Medical Center, July 1990.
Bradley, J.E., Powell, A.E., Niermann, W., et al. (1956) "The Incidence of Abnormal Blood
Levels of Lead in a Metropolitan Pediatric Clinic: With Observation on the Value of
Coproporphyrinuria as a Screening Test." Journal of Pediatrics. 49:1-6.
Bradley, I.E., Baumgartner, R.J. (1958) "Subsequent Mental Development of Children with
Lead Encephalopathy, as Related to Type of Treatment." Journal of Pediatrics. 53:311-315.
Braunstein, G.D., Dahlgren, J. Loriaux, D.L. (1978) "Hypogonadism in Chronically Lead-
poisoned Men." Infertility. 1:33-51.
Brody, D. J., Pirkle, J.L., Kramer, R. A., Flegal, K.M., Matte, T. D., Gunter, E.W., Paschal,
D.C. (1994) "Blood Lead Levels in the US Population: Phase I of the Third National Health
and Nutritional Examination Survey (NHANES m, 1988 to 1991)," Journal of the American
Medical Association. 272(4):277-283.
Buc, H.A., Kaplan, J.C. (1978) "Red-cell Pyrimidine S'-nucleotidase and Lead Poisoning."
Clin ChimActa. 87:49-55.
Bureau of the Census, HUD (1995). American Housing Survey for the United States in 1993.
Current Housing Reports, No. HI 50/93. Bureau of the Census, U.S. Department of Commerce;
and the Office of Policy Development and Research, U.S. Department of Housing and Urban
Development, February, 1995.
Bureau of the Census, HUD (1996). Housing Completions, December 1995/January 1996.
C22/96-1, March 1996.
Butte-Silver Bow Department of Health, and Department of Environmental Health, University
of Cincinnati. (1991) "The Butte-Silver Bow Environmental Health Lead Study." Draft Final
Report, 10 June 1991.
Campara, P., D'Andrea, F., Micciolo, R., et al. (1984) "Psychological Performance of Workers
With Blood-lead Concentration Below the Current Threshold Limit Value." International
Archives of Occupational Environmental Health. 53:233-246.
Centers for Disease Control (1985) "Preventing Lead Poisoning in Young Children." Altanta,
GA: U.S. Department of Health and Human Services, Public Health Service, Centers for
Disease Control. Publication No. 99-2230,7-19.
Draft - Do Not Cite or Quote 238 September 27, 1996
-------
Centers for Disease Control (1991) "Preventing Lead Poisoning in Young Children: A
Statement by the Centers for Disease Control." Public Health Service, Centers for Disease
Control and Prevention, U.S. Department of Health and Human Services, October 1991.
Centers for Disease Control (1992) "Sample Design: Third National Health and Nutrition
Examination Survey. Series 2: Data Evaluation and Methods Research." National Center for
Health Statistics, Centers for Disease Control and Prevention, U.S. Department of Health and
Human Services, DHHS Publication No. (PHS) 92-1887.
Centers for Disease Control (1994) "Plan and Operation of the Third National Health and
Nutrition Examination Survey, 1988-94." National Center for Health Statistics, Centers for
Disease Control and Prevention, U.S. Department of Health and Human Services, DHHS
Publication No. (PHS) 94-1308.
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." New England
Journal of Medicine. 309:1089-1093.
Chisolm, J.J., Jr., Harrison, H.E. (1956) "The Exposure of Children to Lead." Pediatrics.
18:943-958.
Chisolm, J.J., Jr. (1962) "Aminoaciduria as a Manifestation of Renal Tubular Injury in Lead
Intoxication and a Comparison with Patterns of Aminoaciduria Seen in Other Diseases."
Journal of Pediatrics. 60:1-17.
Chisolm, J.J., Jr. (1965) "Chronic Lead Intoxication in Children." Dev Med Child Neural.
7:529-536.
Chisolm, J.J., Thomas, D.J., Hammill, T.G. (1985) "Erythrocyte Porphobilinogen Synthase
Activity as an Indicator of Lead Exposure to Children." Clin Chem. 31:601-605.
Chowdhury, A.R., Chinoy, N.J., Gautam, A.K., et al. (1986) "Effect of Lead on Human
Semen." Adv Contracept Deliv Syst. 2:208-211.
Christoffersson, J.O., Ahlgren, L., Schutz, A., Skerfving, S., Mattsson, S. (1986) "Decrease of
Skeletal Lead Levels in Man after End of Occupational Exposure." Archives of Environmental
Health. 41:5:312-318.
Clark, C.S., Bornschein, R.L., Succop, P., Que Hee, S.S., Hammond, P.B., Peace, B. (1985)
"Condition and Type of Housing as an Indicator of Potential Lead Exposure and Pediatric
Blood Lead Levels." Environmental Research. 38:46-53.
Coate, D., Fowles, R. (1989) "Is There Statistical Evidence for a Blood Lead-Blood Pressure
Relationship?" Journal of Economics. 8:173-184.
Draft - Do Not Cite or Quote 239 September 27, 1996
-------
Cohen, N., Jaakkola, T., Wrenn, M.E. (1973) "Lead-210 Concentrations in the Bone, Blood,
and Excreta of a Former Uranium Miner." Health Physics. 24:601-609.
Colorado Department of Health, University of Colorado at Denver, and ATSDR (1990)
"Leadville Metals Exposure Study." Final Report by the Colorado Department of Health,
Division of Disease Control and Environmental Epidemiology; University of Colorado at
Denver, Center for Environmental Sciences; and the Agency for Toxic Substances and Disease
Registry, Public Health Service, U.S. Department of Health and Human Services, April, 1990.
Cooney, G.H., Bell, A., McBride, W., et al. (1989a) "Low-Level Exposures to Lead: The
Sydney Lead Study." Dev Med Child Neurol. 31:640-649.
Cooney, G., Bell, A., Stavron, C. (1991) "Low Level Exposures to Lead and Neurobehavioural
Development: the Sydney Study at Seven Years." Heavy Metals in the Environment.
Edinburgh: CEP Consultants, 1991:16-9.
Cooper, W.C., Gaffey, W.R. (1975) "Mortality of Lead Workers." Journal of Occupational
Medicine. 17:100-107.
Cooper, W.C. (1976) "Cancer Mortality Patterns in the Lead Industry." Ann, NY AcadSci.
271:250-259.
Cooper, W.C., Wong., O., Kheifets., L. (1985) "Mortality Among Employees of Lead Battery
Plants and Lead Producing Plants, 1947-1980." ScandJ Work Environ Health. 11:331-345.
Cooper, W.C. (1988) "Deaths from Chronic Renal Disease in U.S. Battery and Lead
Production Workers." Environmental Health Perspectives. 78:61-63.
Copley, C. G. (1983) "The Effect of Lead Hazard Source Abatement and Clinic Appointment
Compliance on the Mean Decrease of Blood Lead and Zinc Protoporphyrin Levels. Mimeo."
City of St. Louis, Department of Health and Hospitals, Division of Health, Office of the Health
Commissioner, St. Louis, MO.
Cramer, K., Goyer, R.A., Jagenburg, R., et al. (1974) "Renal Ultrastructure, Renal Function,
and Parameters of Lead Toxicity in Workers with Different Periods of Lead Exposure." British
Journal of Industrial Medicine. 31:113-127.
Cristy, M., Leggett, R.W., Dunning, D.E., Jr., Eckerman, K.F. (1986) "Relative Age-Specific
Radiation Dose Commitment Factors for Major Radionuclides Released from Nuclear Fuel
Facilities." NUREG/CR-4628 ORNL/TM-9890.
Cullen, M.R., Kayne, R.D., Robins, J.M. (1984) "Endocrine and Reproductive Dysfunction in
Men Associated with Occupational Inorganic Lead Intoxication." Archives of Environmental
Health. 39:431-440.
Draft - Do Not Cite or Quote 240 September 27, 1996
-------
Dalpra, L., Tibiletti, M.G., Nocera, G., et al. (1983) "SCE Analysis in Children Exposed to
Lead Emission from a Smelting Plant." Mut Res. 120:249-256.
Davies, D.J., Watt, J.M., Thornton, I. (1987) "Lead Levels in Birmingham Dusts and Soils."
The Science of the Total Environment. 87(2-3): 177-185.
Davies, D.J.A, Thornton, I., Watt, J.M., Culbard, E.B., Harvey, P.O., Delves, H.T., Sherlock,
J.C., Smart, G.A., Thomas, J.F.A., Quinn, M.J. (1990) "Lead Intake and Blood Lead in Two-
Year-Old U.K. Urban Children." The Science of the Total Environment. 90:13-29.
