DRAFT- Do Not Cite, Quote, or Release
&ERA
OSWER 9285.7-77
December 2004
ESTIMATION OF
RELATIVE BIOAVAILABILITY 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
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
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 (USEPA) support for the development of this report was
provided by Michael Beringer, USEPA Region 7, Kansas City, KS; Jim Luey, USEPA Region 8,
Denver, CO; and Richard Troast, USEPA OSRTI, Washington, DC. Contractor support to
USEPA was provided by Syracuse Research Corporation.
IN VIVO STUDIES
All of the in vivo studies described in this report were planned and sponsored by USEPA, 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, DACVIM, 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. Mike Ruby at
Exponent. Dr. Drexler also performed all of the electron microprobe and particle size analyses
of the test materials evaluated in these studies.
-------
DRAFT- Do Not Cite, Quote, or Release
STATISTICAL ANALYSIS
Dr. Timothy Barry, USEPA 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 USEPA's National Risk
Management Research Laboratory provided several rounds of valuable review comments and
constructive discussions regarding statistical methodology.
REVIEWERS
A draft of this report was provided to three independent experts for review and comment. 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
-------
DRAFT- Do Not Cite, Quote, or Release
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 (|j,g) of lead
dissolved in drinking water were ingested and a total of 50 ng were absorbed into the body, the
ABA would be 0.50 (50%). Likewise, if 100 ng of lead contained in soil were ingested and 30
|xg 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 (USEPA) 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 USEPA
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.
ES-1
-------
DRAFT- Do Not Cite, Quote, or Release
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 four different tissues.
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:
ES-2
-------
DRAFT- Do Not Cite, Quote, or Release
Liver. Kidney. Bone
C(tissue) = a + b-Dose
Blood AUC
AUC = a + b [l - exp(-c-Dose)]
Based on these models, RBA is calculated from the best model fits as follows:
RBA(liver, kidney, bone) = b(test material) / preference material)
RBA (blood AUC) = c(test material) / preference 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). 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 RBA 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 105% (Fe/Mn lead oxide sample). This wide variability
highlights the importance of obtaining and applying reliable RBA data in order help to improve
risk assessments for lead exposure.
Correlation of RBA with Mineral Phase
Available data are not yet sufficient to establish reliable quantitative estimates of RBA for each
of the different mineral phases of lead that are observed to occur in the test materials. However,
multi-variate regression analysis between point estimate RBA values and mineral phase content
ES-3
-------
DRAFT- Do Not Cite, Quote, or Release
of the different test materials allows a tentative rank ordering of the phases into three semi-
quantitative tiers (low, medium, or high RBA), as follows:
Low Bioavailability
Medium Bioavailability
High Bioavailability
Fe(M) Sulfate
Anglesite
Galena
Pb(M) Oxide
Fe(M) Oxide
Lead Phosphate
Lead Oxide
Cerussite
Mn(M) Oxide
3.0 IN VITRO STUDIES
Measurement of lead RBA 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 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 Eire 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).
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
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.
ES-4
-------
DRAFT- Do Not Cite, Quote, or Release
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 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. Figure ES-1 shows the best fit
linear regression correlation between the in vivo RBA estimates and the in vitro bioaccessibility
estimates for each of the 19 test materials investigated during this program. The equation of the
line is:
RBA = 1.03 IVBA- 0.06
Non-linear models yield a slightly better fit to the data, but this is not thought to be meaningful.
These results indicate that the in vivo RBA of 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 shown by the 5% and 95%
prediction limits 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
ES-5
-------
DRAFT- Do Not Cite, Quote, or Release
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 constitute a useful and stable animal model for measuring in vivo lead
absorption from a variety of test materials. The model is most useful for estimating the
RELATIVE bioavailability of a test material in comparison to some reference material
(usually lead acetate).
2. Each of the four different endpoints employed in these studies (blood AUC, liver, kidney,
bone) to estimate RB A in vivo yield reasonable data, and the best estimate of the RBA
value for any particular sample is the average across all four endpoint-specific RBA
values.
3. There are clear differences in the in vivo RBA of lead between different types of test
material, ranging from near zero to close to 100%. Thus, knowledge of the RBA value
for different types of test 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 RBA 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 RBA, 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.
ES-6
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE ES-1. SUMMARY OF ESTIMATED RBA VALUES FOR TEST MATERIALS
Experiment
Test Material
Blood AUC
Liver
Kidney
Femur
Point
Estimate
2
Bingham Creek Residential
0.34
0.28
0.22
0.24
0.27
Bingham Creek Channel Soil
0.30
0.24
0.27
0.26
0.27
3
Jasper County High Lead Smelter
0.65
0.56
0.58
0.65
0.61
Jasper County Low Lead Yard
0.94
1.00
0.91
0.75
0.90
4
Murray Smelter Slag
0.47
0.51
0.31
0.31
0.40
Jasper County High Lead Mill
0.84
0.86
0.70
0.89
0.82
5
Aspen Berm
0.69
0.87
0.73
0.67
0.74
Aspen Residential
0.72
0.77
0.78
0.73
0.75
6
Midvale Slag
0.21
0.13
0.12
0.11
0.14
Butte Soil
0.19
0.13
0.15
0.10
0.14
7
California Gulch Phase I Residential Soil
0.88
0.75
0.73
0.53
0.72
California Gulch Fe/Mn PbO
1.16
0.99
1.25
0.80
1.05
8
California Gulch AV Slag
0.26
0.19
0.14
0.20
0.20
9
Palmerton Location 2
0.82
0.60
0.51
0.47
0.60
Palmerton Location 4
0.62
0.53
0.41
0.40
0.49
11
Murray Smelter Soil
0.70
0.58
0.36
0.39
0.51
NIST Paint
0.86
0.73
0.55
0.74
0.72
12
Galena-enriched Soil
0.01
0.02
0.01
0.01
0.01
California Gulch Oregon Gulch Tailings
0.07
0.11
0.05
0.01
0.06
Tables.xls (ES-1_RBA)
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE ES-2. IN VITRO BIOACCESSIBILITY VALUES
Experiment
Sample
In Vitro Bioaccessibility
(Mean % ± Standard Deviation)
2
Bingham Creek Residential
47.0 ± 1.2
2
Bingham Creek Channel Soil
37.8 ± 0.7
3
Jasper County High Lead Smelter
69.3 ± 5.5
3
Jasper County Low Lead Yard
79.0 ± 5.6
4
Murray Smelter Slag
65.5 ± 7.5
4
Jasper County High Lead Mill
80.4 ± 4.2
5
Aspen Berm
64.9 ± 1.6
5
Aspen Residential
71.4 ± 1.9
6
Midvale Slag
17.9 ± 1.0
6
Butte Soil
22.1 ± 0.6
7
California Gulch Phase I Residential Soil
65.1 ± 1.5
7
California Gulch Fe/Mn PbO
87.2 ± 0.5
8
California Gulch AV Slag
9.4 ± 1.6
9
Palmerton Location 2
63.6 ± 0.4
9
Palmerton Location 4
69.7 ± 2.7
11
Murray Smelter Soil
74.7 ± 6.8
11
NIST Paint
72.5 ± 2.0
12
Galena-enriched Soil
4.5 ± 1.2
12
California Gulch Oregon Gulch Tailings
11.2 ±0.9
Tbl 3-1, ES-2 IVBA Data.xls (Table ES-2)
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE ES-1. RELATION BETWEEN RBA AND IVBA
Fig 3-6, D-8_Prediction lntervals.xls (Fig ES-1)
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE OF CONTENTS
1.0 INTRODUCTION 14
1.1 Overview 14.
1.2 Using Bioavailability Data to Improve Exposure Calculations for Lead L2
1.3 Overview of USEPA's Program to Study Lead Bioavailability in Animals ... 1-3
1.4 Overview of Methods for Estimating Lead RBA In Vitro K3
2.0 IN VIVO STUDIES 24
2.1 Basic Approach for Measuring RBA In Vivo 24
2.2 Animal Exposure and Sample Collection 24
2.3 Preparation of Biological Samples for Analysis 2zL
2.4 Data Reduction 2^3
2.5 Results and Discussion 2^3
2.5.1 Effect of Dosing on Animal Health and Weight 2-3
2.5.2 Time Course of Blood Lead Response 2^4
2.5.3 Dose-Response Patterns 2^4
2.5.4 Estimation of ABA for Lead Acetate 2-5
2.5.5 Estimation of RBA for Lead in Test Materials 2-6
2.5.6 Effect of Food 2£7
2.5.7 Correlation of RBA with Mineral Phase 2-9
2.5.8 Quality Assurance 2-11
3.0 IN VITRO STUDIES 34
3.1 Introduction 34,
3.2 In Vitro Method 34
3.2.1 Sample Preparation 34
3.2.2 Apparatus 3^2
3.2.3 Selection of IVBA Test Conditions 3^2
3.2.4 Summary of Final Leaching Protocol 3^4
3.2.5 Extraction Fluid Analysis 3^5
3.2.6 Quality Control/Quality Assurance 3^5
3.3 Results and Discussion 3^6
3.3.1 IVBA Values M
3.3.2 Comparison with In Vivo Results 3£7
4.0 REFERENCES 4-1
-------
DRAFT- Do Not Cite, Quote, or Release
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 CDCP 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
ii
-------
DRAFT- Do Not Cite, Quote, or Release
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
iii
-------
DRAFT- Do Not Cite, Quote, or Release
°C
Hg
[xm
ABA
AF0
AIC
AUC
cc
CDCP
dL
g
GLP
HC1
HDPE
ICP-AES
ICP-MS
IV
IVBA
kg
L
M
MDL
mg
mL
mm
NIST
Pb
PbAc
PbB
PPb
ppm
ACRONYMS AND ABBREVIATIONS
Degrees Celsius
Microgram
Micrometer
Absolute bioavailability
Oral absorption fraction
Akaike's Information Criterion
Area under the curve
Cubic centimeter
Centers for Disease Control and Prevention
Deciliter
Gram
Good Laboratory Practices
Hydrochloric acid
High density polyethylene
Inductively Coupled Plasma-Atomic Emission Spectrometry
Inductively Coupled Plasma-Mass Spectrometry
Intravenous
In vitro bioaccessibility
Kilogram
Liter
Molar
Method detection limit
Milligram
Milliliter
Millimeter
National Institute of Standards and Testing
Lead
Lead acetate
Blood lead
Parts per billion
Parts per million
iv
-------
DRAFT- Do Not Cite, Quote, or Release
ACRONYMS AND ABBREVIATIONS
(Continued)
RBA
Relative bioavailability
RLM
Relative lead mass
rpm
Revolutions per minute
SOP
Standard operating procedure
SRM
Standard Reference Material
TAL
Target Analyte List
TCLP
Toxicity Characteristic Leaching Procedure
USEPA
U.S. Environmental Protection Agency
v
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
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 (soil, dust, water, food, air, paint, etc.), 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 CAB A") is the ratio of the amount of lead absorbed compared to
the amount ingested:
ABA = (Absorbed Dose) / (Ingested Dose)
This ratio is also referred to as the oral absorption fraction (AF0).
