OSWER 9285.7-77
May 2007
ESTIMATION OF
RELATIVE BIO AVAILABILITY OF LEAD
IN SOIL AND SOIL-LIKE MATERIALS USING
IN VIVO AND IN VITRO METHODS
Office of Solid Waste and Emergency Response
U.S. Environmental Protection Agency
Washington, DC 20460
-------
ACKNOWLEDGMENTS
The work described in this report is the product of a team effort involving a large number
of people. In particular, the following individuals contributed significantly to the findings
reported here and the preparation of this report:
PROGRAM SUPPORT
U.S. Environmental Protection Agency (U.S. EPA) support for the development of this
report was provided by Michael Beringer, U.S. EPA Region 7, Kansas City, KS; Jim Luey, U.S.
EPA Region 8, Denver, CO; and Richard Troast, formerly with U.S. EPA Office of Superfund
Remediation and Technology Innovation, Washington, DC. Contractor support to U.S. EPA was
provided by Syracuse Research Corporation.
IN VIVO STUDIES
All of the in vivo studies described in this report were planned and sponsored by U.S.
EPA, Region 8. The technical direction for all aspects of the in vivo portion of this project was
provided by Christopher P. Weis, PhD, DABT, and Gerry M. Henningsen, DVM, PhD,
DABT/DABVT. Mr. Stan Christensen provided oversight and quality assurance support for
analyses of blood during the later studies performed in this program.
All of the in vivo studies described in this report were performed by Stan W. Casteel,
DVM, PhD, DABVT, at the Veterinary Medical Diagnostic Laboratory, College of Veterinary
Medicine, University of Missouri, Columbia, Missouri. Dr. Casteel was supported by Larry D.
Brown, DVM, MPH, Ross P. Cowart, DVM, MS, DACVEVI, James R. Turk, DVM, PhD,
DACVP, John T. Payne, DVM, MS, DACVS, Steven L. Stockham, DVM, MS, DACVP, and
Roberto E. Guzman, DVM, MS. Analysis of biological samples (blood, tissues) was performed
by Dr. Edward Hindenberger, of L.E.T., Inc, Columbia, Missouri.
IN VITRO STUDIES
Development of the method used to estimate in vitro bioaccessibility was performed
primarily by John Drexler, PhD, at the University of Colorado, Boulder, with input and
suggestions from a consortium of industry, academic, and governmental personnel organized by
Mr. Michael V. Ruby at Exponent. Dr. Drexler also performed all of the electron microprobe
and particle size analyses of the test materials evaluated in these studies.
STATISTICAL ANALYSIS
Dr. Timothy Barry, U.S. EPA National Center for Environmental Economics, provided
on going support in the selection and application of the statistical methods used in dose-response
curve-fitting and data reduction. In addition, Glenn Shaul and Lauren Drees at U.S. EPA's
National Risk Management Research Laboratory provided several rounds of valuable review
comments and constructive discussions regarding statistical methodology.
OSWER 9285.7-77
-------
REVIEWERS
A draft of this report was provided to three independent experts for external peer review
and comment. This satisfies the Agency's requirements for peer review. These reviewers were:
Paul Mushak, PB Associates, Durham, NC
Michael Rabinowitz, Marine Biological Laboratory, Woods Hole, MA
Rosalind Schoof, Integral Consulting, Inc., Mercer Island, WA
The Agency has responded to the peer review comments, as appropriate. The comments and
Agency responses are contained in a responsiveness summary that has been placed in the
Administrative Record.
OSWER 9285.7-77
-------
This page intentionally left blank to facilitate double-sided printing.
-------
EXECUTIVE SUMMARY
1.0 INTRODUCTION
Reliable analysis of the potential hazard to children from ingestion of lead in
environmental media depends on accurate information on a number of key parameters, including
the rate and extent of lead absorption from each medium ("bioavailability"). Bioavailability of
lead in a particular medium may be expressed either in absolute terms (absolute bioavailability,
ABA) or in relative terms (relative bioavailability, RBA). For example, if 100 micrograms (ug)
of lead dissolved in drinking water were ingested and a total of 50 ug were absorbed into the
body, the ABA would be 0.50 (50%). Likewise, if 100 ug of lead contained in soil were
ingested and 30 ug were absorbed into the body, the ABA for soil would be 0.30 (30%). If the
lead dissolved in water was used as the frame of reference for describing the relative amount of
lead absorbed from soil, the RBA would be 0.30/0.50, or 0.60 (60%).
When reliable data are available on the absolute or relative bioavailability of lead in soil,
dust, or other soil-like waste material at a site, this information can be used to improve the
accuracy of exposure and risk calculations at that site. Based on available information in the
literature on lead absorption in humans, the U.S. Environmental Protection Agency (U.S. EPA)
estimates that relative bioavailability of lead in soil compared to water and food is about 60%.
Thus, when the measured RBA in soil or dust at a site is found to be less than 60%, it may be
concluded that exposures to and hazards from lead in these media at that site are probably lower
than typical default assumptions. Conversely, if the measured RBA is higher than 60%,
absorption of and hazards from lead in these media may be higher than usually assumed.
This report summarizes the results of a series of studies performed by scientists in U.S.
EPA Region 8 to measure the RBA of lead in a variety of soil and soil-like test materials using
both in vivo and in vitro techniques.
2.0 IN VIVO STUDIES
Basic Approach for Measuring RBA In Vivo
The in vivo method used to estimate the RBA of lead in a particular test material
compared to lead in a reference material (lead acetate) is based on the principle that equal
absorbed doses of lead will produce equal increases in lead concentration in the tissues of
exposed animals. Stated another way, RBA is the ratio of oral doses that produce equal
increases in tissue burden of lead.
Based on this, the technique for estimating lead RBA in a test material is to administer a
series of oral doses of reference material (lead acetate) and test material (site soil) to groups of
experimental animals, and to measure the increase in lead concentration in one or more tissues in
the animals. For each tissue, the RBA is calculated by fitting an appropriate dose-response
model to the data, and then solving the equations to find the ratio of doses that produce equal
responses. The final estimate of RBA for the test material then combines the RBA estimates
across the different tissues.
OSWER 9285.7-77 ES-1
-------
Animal Exposure and Sample Collection
All animals used in this program were intact male swine approximately 5 to 6 weeks of
age. In general, exposure occurred twice a day for 15 days. Most groups were exposed by oral
administration, with one group usually exposed to lead acetate by intravenous injection.
Lead concentrations were measured in four different tissues: blood, liver, kidney, and
bone. For blood, samples were collected from each animal at multiple times during the course of
the study (e.g., days 0, 1, 2, 3, 4, 6, 9, 12, and 15), and the blood concentration integrated over
time (commonly referred to as "area under the curve" or AUC) was used as the measure of blood
lead response. For liver, kidney, and bone, the measure of response was the concentration of
lead in these tissues on day 15.
Calculation of RBA
Based on testing several different types of dose-response models to the data, it was
concluded that most dose-response curves for liver, kidney, and bone lead were well described
by a linear model, and that most blood lead AUC data sets were well described by an exponential
model:
Liver, Kidney, Bone
Blood AUC
AUC = a + b-[l- exp(-c Dose)]
where Ctissue is the concentration of lead in a given tissue; a, b, and c are the terms of the
mathematic equation used to describe the shape of the curve; and Dose is the total daily
administered dose of lead (ug/kg-day).
Based on these models, RBA is calculated from the best model fits as follows:
bt.
test material
iver, kidney, bone
b f . . i
reference material
test material
reference material
Results and Discussion
RBA Values for Various Test Materials
Table ES-1 lists the 19 different materials tested in this program and shows the RBA
values estimated using each of the four alternative endpoints (blood AUC, liver, kidney, bone).
OSWER 9285.7-77 ES-2
-------
Based on an analysis that indicated that each endpoint has approximately equal reliability, the
point estimate for each test material is the mean of the four endpoint-specific values.
Inspection of these KB A point estimates for the different test materials reveals that there
is a wide range of values across different samples, both within and across sites. For example, at
the California Gulch site in Colorado, KB A estimates for different types of material range from
about 6% (Oregon Gulch tailings) to 105% (Fe/Mn lead oxide sample). This wide variability
highlights the importance of obtaining and applying reliable KB A data in order help to improve
risk assessments for lead exposure.
Correlation ofRBA with Mineral Phase
Available data are not yet sufficient to establish reliable quantitative estimates of KB A
for each of the different mineral phases of lead that are observed to occur in the test materials.
However, multivariate regression analysis between point estimate KBA values and mineral phase
content of the different test materials allows a tentative rank ordering of the phases into three
semi-quantitative tiers (low, medium, or high KBA), as follows:
Low Bioavailability
Fe(M) Sulfate
Anglesite
Galena
Pb(M) Oxide
Fe(M) Oxide
Medium Bioavailability
Lead Phosphate
Lead Oxide
High Bioavailability
Cerussite
Mn(M) Oxide
(M) = Metal
3.0 IN VITRO STUDIES
Measurement of lead KBA in animals has a number of potential benefits, but is also
rather slow and costly and may not be feasible in all cases. It is mainly for this reason that a
number of scientists have been working to develop alternative in vitro procedures that may
provide a faster and less costly alternative for estimating the KBA of lead in soil or soil-like
samples. These methods are based on the concept that the rate and/or extent of lead
solubilization in gastrointestinal fluid is likely to be an important determinant of lead
bioavailability in vivo, and most in vitro tests are aimed at measuring the rate or extent of lead
solubilization in an extraction solvent that resembles gastric fluid. The fraction of lead which
solubilizes in an in vitro system is referred to as in vitro bioaccessibility (IVBA).
Description of the Method
The IVBA extraction procedure is begun by placing 1.0 g of test substrate into a bottle
and adding 100 mL of extraction fluid (0.4 M glycine, pH 1.5). This pH is selected because it is
similar to the pH in the stomach of a fasting human. Each bottle is placed into a water bath
adjusted to 37°C, and samples are extracted by rotating the samples end-over-end for 1 hour.
After 1 hour, the bottles are removed, dried, and placed upright on the bench top to allow the soil
OSWER 9285.7-77
-------
to settle to the bottom. A sample of supernatant fluid is removed directly from the extraction
bottle into a disposable syringe and is filtered to remove any particulate matter. This filtered
sample of extraction fluid is then analyzed for lead.
Results
Table ES-2 summarizes the in vitro bioaccessibility results for the set of 19 different test
materials evaluated under the Phase II program. As seen, IVBA values span a considerable
range (min of 4.5%, max of 87%), with a mean of about 55%. This variability among test
materials indicates that the rate and extent of solubilization of lead from the solid test material
into the extraction fluid do depend on the attributes of the test material, and that IVBA may be a
useful indication of absorption in vivo (see below).
Comparison of In Vivo and In Vitro Results
In order for an in vitro bioaccessibility test system to be useful in predicting the in vivo
KB A of a test material, it is necessary to establish empirically that a strong correlation exists
between the in vivo and the in vitro results across many different samples. Figure ES-1 shows
the best fit weighted linear regression correlation between the in vivo lead RBA estimates and
the in vitro lead bioaccessibility estimates for each of the 19 test materials investigated during
this program. The equation of the line is:
RBA = 0.878-IVBA -0.028 (r2 = 0.924)
These results indicate that the in vivo RBA of lead in soil-like materials can be estimated
by measuring the IVBA and using the equation above to calculate the expected in vivo RBA.
Actual RBA values may be either higher or lower than the expected value, as indicated by the
95% prediction interval shown in Figure ES-1.
At present, it appears that this equation is likely to be widely applicable, having been
found to hold true for a wide range of different soil types and lead phases from a variety of
different sites. However, most of the samples tested have been collected from mining and
milling sites, and it is plausible that some forms of lead that do not occur at this type of site
might not follow the observed correlation. Thus, whenever a sample that contains an unusual
and/or untested lead phase is evaluated by the in vitro bioaccessibility protocol, this should be
identified as a potential source of uncertainty. In the future, as additional samples with a variety
of new and different lead forms are tested by both in vivo and in vitro methods, the applicability
of the method will be more clearly defined.
4.0 CONCLUSIONS
The data from the investigations performed under this program support the following
main conclusions:
1. Juvenile swine are believed to be a useful model for the evaluation of lead absorption in
children and provide a reliable system for measuring the RBA of lead in a variety of soil
and soil-like materials.
OSWER 9285.7-77 ES-4
-------
2. Each of the four different endpoints employed in these studies (blood AUC, liver, kidney,
bone) to estimate KB A in vivo yield reasonable data, and the best estimate of the KB A
value for any particular sample is the average across all four endpoint-specific KBA
values.
3. There are clear differences in the in vivo KBA of lead between different types of test
material, ranging from near zero to close to 100%. Thus, knowledge of the KBA value
for different types of materials at a site can be very important in improving lead risk
assessments at a site.
4. Available data support the view that certain types of lead minerals are well-absorbed
(e.g., cerussite, manganese lead oxide), while other forms are poorly absorbed (e.g.,
galena, anglesite). However, the data are not yet sufficient to allow reliable quantitative
calculation or prediction of the KBA for a test material based on knowledge of the lead
mineral content alone.
5. In vitro measurements of bioaccessibility performed using the protocol described in this
report correlate well with in vivo measurements of KBA, at least for 19 materials tested
under this program. At present, the results appear to be broadly applicable, although
further testing of a variety of different lead forms is required to determine if there are
exceptions to the apparent correlation.