Davis, J.M., Svendsgaard, D.J. (1987) "Lead and Child Development." Nature. 329:297-300.
Day, J.C. (1993) Population Projections of the United States, by Age, Sex, Race, and Hispanic
Origin: 1993 to 2050. U.S. Bureau of the Census, Current Population Reports P25-1104.
de la Burde, B., Choate, M.S., Jr. (1972) "Does Asymptomatic Lead Exposure in Children
Have Latent Sequelae?" Journal of Pediatrics. 81:1088-1091.
de la Burde, B., Choate, M.S., Jr. (1975) "Early Asymptomatic Lead Exposure and
Development at School Age." Journal of Pediatrics. 87:638-642.
deKort, W.L.A.M., Verschoor, M.A., Wibowo, A.A.E., et al. (1987) "Occupational Exposure
to Lead and Blood Pressure: A Study of 105 Workers." American Journal of Industrial
Medicine. 11:145-156.
Dietrich, K.N., Krafft, K.M., Bier, M., et al. (1986) "Early Effects of Fetal Lead Exposure:
Neurobehavioral Findings at 6 Months." International Journal ofBiosocial and Medical
Record. 8:151-168.
Dietrich, K.N., Krafft, K.M., Bornschein, R.L., et al. (1987a) "Low-Level Fetal Lead Exposure
Effect on Neurobehavioral Development in Early Infancy." Pediatrics. 80:721-730.
Dietrich, K.N., Krafft, K.M., Shukla, R., et al. (1987b) "The Neurobehavioral Effects of Early
Lead Exposure." Monogr Am Assoc Ment Defic. 8:71-95.
Dietrich, K.N., Berger, O.G., Succop, P.A., Hammond, P.B., Bornschein, I. (1993) "The
Developmental Consequences of Low To Moderate Prenatal and Postnatal Lead Exposure:
Intellectual Attainment in the Cincinnati Lead Study Cohort Following School Entry."
Neurotoxicology and Teratology. 15:37-44.
Drasch, G.A., Kretschmer, E., Lochner, C. (1988) "Lead and Sudden Infant Death:
Investigations on Blood Samples of SID Babies." EurJPediatr. 147:79-84.
Draft - Do Not Cite or Quote 241 September 27, 1996
-------
Elwood:, P.C., Davey-Smith, G., Oldham, P.O., et al. (1988) "Two Welsh Surveys of Blood
Lead and Blood Pressure." Environmental Health Perspectives. 78:119-121.
Erenberg, G., Rinsler, S.S., Fish, B.G. (1974) "Lead Neuropathy and Sickle Cell Disease."
Pediatrics. 54:438-441.
Ernhart, C.B., Landa, B., Schell, N.B. (1981) "Subclinical Levels of Lead and Developmental
Deficit — A Multivariate Follow-up Reassessment." Pediatrics. 67:911-919.
Ernhart, C.B., Wolf, A.W., Kennard, M.J., et al. (1985) "Intrauterine Lead Exposure and the
Status of the Neonate." In: Lekkas TD, ed. International Conference on Heavy Metals in the
Environment, Athens, Greece, September. Vol. 1. Edinburgh, United Kingdom: CEP
Consultants, Ltd. 35-37.
Ernhart, C.B., Wolf, A.W., Kennard, M.J., et al. (1986) "Intrauterine Exposure to Low Levels
of Lead: The Status of the Neonate." Archives of Environmental Health. 41:287-291.
Ernhart, C.B., Morrow-Tlucak, M., Marler, M.R., et al. (1987) "Low Level Lead Exposure in
the Prenatal and Early Preschool Periods: Early Preschool Development." Neurotoxicol
Teratol. 9:259-270.
Ernhart, C.B. (1988) "Cofactors in Research on the Environmental Toxicology of Childhood:
Issues and Examples from Lead Effects Studies". In: Environmental Toxicology of Childhood,
University of Nebraska, Children and the Law Series.
Ernhart, C.B., Morrow-Tlucak, M., Worf, A.W., Super, D., Drotar, D. (1989) "Low Level Lead
Exposure in the Prenatal and Early Preschool Periods; Intelligence Prior to School Entry."
Neurotoxicology and Teratology. 11:161 -70.
Ernhart, C.B., Green, T. (1990) "Low-Level Lead Exposure in Prenatal and Early Preschool
Periods: Language Development." Archives of Environmental Health. 45:342-354.
Ewers, U., Stiller-Winkler, R., Idel, H. (1982) "Serum Immunoglobulin, Complement C3, and
Salivary IgA Level in Lead Workers." Environmental Research. 29:351-357.
Factor-Litvak, P., Graziano, J.H., Kline, J.K., et al. (1991) "A Prospective Study of
Birthweight and Length of Gestation in Population Surrounding a Lead Smelter in Kosovo,
Yugoslavia." Int JEpidemiol. 20:772-728.
Fanning, D. (1988) "A Mortality Study of Lead Workers, 1926-1985." ,4rc/Hvej of
Environmental Health. 43:247-251.
Draft - Do Not Cite or Quote 242 September 27, 1996
-------
Farfel, M.R., Chisolm, J.J., Jr. (1990) "Health and Environmental Outcomes of Traditional and
Modified Practices for Abatement of Residential Lead-Based Paint." American Journal of
Public Health. 80(10): 1240-1245.
Farfel, M.R., Lira, B.S. (1995) "The Lead Paint Abatement and Repair and Maintenance Study
in Baltimore." In: 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. Philadelphia: American Society for Testing and Materials, 107-118.
Fayerweather, W.E., Karns, M.E., Nuwayhid, I.A., et al. (1991) "An Epidemiologic Study of
Cancer Risk Following Exposure to Organic Lead Among the DuPont Company's Chamber
Works Employees." DuPont Company, Human Resources, Epidemiology Section Medical
Division, Wilmington, DE.
Fulton, M., Raab, G., Thomson, G., Laxen, D., Hunter, R., Hepburn, W. (1987) "Influence of
Blood Lead on the Ability and Attainment of Children in Edinburgh." Lancet. 1:1221-1226.
Gant, V.A. (1938) "Lead Poisoning." Industial Medicine. 7:679-699.
Gartside, P.S. (1988) "The Relationship of Blood Lead Levels and Blood Pressure in
NHANES: II. Additional Getzcalculations." Environmental Health Perspectives. 78:31-34.
Gerhardsson., L., Lundstrom, N.G., Nordberg, G., et al. (1986b) "Mortality and Lead
Exposure: A Retrospective Cohort Study of Swedish Smelter Workers." British Journal of
Industrial Medicine. 43:707-712.
Glickman, L., Valciukas, J.A., Lilis, R., et al. (1984) "Occupational Lead Exposure: Effects on
Saccadic Eye Movements." International Archives of Occupational Environmental Health.
54:115-125.
Goyer, R.A. (1993) "Lead Toxicity: Current Concerns." Environmental Health Perspectives.
100:177-187.
Grandjean, P., Lintrup, J. (1978) "Erythrocyte-Zn-protoporphyrin as an Indicator of Lead
Exposure." ScandJClin Lab Invest. 38:669-675.
Grandjean, P. (1979) "Occupational Lead Exposure in Denmark: Screening with the
Haematofluorometer." British Journal of Industrial Medicine. 36:52-58.
Grandjean, P., Wulf, H.C., Neibuhr, E. (1983) "Sister Chromatid Exchange in Response to
Variations in Occupational Lead Exposure." Environmental Research. 32:199-204.
Grandjean, P., Hollnagel, H., Hedegaard, L., et al. (1989) "Blood Lead-Blood Pressure
Relation: Alcohol Intake and Hemoglobin as Confounders." Am JEpidemiol. 129:732-739.
Draft - Do Not Cite or Quote 243 September 27, 1996
-------
Greene, T., Ernhart, C.B. (1991) "Prenatal and Preschool Age Lead Exposure: Relationship
with Size." Neurotoxicology and Teratology. 13:417-427.
Griffin, T.B., Coulston, F., Wills, H. (1975b) "Biological and Clinical Effects of Continuous
Exposure to Airborne Particulate Lead.'Mr/i Hig Toksikol. 26:191-208. (Yugoslavian)
Gulson, B.L., Mahaffey, K.R., Mizon, K.J., Korsch, M.J., Cameron, M.A., Vimpani, G. (1995)
"Contribution of Tissue Lead to Blood Lead in Adult Female Subjects Based on Stable Lead
Isotope Methods." Journal of Lab and Clinical Medicine. 125:703.