Relative Bioavailability fRBA) is the ratio of the absolute bioavailability of lead present
in some test material compared the absolute bioavailability of lead in some appropriate
reference material:
1-1
-------
DRAFT- Do Not Cite, Quote, or Release
RBA = ABA(test) / ABA(reference)
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 (ng) of lead dissolved in drinking water were ingested and a
total of 50 |ig entered the body, the ABA would be 50/100, or 0.50 (50%). Likewise, if 100 |ig
of lead contained in soil were ingested and 30 |ig 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 et al. (1990), Mushak (1991), and/or Klaassen et al. (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:
ABAsoii = ABAsoluble-RBAsoil
where:
ABAsoil = Absolute bioavailability of lead in soil ingested by a child
ABAsolublc = Absolute bioavailability in children of some dissolved or fully
soluble form of lead
RBAsoi, = Relative bioavailability of lead in soil
Based on available information in the literature on lead absorption in humans, the U.S.
Environmental Protection Agency (USEPA) estimates that the absolute bioavailability of lead
from water and the diet is usually about 50% in children (USEPA, 1994). Thus, when a reliable
1-2
-------
DRAFT- Do Not Cite, Quote, or Release
site-specific RBA value for soil is available, it may be used to estimate a site-specific absolute
bioavailability in that soil, as follows:
ABAsoil = 50% • RBAsoil
In the absence of site-specific data, the absolute absorption of lead from soil, dust, and other
similar media is estimated by USEPA to be about 30% (USEPA, 1994). Thus, the default RBA
used by USEPA 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.
1.3 Overview of USEPA's Program to Study Lead Bioavailability in Animals
Scientists in USEPA 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 al. 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
1-3
-------
DRAFT- Do Not Cite, Quote, or Release
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 et al., 1993; Ruby et al., 1996; Medlin, 1997; Rodriguez et
al., 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.
1-4
-------
DRAFT- Do Not Cite, Quote, or Release
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 Health-National Research Council. 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
|j.g/kg, corresponding to a daily intake of less than 2.5 |ig/kg-day.
2-1
-------
DRAFT- Do Not Cite, Quote, or Release
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 |ig/L, corresponding to a daily intake of less than 0.2 [ig/kg-
day.
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 (CDCP) for analysis of blood samples for lead. 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 |xg Pb/L of prepared sample. The quantitation 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 jxg/L).
2-2
-------
DRAFT- Do Not Cite, Quote, or Release
2.4 Data Reduction
The basic data reduction task required to calculate an RB A 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- [1 - exp(-c-Dose)] (2)
Appendix D presents a detailed description of the curve-fitting methods and rationale, along with
the methods used to quantify uncertainty in the RBA 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
2-3
-------
DRAFT- Do Not Cite, Quote, or Release
typically gained about 0.3 to 0.5 kg/day, and the rate of weight gain was normally 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 al., 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 quantitation limit (usually about 1 |ig/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 within about 7 to 10 days.
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.
1 Most studies in Phase II utilized only one IV dose level (100 ng/kg-day), and hence do not provide dose-response
data. Study 8 included three IV exposure levels (25, 50, and 100 ng/kg-day), and the data from this study are shown
in Figures 2-3 to 2-6.
2-4
-------
DRAFT- Do Not Cite, Quote, or Release
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, it is apparent 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. 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 reveal 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
Estimated ABA of PbAc
Blood AUC
0.10 ±0.02
Liver
0.16 ±0.05
Kidney
0.19 ±0.05
Femur
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 al., 1974; Ziegler et al.,
1978). If so, lead absorption in juvenile swine is apparently lower than for young 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
-------
DRAFT- Do Not Cite, Quote, or Release
2.5.5 Estimation of RBA 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 |im) were used. This is because it is believed
that soil particles less than about 250 jxm are most likely to adhere to the hands and be ingested
by hand-to-mouth contact, especially in young children.
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
liberated (partially or entirely) or included 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.
2-6
-------
DRAFT- Do Not Cite, Quote, or Release
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
range of particle sizes, often with the majority of the particles being less than 50 |j,m.
(Remember that all samples were sieved to isolate particles less than 250 |j,m 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; USEPA, 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
2-7
-------
DRAFT- Do Not Cite, Quote, or Release
for active and/or passive uptake sites in the gastrointestinal epithelium (Diamond, 2002).
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:
Measurement
Endpoint
Ratio of PbAc Absorption Given With Food or After Feeding
Compared to PbAc Given Without Food
PbAc Given with Food
PbAc Given 2 hrs after Food
Blood Lead AUC
0.39 ± 0.05
0.40 ± 0.06
Liver Lead
0.86 ± 0.24
0.58 ±0.16
Kidney Lead
0.72 ± 0.26
0.73 ± 0.27
Bone Lead
0.35 ± 0.05
0.33 ± 0.05
Point Estimate
0.58 ± 0.28
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 holding time study in juvenile swine
which indicated that food is held in the stomach for up to four hours after eating.
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.
2-8
-------
DRAFT- Do Not Cite, Quote, or Release
2.5.7 Correlation of RBA with Mineral Phase
In principle, each unique combination of phase, size, and matrix association constitutes a unique
mineralogical 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:
p s m
RBAsampie ~
i=i j=i k=i
where:
RBAsample = Observed RBA of lead in a sample
Qj>k = Fraction of total lead in phase "i" of size "j" and matrix association
"k"
RBAij k = Relative bioavailability of lead in phase "i" of size "j" and matrix
association "k"
p = Number of different lead phase categories
s = Number of different size categories
m = Number of different matrix association categories
If the number of different lead phases which may 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, liberated), then the total number of different "types" of lead is on the
order of 200. Because measured RBA 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:
p
RBA.sclmple ^i,liberated ^^^i,liberated
i=1
2-9
-------
DRAFT- Do Not Cite, Quote, or Release
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 10 categories as shown in Table
2-8. These groups were based on professional judgement 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:
2-10
-------
DRAFT- Do Not Cite, Quote, or Release
Low Bioavailability
(RBA < 0.25)
Medium Bioavailability
(RBA = 0.25-0.75)
High Bioavailability
(RBA >0.75)
Fe(M) Sulfate
Anglesite
Galena
Fe(M) Oxide
Pb(M) Oxide
Lead Oxide
Lead Phosphate
Cerussite
Mn(M) Oxide
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 Centers for Disease Control and Prevention (CDCP) 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 |ig/dL, 4.8 |ig/dL and 14.9 |J.g/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 |ig/L
for the low standards (nominal concentration of 1.7 ng/L), 4.3 |ig/L for the middle standard
2-11
-------
DRAFT- Do Not Cite, Quote, or Release
(nominal concentration of 4.8 |J.g/L), and 14.5 |j,g/L for the high standards (nominal
concentration of 14.9 ng/L).
Interlaboratorv 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 CDCP for blind
independent preparation and analysis. The results are plotted in Figure 2-10. As seen, the
results of analyses by USEPA's laboratory are generally similar to those of CDCP, with a mean
inter-sample difference (USEPA minus CDCP) of 0.07 ng/dL. The slope of the best-fit straight
line through the data is 0.84, indicating that the concentration values estimated by the USEPA
laboratories tended to be about 15% lower than those estimated by CDCP. The reason for this
apparent discrepancy between the USEPA laboratory and the CDCP 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
RBA 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 Across Studies
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 parameters (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.) can not be excluded.
2-12
-------
DRAFT- Do Not Cite, Quote, or Release
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 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.
2-13
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
3.0 IN VITRO STUDIES
3.1 Introduction
Measurement of lead RBA 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 test systems for estimating lead
bioaccessibility can be found in Imber (1993), Ruby et al. (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. A
detailed standard operating procedure (SOP) may be downloaded from
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 |im. The <250-|im 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.
3-1
-------
DRAFT- Do Not Cite, Quote, or Release
3.2.2 Apparatus
The main piece of equipment used in these studies is 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-cm 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. The 125-mL
HDPE bottles had air-tight screw-cap seals, and care 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. Further details on the extraction apparatus can be
obtained from Dr. John Drexler at (303) 492-5251 or drexlerj@spot.colorado.edu.