OSWER 9285.7-77 ES-5
-------
TABLE ES-1. SUMMARY OF ESTIMATED RBA VALUES FOR TEST MATERIALS
Experiment
2
3
4
5
6
7
8
9
11
12
Test Material
Bingham Creek Residential
Bingham Creek Channel Soil
Jasper County High Lead Smelter
Jasper County Low Lead Yard
Murray Smelter Slag
Jasper County High Lead Mill
Aspen Berm
Aspen Residential
Midvale Slag
Butte Soil
California Gulch Phase I Residential
Soil
California Gulch Fe/Mn PbO
California Gulch AV Slag
Palmerton Location 2
Palmerton Location 4
Murray Smelter Soil
NIST Paint
Galena-enriched Soil
California Gulch Oregon Gulch
Tailings
Blood AUC
RBA
0.34
0.30
0.65
0.94
0.47
0.84
0.69
0.72
0.21
0.19
0.88
1.16
0.26
0.82
0.62
0.70
0.86
0.01
0.07
LB
0.23
0.20
0.47
0.66
0.33
0.58
0.54
0.56
0.15
0.14
0.62
0.83
0.19
0.61
0.47
0.54
0.66
0.00
0.04
UB
0.50
0.45
0.89
1.30
0.67
1.21
0.87
0.91
0.31
0.29
1.34
1.76
0.36
1.05
0.80
0.89
1.09
0.02
0.13
Liver
RBA
0.28
0.24
0.56
1.00
0.51
0.86
0.87
0.77
0.13
0.13
0.75
0.99
0.19
0.60
0.53
0.58
0.73
0.02
0.11
LB
0.20
0.17
0.42
0.75
0.33
0.54
0.58
0.50
0.09
0.09
0.53
0.69
0.11
0.41
0.37
0.42
0.52
0.00
0.04
UB
0.39
0.34
0.75
1.34
0.88
1.47
1.39
1.21
0.17
0.19
1.12
1.46
0.32
0.91
0.79
0.80
1.03
0.04
0.21
Kidney
RBA
0.22
0.27
0.58
0.91
0.31
0.70
0.73
0.78
0.12
0.15
0.73
1.25
0.14
0.51
0.41
0.36
0.55
0.01
0.05
LB
0.15
0.19
0.43
0.68
0.22
0.50
0.46
0.49
0.08
0.09
0.50
0.88
0.08
0.30
0.25
0.25
0.38
0.00
0.02
UB
0.31
0.37
0.79
1.24
0.46
1.02
1.26
1.33
0.18
0.22
1.12
1.91
0.25
0.91
0.72
0.52
0.78
0.02
0.09
Femur
RBA
0.24
0.26
0.65
0.75
0.31
0.89
0.67
0.73
0.11
0.10
0.53
0.80
0.20
0.47
0.40
0.39
0.74
0.01
0.01
LB
0.19
0.21
0.52
0.60
0.23
0.69
0.51
0.56
0.06
0.04
0.33
0.51
0.13
0.37
0.32
0.31
0.59
-0.01
-0.04
UB
0.29
0.31
0.82
0.95
0.41
1.18
0.89
0.97
0.18
0.19
0.93
1.40
0.30
0.60
0.52
0.49
0.93
0.03
0.06
Point Estimate
RBA
0.27
0.27
0.61
0.90
0.40
0.82
0.74
0.75
0.14
0.14
0.72
1.05
0.20
0.60
0.49
0.51
0.72
0.01
0.06
LB
0.17
0.19
0.43
0.63
0.23
0.51
0.48
0.50
0.07
0.06
0.38
0.57
0.09
0.34
0.29
0.29
0.44
0.00
-0.01
UB
0.40
0.36
0.79
1.20
0.64
1.14
1.08
1.04
0.24
0.23
1.07
1.56
0.31
0.93
0.72
0.79
0.98
0.03
0.15
LB = 5% Lower Confidence Bound
UB = 95% Upper Confidence Bound
-------
TABLE ES-2 IN VITRO BIOACCESSIBILITY VALUES
Experiment
2
2
3
3
4
4
5
5
6
6
7
7
8
9
9
11
11
12
12
Test
Material
1
2
1
2
1
2
1
2
1
2
1
2
1
1
2
1
2
1
3
Sample
Bingham Creek Residential
Bingham Creek Channel Soil
Jasper County High Lead Smelter
Jasper County Low Lead Yard
Murray Smelter Slag
Jasper County High Lead Mill
Aspen Berm
Aspen Residential
Midvale Slag
Butte Soil
California Gulch Phase I Residential Soil
California Gulch Fe/Mn PbO
California Gulch AV Slag
Palmerton Location 2
Palmerton Location 4
Murray Smelter Soil
NIST Paint
Galena-enriched Soil
California Gulch Oregon Gulch Tailings
In Vitro Bioaccessibility (%)
(Mean ± Standard Deviation)
47.0 ±1.2
37.8 ±0.7
69.3 ±5. 5
79.0 ±5.6
64.3 ±7.3
85.3 ±0.2
64.9 ±1.6
71. 4 ±2.0
17.4 ±0.9
22.3 ±0.6
65.1 ±1.5
87.2 ±0.5
9.4 ±1.6
63.6 ±0.4
69.7 ±2. 7
74.7 ±6. 8
72.5 ±2.0
4.5 ±1.2
11. 2 ±0.9
-------
FIGURE ES-1. RELATION BETWEEN RBA AND IVBA
1.4
00
QL
1.2 -
1.0
0.8
0.6 -
0.4
0.2
0.0
95% Prediction Interval
0.8781VBA - 0.028
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
IVBA
-------
TABLE OF CONTENTS
1.0 INTRODUCTION 1
1.1 Overview 1
1.2 Using Bioavailability Data to Improve Exposure Calculations for Lead 2
1.3 Overview of U.S. EPA's Program to Study Lead Bioavailability in Animals 3
1.4 Overview of Methods for Estimating LeadRBA/w Vitro 3
2.0 IN VIVO STUDIES 4
2.1 Basic Approach for Measuring RBA In Vivo 4
2.2 Animal Exposure and Sample Collection 4
2.3 Preparation of Biological Samples for Analysis 5
2.4 Data Reduction 5
2.5 Results and Discussion 6
2.5.1 Effect of Dosing on Animal Health and Weight 6
2.5.2 Time Course of Blood Lead Response 6
2.5.3 Dose-Response Patterns 7
2.5.4 Estimation of ABA for Lead Acetate 7
2.5.5 Estimation of RBA for Lead in Test Materials 8
2.5.6 Effect of Food 9
2.5.7 Correlation of RBA with Mineral Phase 10
2.5.8 Quality Assurance 12
3.0 IN VITRO STUDIES 14
3.1 Introduction 14
3.2 In Vitro Method 14
3.2.1 Sample Preparation 14
3.2.2 Apparatus 14
3.2.3 Select!on of IVBA Test Conditions 15
3.2.4 Summary of Final Leaching Protocol 16
3.2.5 Analysis of Extraction Fluid for Lead 17
3.2.6 Quality Control/Quality Assurance 17
3.3 Results and Discussion 18
3.3.1 IVBA Values 18
3.3.2 Comparison with In Vivo Results 19
4.0 REFERENCES 21
OSWER 9285.7-77
-------
LIST OF TABLES
TABLE TITLE
2-1 Typical Feed Composition
2-2 Typical In Vivo Study Design
2-3 Description of Phase II Test Materials
2-4 Relative Lead Mass of Mineral Phases Observed in Test Materials
2-5 Matrix Associations for Test Materials
2-6 Particle Size Distributions for Test Materials
2-7 Estimated RBA Values for Test Materials
2-8 Grouped Lead Phases
2-9 Curve Fitting Parameters for Oral Lead Acetate Dose-Response Curves
2-10 Reproducibility of RBA Measurements
3-1 In Vitro Bioaccessibility Values
LIST OF FIGURES
FIGURE TITLE
2-1 Average Rate of Body Weight Gain in Test Animals
2-2 Example Time Course of Blood Lead Response
2-3 Dose Response Curve for Blood Lead AUC
2-4 Dose Response Curve for Liver Lead Concentration
2-5 Dose Response Curve for Kidney Lead Concentration
2-6 Dose Response Curve for Femur Lead Concentration
2-7 Estimated Group-Specific RBA Values
2-8 Correlation of Duplicate Analyses
2-9 Results for CDC Blood Lead Check Samples
2-10 Interlaboratory Comparison of Blood Lead Results
3-1 In Vitro Bioaccessibility Extraction Apparatus
3-2 Effect of Temperature, Time, and pH on IVBA
3-3 Precision of In Vitro Bioaccessibility Measurements
3-4 Reproducibility of In Vitro Bioaccessibility Measurements
3-5 RBA vs. IVBA
3-6 Prediction Interval for RBA Based on Measured IVBA
OSWER 9285.7-77
11
-------
LIST OF APPENDICES
APPENDIX TITLE
A Evaluation of Juvenile Swine as a Model for Gastrointestinal Absorption in
Young Children
B Detailed Description of Animal Exposure
C Detailed Methods of Sample Collection and Analysis
D Detailed Methods for Data Reduction and Statistical Analysis
E Detailed Dose-Response Data and Model Fitting Results
F Detailed Lead Speciation Data for Test Materials
OSWER 9285.7-77 111
-------
ACRONYMS AND ABBREVIATIONS
°C Degrees Celsius
ug Microgram
um Micrometer
ABA Absolute bioavailability
AF0 Oral absorption fraction
AIC Akaike's Information Criterion
AUC Area under the curve
cc Cubic centimeter
CDC Centers for Disease Control and Prevention
dL Deciliter
g Gram
GLP Good Laboratory Practices
HC1 Hydrochloric acid
HOPE High density polyethylene
ICP-AES Inductively Coupled Plasma-Atomic Emission Spectrometry
ICP-MS Inductively Coupled Plasma-Mass Spectrometry
IV Intravenous
IVBA In vitro bioaccessibility
kg Kilogram
L Liter
M Molar
(M) Metal
MDL Method detection limit
mg Milligram
mL Milliliter
mm Millimeter
NIST National Institute of Standards and Testing
Pb Lead
PbAc Lead acetate
ppm Parts per million
RBA Relative bioavailability
RLM Relative lead mass
OSWER 9285.7-77
IV
-------
ACRONYMS AND ABBREVIATIONS
(CONTINUED)
rpm Revolutions per minute
SOP Standard operating procedure
SRM Standard Reference Material
TAL Target Analyte List
TCLP Toxicity Characteristic Leaching Procedure
U.S. EPA U.S. Environmental Protection Agency
OSWER 9285.7-77
-------
This page intentionally left blank to facilitate double-sided printing.
-------
ESTIMATION OF RELATIVE BIOAVAILABILITY
OF LEAD IN SOIL AND SOIL-LIKE MATERIALS
USING IN VIVO AND IN VITRO METHODS
1.0 INTRODUCTION
1.1 Overview
Reliable analysis of the potential hazard to children from ingestion of lead in the
environment depends on accurate information on a number of key parameters, including 1) lead
concentration in environmental media (e.g., soil, dust, water, food, air, paint), 2) childhood
intake rates of each medium, and 3) the rate and extent of lead absorption from each medium
("bioavailability"). Knowledge of lead bioavailability is important because the amount of lead
which actually enters the body from an ingested medium depends on the physical-chemical
properties of the lead and of the medium. For example, lead in soil may exist, at least in part, as
poorly water-soluble minerals, and may also exist inside particles of inert matrix such as rock or
slag of variable size, shape, and association. These chemical and physical properties may tend to
influence (usually decrease) the absorption (bioavailability) of lead when ingested. Thus, equal
ingested doses of different forms of lead in different media may not be of equal health concern.
Bioavailability of lead in a particular medium may be expressed either in absolute terms
(absolute bioavailability) or in relative terms (relative bioavailability).
Absolute Bioavailability (ABA) is the ratio of the amount of lead absorbed compared to
the amount ingested:
. . Absorbed Dose
ABA =
Ingested Dose
This ratio is also referred to as the oral absorption fraction (AF0).
Relative Bioavailability (RBA) is the ratio of the absolute bioavailability of lead present
in some test material compared to the absolute bioavailability of lead in some appropriate
reference material:
ARA
r>n A _ test material
ABA
reference material
Usually the form of lead used as reference material is a soluble compound such as lead
acetate that is expected to completely dissolve when ingested.
For example, if 100 micrograms (ug) of lead dissolved in drinking water were ingested
and a total of 50 ug entered the body, the ABA would be 50/100, or 0.50 (50%). Likewise, if
OSWER 9285.7-77
-------
100 ug of lead contained in soil were ingested and 30 ug entered the body, the ABA for soil
would be 30/100, or 0.30 (30%). If the lead dissolved in water were used as the frame of
reference for describing the relative amount of lead absorbed from soil, the RBA would be
0.30/0.50, or 0.60 (60%).
For additional discussion about the concept and application of bioavailability, see Gibaldi
and Perrier (1982), Goodman etal. (1990), Mushak (1991), and/or Klaassen etal. (1996).
1.2 Using Bioavailability Data to Improve Exposure Calculations for Lead
When reliable data are available on the bioavailability of lead in soil, dust, or other soil-
like waste material at a site, this information can be used to improve the accuracy of exposure
and risk calculations at that site. For example, the basic equation for estimating the site-specific
ABA of a test soil is as follows:
ABAsoa=ABAsolMe-RBAsal
where:
ABASOii = Absolute bioavailability of lead in soil ingested by a child
ABAsoiubie = Absolute bioavailability in children of some dissolved or fully soluble form
of lead
RBAson = Relative bioavailability of lead in soil
Based on available information in the literature on lead absorption in humans, the U.S.