Haas, T., Wieck, A.G., Schaller, K.H., et al. (1972) "The Usual Lead Load in New-born Infants
and Their Mothers." Zentralblattjur Bakteriologie [B]. 155:341-349. (German)
Haenninen, H., Hernberg, S., Mantere, P., et al. (1978) "Psychological Performance of Subjects
with Low Exposure to Lead." Journal of Occupational Medicine. 20:683-689.
Haenninen, H., Mantere, P., Hemberg, S., et al. (1979) "Subjective Symptoms in Low-Level
Exposure to Lead." Neurotoxicology. 1:333-347.
Hammond, P.B., Bomschein, R.L., Succop, P. (1985) "Dose-effect and Dose-response
Relationships of Blood Lead to Erythrocytic Protoporphyrin in Young Children." In:
Bomschein, R.L., Rabinowitz, M.B., eds. The Second International Conference on Prospective
Studies of Lead, Cincinnati, OH: April, 1984. Environmental Research. 38:187-196.
Hansen, O.N., Trillingsgaard, A., Beese, I., et al. (1989) "A Neuropsychological Study of
Children with Elevated Dentine Lead Level: Assessment of the Effect of Lead in Different
Socio-economic Groups." Neurotoxicol Teratol. 11:205-213.
Harlan, W.R. (1988) "The Relationship of Blood Lead Levels to Blood Pressure in the US
Population. Environmental Health Perspectives. 78:9-13.
Harlan, W.R., Landis, J.R., Schmouder, R.L., et al. (1988) "Blood Lead and Blood Pressure:
Relationship in the Adolescent and Adult US Population." Journal of the American Medical
Association. 253:530-534.
Harvey, P.O., Hamlin, M.W., Kumar, R., et al. (1984) "Blood Lead, Behavior and Intelligence
Test Performance in Preschool Children." Sci Total Environ. 40:45-60.
Harvey, P.O., Hamlin, M.W., Kumar, R,. Morgan, G., Spurgeon, A., Delves, H.T. (1988)
"Relationship Between Blood Lead, Behaviour, Psychometric and Neuropsychological Test
Performance in Young Children." British Journal of Develomental Psychology. 29:43-52.
Draft - Do Not Cite or Quote 244 September 27, 1996
-------
Hatzakis, A., Kolevi, A., Katsouyanni, K., et al. (1987) "Psychometric Intelligence and
Attentional Performance Deficits in Lead Exposed Children." International Conference on
Heavy Metals in the Environment. Edinburgh, Scotland; CEP Consultants. 16:57-67.
Hatzakis, A., Kokkeni, A., Maranelias, C., Katsouyanni, K., Salaminios, F., Kalandidi, A., et
al. (1989) "Psychometric Intelligence Deficits in Lead-exposed Children." In: Smith MA.,
Grant, I.D., Sora Al, eds. Lead Exposure and Child Development. London: Kluwer, 1989:211-
23.
Hawk, B.A., Schroeder, S.R., Robinson, G., et al. (1986) "Relation of Lead and Social Factors
to IQ of Low-SES Children: A Partial Replication." AmJMent Defic. 91:178-183.
Heard, M.J., Chamberlain, A.C. (1984) "Uptake of Pb by Human Skeleton and Comparative
Metabolism of Pb and Alkaline Earth Elements." Health Physics. 47:857-865.
Herber, R.F.M. (1980) "Estimation of Blood Lead Values from Blood Porphyrin and Urinary
5-aminolevulinic Acid Levels in Workers." International Archives of Occupational
Environmental Health. 45:169-179.
Hemberg, S., Nikkanen, J., Mellin, G., et al. (1970) "6-Aminolevulinic Acid Dehydrase as a
Measure of Lead Exposure." Archives of Environmental Health. 21:140-145.
Hogstedt, C., Hane, M., Agrell, A., et al. (1983) "Neuropsychological Test Results and
Symptoms Among Workers with Well-defined Long-term Exposure to Lead." British Journal
of Industrial Medicine. 40:99-105.
Holness, D.L., Nethercott, J.R. (1988) "Acute Lead Intoxication in a Group of Demolition
Workers." Applied Industrial Hygiene. 3:338-341.
Hu, H., Pepper, L., Goldman, R. (1991) "Effect of Repeated Occupational Exposure to Lead,
Cessation of Exposure, and Chelation on Levels of Lead in Bone." American Journal of
Industrial Medicine. 20-723-735.
Huang, X.P., Feng, Z.Y., Zhai, W.L., et al. (1988b) "Chromosomal Aberrations and Sister
Chromatid Exchanges in Workers Exposed to Lead." Biomed Environ Sci. 1:382-387.
Hyrhorczuk, D., Rabinowitz, M., Hessel, S., Hoffman, D., Hogan, M., Mallin, K., Finch, H.,
Orris, P., Berman, E. (1985) "Elimination Kinetics of Blood Lead in Workers with Chronic
Lead Intoxication." American Journal of Industrial Medicine. 8:33-42.
Draft - Do Not Cite or Quote 245 September 27, 1996
-------
ICF (1995) "Regulatory Impact Analysis of the Proposed Rule on Lead-Based Paint:
Requirements for Notification, Evaluation and Reduction of Lead-Based Paint Hazards in
Federally-Owned Residential Property and Housing Receiving Federal Assistance." Draft
report for the Office of Lead-Based Paint Abatement and Poisoning Prevention, U.S.
Department of Housing and Urban Development.
Janin, Y., Couinaud, C., Stone, A., et al. (1985) "The 'Lead-Induced Colic' Syndrome in Lead
Intoxication." SurgAnn. 17:287-307.
Johnson, N.E., Renuta, K. (1979) "Diets and Lead Blood Levels of Children Who Practice
Pica." Environmental Research. 18:869-376.
Kang, H.K., Infante, P.P., Carra, J.S. (1980) "Occupational Lead Exposure and Cancer
(Letter)." Science. 207:935-936.
Kehoe, R.A. (1961) "The Metabolism of Lead in Man in Health and Disease: Present Hygienic
Problems Relating to the Absorption of Lead: The Harben Lectures, 1960." J R Inst Public
Health Hyg. 24:177-203.
Khera, A.K., Wibberley, D.G., Edwards, K.W., et al. (1980b) "Cadmium and Lead Levels in
Blood and Urine in a Series of Cardiovascular and Normotensive Patients." International
Journal of Environmental Studies. 14:309-312.
Kimbrough, R.D., LeVois, M., Webb, D.R. (1994) "Management of Children with Slightly
Elevated Blood Lead Levels." Pediatrics. 93(2):188-191.
Kirkby, H., Gyntelberg, F. (1985) "Blood Pressure and Other Cardiovascular Risk Factors of
Long-term Exposure to Lead." Scand J Work Environ Health. 11:15-19.
Koo, W.W.R., Succop, P.A., Bomschein, R.L., et al. (1991) "Serum Vitamin D Metabolites
and Bone Mineralization in Young Children with Chronic Low to Moderate Lead Exposure."
Pediatrics. 87:680-687.
Kosmider, S., Petelenz, T. (1962) "Electrocardiographic Changes in Elderly Patients with
Chronic Professional Lead Poisoning." Pol Arch Med Wewn. 32:437-442. (Polish)
Kotok, D. (1972) "Development of Children with Elevated Blood Levels: A Controlled Study."
Journal of Pediatrics. 80:57-61.
Kotok, D., Kotok, R., Heriot, T. (1977) "Cognitive Evaluation of Children with Elevated Blood
Lead Levels." Am JDis Child. 131:791-793.
Kuhnert, P.M., Erhard, P., Kuhnert, B.R. (1977) "Lead and 5-aminoluvulinic Acid Dehydratase
in RBC's of Urban Mothers and Fetuses." Environmental Research. 14:73-80.
Draft - Do Not Cite or Quote 246 September 27, 1996
-------
Kumar, S., Jain, S., Aggarwal, C.S., et al. (1987) "Encephalopathy Due to Inorganic Lead
Exposure in an Adult." Jpn JMed. 26:253-254.
Lancranjan, I., Popescu, H.I., Gavanescu, O., et al. (1975) "Reproductive Ability of Workmen
Occupationally Exposed to Lead." Archives of Environmental Health. 30:396-401.
Landis, J.R., Flegal, K.M. (1988) "A Generalized Mantel-Haenszel Analysis of the Regression
of Blood Pressure on Blood Lead Using NHANESII Data." Environmental Health
Perspectives. 78:35-41.