3.2.3 Selection of IVBA 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.
Extraction Fluid. The extraction fluid selected for this procedure is 0.4 M glycine, adjusted to a
pH of 1.5 with 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 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
3-2
-------
DRAFT- Do Not Cite, Quote, or Release
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. 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.
pH. Pediatric gastric pH values tend to range from about 1 to 4 during fasting, and may be
elevated to about 5 for a few hours after ingestion of food. Previous authors have used stomach
phase pH values between 1.3 and 2.5 for their in vitro experiments (Ruby et al., 1993; Miller and
Schricker, 1982; Medlin, 1997). 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 in accordance with the reasoning of Ruby et al. (1996). Tests using
3-3
-------
DRAFT- Do Not Cite, Quote, or Release
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 +/- O.Olg.
3.2.4 Summary of Final Leaching Protocol
The extraction procedure begins by placing 1.00 ± 0.05 g of test substrate into a 125-mL wide-
mouth HDPE bottle. Care should be taken to ensure that static electricity does not cause soil
particles to adhere to the lip or outside threads of the bottle. To this is added 100 ± 0.5 mL of the
extraction fluid (0.4 M glycine, pH 1.5). The bottle is tightly sealed and then shaken or inverted
to ensure that there is no leakage and that no soil is caked on the bottom of the bottle.
Each bottle is placed into the modified TCLP extractor (water temperature = 37±2°C). Samples
are extracted by rotating the samples end-over-end at 30±2 rpm for 1 hour. After 1 hour, the
bottles are 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 an 0.45-|im 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.
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
3-4
-------
DRAFT- Do Not Cite, Quote, or Release
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, the test was repeated, stopping the extraction at 5,10,15, and 30
minutes and manually adjusting the pH down to pH 1.5 at each interval by dropwise addition of
HC1.
3.2.5 Extraction Fluid Analysis
Filtered samples of extraction fluid were stored in a refrigerator at 4°C until they were analyzed
(within 1 week of extraction). Filtered samples were analyzed for lead by ICP-AES or ICP-MS
(USEPA Method 6010 or 6020). Method detection limits (MDL) in extraction fluid were
calculated to be 19 and 0.1 jxg/L for Methods 6010 and 6020, respectively.
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.
Blank Spike — extraction fluid spiked at 10 mg/L lead, and run through the complete
procedure at a frequency of 1 in 20 samples.
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.
Duplicate Sample — duplicate sample extractions were performed on 1 in 10 samples.
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
in triplicate.
3-5
-------
DRAFT- Do Not Cite, Quote, or Release
Control limits for these quality control samples were as follows:
Analysis
Frequency
Control Limits
Reagent blank
once per batch
<25 fig/L lead
Bottle blank
5%
<50 fig/L lead
Blank spike (10 mg/L)
5%
85-115% recovery
Matrix spike (10 mg/L)
10%
75-125% recovery
Duplicate sample
10%
+/- 20% RPD*
Control soil (NIST 2711)
5%
+/- 10% RPD
*RPD = Relative percent difference
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%).
3-6
-------
DRAFT- Do Not Cite, Quote, or Release
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 0.874 (p <
0.001), and the Pearson product moment correlation coefficient is 0.915 (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. The details are presented in Appendix D, and
the results are summarized below:
Model
R2
AIC
Linear (RBA = a + b-IVBA)
0.837
-72.75
Power (RBA = a + bTVBA")
0.881
-75.35
2-Parameter Exponential (RBA = a + bexp(IVBA))
0.866
-73.16
3-Parameter Exponential (RBA = a + bexp(cTVBA))
0.883
-75.74
As seen, all of the models fit the data reasonably well, with the non-linear models (power,
exponential) fitting somewhat better than the linear model. However, as discussed in Appendix
D, the difference in quality of fit between linear and non-linear 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.
The process of fitting a linear model to the data is complicated by the fact that there are random
measurement errors in both the IVBA and the in vivo RBA estimates. However, as discussed in
Appendix D, measurement errors in IVBA are small compared to measurement errors in RBA,
so that a fit derived by ordinary linear regression appears to be reasonable. Based on this, the
currently preferred model is:
3-7
-------
DRAFT- Do Not Cite, Quote, or Release
RBA = 1.03IVBA - 0.06
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
that the true RBA may be somewhat different (either higher or lower). The distribution of
possible values of RBA that may be observed at any specified value of IVBA may be
characterized as a t-distribution, calculated as detailed in Appendix D (Section 5.0). The best fit
line and the 90% prediction interval for this data set are shown in Figure 3-6. For example, if the
measured IVBA for a test material were 0.60, the RBA value is expected to be about 0.56, with
90% of all future RBA values observed in conjunction with an IVBA of 0.60 expected to be
greater than 0.34 and less than 0.79.
Applicability of the IVBA-RBA Methodology
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 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.
3-8
-------
DRAFT- Do Not Cite, Quote, or Release
4.0 REFERENCES
Alexander, F. W., B. E. Clayton, and H. T. Delves. 1974. Mineral and trace-metal balances in
children receiving normal and synthetic diets. QJ Med 43:89-111.
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 kidney. In:
Molecular Biology and Toxicology of Metals. R. K. Zalups and J. Koropatnick (eds). Taylor &
Francis, London.
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.
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.
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.
4-1
-------
DRAFT- Do Not Cite, Quote, or Release
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.
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.
Mushak, P. 1998. Uses and limits of empirical data in measuring and modeling human lead
exposure. Env. Heath Perspect. 106 (Suppl. 6): 1467-1484.
Nernst, W., and E. Brunner. 1904. Theorie der reaktionsgeschwindigkeit in heterogenen
systemen. Z. Phys. Chem. 47:52.
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. W., 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. W., 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.
USEPA. 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.
4-2
-------
DRAFT- Do Not Cite, Quote, or Release
USEPA. 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.
Weis, C. P., G. M. Henningsen, R. H. Poppenga, and B. J. 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, B. J. 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 S. J. Fomon. 1978. Absorption
and retention of lead by infants. Pediatr. Res. 12:29-34.
4-3
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
OSWER 9285.7-77
TABLES
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE 2-1. TYPICAL FEED COMPOSITION
Nutrient Name
Amount
Protein
20.1021%
Arginine
1.2070%
Lysine
1.4690%
Methionine
0.8370%
Met+Cys
0.5876%
Tryptophan
0.2770%
Histidine
0.5580%
Leucine
1.8160%
Isoleucine
1.1310%
Phenylalanine
1.1050%
Phe+Tyr
2.0500%
Threonine
0.8200%
Valine
1.1910%
Fat
4.4440%
Saturated Fat
0.5590%
Unsaturated Fat
3.7410%
Linoleic 18:2:6
1.9350%
Linoleic 18:3:3
0.0430%
Crude Fiber
3.8035%
Ash
4.3347%
Calcium
0.8675%
Phos Total
0.7736%
Available Phosphorous
0.7005%
Sodium
0.2448%
Potassium
0.3733%
Nutrient Name
Amount
Chlorine
0.1911%
Magnesium
0.0533%
Sulfur
0.0339%
Manganese
20.4719 ppm
Zinc
118.0608 ppm
Iron
135.3710 ppm
Copper
8.1062 ppm
Cobalt
0.0110 ppm
Iodine
0.2075 ppm
Selenium
0.3196 ppm
Nitrogen Free Extract
60.2340%
Vitamin A
5.1892 klU/kg
Vitamin D3
0.6486 klU/kg
Vitamin E
87.2080 lU/kg
Vitamin K
0.9089 ppm
Thiamine
9.1681 ppm
Riboflavin
10.2290 ppm
Niacin
30.1147 ppm
Pantothenic Acid
19.1250 ppm
Choline
1019.8600 ppm
Pyridoxine
8.2302 ppm
Folacin
2.0476 ppm
Biotin
0.2038 ppm
Vitamin B12
23.4416 ppm
Feed obtained from and nutritional values provided by Zeigler Bros., Inc
Tables.xls (2-1_Feed)
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE 2-2. TYPICAL IN VIVO STUDY DESIGN
Dose
Dose
Exposure
Target Dose
Number of
Group
Material
Route
pg Pb/kg-day
Animals
1
None
Oral
--
2-5
2
Lead Acetate
Oral
25
5
3
75
5
4
225
5
5
Test Material 1
Oral
75
5
6
225
5
7
625
5
8
Test Material 2
Oral
75
5
9
225
5
10
625
5
11
Lead Acetate
Intravenous
100
5-8
Tables.xls (2-2_Design)
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE 2-3. DESCRIPTION OF PHASE II TEST MATERIALS
Experiment
Sample Designation
Site
Sample Description
Lead
Concentration
(DDm)1
2
Bingham Creek Residential
Kennecott NPL Site, Salt Lake City,
Utah
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.
1,590
Bingham Creek Channel Soil
Kennecott NPL Site, Salt Lake City,
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.
6,330
3
Jasper County High Lead Smelter
Jasper County, Missouri Superfund
Site
Soil composite collected from an on-site location.
10,800
Jasper County Low Lead Yard
Jasper County, Missouri Superfund
Site
Soil composite collected from an on-site location.
4,050
4
Murray Smelter Slag
Murray Smelter Superfund Site,
Murray City, Utah
Composite of samples collected from areas where exposed slag
existed on site.
11,700
Jasper County High Lead Mill
Jasper County, Missouri Superfund
Site
Soil composite collected from an on-site location.
6,940
5
Aspen Berm
Smuggler Mountain NPL Site, Aspen,
Colorado
Composite of samples collected from the Racquet Club property
(including a parking lot and a vacant lot).
14,200
Aspen Residential
Smuggler Mountain NPL Site, Aspen,
Colorado
Composite of samples collected from residential properties within
the study area.