EPA estimates that the absolute bioavailability of lead from water and the diet is usually about
50% in children (U.S. EPA, 1994). Thus, when a reliable site-specific RBA value for soil is
available, it may be used to estimate a site-specific absolute bioavailability in that soil, as
follows:
ABAsoa=50%-RBAsml
In the absence of site-specific data, the absolute absorption of lead from soil, dust, and
other similar media is estimated by U.S. EPA to be about 30% (U.S. EPA, 1994). Thus, the
default RBA used by U.S. EPA for lead in soil and dust compared to lead in water is 30%/50%,
or 60%. When the measured RBA in soil or dust at a site is found to be less than 60% compared
to some fully soluble form of lead, it may be concluded that exposures to and hazards from lead
in these media at that site are probably lower than typical default assumptions. If the measured
RBA is higher than 60%, absorption of and hazards from lead in these media may be higher than
usually assumed.
OSWER 9285.7-77
-------
1.3 Overview of U.S. EPA's Program to Study Lead Unavailability in Animals
Scientists in U.S. EPA Region 8 have been engaged in a multi-year investigation of lead
absorption from a variety of different environmental media, especially soils and solid wastes
associated with mining, milling, and smelting sites. All studies in this program employed
juvenile swine as the animal model. Juvenile swine were selected for use in these studies
because they are considered to be a good physiological model for gastrointestinal absorption in
children (see Appendix A).
Initial studies in the program (referred to as "Phase I") were performed by Dr. Robert
Poppenga and Dr. Brad Thacker at Michigan State University (Weis et a/., 1995). The Phase I
study designs and protocols were refined and standardized by Dr. Stan Casteel and his
colleagues at the University of Missouri, Columbia, and this group has performed a large number
of studies (collectively referred to as "Phase II") designed to further characterize the swine
model and to quantify lead absorption from a variety of different test materials. Section 2 of this
report summarizes the Phase II work performed at the University of Missouri.
1.4 Overview of Methods for Estimating Lead RBA In Vitro
Measurement of lead RBA in animals has a number of potential benefits, but is also
rather slow and costly and may not be a feasible option in all cases. It is mainly for these
reasons that a number of scientists have been working to develop in vitro procedures that may
provide faster and less costly alternatives for estimating the RBA of lead in soil or soil-like
samples (Miller and Schricker, 1982; Imber, 1993; Ruby etal, 1993, 1996; Medlin, 1997;
Rodriguez et a/., 1999). These methods are based on the concept that the rate and/or extent of
lead solubilization in the gastrointestinal fluid are likely to be important determinants of lead
bioavailability in vivo, and most in vitro tests are aimed at measuring the rate or extent of lead
solubilization from soil into an extraction solvent that resembles gastric fluid. To help avoid
confusion in nomenclature, the fraction of lead which solubilizes in an in vitro system is referred
to as bioaccessibility, while the fraction that is absorbed in vivo is referred to as bioavailability.
More recently, development and testing of a simplified in vitro method for estimating
lead bioaccessibility has been performed by Dr. John Drexler at the University of Colorado.
Section 3 of this report describes this in vitro method and presents the results.
OSWER 9285.7-77
-------
2.0 IN VIVO STUDIES
2.1 Basic Approach for Measuring RBA In Vivo
The basic approach for measuring lead absorption in vivo is to administer an oral dose of
lead to test animals and measure the increase in lead level in one or more body compartments
(blood, soft tissue, bone). In order to calculate the RBA value of a test material, the increase in
lead in a body compartment is measured both for that test material and a reference material (lead
acetate). Equal absorbed doses of lead (as Pb+2) are expected to produce approximately equal
increases in concentration in tissues regardless of the source or nature of the ingested lead, so the
RBA of a test material is calculated as the ratio of doses (test material and reference material)
that produce equal increases in lead concentration in the body compartment. Note that this
approach is general and yields reliable results for both non-linear and linear responses.
2.2 Animal Exposure and Sample Collection
All in vivo studies carried out during this program were performed as nearly as possible
within the spirit and guidelines of Good Laboratory Practices (GLP: 40 CFR 792). Standard
Operating Procedures (SOPs) for all of the methods are documented in a project notebook that is
available through the administrative record.
Experimental Animals
All animals used in this program were intact male swine approximately 5 to 6 weeks of
age. All animals were monitored to ensure they were in good health throughout the study.
Diet
In order to minimize lead exposure from the diet, animals were fed a special low-lead
diet purchased from Zeigler Brothers, Inc. (Gardners, PA). The amount of feed provided was
equal to 5% of the average body weight of animals on study. The feed was nutritionally
complete and met all requirements of the National Institutes of HealthNational Research
Council (NRC, 1988). The typical nutritional components and chemical analysis of the feed are
presented in Table 2-1. Periodic analysis of feed samples during this program indicated the
mean lead level was less than 50 ug/kg, corresponding to a daily intake of less than 2.5 ug/kg-
day.
Drinking water was provided ad libitum via self-activated watering nozzles within each
cage. Periodic analysis of samples from randomly selected drinking water nozzles indicated the
mean lead concentration was less than 2 ug/L, corresponding to a daily intake of less than 0.2
ug/kg-day.
OSWER 9285.7-77
-------
Exposure
Appendix B provides the details of animal exposure, including the design (number of
dose groups, number of animals, dosing material, and dose levels) for all of the Phase II studies.
A typical study design is summarized in Table 2-2. In general, groups of animals were exposed
to a series of doses of either lead acetate or test material. For convenience, in this report, lead
acetate is abbreviated as "PbAc." Exposure occurred twice a day for 15 days. Most groups were
exposed by oral administration, with one group usually exposed to lead acetate by intravenous
(IV) injection via an indwelling venous catheter.
2.3 Preparation of Biological Samples for Analysis
Samples of blood were collected from each animal at multiple times during the course of
a study (e.g., days 0, 1, 2, 3, 4, 6, 9, 12, and 15). On day 15, the animals were sacrificed and
samples of liver, kidney, and bone (femur) were collected.
Appendix C presents details of biological sample collection, preparation, and analysis. In
brief, samples of blood were diluted in "matrix modifier," a solution recommended by the
Centers for Disease Control and Prevention (CDC) for analysis of blood samples for lead (CDC,
2001). Samples of soft tissue (kidney, liver) were digested in hot acid, while samples of bone
were ashed and then dissolved in acid.
Prepared samples were analyzed for lead using a Perkin Elmer Model 5100 graphite
furnace atomic absorption spectrophotometer. All results from the analytical laboratory were
reported in units of ug Pb/L of prepared sample. The detection limit was defined as three-times
the standard deviation of a set of seven replicates of a low-lead sample (typically about 2 to 5
2.4 Data Reduction
The basic data reduction task required to calculate an RBA for a test material is to fit
mathematical equations to the dose-response data for both the test material and the reference
material, and then solve the equations to find the ratio of doses that would be expected to yield
equal responses. After testing a variety of different equations, it was found that nearly all blood
lead AUC data sets could be well-fit using an exponential equation, while most data sets for
liver, kidney, and bone lead could be well-fit using a linear equation:
Linear: Response = a + b Dose (1)
Exponential: Response = a + b-\\- exp(-c Dose)] (2)
where a, b, and c are the parameters of the models, and Dose is the total daily administered dose
of lead (ug/kg-day).
OSWER 9285.7-77
-------
Appendix D presents a detailed description of the curve-fitting methods and rationale,
along with the methods used to quantify uncertainty in the KB A estimates for each test material.
Detailed dose-response data and curve-fitting results are presented in Appendix E.
2.5 Results and Discussion
2.5.1 Effect of Dosing on Animal Health and Weight
Lead exposure levels employed in this program are substantially below those which
cause clinical symptoms in swine, and no evidence of treatment-related toxicity was observed in
any dose group. All animals exposed to lead by the oral route remained in good health
throughout each study, and the only clinical signs observed were characteristic of normal swine.
However, animals implanted with indwelling venous catheters (used for intravenous injections)
were subject to infection, and a few animals became quite ill. This was a problem mainly at the
start of the program and tended to diminish as experience was gained on the best surgical and
prophylactic techniques for catheter implantation. When an animal became ill, if good health
could not be restored by administration of antibiotics, the animal was promptly removed from
the study.
All animals were weighed every three days during the course of each study. The rate of
weight gain (kg/day) averaged across all Phase II studies is illustrated in Figure 2-1. As shown,
animals typically gained about 0.3 to 0.5 kg/day, and the rate of weight gain was generally
comparable in all groups.
2.5.2 Time Course of Blood Lead Response
The time course of the blood lead response to oral or intravenous exposure may be
thought of on two different time scales: the short-term "spike" that occurs immediately
following an exposure, and the longer-term trend toward "steady-state" blood lead following
repeated exposures.
Initial studies performed during Phase I of this program revealed that a single oral dose
of lead acetate causes blood lead levels rise to a peak about two hours post-ingestion, and then
decrease over the course of 12 to 24 hours to a near steady-state value (Weis et a/., 1993).
Although knowledge of these rapid kinetics is important in fully understanding the toxicokinetics
of lead, investigations in Phase II of this program focused mainly on quantifying the slower rise
in "steady-state" blood lead following repeated exposures. To achieve this goal, all blood lead
samples were collected 17 hours after lead exposure, at a time when the rate of change in blood
lead due to the preceding dose is minimal.
Figure 2-2 presents an example graph of the time course of "steady-state" blood lead
levels following repeated oral and intravenous exposure to lead acetate. As seen, blood lead
levels begin below the detection limit (usually about 1 ug/dL) and stay very low in control
animals throughout the course of the study. In animals exposed to lead acetate, blood lead
values begin to rise within 1 to 2 days and tend to flatten out to a near steady-state in about 7 to
10 days.
OSWER 9285.7-77
-------
2.5.3 Dose-Response Patterns
Figures 2-3 to 2-6 present the dose response patterns observed for blood, liver, kidney,
and bone (femur) following repeated oral or intravenous exposure to lead acetate. For blood, the
endpoint is the area under the blood lead vs time curve (AUC). For femur, kidney, and liver, the
endpoint is the concentration in the tissue at the time of sacrifice. The data for intravenous
exposure are based on a single study1, while the patterns for oral exposure are based on the
combined results across all studies performed during Phase II.
As seen, there is substantial variability in response between individuals (both within and
between studies), and this variability tends to increase as dose (and response) increases. This
pattern of increasing variance in response is referred to as heteroscedasticity, and is accounted
for in the model-fitting procedure through the use of weighted least squares regression (see
Appendix D). Despite the variability in response, the regression analyses indicate that the dose
response pattern is typically non-linear for blood lead AUC following both oral and intravenous
exposure, but is approximately linear in both cases for liver, kidney, and bone lead (see Table
Dl). This pattern of dose-response relationships suggests that, at least over the dose range tested
in this program, absorption of lead from the gastrointestinal tract of swine is linear, and that the
non-linearity observed in blood lead AUC response is due to some sort of saturable binding in
the blood.
2.5.4 Estimation of ABA for Lead Acetate
Inspection of Figures 2-3 to 2-6 reveals that each of the measured responses to ingested
lead acetate is smaller than the response for intravenously injected lead acetate. These data were
used to calculate the absolute bioavailability of ingested lead acetate using the data reduction
approach described in Section 2.4. The results are summarized below:
Measurement Endpoint
Blood AUC
Liver
Kidney
Femur
Estimated ABA of PbAc
0.10 ±0.02
0.16 ±0.05
0.19 ±0.05
0.14 ±0.03
Although the four different measurement endpoints do not agree precisely, it seems clear
that the absolute bioavailability of lead acetate in juvenile swine is about 15% ± 4%. Although
data are limited, results from balance studies in infants and young children (age 2 weeks to 8
years) suggest that lead absorption is probably about 42% to 53% (Alexander et a/., 1974;
Ziegler et a/., 1978). If so, lead absorption in juvenile swine is apparently lower than for young
1 Most studies in Phase II utilized only one intravenous dose level (100 ug/kg-day) and, hence, do not provide dose-
response data. Study 8 included three intravenous exposure levels (25, 50, and 100 ug/kg-day); the data from this
study are shown in Figures 2-3 to 2-6.
OSWER 9285.7-77
-------
humans. Although the reason for this apparent difference is not known, it is important to note
that even if swine do absorb less lead than children under similar dosing conditions, this does not
invalidate the swine as an animal model for estimating relative bioavailability of lead in different
test materials.
2.5.5 Estimation ofRBA for Lead in Test Materials
Characterization of Test Materials
Table 2-3 describes the Phase II test materials for which RBA was measured in this
program and provides the analytical results for lead. Data on other Target Analyte List (TAL)
metals, if available, are provided in Appendix F. As seen, 17 different samples from eight
different sites were investigated, along with one sample of paint flakes mixed with clean soil and
one sample of finely-ground native galena mixed with clean soil. Prior to analysis and dosing,
all samples were dried (<40°C) and sieved, and only materials which passed through a 60-mesh
screen (corresponding to particles smaller than about 250 um) were used. This range of particle
sizes was selected because the U.S. EPA considers particles less than about 250 um to be the
most likely to adhere to the hands and be ingested by hand-to-mouth contact, especially in young
children (U.S. EPA, 2000).