Landrigan, P.J., Gehlbach, S.H., Rosenblum, B.F., Shoults, J.M., Candelaria, R.M., Barthel,
W.F., Liddle, J.A., Smrek, A.L., Staehling, N.W., Sanders, J.F. (1975). "Epidemic Lead
Absorption Near an Ore Smelter: The Role of Participate Lead." New England Journal of
Medicine. 292(3): 123-129.
Landrigan, P.J., Baker, E.L., Jr., Feldman, R.G., et al. (1976) "Increased Lead Absorption with
Anemia and Slowed Nerve Conduction in Children Near a Lead Smelter." Journal of
Pediatrics. 89:904-910.
Lanphear, B.P., Emond, M., Jacobs, D.E., Weitzman, M., Tanner, M, Winter, N.L., Yakir, B.,
Eberly, S. (1995) "A Side-by-Side Comparison of Dust Collection Methods for Sampling
Lead-Contaminated House Dust." Environmental Research. 68(2): 114-123.
Lansdown, R., Yule, W., Urbanowicz, M.A., et al. (1986) "The Relationship Between Blood
Lead Concentrations, Intelligence, Attainment and Behavior in a School Population: The
Second London Study." International Archives of Occupational Environmental Health.
57:225-235.
Lauwers, M.C., Hauspie, R.C., Susanne, C., et al. (1986) "Comparison of Biometric Data of
Children with High and Low Levels of Lead in the Blood." Am J Phys Anthropol. 69:107-116.
Lauwerys, R., Buchet, J-P, Roels, H.A., et al. (1974) "Relationship Between Urinary 6-
aminolevulinic Acid Excretion and the Inhibition of Red Cell 6-aminolevulinate Dehydratase
by Lead." Clin Toxicol. 7:383-388.
Lauwerys, R., Buchet, J-P, Roels, H.A., et al. (1978) "Placental Transfer of Lead, Mercury,
Cadmium, and Carbon Monoxide in Women: I. Comparison of the Frequency Distributions of
the Biological Indices in Maternal and Umbilical Cord Blood." Environmental Research.
15:278-289.
Leal-Garza, C., Moates, De Oca R, Cerda-Flores, R.M., et al. (1986) "Frequency of Sister-
Chromatid Exchanges (SCE) in Lead Exposed Workers." Arch Invest Med. 17:267-276.
Draft - Do Not Cite or Quote 247 September 27, 1996
-------
Leggett, R.W., Eckerman, K.F., and Williams, L.R. (1982) "Strontium-90 in Bone: A Case
Study in Age-Dependent Dosimetric Modeling." Health Physics. 43(3):307-322.
Lewis and Clark County Health Department, Montana Department of Health and
Environmental Sciences, Centers for Disease Control, U.S. Department of Health and Human
Services, and U.S. EPA (1986) "East Helena, Montana: Child Lead Study, Summer 1983."
Final Report, July 1986.
Lilis, R., Gavrilescu, N., Nestorescu, B., et al. (1968) "Nephropathy in Chronic Lead
Poisoning." British Journal of Industrial Medicine. 25:196-202.
Lilis, R., Eisinger, J., Blumberg, W., et al. (1978) "Hemoglobin, Serum Iron, and Zinc
Protoporphyrin in Lead-exposed Workers." Environmental Health Perspectives. 25:97-102.
Lyngbye, T., Hansen, O.N., Grandjean, P. (1987) "The Influence of Environmental Factors on
Physical Growth in School Age: A Study of Low Level Lead Exposure." In: Lindberg S.E.,
Hutchinson, T.C., eds. International Conference on Heavy Metals in the Environment, Vol. 2,
New Orleans, LA, September. Edinburgh, UK: CEP Consultants, Ltd. 210-212.
Mahaffey, K.R. (1977) "Quantities of Lead Producing Health Effects in Humans: Sources and
Bioavailability." Environmental Health Perspectives. 19:285-295.
Mahaffey, K.R., Rosen, J.F., Chesney, R.W., et al. (1982) "Association Between Age, Blood
Lead Concentration, and Serum 1,25-dihydroxycholecalciferol Levels in Children." Am JClin
Nutr. 35:1327-1331.
Mahaffey, K.R., Annest, J.L. (1986) "Association of Erythrocyte Protoporphyrin with Blood
Lead Level and Iron Status in the Second National Health and Nutrition Examination Survey,
1976-1980." Environmental Research. 41-327-338.
Mahaffey, K.R. (1994) "Analysis of Epidemiological Data on Association between Blood Lead
and Lead in Soil and Dust for OPPTS/EPA." February 1994.
Maki-Paakkanen, J., Sorsa, M., Vainio, H. (1981) "Chromosome Aberrations and Sister
Chromatid Exchanges in Lead-Exposed Workers." Hereditas. 94:269-275.
Malcolm, D., Barnett, H.A.R. (1982) "A Mortality Study of Lead Workers: 1925-76." British
Journal of Industrial Medicine. 39:404-410.
Mantere, P., Haenninen, H., Hemberg, S. (1982) "Subclinical Neurotoxic Lead Effects: Two-
Year Follow-up Studies with Psychological Test Methods." Neurobehav Toxicol Teratol.
4:725-727.
Draft - Do Not Cite or Quote 248 September 27, 1996
-------
Maranelli, G., Apostoli, P. (1987) "Assessment of Renal Function in Lead Poisoned Workers."
Occup Environ Chem Hazards. 344-348.
Marcus, A.H. (1985) "Multicompartment Kinetics Models for Lead. I. Bone Diffusion Models
for Long-term Retention." Environmental Research. 36:441-458.
Marcus, A.H., Schwartz, J. (1987) "Dose-response Curves for Erythrocyte Protoporphyrin Vs
Blood Lead: Effects of Iron Status." Environmental Research. 44:221-227.
Marcus, A.M., Elias, R.W. (1994) "Estimates of Soil and Dust Lead Slope Factors for U. S.
Children, Ages 12 to 84 Months, Based on Consistent Analyses of Studies Since 1980."
January 1994. DO NOT CITE OR QUOTE.
Marino, P.E., Franzblau, A., Lilis, R., et al. (1989) "Acute Lead Poisoning in Construction
Workers: The Failure of Current Protective Standards." Archives of Environmental Health.
44:140-145.
Markowitz, M.E., Bijur, P.E., Ruff, H.A., Rosen, J.F. (1993) "Effects of Calcium Disodium
Versenate (Cana2edta) Chelation in Moderate Childhood Lead Poisoning." Pediatrics.
92(2):265-271.
Matte, T.D., Figueroa, J.P., Burr, G., et al. (1989) "Lead Exposure Among Lead-acid Battery
Workers in Jamaica." American Journal of Industrial Medicine. 16:167-177.
McBride, W.G., Black, B.P., English, B.J. (1982) "Blood Lead Levels and Behavior of 400
Preschool Children." MedJAust. 10:2(l):26-29.
McMichael, A.J., Vimpani, G.V., Robertson, E.F., et al. (1986) "The Port Pine Cohort Study:
Maternal Blood Lead and Pregnancy Outcome." JEpidemiol Community. 40:18-25.
McMichael, A.J., Baghurst, P.A., Wigg, N.R., et al. (1988) "Port Pine Cohort Study:
Environmental Exposure to Lead and Children's Abilities at the Age of Four Years." New
England Journal of Medicine. 319:468-476.
Menton, R.G., Burgeon, D.A., Marcus, A.H. (1995) "Pathways of Lead Contamination for the
Brugham and Women's Hospital Longitudinal Lead Study." In: 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. Philadelphia:
American Society for Testing and Materials, 92-106.
Meredith, P.A., Moore, M.R., Campbell, B.C., et al. (1978) "6-aminolaevulinic Acid
Metabolism in Normal and Lead-exposed Humans." Toxicology. 9:1-9.
Draft - Do Not Cite or Quote 249 September 27, 1996
-------
Michaels, D., Zoloth, S.R., Stern, F.B. (1991) "Does Low-level Lead Exposure Increase Risk
of Death?: A Mortality Study of Newspaper Printers." IntJEpidemiol. 20:978-983.
Milbum, H., Mitran, E., Crockford, G.W. (1976) "An Investigation of Lead Workers for
Subclinical Effects of Lead Using Three Performance Tests." Arm Occup Hyg. 19:239-249.
Moore, M.R., Goldberg, A., Pocock, S.J., et al. (1982) "Some Studies of Maternal and Infant
Lead Exposure in Glasgow." Scott MedJ. 27:113-122.
Muijser, H., Hoogendijk, E.M., Hoosma, J., et al. (1987) "Lead Exposure During Demolition
of a Steel Structure Coated with Lead-based Paints. II. Reversible Changes in the Conduction
Velocity of the Motor Nerves in Transiently Exposed Workers." ScandJ Work Environ Health.