3,870
6
Midvale Slag
Midvale Slag NPL Site, Midvale, Utah
Composite of samples collected from a water-quenched slag pile
in Midvale Slag Operable Unit 2.
8,170
Butte Soil
Silver Bow Creek/Butte Area NPL
Site, Butte, Montana
Soil composite collected from waste rock dumps in Butte Priority
Soils Operable Unit (BPSOU).
8,530
7
California Gulch Phase I Residential
Soil
California Gulch NPL Site, Leadville,
Colorado
Soil composite collected from residential properties within
Leadville.
7,510
California Gulch Fe/Mn PbO
California Gulch NPL Site, Leadville,
Colorado
Soil composite collected from near the Lake Fork Trailer Park
located southwest of Leadville near the Arkansas River.
4,320
8
California Gulch AV Slag
California Gulch NPL Site, Leadville,
Colorado
Sample collected from a water-quenched slag pile on the property
of the former Arkansas Valley (AV) Smelter, located just west of
Leadville.
10,600
9
Palmerton Location 2
New Jersey Zinc NPL Site,
Palmerton, Pennsylvania
Soil composite collected from on-site.
3,230
Palmerton Location 4
New Jersey Zinc NPL Site,
Palmerton, Pennsylvania
Soil composite collected from on-site.
2,150
11
Murray Smelter Soil
Murray Smelter Superfund Site,
Murray City, Utah
Soil composite collected from on-site.
3,200
NIST Paint
-
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 100 um in size.
8,350
12
Galena-enriched Soil
-
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.
11,200
California Gulch Oregon Gulch Tailing!
California Gulch NPL Site, Leadville,
Colorado
A composite of tailings samples collected from the Oregon Gulch
tailings impoundment.
1,270
1 Samples were analyzed for lead by inductively coupled plasma-atomic emission spectrometry (ICP-AES) in accord with USEPA Method 200.7
Tables.xls (2-3_TMs)
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE 2-4. RELATIVE LEAD MASS OF MINERAL PHASES OBSERVED IN TEST MATERIALS
Experiment:
2
3
4
5
6
7
8
9
11
12
Phase
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
Cal. Gulch
Phase I
Residential
Soil
Cal. Gulch
Fe/Mn PbO
Cal. Gulch
AV Slag
Palmerton
Location 2
Palmerton
Location 4
Murray
Smelter Soil
NIST Paint
Galena-
enriched
Soil
Cal. Gulch
Oregon
Gulch
Tailings
Anglesite
28%
1%
0.5%
1.0%
2%
7%
1%
36%
10%
2%
6%
4%
1%
As(M)0
0.003%
Calcite
0.2%
0.1%
Cerussite
2%
0.3%
32%
81%
1.1%
57%
62%
64%
4%
0.3%
20%
1%
14%
55%
Clay
0.018%
0.003%
0.017%
0.1%
0.1%
0.01%
0.03%
0.13%
Fe-Pb Oxide
6%
3%
14%
2%
2%
10%
9%
7%
0.3%
7%
6%
8%
51%
2%
2%
0.13%
Fe-Pb Sulfate
22%
30%
3%
1%
0.3%
1%
5%
5%
0.1%
20%
6%
3%
0.3%
1%
0.6%
Galena
9%
8%
9%
3%
12%
17%
6%
12%
2%
3%
20%
100%
100%
Lead Barite
0.04%
0.01%
0.06%
0.007%
0.15%
0.14%
1%
0.1%
Lead Organic
0.3%
0.03%
0.03%
0.11%
0.11%
1%
Lead Oxide
0.09%
69%
7%
27%
44%
Lead Phosphate
50%
26%
21%
6%
7%
1%
1%
3.6%
30%
15%
24%
1%
Lead Silicate
0.04%
0.5%
1.9%
0.8%
1.4%
Lead Vanidate
0.1%
0.4%
18%
Mn-Pb Oxide
18%
2%
2%
2%
0.8%
9%
4%
5%
20.2%
22%
72%
66%
66%
Native Lead
22%
0.7%
2%
15%
Pb(M)0
4%
26%
7%
3%
Pb-As Oxide
2%
1%
0.15%
6%
33%
0.1%
31%
29%
PbO-Cemssite
1%
Slag
4%
7%
1%
16%
1%
10%
6%
Sulfosalts
0.4%
Zn-Pb Silicate
0.03%
2%
Tbl 2-4_Lead Phases.xls (Table 2-4)
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE 2-5. MATRIX ASSOCIATIONS FOR TEST MATERIALS
Experiment
Test Material
Particle Frequency
Relative Lead Mass
Liberated
Included
Liberated
Included
2
Bingham Creek Residential
100%
0%
100%
0%
Bingham Creek Channel Soil
100%
0%
100%
0%
3
Jasper County High Lead Smelter
81%
19%
76%
24%
Jasper County Low Lead Yard
100%
0%
94%
6%
4
Murray Smelter Slag
87%
13%
77%
23%
Jasper County High Lead Mill
96%
4%
93%
7%
5
Aspen Berm
86%
14%
93%
8%
Aspen Residential
98%
2%
94%
6%
6
Midvale Slag
91%
9%
77%
23%
Butte Soil
91%
9%
91%
9%
7
California Gulch Phase I Residential Soil
79%
21%
65%
35%
California Gulch Fe/Mn PbO
98%
2%
100%
0%
8
California Gulch AV Slag
78%
22%
80%
20%
g
Palmerton Location 2
100%
0%
100%
0%
Palmerton Location 4
79%
21%
89%
11%
11
Murray Smelter Soil
80%
20%
70%
30%
NIST Paint
100%
0%
100%
0%
12
Galena-enriched Soil
100%
0%
100%
0%
California Gulch Oregon Gulch Tailings
2%
98%
5%
95%
Tables.xls (2-5_Matrix)
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE 2-6. PARTICLE SIZE DISTRIBUTIONS FOR TEST MATERIALS
Experiment
Test Material
Particle Size (|jm)
<5
5-9
10-19
20-49
50-99
100-149
150-199
200-249
>250
2
Bingham Creek Residential
38%
22%
19%
16%
4%
2%
0%
0%
0%
Bingham Creek Channel Soil
66%
13.6%
10%
6.1%
3%
1%
0%
0%
0%
3
Jasper County High Lead Smelter
44%
19%
8%
8%
9%
9%
2%
1%
1%
Jasper County Low Lead Yard
29%
20%
21%
20%
8%
3%
0%
0%
0%
4
Murray Smelter Slag
14%
13%
15%
6%
20%
24%
4%
3%
0%
Jasper County High Lead Mill
23%
21%
22%
19%
9%
6%
1%
1%
0%
5
Aspen Berm
27%
19%
22%
17%
8%
6%
1%
1%
0%
Aspen Residential
38%
35%
12%
8%
4%
2%
0%
0%
0%
6
Midvale Slag
6%
1%
3%
4%
20%
29%
18%
13%
5%
Butte Soil
23%
15%
14%
23%
14%
9%
2%
1%
0%
7
California Gulch Phase I Residential Soil
24%
9%
18%
22%
15%
9%
1%
1%
1%
California Gulch Fe/Mn PbO
26%
19%
24%
17%
10%
4%
0%
0%
0%
8
California Gulch AV Slag
19%
8%
8%
5%
9%
19%
10%
13%
9%
9
Palmerton Location 2
26%
23%
25%
18%
6%
1%
0%
0%
0%
Palmerton Location 4
25%
15%
21%
25%
13%
2%
0%
0%
0%
11
Murray Smelter Soil
23%
10%
29%
17%
6%
8%
3%
3%
1%
NIST Paint
76%
4%
6%
8%
6%
0%
0%
0%
0%
12
Galena-enriched Soil
48%
2%
4%
41%
4%
0%
0%
0%
0%
California Gulch Oregon Gulch Tailings
85%
8%
6%
0%
0%
0%
0%
0%
0%
Tables.xls (2-6_Size)
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE 2-7. ESTIMATED RBA VALUES FOR TEST MATERIALS
Experiment
Test Material
Blood AUC
Liver
Kidney
Femur
Point Estimate
RBA
LB
UB
RBA
LB
UB
RBA
LB
UB
RBA
LB
UB
RBA
LB
UB
2
Bingham Creek Residential
0.34
0.23
0.50
0.28
0.20
0.39
0.22
0.15
0.31
0.24
0.19
0.29
0.27
0.17
0.40
Bingham Creek Channel Soil
0.30
0.20
0.45
0.24
0.17
0.34
0.27
0.19
0.37
0.26
0.21
0.31
0.27
0.19
0.36
3
Jasper County High Lead Smelter
0.65
0.47
0.89
0.56
0.42
0.75
0.58
0.43
0.79
0.65
0.52
0.82
0.61
0.43
0.79
Jasper County Low Lead Yard
0.94
0.66
1.30
1.00
0.75
1.34
0.91
0.68
1.24
0.75
0.60
0.95
0.90