Each sample of test material that was evaluated in the swine bioassay program was
thoroughly characterized with regard to mineral phase, particle size distribution, and matrix
association using electron microprobe analysis. Detailed results for each test material are
presented in Appendix F, and the results are summarized in Tables 2-4 to 2-6.
Table 2-4 lists the different lead phases observed in the test materials, and gives the
relative lead mass (RLM) for each phase in each test material. The RLM is the estimated
percentage of the total lead in a sample that is present in a particular phase. Of the 22 different
phases detected in one or more samples, 9 are very minor, with RLM values no higher than 2%
in any sample. However, 13 of the phases occur at concentrations that could contribute
significantly to the overall bioavailability of the sample (RLM >10%). It should be noted that a
particle is classified as "slag" only if the particle is glassy or vitreous in nature. Inclusions or
other non-vitreous grains of lead-bearing material are classified according to their mineral
content and are not classified as slag particles (even if they are observed in bulk samples that are
referred to as "slag").
Table 2-5 summarizes information on the degree to which lead-bearing grains in each
sample are partially or entirely liberated (i.e., exposed to gastric fluids when ingested) or
included (i.e., fully enclosed or encased in mineral or vitreous matrices). Data are presented
both on a particle frequency basis and on the basis of relative lead mass. As seen, the majority
of lead-bearing particles in most samples are partially or entirely liberated, although the tailings
sample from Oregon Gulch is a clear exception.
Table 2-6 summarizes data on the distribution (frequency) of particle sizes (measured as
the longest dimension) in each sample. For convenience, the data presented are for liberated
particles only (Appendix F contains the data for all particles). As seen, most samples contain a
OSWER 9285.7-77
-------
range of particle sizes, often with the majority of the particles being less than 50 um.
(Remember that all samples were sieved to isolate particles less than 250 um before analysis.)
RBA Results for Test Materials
Detailed model fitting results and RBA calculations for each test material are presented
in Appendix E and are summarized in Table 2-7.
As shown in Table 2-7, there are four independent estimates of RBA (based on blood
AUC, liver, kidney, and bone) for each test material. Conceptually, each of these four values is
an independent estimate of the RBA for the test material, so the estimates from all four endpoints
need to be combined to yield a final point estimate for each test material. As discussed in
Appendix D (Section 4.7), an analysis of the relative statistical reliability of each endpoint (as
reflected in the average coefficient of variation in RBA values derived from each endpoint)
suggests that the four endpoint-specific RBA values are all approximately equally reliable.
Based on this, the point estimate for a test material is the simple average across the four
endpoint-specific RBA values. The resulting point estimate values are presented in the far right
portion of Table 2-7. Uncertainty bounds around the point estimates were derived as described
in Appendix D (Section 4.7).
Inspection of these point estimates for the different test materials reveals that there is a
wide range of values across different samples, both within and across sites. For example, at the
California Gulch site in Colorado, RBA estimates for different types of material range from
about 6% (Oregon Gulch tailings) to about 105% (Fe/Mn lead oxide sample). This wide
variability highlights the importance of obtaining and applying reliable RBA data to site-specific
samples in order help to improve risk assessments for lead exposure.
2.5.6 Effect of Food
Studies in humans indicate that lead absorption is reduced by the presence of food in the
stomach (Garber and Wei, 1974; U.S. EPA, 1996). The mechanism by which the presence of
food leads to decreased absorption is not certain, but may be related to competition between lead
and calcium for active and/or passive uptake sites in the gastrointestinal epithelium (Diamond,
2000). Because of the potential inhibitory effects of food, all of the studies performed during
this program were designed to estimate the RBA of lead associated with a fasting state, each
dose being administered to animals no less than six hours after the last feeding. In order to
investigate how the presence of food in the stomach might influence absorption, a study was
performed to measure the absorption of lead acetate given two hours before feeding and compare
that to the absorption of lead acetate given either at the time of feeding or two hours after
feeding. The results, expressed using the absorption two hours before feeding as the frame of
reference, are summarized below:
OSWER 9285.7-77
-------
Measurement
Endpoint
Blood Lead AUC
Liver Lead
Kidney Lead
Bone Lead
Point Estimate
Ratio of PbAc Absorption
Given With Food
Compared to PbAc Given
Without Food
0.39 ±0.05
0.86 ±0.24
0.72 ±0.26
0.35 ±0.05
0.58 ±0.28
Ratio of PbAc Absorption
Given 2 Hours After Feeding
Compared to PbAc Given
Without Food
0.40 ±0.06
0.58 ±0.16
0.73 ±0.27
0.33 ±0.05
0.51 ±0.22
These findings indicate that uptake of lead is reduced by close to half (RBA point
estimates are 51% and 58%) when the lead is administered to animals along with food compared
to when it is administered on an empty stomach. This effect appears to endure for at least two
hours after feeding, which is consistent with the results of a gastric emptying time study in
juvenile swine that indicated that food is held in the stomach for up to four hours after eating
(Casteeletal., 1998).
This study, which utilized lead acetate only, does not provide information about the
effect of food on the absorption of lead ingested in a solid form such as soil. However, it is
suspected that the magnitude of the decrease in absorption caused by food is likely to be at least
as large as that observed for lead acetate, and perhaps even larger. This is because food may
influence not only the absorption of soluble lead ions, but might also tend to decrease the rate
and extent of lead solubilization from soil by tending to increase the pH of gastric fluids.
2.5.7 Correlation of RBA with Mineral Phase
In principle, each unique combination of phase, size, and matrix association constitutes a
unique mineral ogical form of lead, and each unique form could be associated with a unique RBA
that is the inherent value for that "type" of lead. If so, then the concentrated-weighted average
RBA value for a sample containing a mixture of different types of lead is given by:
n s m
(3)
where:
t\tjA sa
mple
n
s
m
Observed RBA of lead in a sample
Fraction of total lead in phase /' of sizey and matrix association k
RBA of lead in phase / of sizey and matrix association k
Number of different lead phase categories
Number of different size categories
Number of different matrix association categories
OSWER 9285.7-77
10
-------
If the number of different lead phases which can exist in the environment is on the order
of 20, the number of size categories is on the order of five, and the number of matrix association
categories is two (included and liberated), then the total number of different "types" of lead is on
the order of 200. Because measured KB A data are available from this study for only 19 different
samples, it is clearly impossible (with the present data set) to estimate type-specific RBA values
for each combination of phase, size, and matrix association. Therefore, in order to simplify the
analysis process, it was assumed that the measured RBA value for a sample was dominated by
the liberated mineral phases present, and the effect of included materials or of particle size were
not considered. That is, the data were analyzed according to the following model:
n
ple ~ / > ^ i .liberated ' -^^^i .liberated VV
z=l
Because 22 different phases were identified and only 19 different samples were analyzed,
it was necessary to reduce the number of phases to a smaller number so that regression analysis
could be performed. Therefore, the different phases were grouped into ten categories as shown
in Table 2-8. These groups were based on professional judgment regarding the expected degree
of similarity between members of a group, along with information on the relative abundance of
each phase (see Table 2-4).
The total lead mass in each group was calculated by summing the relative lead mass for
each individual component in the group. As noted above, only the lead mass in partially or
entirely liberated particles was included in the sum.
Group-specific RBA values were estimated by fitting the grouped data to the model
(equation 4) using minimization of squared errors. Two different options were employed. In the
first option, each parameter (group-specific RBA) was fully constrained to be between zero and
one, inclusive. In the second option, each parameter was partially constrained to be greater than
or equal to zero. Because Group 10 contains only phases which are present in relatively low
levels, an arbitrary coefficient of 0.5 was assumed for this group and the coefficient was not
treated as a fitting parameter.
The resulting estimates of the group-specific RBA values are shown in Figure 2-7. As
seen, there is a wide range of group-specific RBA values, with equal results being obtained by
both methods of constraint. It is important to stress that these group-specific RBA estimates are
derived from a very limited data set (nine independent parameter estimates based on only
19 different measurements), so the group-specific RBA estimates are inherently uncertain. In
addition, both the measured sample RBA values and the relative lead mass in each phase are
subject to additional uncertainty. Therefore, the group-specific RBA estimates should not be
considered to be highly precise, and calculation of a quantitative sample-specific RBA value
from these estimates is not appropriate. Rather, it is more appropriate to consider the results of
this study as sufficient to support only semi-quantitative rank-order classification of phase-
specific RBA values, as follows:
OSWER 9285.7-77 11
-------
Low Bioavailability
(RBA <0.25)
Fe(M) Sulfate
Anglesite
Galena
Fe(M) Oxide
Pb(M) Oxide
Medium Bioavailability
(RBA = 0.25-0.75)
Lead Oxide
Lead Phosphate
High Bioavailability
(RBA >0.75)
Cerussite
Mn(M) Oxide
(M) = Metal
As noted above, the estimates apply only to particles that are liberated, not those that are
included.
2.5.8 Quality Assurance
A number of steps were taken throughout each of the studies in this program to assess
and document the quality of the data that were collected. These steps are summarized below.
Duplicates
A randomly selected set of about 5% of all blood and tissue samples generated during
each study were submitted to the laboratory in a blind fashion for duplicate analysis. Figure 2-8
plots the results for blood (Panel A) and for liver, kidney, and bone (Panel B). As seen, there
was good intra-laboratory reproducibility between duplicate samples for both blood and tissues,
with both linear regression lines having a slope near 1.0, an intercept near zero, and an R2 value
near 1.00.
Standards
The CDC provides blood lead "check samples" that may be used for use in quality
assurance programs for blood lead studies. Three types of check samples (nominal
concentrations of 1.7 ug/dL, 4.8 ug/dL, and 14.9 ug/dL) were used in these studies. Each day
that blood samples were collected from experimental animals, several check samples of different
concentrations were also prepared and submitted for analysis in random order and in a blind
fashion. The results (averaged across all studies) are plotted in Figure 2-9. As seen, the
analytical results obtained for the check samples were generally in good agreement with the
expected value at all three concentrations, with an overall mean of 1.4 ug/L for the low standards
(nominal concentration of 1.7 ug/L), 4.3 ug/L for the middle standard (nominal concentration of
4.8 ug/L), and 14.5 ug/L for the high standards (nominal concentration of 14.9 ug/L).
Interlaboratory Comparison
In each study, an interlaboratory comparison of blood lead analytical results was
performed by sending a set of about 15 to 20 randomly selected whole blood samples to CDC for
blind independent preparation and analysis. The results are plotted in Figure 2-10. As seen, the
results of analyses by U.S. EPA's laboratory are generally similar to those of CDC, with a mean
inter-sample difference (U.S. EPA minus CDC) of 0.07 ug/dL. The slope of the best-fit straight
line through the data is 0.84, indicating that the concentration values estimated by the U.S. EPA
OSWER 9285.7-77 12
-------
laboratories tended to be about 15% lower than those estimated by CDC. The reason for this
apparent discrepancy between the U.S. EPA laboratory and the CDC laboratory is not clear, but
might be related to differences in sample preparation techniques. Regardless of the reason, the
differences are sufficiently small that they are likely to have no significant effect on calculated
KB A values. In particular, it is important to realize that if both the lead acetate and test material
dose-response curves are biased by the same factor, then the biases cancel in the calculation of
the ratio.
Reproducibility of RB A Estimates
As with any study involving animals, there may be substantial variability between
animals within each dose group, and there may also be variability in observed responses to
exposure across different studies. Because each study involved administration of a standard
series of doses of lead acetate, the data for lead acetate can be used to assess the stability and
reproducibility of the swine model. Table 2-9 lists the best-fit parameters for the best-fit curves
for oral lead acetate dose responses for blood AUC, liver, kidney, and bone in each study, and
for all studies combined. As seen, the variability (expressed as the between-study coefficient of
variation) is generally on the order of 25 to 50% for the b and c parameters, with somewhat
higher variability in the intercept parameter (a). This degree of between-study variability is not
unexpected for a study in animals and emphasizes the need for generating the dose-response
curve for the reference material within each study. The source of the between-study variation is
likely to be mainly a consequence of variation in animals between different groups (different
dams, different ages, different weights), although a possible contribution from other variables
(time of year, laboratory personnel, etc.) cannot be excluded.
Because RBA calculations are based on the within-study ratio of responses between a test
material and reference material, the variability in response between studies may be at least partly
cancelled in the calculation of the RBA. The most direct way to test this hypothesis is to
compare RBA estimates for the same material that has been tested in two different studies. To
date, only two test materials have been tested more than once. The results are shown in
Table 2-10 and are summarized below.
For the Palmerton Location 2 sample (tested twice in Phase II), agreement is moderately
good between the two studies for the blood AUC and kidney endpoints and for the point
estimate, although there is relatively low agreement for the liver and bone endpoints. For the
Residential Soil Composite from the California Gulch Superfund site (tested once by the
University of Michigan during Phase I and again by the University of Missouri during Phase II),
agreement is good for all four endpoints, with between-study differences of less than 20%.
These differences are generally similar to the within-study confidence bounds, which are
typically in the 10% to 20% range. Taken together, these studies support the view that the in
vivo RBA assay has acceptable inter-study and inter-laboratory reproducibility.