13:56-61.
Murphy, M.J., Graziano, J.H., Popovac, D., et al. (1990) "Past Pregnancy Outcomes Among
Women Living in the Vicinity of a Lead Smelter in Kosovo, Yugoslavia." American Journal of
Public Health. 80:33-35.
National Academy of Sciences. (1972) "Lead: Airborne Lead in Perspective: Biologic Effects
of Atmospheric Pollutants." Washington, DC: National Academy of Sciences. 71-177,281-
313.
Needleman, H.L., Gunnoe, C., Leviton, A., et al. (1979) "Deficits in Psychologic and
Classroom Performance of Children with Elevated Dentine Lead Levels." New England
Journal of Medicine. 300:689-695.
Needleman, H.L., Rabinowitz, M., Leviton, A., et al. (1984) "The Relationship Between
Prenatal Exposure to Lead and Congenital Anomalies." Journal of the American Medical
Association. 251:2956-2959.
Needleman, H.L., Geiger, S.K., Frank, R. (1985) "Lead and IQ Scores: A Reanalysis (letter)."
Science. 227:701-704.
Needleman, H., Gatsonis, C. (1990) "Low-Level Lead Exposure and the IQ of Children."
Journal of the American Medical Association. 263:673-678.
Needleman, H.L., Schell, A., Bellinger, D., Leviton, A., Allred, E.N. (1990) "The Long-Term
Effects of Exposure to Low Doses of Lead in Childhood: An 11-Year Follow-up Report." The
New England Journal of Medicine. 322(2):83-88.
Neri, L.C., Hewitt, D., Orser, B. (1988) "Blood Lead and Blood Pressure: Analysis of Cross-
sectional and Longitudinal Data from Canada." Environmental Health Perspectives. 78:123-
126.
Draft - Do Not Cite or Quote 250 September 27, 1996
-------
Neter, J. and Wasserman, W. (1974) "Applied Linear Statistical Models." Homewood, IL:
Richard D. Irwin, Inc.
Nilsson, U., Attewell, R., Christoffersson, J.O., Schutz, A., Ahlgren, S., Skerfving, S., Mattson,
S. (1991) "Kinetics of Lead in Bone and Blood After End of Occupational Exposure."
Pharmacology and Toxicology. 69:477-484.
Nordberg, G.R., Mahaffey, K.R., and Fowler, B.A. (1991) "Introduction and Summary."
International Workshop on Lead in Bone: Implications for Dosimetry and Toxicology.
Environmental Health Perspectives. 91:3-7.
Nordenson, I., Beckman, G., Beckman, L., et al. (1978) "Occupational and Environmental
Risks In and Around a Smelter in Northern Sweden: IV. Chromosomal Aberrations in Workers
Exposed to Lead." Hereditas. 88:263-267.
Nordstrom, S., Beckman, L., Nordenson, I. (1979) "Occupational and Environmental Risks in
and Around a Smelter in Northern Sweden: V. Spontaneous Abortion Among Female
Employees and Decreased Birth Weight in Their Offspring." Hereditas. 90:291-296.
Nye, L.J.J. (1929) "An Investigation of the Extraordinary Incidence of Chronic Nephritis in
Young People in Queensland." MedJAust. 2:145-159.
Ong, C.N., Endo, G., Chia, K.S., et al. (1987) "Evaluation of Renal Function in Workers with
Low Blood Lead Levels." In: Fao V., Emmett, E.A., Maroni, M., et al., eds. Occupational and
Environmental Chemical Hazards. Chichester: Ellis Horwood Limited, 327-333.
O'Riordan, M.L., Evans, H.J. (1974) "Absence of Significant Chromosome Damage in Males
Occupationally Exposed to Lead." Nature. 247:50-53.
Paglia, D.E., Valentine, W.N., Dahlgren, J.G. (1975) "Effects of Low-level Lead Exposure on
Pyrimidine 5'-nucleotidase and Other Erythrocyte Enzymes: Possible Role of Pyrimidine 5'-
nucleotidase in the Pathogenesis of Lead-induced Anemia." J Clin Invest. 56:1164-1169.
Paglia, D.E., Valentine, W.N., Fink, K. (1977) "Lead Poisoning: Further Observations on
Erythrocyte Pyrimidine-nucleotidase Deficiency and Intracellular Accumulation of Pyrimidine
Nucleotides." J Clin Invest. 60:1362-1366.
Pagliuca, A., Mufti, G.J., Baldwin, D., et al. (1990) "Lead Poisoning: Clinical, Biochemical,
and Hematological Aspects of a Recent Outbreak." JClin Path. 43:277-281.
Panhandle District Health Department, Idaho Department of Health and Welfare, Centers for
Disease Control, and U.S. EPA. (1986) "Kellogg Revisited -1983: Childhood Blood Lead and
Environmental Status Report." Final Report of the U.S. Public Health Service, July 1986.
Draft'- Do Not Cite or Quote 251 September 27, 1996
-------
Parkinson, O.K., Ryan, C, Boimet, J., et al. (1986) "A Psychiatric Epidemic-logic Study of
Occupational Lead Exposure." Am JEpidemiol. 123:261-269.
Parkinson, D.K., Hodgson, M.J., Bromet, E.J., et al. (1987) "Occupational Lead Exposure and
Blood Pressure." British Journal of Industrial Medicine. 44:744-748.
Pasternak, G., Becker, C.E., Lash, A., et al. (1989) "Cross-sectional Neurotoxicology Study of
Lead-exposed Cohort." Clin ToxicoL 27:37-51.
Piomelli, S., Seaman, C., Zullow, D., et al. (1982) "Threshold for Lead Damage to Heme
Synthesis in Urban Children." Proc NatlAcadSci. 7:3335-3339.
Pirkle, J.L., Schwartz, J., Landis, J.R., et al. (1985) "The Relationship Between Blood Lead
Levels and Blood Pressure and its Cardiovascular Risk Implications." Am JEpidemiol.
121:246-258.
Pirkle, J.L., Brody, D.J., Gunter, E.W., Kramer, R.A., Paschal, D.C., Flegal, K.M., Matte, T.D.
(1994) "The Decline in Blood Lead Levels in the United States: The National Health and
Nutrition Examination Surveys (NHANES)." Journal of the American Medical Association.
272(4):284-291.
Pocock, S.J., Shaper, A.G., Ashby, D., et al. (1985) "Blood Lead and Blood Pressure in
Middle-aged Men." In: Lekkas, T.D., ed. International Conference on Heavy Metals in the
Environment, Vol. 1, Athens, Greece, Sepember. Edinburgh, United Kingdom: CEP
Consultants, Ltd., 303-305.
Pocock, S.J., Shaper, A.G., Ashby, D., et al. (1988) "The Relationship Between Blood Lead,
Blood Pressure, Stroke, and Heart Attacks in Middle-aged British Men." Environmental Health
Perspectives. 78:23-30.
Pocock, S.J., Ashby, D., Smith, M.A. (1989) "Lead Exposure and Children's Intellectual
Performance: The Institute of Child Health/Southampton Study." In: Smith, M., Grant, L.D.,
Sors, A., eds. Lead Exposure and Child Development: An International Assessment. Lancaster,
UK: Kluwer Academic Publishers.
Pocock, S.J., Smith, M., Baghurst, P. (1994) "Environmental Lead and Children's Intelligence:
A Systematic Review of the Epidemiological Evidence." BMJ. 309:1189-1197.
Pollock, C.A., Ibels, L.S. (1986) "Lead Intoxication in Paint Removal Workers on the Sidney
Harbour Bridge." MedJAust. 145:635-639.
Popcock, S.J., Shaper, A.G., Ashby, D., et al. (1984) "Blood Lead Concentration, Blood
Pressure, and Renal Function." BrMedJ. 289:872-874.
Draft - Do Not Cite or Quote 252 September 27, 1996
-------
Pueschel, S.M., Kopito, L., Schwachman, H. (1972) "Children with an Increased Lead Burden:
A Screening and Follow-up Study." Journal of the American Medical Association. 222:462-
466.
Que Hee, S.S., Peace, B., Clark, S., Boyle, J.R., Bomschein, R.L., Hammond, P.B. (1985)
"Evolution of Efficient Methods to Sample Lead Sources, Such as House Dust and Hand Dust,
in the Homes of Children." Environmental Research 38:77-95.
Rabinowitz, M.B., Wetherill, G.W., Kopple, J.D. (1973) "Lead Metabolism in the Normal
Human: Stable Isotope Studies." Science. 182:725-727.