0.63
1.20
4
Murray Smelter Slag
0.47
0.33
0.67
0.51
0.33
0.88
0.31
0.22
0.46
0.31
0.23
0.41
0.40
0.23
0.64
Jasper County High Lead Mill
0.84
0.58
1.21
0.86
0.54
1.47
0.70
0.50
1.02
0.89
0.69
1.18
0.82
0.51
1.14
5
Aspen Berm
0.69
0.54
0.87
0.87
0.58
1.39
0.73
0.46
1.26
0.67
0.51
0.89
0.74
0.48
1.08
Aspen Residential
0.72
0.56
0.91
0.77
0.50
1.21
0.78
0.49
1.33
0.73
0.56
0.97
0.75
0.50
1.04
6
Midvale Slag
0.21
0.15
0.31
0.13
0.09
0.17
0.12
0.08
0.18
0.11
0.06
0.18
0.14
0.07
0.24
Butte Soil
0.19
0.14
0.29
0.13
0.09
0.19
0.15
0.09
0.22
0.10
0.04
0.19
0.14
0.06
0.23
7
California Gulch Phase I Residential
Soil
0.88
0.62
1.34
0.75
0.53
1.12
0.73
0.50
1.12
0.53
0.33
0.93
0.72
0.38
1.07
California Gulch Fe/Mn PbO
1.16
0.83
1.76
0.99
0.69
1.46
1.25
0.88
1.91
0.80
0.51
1.40
1.05
0.57
1.56
8
California Gulch AV Slag
0.26
0.19
0.36
0.19
0.11
0.32
0.14
0.08
0.25
0.20
0.13
0.30
0.20
0.09
0.31
9
Palmerton Location 2
0.82
0.61
1.05
0.60
0.41
0.91
0.51
0.30
0.91
0.47
0.37
0.60
0.60
0.34
0.93
Palmerton Location 4
0.62
0.47
0.80
0.53
0.37
0.79
0.41
0.25
0.72
0.40
0.32
0.52
0.49
0.29
0.72
11
Murray Smelter Soil
0.70
0.54
0.89
0.58
0.42
0.80
0.36
0.25
0.52
0.39
0.31
0.49
0.51
0.29
0.79
NIST Paint
0.86
0.66
1.09
0.73
0.52
1.03
0.55
0.38
0.78
0.74
0.59
0.93
0.72
0.44
0.98
12
Galena-enriched Soil
0.01
0.00
0.02
0.02
0.00
0.04
0.01
0.00
0.02
0.01
-0.01
0.03
0.01
0.00
0.03
California Gulch Oregon Gulch
Tailings
0.07
0.04
0.13
0.11
0.04
0.21
0.05
0.02
0.09
0.01
-0.04
0.06
0.06
-0.01
0.15
LB = 5% Lower Confidence Bound
UB = 95% Upper Confidence Bound
Tables.xls (2-7_RBA)
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE 2-8. GROUPED LEAD PHASES
Group
Group Name
Phase Constituents
1
Galena
Galena (PbS)
2
Cerussite
Cerussite
3
Mn(M) Oxide
Mn-Pb Oxide
4
Lead Oxide
Lead Oxide
5
Fe(M) Oxide
Fe-Pb Oxide (including Fe-Pb Silicate)
Zn-Pb Silicate
6
Lead Phosphate
Lead Phosphate
7
Anglesite
Anglesite
8
Pb(M) Oxide
As(M)0
Lead Silicate
Lead Vanidate
Pb(M)0
Pb-As Oxide
9
Fe(M) Sulfate
Fe-Pb Sulfate
Sulfosalts
10
Minor Constituents
Calcite
Clay
Lead Barite
Lead Organic
Native Lead
PbO-Cerussite
Slag
Tables.xls (2-8_Phases)
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE 2-9. CURVE FITTING PARAMETERS FOR ORAL LEAD ACETATE DOSE-RESPONSE CURVES
Experiment
Blood AUC
Liver Lead
Kidney Lead
Bone Lead
a
b
c
a
b
a
b
a
b
2
13.6
116
0.0084
63
2.0
44
2.4
0.7
0.084
3
8.3
163
0.0040
10
2.3
10
2.2
1.8
0.062
4
8.5
144
0.0064
57
1.7
68
2.8
0.5
0.076
5
8.0
163
0.0038
62
2.0
60
1.8
0.5
0.062
6
8.4
85
0.0101
23
2.0
15
2.1
0.4
0.043
7
a
a
a
10
1.7
10
1.4
0.8
0.059
8
8.0
159
0.0032
11
2.1
17
2.4
0.8
0.065
g
7.5
96
0.0087
11
2.3
14
2.3
0.6
0.071
11
7.2
160
0.0035
14
1.3
20
1.7
0.7
0.053
12
7.6
169
0.0040
9
0.7
8
1.1
0.6
0.032
Mean
8.6
140
0.0058
27
1.8
27
2.0
0.7
0.061
Standard Deviation
1.9
32
0.0026
24
0.5
22
0.5
0.4
0.015
Coefficient of Variation
23%
23%
46%
88%
27%
84%
26%
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)
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE 2-10. REPRODUCIBILITY OF RBA MEASUREMENTS
RBA
Palmerton
Location 2
California Gulch
Phase I Residential Soil
Estimate
Test 1
(Phase 2 Study 9)
Test 2
(Phase 2 Study 12)
Test 1*
(Phase 1 Studv 2)
Test 2
(Phase 2 Study 7)
Blood AUC
0.82 ± 0.12
0.71 ± 0.09
0.69
0.88 ± 0.19
Liver
0.60 ± 0.14
1.25 ± 0.32
0.58
0.75 ± 0.16
Kidney
0.51 ± 0.16
0.54 ± 0.13
0.62
0.73 ± 0.17
Bone
0.47 ± 0.07
0.95 ± 0.18
0.50
0.53 ± 0.15
Point Estimate
0.60 ± 0.18
0.86 ± 0.33
0.60
0.72 ± 0.21
"Calculated using ordinary least squares.
Tables.xls (2-10_Reprod)
-------
DRAFT- Do Not Cite, Quote, or Release
TABLE 3-1. IN VITRO BIOACCESSIBILITY VALUES
Experiment
Sample
In Vitro Bioaccessibility (%)
(Mean ± Standard Deviation)
2
Bingham Creek Residential
47.0 ± 1.2
2
Bingham Creek Channel Soil
37.8 ± 0.7
3
Jasper County High Lead Smelter
69.3 ± 5.5
3
Jasper County Low Lead Yard
79.0 ± 5.6
4
Murray Smelter Slag
65.5 ± 7.5
4
Jasper County High Lead Mill
80.4 ± 4.2
5
Aspen Berm
64.9 ± 1.6
5
Aspen Residential
71.4 ± 1.9
6
Midvale Slag
17.9 ± 1.0
6
Butte Soil
22.1 ± 0.6
7
California Gulch Phase I Residential Soil
65.1 ± 1.5
7
California Gulch Fe/Mn PbO
87.2 ± 0.5
8
California Gulch AV Slag
9.4 ± 1.6
9
Palmerton Location 2
63.6 ± 0.4
9
Palmerton Location 4
69.7 ± 2.7
11
Murray Smelter Soil
74.7 ± 6.8
11
NIST Paint
72.5 ± 2.0
12
Galena-enriched Soil
4.5 ± 1.2
12
California Gulch Oregon Gulch Tailings
11.2 ±0.9
Tbl 3-1, ES-2 IVBA Data.xls (Table 3-1)
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
OSWER 9285.7-77
FIGURES
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 2-1. AVERAGE RATE OF BODY WEIGHT GAIN IN TEST ANIMALS
0.7
0.6 --
CO
¦O
0.5 -
CB
0
•*-»
-C
O)
"
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 2-2. EXAMPLE TIME COURSE OF BLOOD LEAD RESPONSE
10
5
o>
TS
CO
0)
—I
"CS
o
o
CO
~ Control
~ PbAc (75)
PbAc (225)
-x—Test Material (75)
-*—Test Material (225)
•—Test Material (675)
4 6
Study Day
Fig 2-2 to 2-6 Dose Resp (outliers excl).xls
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 2-3. DOSE RESPONSE CURVE FOR BLOOD LEAD AUC
Panel A: Lead Acetate - IV
250
>»
(0
X!
200
100 150 200 250
Lead Dose (|jg Pb/kg-day)
300
350
Panel B: Lead Acetate - Oral
Lead Dose (pg Pb/kg-day)
Fig 2-2 to 2-6 Dose Resp (outliers excl).xls (Fig 2-3_AUC)
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 2-4. DOSE RESPONSE CURVE FOR LIVER LEAD
CONCENTRATION
Pane! A: Lead Acetate - IV
Lead Dose ((jg Pb/kg-day)
1200
Panel B: Lead Acetate - Oral
1000 -
o
o
800
.EP
I
-i—•
I
o>
o> 600 -
"O
CO
200
%
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)
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 2-5. DOSE RESPONSE CURVE FOR KIDNEY LEAD
CONCENTRATION
Pane! A: Lead Acetate - IV
Lead Dose ((jg Pb/kg-day)
Fig 2-2 to 2-6 Dose Resp (outliers excl).xls (Fig 2-5_Kidney)
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 2-6. DOSE RESPONSE CURVE FOR FEMUR LEAD
CONCENTRATION
70
60 -
O)
'
-------
siissruao
9P!*0 (SAl)uiAI
apjXQ pean
3)Bl|dS0Md PB91
"O
CD
TJ
C
0
C
e
P
CO
c
o
(/)
C"
o
o
o
m
>
•e
"3
(0
LL
r_
~
~
Q.