OSWER 9285.7-77 13
-------
3.0 IN VITRO STUDIES
3.1 Introduction
Measurement of lead KB A in animals using the approach described above has a number
of potential benefits, but is also rather slow and costly and may not be feasible in all cases. It is
mainly for this reason that a number of scientists have been working to develop alternative
in vitro procedures that may provide a faster and less costly alternative for estimating the RBA
of lead in soil or soil-like samples. These methods are based on the concept that the rate and/or
extent of lead solubilization in gastrointestinal fluid is likely to be an important determinant of
lead bioavailability in vivo, and most in vitro tests are aimed at measurement of the rate or extent
of lead solubilization in an extraction solvent that resembles gastric fluid. The fraction of lead
which solubilizes in an in vitro system is referred to as in vitro bioaccessibility (IVBA), which
may then be used as an indicator of in vivo RBA.
Background on the development and validation of in vitro methods for estimating lead
bioaccessibility can be found in Imber (1993), Ruby etal. (1993, 1996), and Medlin (1997).
3.2 In Vitro Method
The method described in this report represents a simplification from most preceding
approaches. The method was designed to be fast, easy, and reproducible, and some test
conditions were adjusted to yield results that best correlated with in vivo measurements of lead
bioavailability. The detailed standard operating procedure (SOP) is presented below; additional
information on this procedure may be obtained from http://www.colorado.edu/geolsci/legs.
3.2.1 Sample Preparation
All test materials tested in the bioaccessibility protocol were identical to the test materials
administered to swine in the in vivo studies described above. As noted previously, soils were
prepared by drying (<40°C) and sieving to <250 um. The <250-um size fraction was used
because this particle size is representative of that which adheres to children's hands. Samples
were thoroughly mixed prior to use to ensure homogenization. All samples were archived after
the study completion and retained for further analysis for a period of six months.
3.2.2 Apparatus
The main piece of equipment used for this procedure is the extraction device shown in
Figure 3-1. An electric motor (the same motor as is used in the Toxicity Characteristic Leaching
Procedure, or TCLP) drives a flywheel, which in turn drives a Plexiglass block situated inside a
temperature-controlled water bath. The Plexiglass block contains ten 5-centimeter holes with
stainless steel screw clamps, each of which is designed to hold a 125-mL wide-mouth high
density polyethylene (HDPE) bottle. The water bath was filled such that the extraction bottles
were completely immersed. Temperature in the water bath was maintained at 37±2 °C using an
immersion circulator heater. The 125-mL HDPE bottles had air-tight screw-cap seals, and care
OSWER 9285.7-77 14
-------
was taken to ensure that the bottles did not leak during the extraction procedure. All equipment
was properly cleaned, acid washed, and rinsed with deionized water prior to use.
3.2.3 Selection of WE A Test Conditions
The dissolution of lead from a test material into the extraction fluid depends on a number
of variables including extraction fluid composition, temperature, time, agitation, solid/fluid ratio,
and pH. These parameters were evaluated to determine the optimum values for maximizing
sensitivity, stability, and the correlation between in vitro and in vivo values.
All reagents were free of lead and the final fluid was tested to confirm that lead
concentrations were less than one-fourth the project required detection limit (PRDL) of 10 ug/L
(i.e., less than 2 ug/L lead in the final fluid). Cleanliness of all materials used to prepare and/or
store the extraction fluid and buffer is essential; all glassware and equipment used to prepare
standards and reagents were properly cleaned, acid washed, and triple-rinsed with deionized
water prior to use.
Extraction Fluid: The extraction fluid selected for this procedure was 0.4 M glycine (free
base, reagent grade glycine in deionized water), adjusted to a pH of 1.50±0.05 at 37°C using
trace metal grade concentrated hydrochloric acid (HC1). Most previous in vitro test systems
have employed a more complex fluid intended to simulate gastric fluid. For example, Medlin
(1997) used a fluid that contained pepsin and a mixture of citric, malic, lactic, acetic, and
hydrochloric acids. When the bioaccessibility of a series of test substances were compared using
0.4 M glycine buffer (pH 1.5) with and without the inclusion of these enzymes and metabolic
acids, no significant difference was observed (p=0.196). This indicates that the simplified buffer
employed in the procedure is appropriate, even though it lacks some constituents known to be
present in gastric fluid.
Temperature: In order to evaluate the effect of the extraction temperature, seventeen
substrates were analyzed (generally in triplicate) at both 37°C and 20°C. The results are shown
in Figure 3-2 (Panel A). In some cases, temperature had little effect, but in three cases the
amount of lead solubilized was more than 20% greater at 37°C than at 20°C, and in two cases it
was more than 20% less. Because the results appeared to depend on temperature in at least some
cases, a temperature of 37°C was selected because this is approximately the temperature of
gastric fluid in vivo.
Extraction Time: The time that ingested material is present in the stomach (i.e., stomach-
emptying time) is about one hour for a child, particularly when a fasted state is assumed (see
Appendix A). To investigate the effect of extraction time on lead solubilization, 11 substrates
were extracted for periods of 1, 2, or 4 hours. The results are shown in Figure 3-2 (Panel B). As
seen, in most cases, the amount of lead solubilized was approximately constant over time, with
only one substrate (test material 6) showing a variation that exceeded the method precision.
Therefore, an extraction time of one hour was selected for the final method. In a subsequent test
(data not shown), it was found that allowing the bottles to stand at room temperature for up to
4 hours after rotation at 37°C caused no significant variation (<10%) in lead concentration.
OSWER 9285.7-77 15
-------
pH: Human gastric pH values tend to range from about 1 to 4 during fasting (see
Appendix A). Previous studies have used stomach phase pH values between 1.0 and 2.5 for their
in vitro experiments (Ruby etal, 1993; CBR, 1993; Gasser etal, 1996; Buckley, 1997; Medlin,
1997; Rodriguez et a/., 1999; Mercier et a/., 2000). To evaluate the effect of pH on lead
bioaccessibility, 24 substrates were analyzed at pH values of 1.5, 2.5, or 3.5. As shown in
Figure 3-2 (Panel C), the amount of lead solubilized is strongly pH-dependent, with the highest
extraction at pH 1.5. For the subset of test materials for which in vivo RBA had been estimated
at that time (N = 13), the empiric correlation between IVBA and in vivo RBA was slightly better
at pH 1.5 (rho = 0.919) than at pH 2.5 (rho = 0.881). Thus, a pH of 1.5 was selected for use in
the final protocol.
Agitation: If the test material is allowed to accumulate at the bottom of the extraction
apparatus, the effective surface area of contact between the extraction fluid and the test material
may be reduced, and this may influence the extent of lead solubilization. Depending on which
theory of dissolution is relevant (Nernst and Brunner, 1904, or Dankwerts, 1951), agitation will
greatly affect either the diffusion layer thickness or the rate of production of fresh surface.
Previous workers have noted problems associated with both stirring and argon bubbling methods
(Medlin and Drexler, 1995; Drexler, 1997). Although no systematic comparison of agitation
methods was performed, an end-over-end method of agitation was chosen to best simulate the
complex peristaltic motion of the gastrointestinal system.
Solid/Fluid Ratio and Mass of Test Material: A solid to fluid ratio of 1/100 (mass per
unit volume) was chosen to reduce the effects of metal dissolution that were noted by Sorenson
et al. (1971) when lower ratios (1/5 and 1/25) were used. Tests using Standard Reference
Materials showed no significant variation (within ± 1% of control means) in the fraction of lead
extracted with soil masses as low as 0.2 gram (g) per 100 mL. However, use of low masses of
test material could introduce variability due to small scale heterogeneity in the sample and/or to
weighing errors. Therefore, the final method employs 1.0 g of test material in 100 mL of
extraction fluid.
In special cases, the mass of test material may need to be less than 1.0 g to avoid the
potential for saturation of the extraction solution. Tests performed using lead acetate, lead oxide,
and lead carbonate indicate that if the bulk concentration of a test material containing these
relatively soluble forms of lead exceeds approximately 50,000 ppm, the extraction fluid becomes
saturated at 37°C and, upon cooling to room temperature and below, lead chloride crystals will
precipitate. To prevent this from occurring, the concentration of lead in the test material should
not exceed 50,000 ppm, or the mass of the test material should be reduced to 0.50±0.01g.
3.2.4 Summary of Final Leaching Protocol
The extraction procedure began by placing 1.00±0.05 g of sieved test material and
100±0.5 mL of the buffered extraction fluid (0.4 M glycine, pH 1.5) into a 125-mL wide-mouth
HDPE bottle. Care was taken to ensure that static electricity did not cause soil particles to
adhere to the lip or outside threads of the bottle; if necessary, an antistatic brush was used to
eliminate static electricity prior to adding the test substrate. The bottle was tightly sealed and
then shaken or inverted to ensure that there was no leakage and that no soil was caked on the
bottom of the bottle.
OSWER 9285.7-77 16
-------
Each bottle was placed into the modified TCLP extractor (water temperature 37±2°C).
Samples were extracted by rotating the samples end-over-end at 30 ± 2 rpm for 1 hour. After 1
hour, the bottles were removed, dried, and placed upright on the bench top to allow the soil to
settle to the bottom. A 15-mL sample of supernatant fluid is removed directly from the
extraction bottle into a disposable 20-cc syringe. After withdrawal of the sample into the
syringe, a Luer-Lok attachment fitted with a 0.45-um cellulose acetate disk filter (25 mm
diameter) is attached, and the 15 mL aliquot of fluid is filtered through the attachment to remove
any particulate matter. This filtered sample of extraction fluid is then analyzed for lead, as
described below. If the total time elapsed for the extraction process exceeds 90 minutes, the test
must be repeated.
As noted above, in some cases (mainly slags), the test material can increase the pH of the
extraction buffer, and this could influence the results of the bioaccessibility measurement. To
guard against this, the pH of the fluid was measured at the end of the extraction step (just after a
sample was withdrawn for filtration and analysis). If the pH was not within 0.5 pH units of the
starting pH (1.5), the sample was re-analyzed. If the second test also resulted in an increase in
pH of greater than 0.5 units, it was apparent that the test material was buffering the solution. In
these cases, the test was repeated using manual pH adjustment during the extraction process,
stopping the extraction at 5, 10, 15, and 30 minutes and manually adjusting the pH down to pH
1.5 at each interval by drop-wise addition of HC1.
3.2.5 Analysis of Extraction Fluid for Lead
The filtered samples of extraction fluid were stored in a refrigerator at 4°C until they
were analyzed (within 1 week of extraction). Once received by the laboratory, all media were
maintained under standard chain-of-custody. The samples were analyzed for lead by ICP-AES
or ICP-MS (U.S. EPA Method 6010 or 6020, U.S. EPA 1986). The method detection limit
(MDL) in extraction fluid was calculated to be around 19 ug/L for Method 6010 and typically
0.1-0.3 ug/L for Method 6020.
3.2.6 Quality Control/Quality Assurance
Quality assurance for the extraction procedure consisted of the following quality control
samples:
Reagent Blank extraction fluid analyzed once per batch.
Bottle Blank extraction fluid only (no test soil) run through the complete
procedure at a frequency of 1 in 20 samples (minimum of 1 per batch).
Blank Spike extraction fluid spiked at 10 mg/L lead, and run through the complete
procedure at a frequency of 1 in 20 samples (minimum of 1 per batch).
Matrix Spikes a subsample of each material used for duplicate analyses was used
as a matrix spike. The spike was prepared at 10 mg/L lead and run through the
extraction procedure at a frequency of 1 in 10 samples (minimum of 1 per batch).
OSWER 9285.7-77 17
-------
Duplicate Sample duplicate sample extractions were performed on 1 in 10 samples
(minimum of 1 per batch).
Control Soil National Institute of Standards and Testing (NIST) Standard
Reference Material (SRM) 2711 (Montana Soil) was used as a control soil. The SRM
was analyzed at a frequency of 1 in 20 samples (minimum 1 per batch).
Control limits for these quality control samples were as follows:
Analysis
Reagent blank
Bottle blank
Blank spike (10 mg/L)
Matrix spike (10 mg/L)
Duplicate sample
Control soil (NIST 27 11)
Frequency
once per batch
5%*
5%*
10%*
10%*
5%*
Control Limits
<25 ug/L lead
<50 ug/L lead
85-1 15% recovery
75-125% recovery
±20% RPD
±10%RPD
RPD = Relative percent difference
* Minimum of once per batch
To evaluate the precision of the in vitro bioaccessibility extraction protocol,
approximately 67 replicate analyses of both NIST SRM 2710 and 2711 were conducted over a
period of several months. Results are shown in Figure 3-3. As seen, both standards yield highly
reproducible results, with a mean coefficient of variation of about 6%.
3.3 Results and Discussion
3.3.1 IVBA Values
Table 3-1 summarizes the in vitro bioaccessibility results for the set of 19 different test
materials evaluated under the Phase II program. Each value is the mean and standard deviation
of three independent measurements performed at the University of Colorado at Boulder.
Figure 3-4 shows the results of an inter-laboratory comparison of results for these test
materials. The participating laboratories included ACZ Laboratories Inc.; University of
Colorado at Boulder; U.S. Bureau of Reclamation Environmental Research Chemistry
Laboratory; and National Exposure Research Laboratory. As seen in the figure, within-
laboratory variability (as shown by the error bars) is quite small (average <2%) and there is very
good agreement between laboratories (average difference of 2 to 3%, range of difference from 1
to 9%).
OSWER 9285.7-77
18
-------
3.3.2 Comparison with In Vivo Results
In order for an in vitro bioaccessibility test system to be useful in predicting the in vivo
RBA of a test material, it is necessary to establish empirically that a strong correlation exists
between the in vivo and the in vitro results across many different samples. A scatter plot of the
in vivo RBA and in vitro bioaccessibility data from this program is shown in Figure 3-5. The
Spearman rank order correlation coefficient between the paired RBA and IVBA point estimates
is 0.896 (p <0.001) and the Pearson product moment correlation coefficient is 0.917 (p <0.001),
indicating that there is a statistically significant positive correlation between IVBA and RBA.