Rabinowitz, M.B., Wetherill, G.W., Kopple, J.D. (1976) "Kinetic Analysis of Lead
Metabolism in Healthy Humans." Journal of Clinical Investigation. 58:260-270.
Rabinowitz, M.B., and Needleman, H.L. (1982) "Temporal Trends in the Lead Concentrations
of Umbilical Cord Blood." Science. 216:1429-1430.
Rabinowitz, M.B., and Needleman, H.L. (1984a) "Environmental, Demographic, and Medical
Factors Related to Cord Blood Lead Levels." Biological Trace Element Research. 6:57-67.
Rabinowitz, M.B., Needleman, H.L., Leviton, A. (1984b) "Variability of Blood Lead During
Normal Infancy." Archives of Environmental Health. 39:74-77
Rabinowitz, M., Leviton, A., Needleman, H., Bellinger, D., Waternaux, C. (1985a).
"Environmental Correlates of Infant Blood Lead Levels in Boston." Environmental Research.
38:96-107.
Rabinowitz, M., Leviton, A., Bellinger, D. (1985b) "Home Refmishing, Lead Paint, and Infant
Blood Lead Levels." American Journal of Public Health. 75(4):403-404.
Rabinowitz, M.B., Leviton, A., Needleman, H. (1986) "Occurrence of Elevated Protoporphyrin
Levels in Relation to Lead Burden in Infants." Environmental Research. 39:253-257.
Rabinowitz, M.B. (1991) "Toxicokinetics of Bone Lead." Environmental Health Perspectives.
91:33-37.
Reigart, J.R., Graber, C.D. (1976) "Evaluation of the Humoral Immine Response of Children
with Low Level Lead Exposure." Bull Environ Contam Toxicol. 16:112-117.
Robinson, G.S., Baumann, S., Kleinbaum, D., et al. (1985) "Effects of Low to Moderate Lead
Exposure on Brainstem Auditory Evoked Potentials in Children: Environmental Health
Document 3." Copenhagen, Denmark: World Health Organization Regional Office for Europe.
177-182.
Draft - Do Not Cite or Quote 253 September 27, 1996
-------
Rodamilans, M., Osaba, M.J., To-Figueras, J., et al. (1988) "Lead Toxicity on Endocrine
Testicular Function in an Occupationally Exposed Population." Hum Toxicol. 7:125-128.
Roels, H.A., Lauwerys, R.R., Buchet, J-P, et al. (1975) "Response of Free Erythrocyte
Porphyrin and Urinary-6-aminolevulinic Acid in Men and Women Moderately Exposed to
Lead." Int Arch Arbeitsmed 34:97-108.
Roels, H.A., Buchet, J-P, Lauwerys, R., et al. (1976) "Impact of Air Pollution by Lead on the
Heme Biosynthetic Pathway in School-age Children." Archives of Environmental Health.
31:310-316.
Roels, H.A., Balis-Jacques, M.N., Buchet, J-P, et al. (1979) "The Influence of Sex and
Chelation Therapy on Erythrocyte Protoporphyrin and Urinary-6-aminolevulinic Acid hi Lead-
exposed Workers." Journal of Occupational Medicine. 21:527-539.
Roels, H.A., Lauwerys, R. (1987) "Evaluation of Dose-effect and Dose-response Relationships
for Lead Exposure in Different Belgian Population Groups (Fetus, Child, Adult Men and
Women)." Trace Elements in Medicine. 4:80-87.
Rosen, J.F., Zarate-Salvador, C., Trinidad, E.E. (1974) "Plasma Lead Levels hi Normal and
Lead-intoxicated Children." Journal of Pediatrics. 84:45-48.
Rosen, J.F., Chesney, R.W., Hamstra, A.J., et al. (1980) "Reduction in 1,25-dihydroxyvitamine
D in Children with Increased Lead Absorption.'Wew England Journal of Medicine. 302:1128-
1131.
Rosen, I., Wildt, K., Gullberg, B., et al. (1983) "Neurophysiological Effects of Lead Exposure."
ScandJ Work Environ Health. 9:431-441.
Rothenberg, S.J., Schnaas, L., Cansino-Ortiz, S., et al. (1989) "Neurobehavioral Deficits After
Low Level Lead Exposure in Neonates: The Mexico City Pilot Study." Neurotoxicol Teratol.
11:85-93.
Rummo, J.H., Routh, O.K., Rummo, N.J., et al. (1979) "Behavioral and Neurological Effects
of Symptomatic and Asymptomatic Lead Exposure in Children." Archives of Environmental
Health. 34:120-125.
Rust, S.W., Burgoon, D.A. (1993) "Development of Health-Based Standards for Lead in
Residential Environments." December 1993.
Rust, S.W., Kumar, P., Burgoon, D.A., Schultz, B. (1996) "Influence of Bone-lead Stores on
the Observed Effectiveness of Lead Hazard Intervention." In press.
Ryan, C.M., Morrow., L., Parkinson, D., et al. (1987) "Low Level Lead Exposure and
Neuropsychological Functioning in Blue Collar Males." Int JNeurosci. 36:29-39.
Draft - Do Not Cite or Quote 254 September 27, 1996
-------
Sachs, H.K., Moel, D.I. (1989) "Height and Weight Following Lead Poisoning in Childhood."
American Journal of Diseases and Children. 143:820-822.
Saenger, P., Markowitz, M.E., Rosen, J.F. (1984) "Depressed Excretion of 6p-hydroxycortisol
in Lead-toxic Children." J Clin Endocrinol Metab. 58:363-367.
Schmid, E., Bauchinger, M., Pietruck, S., et al. (1972) "Cytogenic Action of Lead in Human
Peripheral Lymphocytes In Vitro and In Vivo." Mat Res. 16:401-406. (German)
Schneitzer, L., Osbom, H.H., Bierman, A., et al. (1990) "Lead Poisoning in Adults from
Renovation of a Older Home." Ann Emerg Med. 19:415-420.
Schroeder, H.A., Tipton, I.H. (1968) "The Human Body Burden of Lead." Archives of
Environmental Health. 17:965-977.
Schroeder, S.R., Hawk, B., Otto, D., Mushak, P., Hicks, R.E. (1985) "Separating the Effects of
Lead and Social Factors on IQ." Environmental Research. 91:178-183.
Schroeder, S.R., Hawk, B. (1987) "Psycho-social Factors, Lead Exposure and IQ." MonogrAm
Assoc Ment Deflc. 8:97-137.
Schutz, A., Skerfving, S., Ranstam, J., Christoffersson, J. (1987) "Kinetics of Lead in Blood
after the End of Occupational Exposure." Scand. J. Work Environ. Health. 13:221-231.
Schwartz, J., Otto, D.A. (1987) "Blood Lead, Hearing Thresholds, and Neurobehavioral
Development in Children and Youth." Archives of Environmental Health. 42:153-160.
Schwartz, J. (1988) "The Relationship Between Blood Lead and Blood Pressure in the
NHANES II Survey." Environmental Health Perspectives. 78:15-22.
Schwartz, J., Landrigan, P.J., Feldman, R.G., et al. (1988) "Threshold Effect in Lead-induced
Peripheral Neuropathy." Journal of Pediatrics. 112:12-17.
Schwartz, J., Landrigan, P.J., Baker, E.L., Jr. (1990) "Lead-induced Anemia: Dose-response
Relationships and Evidence for a Threshold." American Journal of Public Health. 80:165-168.
Schwartz, J. (1993) "Beyond LOEL's, p Values, and Vote Counting: Methods for Looking at
the Shapes and Strengths of Association." Neurotoxicology and Teratology. 14(2-3):237-246.
Schwartz, J. (1994) "Low-Level Lead Exposure and Children's IQ: A Meta-Analysis and
Search for a Threshold." Environmental Research. 65:42-55.
Draft - Do Not Cite or Quote 255 September 27, 1996
-------
Secchi, G.C., Erba, L., Cambiaghi, G. (1974) "6-aminolevulinic Acid Dehydratase Activity of
Erythrocytes and Liver Tissue in Man: Relationship to Lead Exposure." Archives of
Environmental Health. 28:130-132.
Selander, S., Cramer, K. (1970) "Interrelationships Between Lead in Blood, Lead in Urine, and
ALA in Urine During Lead Work." British Journal of Industrial Medicine. 27:28-39.
Seppalainen, A.M., Hernberg, S., Vesanto, R., et al. (1983) "Early Neurotoxic Effects of
Occupational Lead Exposure: A Prospective Study." Neurotoxicology. 4:181-192.