3
SP!X0 (iAl)9d |
o
3)!S3|5uV
— BU9|B9
aiBiins (|Al)3d
aP!*0 (lAl)qd
vay oypads-dnojg pajeunjsg
S-
CN
O)
in
X
CO
<
OQ
or
0
w
CS
k
eg
o>
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 2-8. CORRELATION OF DUPLICATE ANALYSES
Panel A: Blood Lead
30
25
5
o>
(D
ZJ
20 --
To 15 +
>
CO
E
10
1—
o
Line of Identity
Observed
0.9645x + 0.0687
0.981
10 15 20
Duplicate Value (pg/dL)
30
Panel B: Tissue Lead
Duplicate Value (pg/dL)
Figures.xls (2-8_Dups)
-------
DRAFT-- Do Not Cite, Quote, or Release
FIGURE 2-9. RESULTS FOR CDCP BLOOD LEAD CHECK SAMPLES
T3
cn
u
03
0)
—I
13
O
o
m
c
so
03
18
16
14 --
12
10
8 --
2 -
-5
~ ~
H
5
Study Day
10
Low Std
~ Med Std
i High Std
15
Figures.xls (2-9_CDCP)
-------
DRAFT-- Do Not Cite, Quote, or Release
FIGURE 2-10. INTERLABORATORY COMPARISON OF BLOOD LEAD RESULTS
5
O)
3
w
CD
DC
U
m
CO
—I
u
o
o
IB
<
a.
ai
co
D
30
25 --
20
15
10
Line of Identity
y = 0.8361 x + 0,7385
R2 = 0.8606
10 15 20
CDCP Blood Lead Results (jjg/dL)
25
30
Figures.xls (2-10_lnterlab)
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 3-1. IN VITRO BIOACCESSIBILITY EXTRACTION APPARATUS
Circulating
Heater Plexiglass Tank
(Set at 37° C)
P
125 ml Nalgene wide mouth bottles
~OdduduD
n
J
Magnetic Flywheel
Gearbox & motor
(28 RPM)
Figures.xls (3-1_Apparatus)
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 3-2. EFFECT OF TEMPERATURE, TIME, AND pH ON IVBA
Panel A: Effect of Temperature
100 "J
90 -
80 -
70 -
60 -
40 -
30 -
20 -
10 -
0 -
~ 37 C
~ 20 C
a
fa
aft
i i —"-r— 1 r— r i ¦ —l~~r— r——n—1 i —i ¦ ¦ ¦ r— r i ¦ ¦
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Test Material
Panel B: Effect of Extraction Time
120 -i
100 -
80 -
g
< 60 -
m
>
40 -
20 -
0 -
~ 1 hour
~ 2 hours
~ 4 hours
J
affi
12 3 4
5 6 7
Test Material
8 9 10 11
Panel C: Effect of pH
100 -r
90 -
80 -
70 -
60 -
0s*
<
50 -
m
>
40 -
30 -
20 -
10 -
0 -
L
U
H
I
Hh.
JlLp,
WCO^lfi(DS(OO)O^-CMCO^lO0SflOO)OT-CN|P3^
t— CMCMCMCMCM
Test Material
Fig 3-2_Effects on IVBA.xls (Fig 3-2_Effects)
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 3-3. PRECISION OF IN VITRO BIOACCESSIBILITY MEASUREMENTS
100
90
80
70
2 60
<
m
50
|
| 40 +
30
20
10
0
MS 2710
Mean = 75.5
Std. Dev. = 4.7
CV = 0.062
N = 68
MS 2711
Mean = 84.4
Std. Dev. = 4.7
CV= 0.055
N = 66
Figures.xls (3-3_Precision)
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 3-4. REPRODUCIBILITY OF IN VITRO BIOACCESSIBILITY MEASUREMENTS
k
c
ra
100
90 -
80 -
70 -
o 60 -
6
s.
< 50
CD
40 -
30 -
20 -
10 ¦
0
IACZ DCUB ¦ ERCL CINERl
J
IB
5
5
5
9 10 11
Test Material
12 13 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 Smetter
9 = Jasper County Low Lead Yard
10 = California Guich AV Slag
11 = California Gulch Fe/Mn PbO
12 = California Gulch Oregon Gulch Tailings
13 = California Gulch Phase I Residential Soil
Laboratories
14 = Midvale Slag
15 = Murray Smelter Slag
16 = Murray Smelter Soil
17 = Palmerton Location 2
18 = Palmerton Location 4
19 = NIST Paint
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)
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 3-5. RBAvs. IVBA
1.2 -
1.1 -
1.0 -
0.9 -
0.8 -
0.7 -
i °-6-
0.5 -
0.4 -
0.3 -
0.2 -
0.1 -
0.0 -
0.0 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)
-------
DRAFT- Do Not Cite, Quote, or Release
FIGURE 3-6. PREDICTION INTERVAL FOR RBA BASED ON MEASURED IVBA
IVBA
Measured
IVBA
Predicted RBA
Best Est.
s(y-hat)
5% P!
95% PI
0.00
-0.057
0.144
-0.31
0.19
0.05
-0.005
0.141
-0.25
0.24
0.10
0.047
0.139
-0.20
0.29
0.15
0.098
0.138
-0.14
0.34
0.20
0.150
0.136
-0.09
0.39
0.25
0.202
0.135
-0.03
0.44
0.30
0.254
0.133
0.02
0.49
0.35
0.305
0.132
0.07
0.54
0.40
0.357
0.132
0.13
0.59
0.45
0.409
0.131
0.18
0.64
0.50
0.460
0.131
0.23
0.69
0.55
0.512
0.131
0.28
0.74
0.60
0.564
0.131
0.34
0.79
0.65
0.616
0.132
0.39
0.84
0.70
0.667
0.132
0.44
0.90
0.75
0.719
0.133
0.49
0.95
0.80
0.771
0.134
0.54
1.00
0.85
0.823
0.136
0.59
1.06
0.90
0.874
0.137
0.64
1.11
0.95
0.926
0.139
0.68
1.17
1.00
0.978
0.141
0.73
0.00
Fig 3-6, D-8_Prediction lntervals.xls (Fig 3-6)
-------
DRAFT- Do Not Cite, Quote, or Release
OSWER 9285.7-77
APPENDIX A
EVALUATION OF JUVENILE SWINE AS A MODEL
FOR GASTROINTESTINAL ABSORPTION
IN YOUNG CHILDREN
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX A
EVALUATION OF JUVENILE SWINE AS A MODEL
FOR GASTROINTESTINAL ABSORPTION IN YOUNG CHILDREN
1.0 INTRODUCTION
Ideally, the reliability of an animal model as a predictor for toxicokinetic responses in humans
would be based on a direct comparison of results in humans and the animal species under
consideration. However, because intentional dosing of children with lead is not feasible, a direct
comparison of lead absorption results in swine with that for children is not possible.
Nevertheless, the relevance of the swine as an animal model for lead absorption can be evaluated
by comparing a number of physiological attributes of the gastrointestinal system that are likely to
be important in influencing the degree to which lead in ingested soil material is released from its
soil or mineral matrix to form soluble compounds that can be absorbed into the body. Factors
that may affect dissolution include gastric acidity and gastric holding time, which determine the
exposure of the ingested material to the acidic environment of the stomach, where dissolution
initially occurs. Morphological and physiological factors in the small intestine, where
absorption of lead is thought to occur, may also affect RBA; however, these are likely to be less
important for those soil materials for which solubility is the limiting factor for RBA.
Weis and LaVelle (1991) and Casteel et al. (1996) determined that gastric function in juvenile
swine is sufficiently similar to that of human children so that juvenile swine could serve as a
model for predicting RBA of soil-borne lead in children. This view is supported by several
reviews on the comparative anatomy and physiology of the human and pig gastrointestinal
systems (Dodds, 1982; Miller and Ullrey, 1987; Moughan et al., 1992; Pond and Houpt, 1978),
and in particular, the following pertinent observations.
2.0 GASTROINTESTINAL TRACT MORPHOLOGY AND HISTOLOGY
The anatomy of the neonatal digestive system in the pig and human are very similar (Moughan et
al., 1992). The body-weight adjusted ratios of intestinal length to stomach volume in the child
and piglet are comparable, as shown below:
A-l
-------
draft-
do Not Cite, Quote, or Release
APPENDIX A
Species
Stomach
Volume
(cm3/kg)
Small Intestine
Length (cm/kg)
Large Intestine
length (cm/kg)
Small intestine
length/stomach
volume
Large intestine
length/stomach
volume
Human
9.6
95.6
19.4
9.96
4.93
Swine
28.9
229.2
59.6
7.93
3.85
Source: Moughan et al.. 1992.
Birth body weights of 3.4 (human) and 1.3 (pig) kg were assumed.
The histology of the small intestine, colon, and rectum in the piglet is similar to that of the
human (Moughan et al., 1992). Small anatomical differences between humans and swine would
not be expected to markedly affect digestion in the neonate (Moughan et al., 1992). The piglet is
considered to be a useful model of the anatomical development of the human neonatal digestive
tract (Moughan et al., 1992; Miller and Ullrey, 1987).
3.0 GASTRIC HOLDING TIMES
Gastric emptying time in humans is highly variable (USEPA, 2001). The rate of emptying of
stomach contents varies depending on the type of food, the volume of the meal, and its caloric
content. High caloric substances such as fat empty more slowly than carbohydrates. The most
important factor effecting liquid gastric emptying is the caloric content of the liquid meal.
Upright positioning and ambulation have been described to speed gastric emptying. Other
factors that are believed to affect gastric emptying include the osmolality, acidity, and chain
length of fatty acids in the meal. Differences in emptying may also exist between males and
females. These factors tend to make direct comparisons of data from different reports difficult.
Nevertheless, the available data do not suggest any substantial differences in gastric holding
times between children and juvenile swine.
In the 4-week old pig, gastric emptying following a meal was rapid, with 30 to 40% passing into
the duodenum within 15 minutes and the remaining portion of gastric contents following about 1
hour later (Pond and Houpt, 1978). Gastric pH did not affect gastric emptying time in juvenile
swine (Pond and Houpt, 1978). In an unpublished study by Casteel (personal communication),
gastric emptying in juvenile swine was shown to be influenced by feeding intervals, both pre-
and post-dosing. The investigators reported rapid clearance of the bolus (complete within 2
A-2
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX A
hours) after an overnight fast; however, feeding 4 hours prior to dosing slowed completion of
gastric emptying to 4 hours. Feeding at two hours post-dosing accelerated the movement of the
residual gastric contents, although most of the bolus had already cleared the stomach.