Several different mathematical models were tested to describe the relation between RBA
and IVBA, including linear, power, and exponential. All fitting was done using weighted least
square regression, as detailed in Appendix D. The results are summarized below:
Model
Linear: RBA=a + b-IVBA
Power: RBA = a + b-IVBAc
2-Parameter Exponential: RBA = a + b-exp(IVBA)
3-Parameter Exponential: RBA = a + b-exp(c-IVBA)
R2
0.924
0.931
0.936
0.936
AIC
-30.46
-29.92
-33.02
-31.11
As seen, all of the models fit the data reasonably well, with the two exponential models
fitting slightly better than the linear model. However, as discussed in Appendix D, the
difference in quality of fit between linear and exponential models is not judged to be meaningful,
and the linear model is selected as the preferred model at present. As more data become
available in the future, the relationship between IVBA and RBA will be reassessed and the best-
fit model form will be reconsidered and revised if needed.
Because there is measurement error not only in RBA but also in IVBA, linear fitting was
also performed taking the error in both RBA and IVBA into account. There was nearly no
difference in fit, so the results of the weighted linear regression were selected for simplicity.
This decision may be revisited as more data become available. Based on this decision, the
currently preferred model is:
RBA = 0.878-IVBA - 0.028
It is important to recognize that use of this equation to calculate RBA from a given IVBA
measurement will yield the "typical" RBA value expected for a test material with that IVBA, and
the true RBA may be somewhat different (either higher or lower). The best fit line and the 95%
prediction interval for this data set are shown in Figure 3-6.
Applicability of the IVBA-RBA Model
At present, it appears that the equation relating IVBA to RBA should be widely
applicable, having been found to hold true for a wide range of different soil types and lead
OSWER 9285.7-77
19
-------
phases from a variety of different sites. However, most of the samples tested have been
collected from mining and milling sites, and it is plausible that some forms of lead that do not
occur at this type of site might not follow the observed correlation. Thus, whenever a sample
containing an unusual and/or untested lead phase is evaluated by the IVBA protocol, this should
be identified as a potential source of uncertainty. In the future, as additional samples with a
variety of new and different lead forms are tested by both in vivo and in vitro methods, the
applicability of the method will be more clearly defined.
OSWER 9285.7-77 20
-------
4.0 REFERENCES
Alexander, F.W., B.E. Clayton, and HT. Delves. 1974. Mineral and trace-metal balances in
children receiving normal and synthetic diets. QJ Med. 43:89-111.
Buckley, B.T. 1997. Estimates of bioavailability of metals in soil with synthetic biofluids: Is
this a replacement for animal studies? In: IBC Conference on Bioavailability, 1998, Scottsdale,
AZ.
Casteel, S.W., L.D. Brown, J. Lattimer, and M.E. Dunsmore. 1998. Fasting and feeding effects
on gastric emptying time in juvenile swine. Contemporary Topics in Laboratory Animal Science
37:106-108.
CBR. 1993. Report: Development of a physiologically relevant extraction procedure. CB
Research International, Sidney, BC, Canada.
CDC. 2001. Laboratory Procedure Manual. Analyte: Cadmium and Lead. Matrix: Blood.
Method: Atomic Absorption Spectroscopy. Method No..: 1090A/02-OD. Revised August 22,
2001. Atlanta, GA:Nutritional Biochemistry Branch, Division of Laboratory Sciences, National
Center for Environmental Health, Centers for Disease Control and Prevention. Available online
at: http://www.cdc.gov/NCHS/data/nhanes/frequency/lab06_met_lead_and_cadmium.pdf.
Dankwerts, P.V. 1951. Significance of liquid-film coefficients in gas absorption. Ind. Eng.
Chem. 43:1460.
Diamond, G.L. 2000. Transport of metals in the gastrointestinal system and kidneys. In:
Molecular Biology and Toxicology of Metals. R.K. Zalups and J. Koropatnick (eds). Taylor &
Francis, London, pp 300-344.
Drexler, J.W. 1997. Validation of an in vitro method: A tandem approach to estimating the
bioavailability of lead and arsenic to humans. IBC Conference on Bioavailability, Scottsdale,
AZ.
Garber, B.T. and E. Wei. 1974. Influence of dietary factors on the gastrointestinal absorption of
lead. Toxicol. App. Pharmacol. 27:685-691.
Gasser, U.G., WJ. Walker, R.S. Borch, and R.G. Burau. 1996. Lead release from smelter and
mine waste impacted materials under simulated gastric conditions and relation to speciation.
Env. Sci. Tech. 30:761-769.
Gibaldi, M., and D. Perrier. 1982. Pharmacokinetics (2nd edition). Marcel Dekker, Inc, NY,
NY, pp 294-297.
Goodman, A.G., T.W. Rail, A.S. Nies, and P. Taylor. 1990. The Pharmacological Basis of
Therapeutics (8th ed.). Pergamon Press, Inc. Elmsford, NY, pp 5-21.
OSWER 9285.7-77 21
-------
Imber, B.D. 1993. Development of a physiologically relevant extraction procedure. Prepared
for BC Ministry of Environment, Lands and Parks, Environmental Protection Division, Victoria,
BC. CB Research International Corporation, Sidney, BC.
Klaassen, C.D., M.O. Amdur, and J. Doull (eds). 1996. Cassarett and Doull's Toxicology: The
Basic Science of Poisons. McGraw-Hill, Inc. NY, NY, pp. 190.
Medlin, E.A. 1997. An in vitro method for estimating the relative bioavailability of lead in
humans. Masters thesis. Department of Geological Sciences, University of Colorado, Boulder.
Medlin, E., and J.W. Drexler. 1995. Development of an in vitro technique for the determination
of bioavailability from metal-bearing solids. International Conference on the Biogeochemistry
of Trace Elements, Paris, France.
Mercier, G., J. Duchesne, and A. Carles-Gibergues. 2000. A new in vitro test to simulate gastric
absorption of copper, lead, tin and zinc from polluted soils. Env. Tech. 23:121-133.
Miller, D.D., and B.R. Schricker. 1982. In vitro estimation of food iron bioavailability. In:
Nutritional Bioavailability of Iron. ACS Symp. Ser. 203:10-25, 1982.
Mushak, P. 1991. Gastro-intestinal absorption of lead in children and adults: Overview of
biological and biophysico-chemical aspects. In: The Proceedings of the International
Symposium on the Bioavailability and Dietary Uptake of Lead. Science and Technology Letters
3:87-104.
Nernst, W., and E. Brunner. 1904. Theorie der reaktionsgeschwindigkeit in heterogenen
systemen. Z. Phys. Chem. 47:52.
NRC. 1988. Nutrient requirements of swine. A report of the Committee on Animal Nutrition.
National Research Council. National Academy Press.
Rodriguez, R.R., N.T. Basta, S.W. Casteel, and L.W. Pace. 1999. An in vitro gastrointestinal
method to estimate bioavailable arsenic in contaminated soils and solid media. Environ. Sci.
Technol. 33:642-649.
Ruby, M.V., A. Davis, T.E. Link, R. Schoof, R.L. Chaney, G.B. Freeman, and P. Bergstrom.
1993. Development of an in vitro screening test to evaluate the in vivo bioaccessibility of
ingested mine-waste lead. Environ. Sci. Technol. 27(13):2870-2877.
Ruby, M.V., A. Davis, R. Schoof, S. Eberle, and C. M. Sellstone. 1996. Estimation of lead and
arsenic bioavailability using a physiologically based extraction test. Environ. Sci. Technol.
30(2):422-430.
Sorenson, R.C., D.D. Oelsligle, and D. Knodsen. 1971. Extraction of zinc, iron, and manganese
from soils with 0.1N hydrochloric acid as affected by soil properties, solution:soil ratio, and
length of extraction period. Soil Sci. 111:352-359.
OSWER 9285.7-77 22
-------
U.S. EPA. 1986. Test methods for evaluating solid waste. Volume 1A and IB; Laboratory
manual physical/chemical methods. SW-846. Third edition. United States Environmental
Protection Agency, Office of Solid Waste and Emergency Response.
U.S. EPA. 1994. Guidance Manual for the Integrated Exposure Uptake Biokinetic Model for
Lead in Children. United States Environmental Protection Agency, Office of Emergency and
Remedial Response. Publication Number 9285.7-15-1. EPA/540/R-93/081.
U.S. EPA. 1996. Recommendations of the Technical Review Workgroup for Lead for an
Interim Approach to Assessing Risks Associated with Adult Exposures to Lead in Soil. United
States Environmental Protection Agency, Technical Review Workgroup for Lead. December,
1996.
U.S. EPA. 2000. Short Sheet: TRW Recommendations for Sampling and Analysis of Soil at
Lead (Pb) Sites. OSWER 9285.7-38. United States Environmental Protection Agency, Office of
Solid Waste and Emergency Response.
Weis, C.P., G.M. Henningsen, R.H. Poppenga, and BJ. Thacker. 1993. Pharmacokinetics of
lead in blood of immature swine following acute oral and intravenous exposure. The
Toxicologist 13(1): 175.
Weis, C.P., R.H. Poppenga, BJ. Thacker, G.M. Henningsen, and A. Curtis. 1995. Design of
pharmacokinetic and bioavailability studies of lead in an immature swine model. In: Lead in
Paint, Soil, and Dust: Health Risks, Exposure Studies, Control Measures, Measurement
Methods, and Quality Assurance. ASTM STP 1226, M. E. Beard and S. D. A. Iske (eds).
American Society for Testing and Materials, Philadelphia, 1995.
Ziegler, E.E., B.B. Edwards, R.L. Jensen, K.R. Mahaffey, and SJ. Fomon. 1978. Absorption
and retention of lead by infants. Pediatr. Res. 12:29-34.
OSWER 9285.7-77 23
-------
This page intentionally left blank to facilitate double-sided printing.
-------
TABLES
OSWER 9285.7-77
-------
This page intentionally left blank to facilitate double-sided printing.
-------
TABLE 2-1. TYPICAL FEED COMPOSITION
Nutrient Name
Protein
Arginine
Lysine
Methionine
Met+Cys
Tryptophan
Histidine
Leucine
Isoleucine
Phenylalanine
Phe+Tyr
Threonine
Valine
Fat
Saturated Fat
Unsaturated Fat
Linoleic 18:2:6
Linoleic 18:3:3
Crude Fiber
Ash
Calcium
Phos Total
Available Phosphorous
Sodium
Potassium
Amount
20.1021%
1.2070%
1 .4690%
0.8370%
0.5876%
0.2770%
0.5580%
1.8160%
1.1310%
1.1050%
2.0500%
0.8200%
1.1910%
4.4440%
0.5590%
3.7410%
1 .9350%
0.0430%
3.8035%
4.3347%
0.8675%
0.7736%
0.7005%
0.2448%
0.3733%
Nutrient Name
Chlorine
Magnesium
Sulfur
Manganese
Zinc
Iron
Copper
Cobalt
Iodine
Selenium
Nitrogen Free Extract
Vitamin A
Vitamin D3
Vitamin E
Vitamin K
Thiamine
Riboflavin
Niacin
Pantothenic Acid
Choline
Pyridoxine
Folacin
Biotin
Vitamin B12
Amount
0.1911%
0.0533%
0.0339%
20.4719 ppm
118.0608 ppm
135.3710 ppm
8.1062 ppm
0.01 10 ppm
0.2075 ppm
0.3196 ppm
60.2340%
5.1892klU/kg
0.6486 klU/kg
87.2080 lU/kg
0.9089 ppm
9.1681 ppm
10.2290 ppm
30.1147 ppm
19.1250 ppm
1019.8600 ppm
8.2302 ppm
2.0476 ppm
0.2038 ppm
23.4416 ppm
Feed obtained from and nutritional values provided by Zeigler Bros., Inc
Tables.xls (2-1_Feed)
-------
TABLE 2-2. TYPICAL IN VIVO STUDY DESIGN
Dose
Group
1
2
3
4
5
6
7
8
9
10
11
Dose
Material
None
Lead Acetate
Test Material 1
Test Material 2
Lead Acetate
Exposure
Route
Oral
Oral
Oral
Oral
Intravenous
Target Dose
ug Pb/kg-day
--
25
75
225
75
225
625
75
225
625
100
Number of
Animals
2-5
5
5
5
5
5
5
5
5
5
5-8
Tables.xls (2-2_Design)
-------
TABLE 2-3. DESCRIPTION OF PHASE II TEST MATERIALS
Experiment
2
3
4
5
6
7
8
9
11
12
Sample Designation
Bingham Creek Residential
Bingham Creek Channel Soil
Jasper County High Lead Smelter
Jasper County Low Lead Yard
Murray Smelter Slag
Jasper County High Lead Mill
Aspen Berm
Aspen Residential
Midvale Slag
Butte Soil
California Gulch Phase I Residential
Soil
California Gulch Fe/Mn PbO
California Gulch AV Slag
Palmerton Location 2
Palmerton Location 4
Murray Smelter Soil
N 1ST Paint
Galena-enriched Soil
California Gulch Oregon Gulch Tailings
Site
Kennecott NPL Site, Salt Lake City,
Utah
Kennecott NPL Site, Salt Lake City,
Utah
Jasper County, Missouri Superfund
Site
Jasper County, Missouri Superfund
Site
Murray Smelter Superfund Site,
Murray City, Utah
Jasper County, Missouri Superfund
Site
Smuggler Mountain NPL Site, Aspen,
Colorado
Smuggler Mountain NPL Site, Aspen,
Colorado
Midvale Slag NPL Site, Midvale, Utah
Silver Bow Creek/Butte Area NPL
Site, Butte, Montana
California Gulch NPL Site, Leadville,
Colorado
California Gulch NPL Site, Leadville,
Colorado
California Gulch NPL Site, Leadville,
Colorado
New Jersey Zinc NPL Site,
Palmerton, Pennsylvania
New Jersey Zinc NPL Site,
Palmerton, Pennsylvania
Murray Smelter Superfund Site,
Murray City, Utah
--
California Gulch NPL Site, Leadville,
Colorado
Sample Description
Soil composite of samples containing less than 2500 ppm lead;
collected from a residential area (Jordan View Estates) located
along Bingham Creek in the community of West Jordan, Utah.