Seto, D.S.Y., Freeman, J.M. (1964) "Lead Neuropathy in Childhood." Am JDis Child.
107:337-342.
Shaheen, S. (1984) "Neuromaturation and Behavior Development: The Case of Childhood
Lead Poisoning." Dev. Psych. 20:542-550.
Shukla., R. Bornschein, R.L., Dietrich, K.N., et al. (1987) "Effects of Fetal and Early Postnatal
Lead Exposure on Child's Growth in Stature — The Cincinnati Lead Study." In: Lindberg, S.E.
Hutchinson, T.C., eds. International Conference on Heavy Metals in the Environment, Vol. 1.
New Orleans, LA, September. Edinburgh, UK: CEP Consultants, Ltd., 210-212.
Shukla, R., Bornschein, R.L., Dietrich, K.N., et al. (1989) "Fetal and Infant Lead Exposure:
Effects on Growth in Stature." Pediatrics. 84:604-612.
Siegel, M., Forsyth, B., Siegel, L., et al. (1989) "The Effect of Lead on Thryoid Function in
Children." Environmental Research. 49:190-196.
Silva, P.A., Hughes, P., Williams, S., Faed, J.M. (1988) "Blood Lead, Intelligence, Reading
Attainment, and Behaviour in Eleven Year Old Children in Dunedin, New Zealand." Journal
of Child Psychology and Psychiatry. 29:43-52.
Silver, W., Rodriguez-Torres, R. (1968) "Electrocardiographic Studies in Children with Lead
Poisoning." Pediatrics. 41:1124-1127.
Smith, M., Delves, T., Lansdown, R., et al. (1983) "The Effects of Lead Exposure on Urban
Children: The Institute of Child Health/Southampton Study." Dev Med Child Neural. 25 (suppl
47).
Smith, F.L., II, Rathmell, T.K., Marcil, G.E. (1938) "The Early Diagnosis of Acute and Latent
Plumbism." AmJClin Pathol. 8:471-508.
Smith, D.R., Osterloh, J.D., Flegal, A.R. (1996) "Use of Endogenous Stable Lead Isotopes to
Determine Release of Lead from the Skeleton." Environmental Health Perspectives. 104(1):60-
66.
Draft - Do Not Cite or Quote 256 September 27. 1996
-------
Spivey, G.H., Baloh, R.W., Brown, C.P., et al. (1980) "Subclinical Effects of Chronic
Increased Lead Absorption—A Prospective Study: III. Neurologic Findings at Follow-up
Examintion." Journal of Occupational Medicine. 22:607-612.
Staes, C., Matte, T., Copley, G., Flanders, D., Binder, S. (1994) "Retrospective Study of the
Impact of Lead-based Paint Hazard Remediation on Children's Blood Lead Levels in St. Louis,
Missouri." American Journal of Epidemiology. 139( 10): 1016-1026.
Staessen, J., Yeoman, W.B., Fletcher, A.E., et al. (1990) "Blood Lead Concentration, Renal
Function, and Blood Pressure in London Civil Servants." British Journal of Industrial
Medicine. 47:442-447.
Staessen, J., Sartor, F., Roels, H., et al. (1991) "The Association Between Blood Pressure,
Calcium and Other Divalent Cations: A Population Study." Journal of Human Hypertension.
5:485-494.
Stark, A.D., Meigs, J.W., Fitch, R.A., DeLouise, E.R. (1978) "Family Operational Cofactors in
the Epidemiology of Childhood Lead Poisoning." Archives of Environmental Health. 33:222-
226.
Stark, A.D., Quah, R.F., Meigs, J.W., DeLouise, E.R. (1982) "The Relationship of
Environmental Lead to Blood Lead Levels in Children." Environmental Research. 27:372-383.
Stollery, B.T., Banks, H.A., Broadbent, D.E., et al. (1989) "Cognitive Functioning in Lead
Workers." British Journal of Industrial Medicine. 46:698-707.
Strauss, W., Buxton, B., Rust, S., Boyd, H. (1996) "Statistical Evaluation of the Relationship
Between Blood-Lead and Dust-Lead Based on Data from the Rochester Lead-in-Dust Study."
February 1996.
Strauss, W., Rust, S., Boyd, H. (1996) "Statistical Evaluation of the Relationship Between
Blood-Lead and Dust-Lead Based on Pre-Intervention Data from the Baltimore Repair and
Maintainence Study." February 1996.
Stuik, E.J. (1974) "Biological Response of Male and Female Volunteers to Inorganic Lead." Int
Arch Arbeitsmed. 33:83-97.
Thornton I, Davies D.J., Watt J.M., Quinn MJ. (1990) "Lead Exposure in Young Children
From Dust and Soil in the United Kingdom." Environmental Health Perspectives. 89:55-60.
Tola, S., Hernberg, S., Asp, S., et al. (1973) "Parameters Indicative of Absorption and
Biological Effect in New Lead Exposure: A Prospective Study." British Journal of Industrial
Medicine. 30:134-141.
Draft - Do Not Cite or Quote 257 September 27, 1996
-------
Triebig, G., Weltle, D., Vanentin, H. (1984) "Investigations on Neurotoxicity of Chemical
Substances at the Workplace: V. Determination of the Motor and Sensory Nerve Conduction
Velocity in Persons Occupationally Exposed to Lead." International Archives of Occupational
Environmental Health. 53:189-204.
Tuppurainen, M., Wagar, G., Kurppa, K. (1988) "Thyroid Function as Assessed by Routine
Laboratory Tests of Workers with Long-term Lead Exposure." ScandJ Work Environ Health.
14:175-180.
U.S. Department of Health and Human Services (1992) "Sample Design: Third National
Health and Nutrition Examination Survey." DHHS Publication No. (PHS) 92-1887.
U.S. Department of Health and Human Services (1994) "Plan and Operation of the Third
National Health and Nutrition Examination Survey." 1988-94 DHHS Publication No. (PHS)
94-1308.
U.S. Department of Housing and Urban Development, Office of Policy Development and
Research. (1990) "Comprehensive and Workable Plan for the Abatement of Lead-Based Paint
in Privately Owned Housing," Report to Congress, December 1990.
U.S. Department of Housing and Urban Development (1995a) "The Relation of Lead-
Contaminated House Dust and Blood Lead Levels Among Urban Children." Volumes I and n.
Final Report to the U.S. Department of Housing and Urdan Development Grant from The
University of Rochester School of Medicine, Rochester, New York, and The National Center
for Lead-Safe Housing, Columbia Maryland.
U.S. Department of Housing and Urban Development (1995b) "Guidelines for the Evaluation
and Control of Lead-Based Paint Hazards in Housing." Office of Lead-Based Paint Abatement
and Poisoning Prevention.
U.S. Environmental Protection Agency (1986) Air Quality Criteria for Lead. Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of Research and Development,
Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office. EPA 600/8-83-028F.
U.S. Environmental Protection Agency (1990a) "Air Quality Criteria for Lead: Supplement to
the 1986 Addendum." Office of Research and Development, Washington, D.C. EPA/600/8-
89/049F.
U.S. Environmental Protection Agency (1990b). "National Air Quality and Emissions Trends
Report, 1988." Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency. EPA 450/4-90-002.
Draft - Do Not Cite or Quote 258 September 27, 1996
-------
U.S. Environmental Protection Agency (1992) "Boston Lead-in-Soil/Lead Free Kids
Demonstration Project." Final Report, July 1992.
U.S. Environmental Protection Agency (1993) "Cincinnati Soil Lead Abatement
Demonstration Project." Final Report, March 1993.
U.S. Environmental Protection Agency (1994a) "Guidance Manual for the Integrated Exposure
Uptake Biokinetic Model for Lead in Children." Washington, DC. EPA/540/R-93/081, PB93-
963510.
U.S. Environmental Protection Agency (1994b) "Validation Strategy for the Integrated
Exposure Uptake Biokinetic Model for Lead in Children." EPA/540/R-94/039.
U.S. Environmental Protection Agency (1994c) "Descriptive Statistics From the Initial
Sampling Campaign of the Lead Paint Abatement and Repair and Maintenance Study in
Baltimore." Draft Final Report to the U.S. Environmental Protection Agency from the
Kennedy Krieger Institute.
U.S. Environmental Protection Agency (1994d) "Reducing Lead Hazards When Remodeling
Your Home." Office of Pollution Prevention and Toxics, U.S. Environmental Protection
Agency. EPA 747-R-94-002, April 1994.
U.S. Environmental Protection Agency (1995a) "Report on the National Survey of Lead-Based
Paint in Housing: Base Report." Office of Pollution Prevention and Toxics, U.S.