In humans, gastric emptying time in neonates and premature infants is typically about 87
minutes, but can be as long as 6 to 8 hours, with adult values (typically about 65 minutes) being
reached at 6 to 8 months of age (FDA, 1998; Balis, 2000).
4.0 GASTRIC ACIDITY
Direct comparisons of gastric acidity as a function of age in humans and swine are not available.
However, available information on gastric acid secretion does not suggest there are any major
differences that would affect extrapolation of RBAs measured in juvenile swine to humans.
Agunod et al. (1969) reported that gastric acid output (corrected for body weight) reached
normal adult levels in swine at 2 to 3 months post partum. In humans, gastric pH is neutral at
birth, but drops to 1 to 3 within hours of birth. Gastric acid secretion then declines on days 10 to
30, and does not approach adult values until approximately 3 months of age (FDA, 1998).
Nagita et al. (1996) reported that the intragastric pH of infants was <4 for only half of the day,
whereas baseline pH in normal adults is <2. The development of maximal acid secretion in the
pig also has some similarities to that of humans (Xu and Cranwell, 1990). In both the pig and
human, maximal acid secretion correlates with age and body weight with pentagastrin,
histamine, and histalog used as secretagogues (Xu and Cranwell, 1990). A limitation of the
available pig data is that all of the studies measure the maturation of gastric acid output rather
than intragastric pH, which Nagita et al. (1996) asserts is a preferable measure of gastric
maturity. Temporal studies of the intragastric pH of juvenile swine are not available.
A-3
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX A
5.0 REFERENCES
Agunod, M., N. Yamaguchi, R. Lopez, A. L. Luhby, and G. B. J. Glass. 1969. Correlative study
of hydrochloric acid, pepsin and intrinsic factor secretion in newborns and infants. Am. J. Dig.
Dis. 14:400-414.
Balis, F. 2000. Module 4: Drug therapy in special populations. In: Principles of Clinical
Pharmacology. Published on-line by NIH: http://www.cc.nih.gov/ccc/principles/.
Casteel, S. W., R. P. Cowart, C. P. Weis, G. M. Henningsen, E. Hoffman, W. J. Brattin, M. F.
Starost, J. T. Payne, S. L. Stockham, S. V. Becker, and J. R. Turk. 1996. A swine model for
determining the bioavailability of lead from contaminated media. In: Advances in Swine in
Biomedical Research. Tumbleson and Schook, eds. Vol 2, Plenum Press, New York. Pp.
637-46.
Cranwell, P. D. 1985. The development of acid and pepsin secretory capacity in the pig: the
effect of age and weaning. 1. Studies in anaesthetized pigs. Br. J. Nutr. 54: 305-20.
Dodds, J. W. 1982. The pig model for biomedical research. Fed. Proc. 41: 247-56.
FDA. 1998. Guidance for Industry: General considerations for pediatric pharmacokinetic
studies for drugs and biological products. U.S. Department of Health and Human Services.
Food and Drug Administration. Center for Drug Evaluation and Research (CDER). Center for
Biologies Evaluation and Research (CBER). Published on-line by the Food and Drug
Administration, November 1998: http://www.fda.gov/cber/guidelines.htm.
Henning, S. J. 1987. Functional development of the gastrointestinal tract. In: Physiology of the
Gastrointestinal Tract. Johnson, L.R., ed. Raven Press, New York. Pp. 285-300.
Miller, E. R., and D. E. Ullrey. 1987. The pig as a model for human nutrition. Ann. Rev. Nutr.
7: 361-82.
A-4
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX A
Moughan, P. J., M. J. Birtles, P. D. Cranwell, W. C. Smith, and M. Pedraza. 1992. The piglet as
a model animal for studying aspects of digestion and absorption in milk-fed human infants.
World Rev Nutr Diet. 1992;67:40-113. PMID: 1557912
Nagita, A., K. Amemoto, A. Yoden, S. Aoki, M. Sakaguchi, K. Ashida, and M. Mino. 1996.
Diurnal variation in intragastric pH in children with and without peptic ulcers. Ped. Res. 40(4):
528-32.
Pond, W. G., and K. A. Houpt. 1978. The Biology of the Pig. Ithaca, NY: Comstock. 371 pp.
Sangild, P. T., P. D. Cranwell, and L. Hilsted. 1992. Ontogeny of gastric function in the pig:
Acid secretion and the synthesis and secretion of gastrin. Bio. Neonate. 62: 363-72.
USEPA. 2001. Exploration of perinatal pharmacokinetic issues. Final report by Versar to
USEPA Risk Assessment Forum, Office of Research and Development. Washington, D.C.
20460. EPA/630/R-01/004.
Waldum, H. L., B. K. Straume, P. G. Burhol, and L. B. Dahl. 1980. Serum group I pepsinogens
in children. Acta Paediatr Scand. 69(2):215-8. PMID: 7368925
Weis, C.P., and LaVelle, J.M. 1991. Characteristics to consider when choosing an animal
model for the study of lead bioavailability. In: Proceedings of the International Symposium on
the Bioavailability and Dietary Uptake of Lead. Sci. Technol. Let. 3:113-119.
WHO. 1986. Principles for evaluating health risks from chemicals during infancy and early
childhood: The need for a special approach. Environmental Health Criteria 59. Geneva,
Switzerland: World Health Organization.
Xu, R. J., and P. D. Cranwell. 1990. Development of gastric acid secretion in pigs from birth to
36 days of age: the response to pentagastrin. J. Dev. Physiol. 13: 315-26.
Xu, C. D., P. Wang, and J. Y. Xu. 1996. Gastric motility study on non ulcer dyspepsia among
children. Chin Natl J New Gastroenterol. 2(Suppl 1): 129.
A-5
-------
DRAFT- Do Not Cite, Quote, or Release
OSWER 9285.7-77
APPENDIX B
DETAILED DESCRIPTION OF ANIMAL EXPOSURE
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B
DETAILED DESCRIPTION OF ANIMAL EXPOSURE
1.0 EXPERIMENTAL ANIMALS
All animals used in this program were young intact males of the Pig Improvement Corporation
(PIC) genetically defined Line 26, and were purchased from Chinn Farms, Clarence, MO. The
number of animals purchased for each study was typically 6 to 8 more than required by the
protocol. These animals were usually purchased at age 4 to 5 weeks (weaning occurs at age 3
weeks), and they were then held under quarantine for one week to observe their health before
beginning exposure to test materials. Any animals which appeared to be in poor health during
this quarantine period were excluded. To minimize weight variations between animals and
groups, extra animals that were most different in body weight on day -4 (either heavier or
lighter) were also excluded from the study. The remaining animals were assigned to dose groups
at random. When exposure began (day zero), the animals were about 5 to 6 weeks old and
weighed an average of about 8 to 11 kg.
All animals were housed in individual lead-free stainless steel cages. Each animal was examined
by a certified veterinary clinician (swine specialist) prior to being placed on study, and all
animals were examined daily by an attending veterinarian while on study. Blood samples were
collected for clinical chemistry and hematological analysis on days -4, 7, and 15 to assist in
clinical health assessments. Any animal that became ill and could not be promptly restored to
good health by appropriate treatment was promptly removed from the study.
2.0 DIET
Animals provided by the supplier were weaned onto standard pig chow purchased from MFA
Inc., Columbia, MO. In order to minimize lead exposure from the diet, the animals were
gradually transitioned from the MFA feed to a special low-lead feed (guaranteed less than 0.2
ppm lead, purchased from Zeigler Brothers, Inc., Gardners, PA) over the time interval from day
-7 to -3, and this feed was then maintained for the duration of the study. The feed was
nutritionally complete and met all requirements of the National Institutes of Health-National
Research Council. The typical nutritional components and chemical analysis of the feed are
B-l
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B
presented in Table 2-1 of the main text. Periodic analysis of feed samples during this program
indicated the mean lead level was less than the detection limit (0.05 ppm).
Each day every animal was given an amount of feed equal to 5% of the mean body weight of all
animals on study. Feed was administered in two equal portions of 2.5% of the mean body
weight at each feeding. Feed was provided at 11:00 AM and 5:00 PM daily. 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 |ig/L.
3.0 DOSING
The dose levels used in these studies were selected to be as low as possible in an effort to make
measurements at the low end of the dose-response curve where saturation of biological systems
is minimal. Based on experience from previous investigations, doses of lead acetate in the range
of 25 to 675 |ig Pb/kg-day were found to give clear and measurable increases in lead levels in all
endpoints measured (blood, liver, kidney, bone), so doses in this range (usually 25 to 225 |j,g
Pb/kg-day) were employed in most studies. The doses of test materials were usually set at the
same level as lead acetate, except that one higher dose was often included in case the test
materials were found to yield very low responses. Depending on the concentration of lead in the
test material and the target dose level for lead, soil intake rates by the swine were in the range of
500 to 2500 mg/day.
Animals were exposed to lead acetate or a test material for 15 days, with the dose for each day
being administered in two equal portions given at 9:00 AM and 3:00 PM (two hours before
feeding). These exposure times were selected so that lead ingestion would occur at a time when
the stomach was largely or entirely empty of food. This is because the presence of food in the
stomach is known to reduce lead absorption (e.g., Chamberlain et al., 1978; Rabinowitz et al.,
1980; Heard and Chamberlain, 1982; Blake et al., 1983; James et al., 1985). Dose calculations
were based on measured group mean body weights and were adjusted every three days to
account for animal growth.