Soil composite of samples containing 3000 ppm or greater of
lead; collected from a residential area (Jordan View Estates)
located along Bingham Creek in the community of West Jordan,
Utah.
Soil composite collected from an on-site location.
Soil composite collected from an on-site location.
Composite of samples collected from areas where exposed slag
existed on site.
Soil composite collected from an on-site location.
Composite of samples collected from the Racquet Club property
(including a parking lot and a vacant lot).
Composite of samples collected from residential properties within
the study area.
Composite of samples collected from a water-quenched slag pile
in Midvale Slag Operable Unit 2.
Soil composite collected from waste rock dumps in Butte Priority
Soils Operable Unit (BPSOU).
Soil composite collected from residential properties within
Leadville.
Soil composite collected from near the Lake Fork Trailer Park
located southwest of Leadville near the Arkansas River.
Sample collected from a water-quenched slag pile on the property
of the former Arkansas Valley (AV) Smelter, located just west of
Leadville.
Soil composite collected from on-site.
Soil composite collected from on-site.
Soil composite collected from on-site.
A mixture of approximately 5.8% NIST Standard Reference
Material (SRM) 2589 and 94.2% low lead soil (< 50 ppm)
collected in Leadville, Colorado. NIST SRM 2589, composed of
paint collected from the interior surfaces of houses in the US,
contains a nominal lead concentration of 10% (100,000 ppm); the
material is powdered with more than 99% of the material being
less than 1 00 um in size.
A mixture of approximately 1 .2% galena and 98.8% low lead soil
(< 50 ppm) that was collected in Leadville, Colorado. The added
galena consisted of a mineralogical (i.e., native) crystal of pure
galena that was ground and sieved to obtain fine particles smaller
than about 65 um.
A composite of tailings samples collected from the Oregon Gulch
tailings impoundment.
Lead
Concentration
(ppm) 1
1,590
6,330
10,800
4,050
11,700
6,940
14,200
3,870
8,170
8,530
7,510
4,320
10,600
3,230
2,150
3,200
8,350
11,200
1,270
Samples were analyzed for lead by inductively coupled plasma-atomic emission spectrometry (ICP-AES) in accord with USEPA Method 200."
Tables.xls (2-3_TMs)
-------
TABLE 2-4. RELATIVE LEAD MASS OF MINERAL PHASES OBSERVED IN TEST MATERIALS
Experiment:
Phase
Anglesite
As(M)O
Calcite
Cerussite
Clay
Fe-Pb Oxide
Fe-Pb Sulfate
Galena
Lead Barite
Lead Organic
Lead Oxide
Lead Phosphate
Lead Silicate
Lead Vanidate
Mn-Pb Oxide
Native Lead
Pb(M)O
Pb-As Oxide
PbO-Cerussite
Slag
Sulfosalts
Zn-Pb Silicate
2
Bingham
Creek
Residential
2%
6%
22%
50%
18%
2%
Bingham
Creek
Channel
Soil
28%
0.3%
3%
30%
9%
0.04%
0.3%
26%
2%
1%
3
Jasper
County
High Lead
Smelter
1%
0.2%
32%
0.018%
14%
3%
0.09%
21%
2%
22%
4%
Jasper
County Low
Lead Yard
0.5%
81%
0.003%
2%
1%
8%
6%
0.04%
2%
0.15%
4
Murray
Smelter
Slag
1.0%
1.1%
2%
0.3%
9%
69%
0.8%
0.7%
4%
6%
7%
0.03%
Jasper
County
High Lead
Mill
2%
0.1%
57%
0.017%
10%
1%
3%
0.01%
7%
7%
0.5%
9%
2%
1%
5
Aspen
Berm
7%
62%
0.1%
9%
5%
12%
0.06%
0.03%
1%
4%
Aspen
Residential
1%
64%
7%
5%
17%
0.03%
1%
5%
6
Midvale
Slag
4%
0.3%
0.1%
6%
15%
26%
33%
16%
0.4%
Butte Soil
36%
0.3%
0.1%
7%
20%
12%
0.007%
3.6%
20.2%
7
Cal. Gulch
Phase I
Residential
Soil
10%
20%
6%
6%
2%
0.15%
0.11%
30%
1.9%
0.1%
22%
0.1%
1%
1%
Cal. Gulch
Fe/Mn PbO
0.01%
8%
3%
0.14%
0.11%
15%
0.8%
0.4%
72%
8
Cal. Gulch
AV Slag
2%
1%
51%
0.3%
3%
1%
31%
10%
9
Palmerton
Location 2
6%
0.03%
2%
1%
1%
24%
66%
Palmerton
Location 4
4%
0.13%
2%
0.1%
1%
1.4%
18%
66%
7%
2%
11
Murray
Smelter Soil
0.003%
14%
0.13%
0.6%
20%
27%
3%
29%
6%
NIST Paint
1%
55%
44%
12
Galena-
enriched
Soil
100%
Cal. Gulch
Oregon
Gulch
Tailings
100%
(M) = Metal
Tbl 2-4_l_ead Phases.xls (Table 2-4)
-------
TABLE 2-5. MATRIX ASSOCIATIONS FOR TEST MATERIALS
Experiment
2
3
4
5
6
7
8
9
11
12
Test Material
Bingham Creek Residential
Bingham Creek Channel Soil
Jasper County High Lead Smelter
Jasper County Low Lead Yard
Murray Smelter Slag
Jasper County High Lead Mill
Aspen Berm
Aspen Residential
Midvale Slag
Butte Soil
California Gulch Phase I Residential Soil
California Gulch Fe/Mn PbO
California Gulch AV Slag
Palmerton Location 2
Palmerton Location 4
Murray Smelter Soil
NIST Paint
Galena-enriched Soil
California Gulch Oregon Gulch Tailings
Particle Frequency
Liberated
100%
100%
81%
100%
87%
96%
86%
98%
91%
91%
79%
98%
78%
100%
79%
80%
100%
100%
2%
Included
0%
0%
19%
0%
13%
4%
14%
2%
9%
9%
21%
2%
22%
0%
21%
20%
0%
0%
98%
Relative Lead Mass
Liberated
100%
100%
76%
94%
77%
93%
93%
94%
77%
91%
65%
100%
80%
100%
89%
70%
100%
100%
5%
Included
0%
0%
24%
6%
23%
7%
8%
6%
23%
9%
35%
0%
20%
0%
11%
30%
0%
0%
95%
Tables.xls (2-5_Matrix)
-------
TABLE 2-6. PARTICLE SIZE DISTRIBUTIONS FOR TEST MATERIALS
Experiment
2
3
4
5
6
7
8
9
11
12
Test Material
Bingham Creek Residential
Bingham Creek Channel Soil
Jasper County High Lead Smelter
Jasper County Low Lead Yard
Murray Smelter Slag
Jasper County High Lead Mill
Aspen Berm
Aspen Residential
Midvale Slag
Butte Soil
California Gulch Phase I Residential Soil
California Gulch Fe/Mn PbO
California Gulch AV Slag
Palmerton Location 2
Palmerton Location 4
Murray Smelter Soil
NIST Paint
Galena-enriched Soil
California Gulch Oregon Gulch Tailings
Particle Size (pm)
<5
38%
66%
44%
29%
14%
23%
27%
38%
6%
23%
24%
26%
19%
26%
25%
23%
76%
48%
85%
5-9
22%
13.6%
19%
20%
13%
21%
19%
35%
1%
15%
9%
19%
8%
23%
15%
10%
4%
2%
8%
10-19
19%
10%
8%
21%
15%
22%
22%
12%
3%
14%
18%
24%
8%
25%
21%
29%
6%
4%
6%
20-49
16%
6.1%
8%
20%
6%
19%
17%
8%
4%
23%
22%
17%
5%
18%
25%
17%
8%
41%
0%
50-99
4%
3%
9%
8%
20%
9%
8%
4%
20%
14%
15%
10%
9%
6%
13%
6%
6%
4%
0%
100-149
2%
1%
9%
3%
24%
6%
6%
2%
29%
9%
9%
4%
19%
1%
2%
8%
0%
0%
0%
150-199
0%
0%
2%
0%
4%
1%
1%
0%
18%
2%
1%
0%
10%
0%
0%
3%
0%
0%
0%
200-249
0%
0%
1%
0%
3%
1%
1%
0%
13%
1%
1%
0%
13%
0%
0%
3%
0%
0%
0%
>250
0%
0%
1%
0%
0%
0%
0%
0%
5%
0%
1%
0%
9%
0%
0%
1%
0%
0%
0%
Tables.xls (2-6_Size)
-------
TABLE 2-7. ESTIMATED RBA VALUES FOR TEST MATERIALS
Experiment
2
3
4
5
6
7
8
9
11
12
Test Material
Bingham Creek Residential
Bingham Creek Channel Soil
Jasper County High Lead Smelter
Jasper County Low Lead Yard
Murray Smelter Slag
Jasper County High Lead Mill
Aspen Berm
Aspen Residential
Midvale Slag
Butte Soil
California Gulch Phase I Residential
Soil
California Gulch Fe/Mn PbO
California Gulch AV Slag
Palmerton Location 2
Palmerton Location 4
Murray Smelter Soil
NIST Paint
Galena-enriched Soil
California Gulch Oregon Gulch
Tailings
Blood AUC
RBA
0.34
0.30
0.65
0.94
0.47
0.84
0.69
0.72
0.21
0.19
0.88
1.16
0.26
0.82
0.62
0.70
0.86
0.01
0.07
LB
0.23
0.20
0.47
0.66
0.33
0.58
0.54
0.56
0.15
0.14
0.62
0.83
0.19
0.61
0.47
0.54
0.66
0.00
0.04
UB
0.50
0.45
0.89
1.30
0.67
1.21
0.87
0.91
0.31
0.29
1.34
1.76
0.36
1.05
0.80
0.89
1.09
0.02
0.13
Liver
RBA
0.28
0.24
0.56
1.00
0.51
0.86
0.87
0.77
0.13
0.13
0.75
0.99
0.19
0.60
0.53
0.58
0.73
0.02
0.11
LB
0.20
0.17
0.42
0.75
0.33
0.54
0.58
0.50
0.09
0.09
0.53
0.69
0.11
0.41
0.37
0.42
0.52
0.00
0.04
UB
0.39
0.34
0.75
1.34
0.88
1.47
1.39
1.21
0.17
0.19
1.12
1.46
0.32
0.91
0.79
0.80
1.03
0.04
0.21
Kidney
RBA
0.22
0.27
0.58
0.91
0.31
0.70
0.73
0.78
0.12
0.15
0.73
1.25
0.14
0.51
0.41
0.36
0.55
0.01
0.05
LB
0.15
0.19
0.43
0.68
0.22
0.50
0.46
0.49
0.08
0.09
0.50
0.88
0.08
0.30
0.25
0.25
0.38
0.00
0.02
UB
0.31
0.37
0.79
1.24
0.46
1.02
1.26
1.33
0.18
0.22
1.12
1.91
0.25
0.91
0.72
0.52
0.78
0.02
0.09
Femur
RBA
0.24
0.26
0.65
0.75
0.31
0.89
0.67
0.73
0.11
0.10
0.53
0.80
0.20
0.47
0.40
0.39
0.74
0.01
0.01
LB
0.19
0.21
0.52
0.60
0.23
0.69
0.51
0.56
0.06
0.04
0.33
0.51
0.13
0.37
0.32
0.31
0.59
-0.01
-0.04
UB
0.29
0.31
0.82
0.95
0.41
1.18
0.89
0.97
0.18
0.19
0.93
1.40
0.30
0.60
0.52
0.49
0.93
0.03
0.06
Point Estimate
RBA
0.27
0.27
0.61
0.90
0.40
0.82
0.74
0.75
0.14
0.14
0.72
1.05
0.20
0.60
0.49
0.51
0.72
0.01
0.06
LB
0.17
0.19
0.43
0.63
0.23
0.51
0.48
0.50
0.07
0.06
0.38
0.57
0.09
0.34
0.29
0.29
0.44
0.00
-0.01
UB
0.40
0.36
0.79
1.20
0.64
1.14
1.08
1.04
0.24
0.23
1.07
1.56
0.31
0.93
0.72
0.79
0.98
0.03
0.15
LB = 5% Lower Confidence Bound
UB = 95% Upper Confidence Bound
Tables.xls (2-7_RBA)
-------
TABLE 2-8. GROUPED LEAD PHASES
Group
1
2
3
4
5
6
7
8
9
10
Group Name
Galena
Cerussite
Mn(M) Oxide
Lead Oxide
Fe(M) Oxide
Lead Phosphate
Anglesite
Pb(M) Oxide
Fe(M) Sulfate
Minor Constituents
Phase Constituents
Galena (PbS)
Cerussite
Mn-Pb Oxide
Lead Oxide
Fe-Pb Oxide (including Fe-Pb Silicate)
Zn-Pb Silicate
Lead Phosphate
Anglesite
As(M)O
Lead Silicate
Lead Vanidate
Pb(M)O
Pb-As Oxide
Fe-Pb Sulfate
Sulfosalts
Calcite
Clay
Lead Barite
Lead Organic
Native Lead
PbO-Cerussite
Slag
(M) = Metal
Tables.xls (2-8_Phases)
-------
TABLE 2-9. CURVE FITTING PARAMETERS FOR ORAL LEAD ACETATE DOSE-RESPONSE CURVES
Experiment
2
3
4
5
6
7
8
9
11
12
Mean
Standard Deviation
Coefficient of Variation
Blood AUC
a b c
13.6 116 0.0084
8.3 163 0.0040
8.5 144 0.0064
8.0 163 0.0038
8.4 85 0.0101
a a a
8.0 159 0.0032
7.5 96 0.0087
7.2 160 0.0035
7.6 169 0.0040
8.6 140 0.0058
1.9 32 0.0026
23% 23% 46%
Liver Lead
a b
63 2.0
10 2.3
57 1.7
62 2.0
23 2.0
10 1.7
11 2.1
11 2.3
14 1.3
9 0.7
27 1.8
24 0.5
88% 27%
Kidney Lead
a b
44 2.4
10 2.2
68 2.8
60 1.8
15 2.1
10 1.4
17 2.4
14 2.3
20 1.7
8 1.1
27 2.0
22 0.5
84% 26%
Bone Lead
a b
0.7 0.084
1.8 0.062
0.5 0.076
0.5 0.062
0.4 0.043
0.8 0.059
0.8 0.065
0.6 0.071
0.7 0.053
0.6 0.032
0.7 0.061
0.4 0.015
55% 25%
Basic Equations:
Blood AUC = a + b*(1-exp(-c*Dose))
a = baseline blood lead value in unexposed animals
b = maximum increase in steady-state blood lead cause by exposure
c = "shape" parameter that determines how steeply the response increases as dose increases
Tissue concentration (bone, liver, kidney) = a + b*Dose
a = baseline blood lead value in unexposed animals
b = slope of the increase in tissue content per unit increase in dose
Coefficient of Variation = Standard Deviation / Mean
a Experiment 7 Blood AUC: No stable solution was obtained using the exponential model.