Environmental Protection Agency, EPA 747-R95-003, April 1995.
U.S. Environmental Protection Agency (1995b) "Review of Studies Addressing Lead
Abatement Effectiveness." Office of Pollution Prevention and Toxics, U.S. Environmental
Protection Agency. EPA 747-R-95-006, July 1995.
U.S. Environmental Protection Agency (1995c) "Sampling House Dust for Lead: Basic
Concepts and Literature Review." Office of Pollution Prevention and Toxics, U.S.
Environmental Protection Agency. EPA 747-R-95-007, September 1995.
U.S. Environmental Protection Agency (1995d) "Laboratory Evaluation of Dust and Dust Lead
Recoveries for Samplers and Vacuum Cleaners: Vol. I and II." Office of Pollution Prevention
and Toxics, U.S. Environmental Protection Agency. EPA 747-R-95-004A, EPA 747-R-95-
004B, March 1995.
U.S. Environmental Protection Agency (1995e) "Technical Support Document: Parameters and
Equations Used in the Integrated Exposure Uptake Biokinetic Model for Lead in Children (v
0.99D)." EPA 540/R-94/040.
Draft - Do Not Cite or Quote 259 September 27, 1996
-------
U.S. Environmental Protection Agency (1995f) "Comprehensive Abatement Performance
Study, Volume I: Summary Report." Final Report prepared by Battelle to the Office of
Pollution Prevention and Toxics, U.S. Environmental Protection Agency, EPA Technical
Report No. 230-R-94-013a.
U.S. Environmental Protection Agency (1995g) "Urban Soil Lead Abatement Demonstration
Project EPA Integrated Report." EPA/600/R-95/139, September 1995.
U.S. Environmental Protection Agency (1995h) "Guidance on Identification of Lead-Based
Paint Hazards." Notice. Federal Register, pp 47248-47257, September 11,1995.
U.S. Environmental Protection Agency (1996) "Urban Soil Lead Abatement Demonstration
Project, Volume I: Integrated Report." Office of Research and Development, U.S.
Environmental Protection Agency. EPA 600-P-93-001aF, April 1996.
Valciukas, J.A., Lilis, R., Eisinger, J., et al. (1978) "Behavioral Indicators of Lead
Neurotoxicity: Results of a Clinical Field Survey." International Archives of Occupational
Environmental Health. 41:217-236.
Vershoor, M., Wibowo, A., Berber, R., et al. (1987) "Influence of Occupational Low-level
Lead Exposure on Renal Parameters." American Journal of Industrial Medicine. 12:341-351.
Vimpani, G.V., Wigg, N.R., Robertson, E.F., et al. (1985) "The Port Pirie Cohort Study: Blood
Lead Concentration and Childhood Developmental Assessment". Presented at: Lead
Environmental Health: Current Issues, May, Duke University, Durham, NC.
Vimpani, G.V., Baghurst, P.A., Wigg, N.R., et al. (1989) "The Port Pirie Cohort Study —
Cumulative Lead Exposure and Neurodevelopmental Status at Age 2 Years: Do HOME Scores
and Maternal IQ Reduce Apparent Effects of Lead on Bayley Mental Scores?" In: Smith, M.,
Grant., L.D., Sors, A., eds. Lead Exposure and Child Development: An International
Assessment. Lancaster, UK: Kluwer Academic Press.
Wada, O., Yano, Y, Ono, T., et al. (1973) "The Diagnosis of Different Degrees of Lead
Absorption in Special References to Choice and Evaluation of Various Parameters Indicative
of an Increased Lead Absorption." Industrial Health. 11:55-67.
Wallsten, T.S., Whitfield, R.G. (1986) "Assessing the Risks to Young Children of Three
Effects Associated with Elevated Blood-lead Levels." Report by Argonne National Laboratory.
Report No. ANL/AA-32. Sponsored by the U.S. EPA Office of Air Quality Planning and
Standards.
Wang, L., Xu S.E., Zhang, G.D, Wang, W.Y. (1989) "Study of Lead Absorption and its Effect
on Children's Development." Biomedical and Environmental Sciences. 2:325-330.
Draft - Do Not Cite or Quote 260 September 27, 1996
-------
Ward, N.I., Watson, R., Bryce-Smith, D. (1987) "Placental Element Levels in Relation to Fetal
Development for Obstetrically Normal Births: A Study of 37 Elements: Evidence for the
Effects of Cadmium, Lead, and Zinc on Fetal Growth and for Smoking as a Source of
Cadmium." IntJBiosoc Res. 9:63-81.
Wedeen, R.P., Mallik, O.K., Batuman, V. (1979) "Detection and Treatment of Occupational
Lead Nephropathy." Archives of Internal Medicine. 139:53-57.
Wedeen, R.P. (1988) "Bone Lead, Hypertension, and Lead Nephropathy." Environmental
Health Perspectives. 78:57-60.
Weiss, ST., Munoz, A., Stein, A., et al. (1986) "The Relationship of Blood Lead to Blood
Pressure in a Longitudinal Study of Working Men." Am JEpidemiol. 123:800-808.
Weiss, ST., Munoz, A., Stein, A., et al. (1988) "The Relationship of Blood Lead to Systolic
Blood Pressure in a Longitudinal Study of Policemen." Environmental Health Perspectives.
78:53-56.
Weitzman, M., Aschengrau, A., Bellinger, D., Jones, R., Hamlin, J.S., Beiser, A. (1993) "Lead-
Contaminated Soil Abatement and Urban Children's Blood Lead Levels." Journal of the
American Medical Association. 269(13): 1647-1654.
Wibberley, D.G., Khera, A.K., Edwards, J.H., et al. (1977) "Lead Levels in Human Placentae
from Normal and Malformed Births." JMed Genet. 14:339-345.
Wigg, N.R., Vimpani, G.V., McMichael, A.J., et al. (1988) "Port Pirie Cohort Study:
Childhood Blood Lead and Neuropsychological Development at Age Two Years." J Epidemiol
Community Health. 42:213-219.
Wildt, K., Eliasson, R., Berlin, M. (1983) "Effects of Occupational Exposure to Lead on Sperm
and Semen." In: Clarkson, T.W., Nordberg, G.F., Sager, P.R., eds. Reproductive and
Developmental Toxicity of Metals. Proceedings of a Joint Meeting, Rochester, NY, May 1982.
New York, NY: Plenum Press. 279-300.
Wilhelm, M., Lombeck, I., Hamer, D., et al. (1989) "Hair Lead Levels in Young Children from
the F.R.G." J Trace Elements and Electrolytes in Health and Disease, 3:165-170.
Williamson, A.M., Teo, R.K.C. (1986) "Neurobehavioral Effects of Occupational Exposure to
Lead." British Journal of Industrial Medicine. 43:374-380.
Winneke, G., Beginn, U., Ewert, T., et al. (1985) "Comparing the Effects of Perinatal and Later
Childhood Lead Exposure on Neurophysiological Outcome." Environmental Research. 38:155-
167.
Draft - Do Not Cite or Quote 261 September 27, 1996
-------
Winneke, G., Brockhaus, A., Collet, W., et al. (1985) "Predictive Value of Different Markers
of Lead-exposure for Neuropsychological Performance." In: Lekkas, T.D., ed. International
Conference on Heavy Metals in the Environment, Athens, Greece, September, Vol. 1.
Edinburgh, United Kingdom: CEP Consultants, Ltd., 44-47.
Wolf, A.W., Ernhart, C.B., White, C.S. (1985) "Intrauterine Lead Exposure and Early
Development." In: Lekkas T.D., ed. International Conference: Heavy Metals in the
Environment, Athens, Greece, September, Vol. 2. Edinburgh, United Kingdom: CEP
Consultants, Ltd. 153-155.
Wrenn, M.E., Cohen, M., Rosen, J.C., Eisenbud, M., Blanchard, R.L. (1972) "In-vivo
Measurements of Lead-210 in Man. Assessment of Radioactive Contamination in Man." IAEA.
129-146.
Yule, W., Lansdown, R., Millar, I., Urbanowiez, M. (1981) "The Relationship Between Blood
Lead Concentration, Intelligence, and Attainment in a School Population: a Pilot Study." Dev
Med Child Neuro. 23:567-576.
Zimmerman-Tansella, C., Campara, P., D' Andrea, F., et al. (1983) "Psychological and Physical
Complaints of Subjects with Low Exposure to Lead." Hum Toxicol. 2:651-623.
Draft - Do Not Cite or Quote 262 September 27, 1996
------- |