Tables.xls (2-9_Fits)
-------
TABLE 2-10. REPRODUCIBILITY OF RBA MEASUREMENTS
RBA
Estimate
Blood AUC
Liver
Kidney
Bone
Point Estimate
Palmerton
Location 2
Testl
(Phase 2 Study 9)
0.82 ± 0.12
0.60 ± 0.14
0.51 ± 0.16
0.47 ± 0.07
0.60 ± 0.18
Test 2
(Phase 2 Study 12)
0.71 ± 0.09
1.25 ± 0.32
0.54 ± 0.13
0.95 ± 0.18
0.86 ± 0.33
California Gulch
Phase I Residential Soil
Testr
(Phase 1 Study 2)
0.69
0.58
0.62
0.50
0.60
Test 2
(Phase 2 Study 7)
0.88 ± 0.19
0.75 ± 0.16
0.73 ± 0.17
0.53 ± 0.15
0.72 ± 0.21
'Calculated using ordinary least squares.
Tables.xls (2-10_Reprod)
-------
TABLE 3-1. IN VITRO BIOACCESSIBILITY VALUES
Experiment
2
2
3
3
4
4
5
5
6
6
7
7
8
9
9
11
11
12
12
Test
Material
1
2
1
2
1
2
1
2
1
2
1
2
1
1
2
1
2
1
3
Sample
Bingham Creek Residential
Bingham Creek Channel Soil
Jasper County High Lead Smelter
Jasper County Low Lead Yard
Murray Smelter Slag
Jasper County High Lead Mill
Aspen Berm
Aspen Residential
Midvale Slag
Butte Soil
California Gulch Phase I Residential Soil
California Gulch Fe/Mn PbO
California Gulch AV Slag
Palmerton Location 2
Palmerton Location 4
Murray Smelter Soil
NIST Paint
Galena-enriched Soil
California Gulch Oregon Gulch Tailings
In Vitro Bioaccessibility (%)
(Mean ± Standard Deviation)
47.0 ±1.2
37.8 ±0.7
69.3 ±5.5
79.0 ±5. 6
64.3 ±7.3
85.3 ±0.2
64.9 ±1.6
71.4 ±2.0
17.4 ±0.9
22.3 ±0.6
65.1 ±1.5
87.2 ±0.5
9.4 ± 1.6
63.6 ±0.4
69.7 ±2.7
74.7 ±6.8
72.5 ±2.0
4.5 ± 1.2
11.2 ±0.9
Tbl 3-1, ES-2JVBA Data.xls (Table 3-1)
-------
This page intentionally left blank to facilitate double-sided printing.
-------
FIGURES
OSWER 9285.7-77
-------
This page intentionally left blank to facilitate double-sided printing.
-------
FIGURE 2-1. AVERAGE RATE OF BODY WEIGHT GAIN IN TEST ANIMALS
0.7 -
0.6 -
ro
ro
£ 0.4 -
D)
§
o o 3
-------
FIGURE 2-2. EXAMPLE TIME COURSE OF BLOOD LEAD RESPONSE
10
O)
6
T3
CC
CD
T3
8 4
CQ
2 -
+ Control
- PbAc(75)
A- PbAc(225)
*Test Material (75)
*Test Material (225)
Test Material (675)
-6
-4
-2
246
Study Day
10
12
14
16
Fig 2-2 to 2-6 Dose Resp (outliers exclj.xls
-------
FIGURE 2-3. DOSE RESPONSE CURVE FOR BLOOD LEAD AUC
250
Panel A: Lead Acetate - IV
50
100 150 200 250
Lead Dose (|jg Pb/kg-day)
300
350
200
Panel B: Lead Acetate - Oral
50
100 150 200 250
Lead Dose (|jg Pb/kg-day)
300
350
Fig 2-2 to 2-6 Dose Resp (outliers excl).xls (Fig 2-3_AUC)
-------
FIGURE 2-4. DOSE RESPONSE CURVE FOR LIVER LEAD
CONCENTRATION
3500
3000-
'5? 2500 ]
2000
Panel A: Lead Acetate - IV
50
100 150 200 250
Lead Dose (|jg Pb/kg-day)
300
350
1200
Panel B: Lead Acetate - Oral
50
100 150 200 250
Lead Dose (|jg Pb/kg-day)
300
350
Fig 2-2 to 2-6 Dose Resp (outliers excl).xls (Fig 2-4_Liver)
-------
FIGURE 2-5. DOSE RESPONSE CURVE FOR KIDNEY LEAD
CONCENTRATION
1800
Panel A: Lead Acetate - IV
50
100 150 200 250
Lead Dose (|jg Pb/kg-day)
300
350
1400
Panel B: Lead Acetate - Oral
50
100 150 200 250
Lead Dose (|jg Pb/kg-day)
300
350
Fig 2-2 to 2-6 Dose Resp (outliers excl).xls (Fig 2-5_ idney)
-------
FIGURE 2-6. DOSE RESPONSE CURVE FOR FEMUR LEAD
CONCENTRATION
50
Panel A: Lead Acetate - IV
100 150 200 250
Lead Dose (|jg Pb/kg-day)
300
350
Panel B: Lead Acetate - Oral
50
100 150 200 250
Lead Dose (|jg Pb/kg-day)
300
350
Fig 2-2 to 2-6 Dose Resp (outliers excl).xls (Fig 2-6_Fe ur)
-------
FIGURE 2-7. ESTIMATED GROUP-SPECIFIC RBA VALUES
Group
I Fully constrained
D Partially constrained
Fig 2-7_Phase RBAs.xls (Fig 2-7)
-------
FIGURE 2-8. CORRELATION OF DUPLICATE ANALYSES
Panel A: Blood Lead
Observed
y = 0.9645x +0.0687
10 15 20
Duplicate Value (|jg/dL)
25
30
350
Panel B: Tissue Lead
Observed
y = 0.9526x+ 1.8423
50 100 150 200 250
Duplicate Value (|jg/dL)
300
350
Figures.xls (2-8_Dups)
-------
FIGURE 2-9. RESULTS FOR CDC BLOOD LEAD CHECK SAMPLES
18 -
16 -
14 -
a 12 -\
T3
1>
0)
T3
8 8
CQ
ca
"1 K
S D -
4
2
0 -
'
"
*
-
A A ^
A A A A
50 5 10 15
A Low Std
* Med Std
High Std
Study Day
Figures.xls (2-9_CDC)
-------
FIGURE 2-10. INTERLABORATORY COMPARISON OF BLOOD LEAD RESULTS
10 15 20
CDCP Blood Lead Results (|jg/dl_)
25
30
Figures.xls (2-10_lnterlab)
-------
FIGURE 3-1. IN VITRO BIOACCESSIBILITY EXTRACTION APPARATUS
Circulating
Heater Plexiglass Tank
(Set at 37° C)
Magnetic Flywheel
125 ml Nalgene wide mouth bottles
Gearbox & motor
(28 RPM)
Figures.xls (3-1_Apparatus)
-------
FIGURE 3-2. EFFECT OF TEMPERATURE, TIME, AND pH ON IVBA
Panel A: Effect of Temperature
1 nn
90 -
80 -
70 -
< 50 -
CQ
> 40-
30 -
20 -
10 -
J
-!
I
fl
Ej
i
n^
i
1
I
1234567
s
p
I
I
I
i
n
f
i
rl
j
i
37 C
20 C
8 9 10 11 12 13 14 15 16 17
Test Material
120
100 -
80 -
g
< 60 -
CQ
40
20
0
Panel B: Effect of Extraction Time
567
Test Material
10
11
Panel C: Effect of pH
90
80
70 -
if 6°-
< 50 -
CQ
>
- 40 -
30 -
20
10
r
l
E
i II
T
F
r
i
n
L
n
i
T
H^
ffi
n
i
Ej
n,
Test Material
T
i
N
[
I
i jfi
E n
n
n
i
i
r co O) o T CM
DpH = 1.5
DpH - 2.5
DpH = 3.5
m
Fig 3-2_Effects on IVBA.xls (Fig 3-2_Effects)
-------
FIGURE 3-3. PRECISION OF IN VITRO BIOACCESSIBILITY MEASUREMENTS
100 n
90 -
80
70
5? 60 -
QQ
tt 50
! 4°-
30
20
10 -
0
0
MS 2710 MS 2711
Mean = 75.5 Mean = 84.4
Std. Dev. = 4.7 Std. Dev. = 4.7
CV = 0.062 CV= 0.055
N = 68 N = 66
Figures.xls (3-3_Precision)
-------
FIGURE 3-4. REPRODUCIBILITY OF IN VITRO BIOACCESSIBILITY MEASUREMENTS
100 -r
90 -
70 --
§ 60-
0
Q.
< 50 -
CD
o:
c 40 -
(0 ^"
0
S
30 --
20 --
10 --
[
fi
T *
T
E
-
fa
IT
12345
M
fl ST'
E
E
ACZ DCUB DERCL DNER
.L
s
f
6 7 8 9 10 11 12
Test Material
j
13
E
E
E
T
T
E
T
r
E
r
-
14 15 16 17 18 19
Test Materials
1 = Aspen Berm
2 = Aspen Residential
3 = Bingham Creek Channel Soil
4 = Bingham Creek Residential
5 = Butte Soil
6 = Galena-enriched Soil
7 = Jasper County High Lead Mill
8 = Jasper County High Lead Smelter
9 = Jasper County Low Lead Yard
10 = California Gulch AV Slag
11 = California Gulch Fe/Mn PbO
12 = California Gulch Oregon Gulch Tailings
13 = California Gulch Phase I Residential Soil
14 = MidvaleSlag
15 = Murray Smelter Slag
16 = Murray Smelter Soil
17 = Palmerton Location 2
18 = Palmerton Location 4
19 = NIST Paint
Laboratories
ACZ = ACZ Laboratories, Inc.
CUB = University of Colorado at Boulder
ERCL = Environmental Research Chemistry Laboratory, U.S. Bureau of Reclamation
NERL = National Exposure Research Laboratory
Fig 3-4_Reproducibility.xls (Fig 3-4)
-------
FIGURE 3-5. RBAvs. IVBA
1.2 n
1.1 -
1.0 -
0.9
0.8 -
0.7 -
I °-6
0.5 -
0.4 -
0.3
0.2 -
0.1 -
0.0
0
*
*
t*
* *
* *
*
*
*
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
IVBA
Fig 3-5_IVBA-RBA.xls (Fig 3-5)
-------
FIGURE 3-6. PREDICTION INTERVAL FOR RBA BASED ON MEASURED IVBA
1.4
00
QL
1.2 -
1.0
0.8
0.6 -
0.4
0.2
0.0
95% Prediction Interval
0.8781VBA - 0.028
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
IVBA
Fig 3-6, D-7_Prediction lntervals_NEW2.xls (Fig 3-6)
------- |