B-2
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B
For animals exposed by the oral route, dose material was placed in the center of a small portion
(about 5 grams) of moistened feed. This "doughball" was administered to the animals by hand.
Most animals consumed the dose promptly, but occasionally some animals delayed ingestion of
the dose for up to two hours (the time the daily feed portion was provided). Random and
intermittent delays of this sort are not considered to be a significant source of error.
Occasionally, some animals did not consume some or all of the dose (usually because the dose
dropped from their mouth while chewing). All missed doses were recorded and the time-
weighted average dose calculation for each animal was adjusted downward accordingly.
For animals exposed by intravenous injection, doses were given via a vascular access port (VAP)
attached to an indwelling venous catheter that had been surgically implanted according to
standard operating procedures by a board-certified veterinary surgeon through the external
jugular vein to the cranial vena cava about 3 to 5 days before exposure began.
4.0 REFERENCES
Blake, K. H. C., G. O. Barbezat, and M. Mann. 1983. Effect of dietary constituents on the
gastrointestinal absorption of 203Pb in man. Environ. Res. 30:182-187.
Chamberlain, A. C., M. J. Heard, P. Little, D. Newton, A. C. Wells, and R. D. Wiffen. 1978.
Investigations into lead from motor vehicles. Harwell, UK: United Kingdom Atomic Energy
Authority. Report No. AERE-9198.
Heard, H. J., and A. C. Chamberlain. 1982. Effect of minerals and food on uptake of lead from
the gastrointestinal tract in humans. Human Toxicol. 1:411-415.
James, H. M., M. E. Hilburn, and J. A. Blair. 1985. Effects of metals and meal times on uptake
of lead from the gastrointestinal tract in humans. Human Toxicol. 4:401-407.
Rabinowitz, M. B., J. D. Kopple, and G. W. Wetherill. 1980. Effect of food intake and fasting
on gastrointestinal lead absorption in humans. Am. J. Clin. Nutrit. 33:1784-1788.
B-3
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B
ATTACHMENT 1
DETAILED STUDY DESIGNS
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B - ATTACHMENT 1
EXPERIMENT 1A STUDY DESIGN
Dose
Pig Number
Group
Material Administered*
(|jg Pb/kg-day)
3
1
Control
0
20
2
2
PbAc
25
22
23
24
27
1
3
PbAc
75
26
29
32
35
9
4
PbAc (-2 hr)
225
14
17
31
34
7
5
PbAc (0 hr)
225
12
19
30
33
5
6
PbAc (+2 hr)
225
18
21
25
36
4
7A
PbAc (IV)
100
15
16
*AII materials administered orally unless designated IV (Intravenously)
Appendix B_Tables.xls (1a)
Page 1 of 11
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B - ATTACHMENT 1
EXPERIMENT 2 STUDY DESIGN
Pig Number
Group
Material Administered*
Dose
(|jg Pb/kg-day)
206
1
Control
0
226
215
2
PbAc
25
220
222
229
251
209
3
PbAc
75
228
244
248
258
204
4
PbAc
225
216
247
252
260
201
5
Bingham Creek
75
207
Residential
221
238
259
236
6
Bingham Creek
225
237
Residential
240
242
249
224
7
Bingham Creek
450
234
Residential
235
243
257
202
8
Bingham Creek
75
217
Channel Soil
219
253
254
203
9
Bingham Creek
225
225
Channel Soil
227
232
250
205
10
Bingham Creek
675
210
Channel Soil
213
218
255
208
11
PbAc (IV)
100
214
230
231
239
241
246
256
*AII materials administered orally unless designated IV (Intravenously)
Appendix B_Tables.xls (2) Page 2 of 11
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B - ATTACHMENT 1
EXPERIMENT 3 STUDY DESIGN
Dose
Pig Number
Group
Material Administered*
(|jg Pb/kg-day)
304
1
Control
0
339
309
2
PbAc
75
312
324
337
340
313
3
PbAc
225
315
342
354
356
305
4
Jasper County
75
311
High Lead Smelter
318
321
331
316
5
Jasper County
225
317
High Lead Smelter
330
352
353
319
6
Jasper County
625
341
High Lead Smelter
344
345
348
325
7
Jasper County
75
329
Low Lead Yard
338
343
351
302
8
Jasper County
225
326
Low Lead Yard
328
332
346
306
9
Jasper County
625
333
Low Lead Yard
334
335
349
307
10
PbAc (IV)
100
320
322
347
350
*AII materials administered orally unless designated IV (intravenously)
Appendix B_Tables.xls (3) Page 3 Of 11
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B - ATTACHMENT 1
EXPERIMENT 4 STUDY DESIGN
Dose
Pig Number
Group
Material Administered*
(|jg Pb/kg-day)
417
1
Control
0
430
409
2
PbAc
75
419
429
443
444
408
3
PbAc
225
410
426
449
455
402
4
Murray Smelter
75
407
Slag
411
423
450
420
5
Murray Smelter
225
431
Slag
432
440
446
412
6
Murray Smelter
625
418
Slag
427
437
442
404
7
Jasper County
75
406
High Lead Mill
416
428
454
401
8
Jasper County
225
433
High Lead Mill
434
435
441
403
9
Jasper County
625
405
High Lead Mill
413
448
453
415
10
PbAc (IV)
100
421
424
425
438
439
445
451
*AII materials administered orally unless designated IV (intravenously)
Appendix B_Tables.xls (4) Page 4 of 11
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B - ATTACHMENT 1
EXPERIMENT 5 STUDY DESIGN
Dose
Pig Number
Group
Material Administered*
(|jg Pb/kg-day)
530
1
Control
0
536
514
2
PbAc
75
518
519
520
524
501
3
PbAc
225
513
529
534
547
503
4
Aspen Berm
75
523
532
549
555
509
5
Aspen Berm
225
512
539
540
550
510
6
Aspen Berm
675
516
525
537
542
502
7
Aspen Residential
75
507
517
522
528
505
8
Aspen Residential
225
506
521
553
554
526
9
Aspen Residential
675
535
541
545
548
504
10
PbAc (IV)
100
508
515
538
543
544
546
551
*AII materials administered orally unless designated IV (intravenously)
Appendix B_Tables.xls (5)
Page 5 of 11
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B - ATTACHMENT 1
EXPERIMENT 6 STUDY DESIGN
Pig Number
Group
Material Administered*
Dose
(|jg Pb/kg-day)
614
1
Control
0
638
613
2
PbAc
75
624
630
639
641
616
3
PbAc
225
644
651
653
654
619
4
Mid vale Slag
75
623
626
631
647
602
5
Midvale Slag
225
605
628
640
650
603
6
Midvale Slag
675
615
629
633
645
610
7
Butte Soil
75
611
617
637
643
601
8
Butte Soil
225
609
618
621
635
620
9
Butte Soil
675
627
634
646
655
604
10
PbAc (IV)
100
606
607
612
625
632
642
648
*AII materials administered orally unless designated IV (intravenously)
Appendix B_Tables.xls (6) Page 6 Of 11
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B - ATTACHMENT 1
EXPERIMENT 7 STUDY DESIGN
Pig Number
Group
Material Administered*
Dose
(|jg Pb/kg-day)
706
1
Control
0
714
718
735
743
703
2
PbAc
25
709
748
750
755
711
3
PbAc
75
715
716
747
752
704
4
California Gulch
25
712
Phase I Residential Soil
736
740
753
702
5
California Gulch
75
708
Phase I Residential Soil
728
739
756
717
6
California Gulch
225
723
Phase I Residential Soil
725
732
737
707
7
California Gulch
25
713
Fe/Mn PbO
730
738
741
733
8
California Gulch
75
742
Fe/Mn PbO
746
749
751
719
9
California Gulch
225
721
Fe/Mn PbO
729
744
745
722
10
PbAc (IV)
100
724
727
734
754
*AII materials administered orally unless designated IV (intravenously)
Appendix B_Tables.xls (7) Page 7 of 11
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B - ATTACHMENT 1
EXPERIMENT 8 STUDY DESIGN
Dose
Pig Number
Group
Material Administered*
(|jg Pb/kg-day)
808
1
PbAc (IV)
0
810
836
805
2
PbAc (IV)
807
812
25
827
834
813
3
PbAc (IV)
815
825
50
845
853
801
4
PbAc (IV)
816
820
100
843
852
809
5
Control
830
841
0
848
855
817
6
PbAc
818
819
25
838
846
804
7
PbAc
840
842
75
844
849
857
8
California Gulch
826
AV Slag
828
25
831
851
806
9
California Gulch
814
AV Slag
823
75
847
854
811
10
California Gulch
822
AV Slag
824
225
837
856
*AII materials administered orally unless designated IV (intravenously)
Appendix B_Tables.xls (8) Page 8 Of 11
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B - ATTACHMENT 1
EXPERIMENT 9 STUDY DESIGN
Pig Number
Group
Material Administered*
Dose
(|jg Pb/kg-day)
907
1
PbAc (IV)
100
912
919
930
942
943
953
901
2
Control
0
902
920
925
928
905
3
PbAc
25
909
927
931
940
923
4
PbAc
75
933
948
950
956
911
5
Palmerton
25
929
Location 2
934
947
954
903
6
Palmerton
75
910
Location 2
938
951
955
906
7
Palmerton
225
908
Location 2
916
918
922
913
8
Palmerton
25
914
Location 4
932
937
946
924
9
Palmerton
75
926
Location 4
944
949
957
917
10
Palmerton
225
921
Location 4
939
941
945
*AII materials administered orally unless designated IV (intravenously)
Appendix B_Tables.xls (9) Page 9 of 11
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B - ATTACHMENT 1
EXPERIMENT 11 STUDY DESIGN
Dose
Pig Number
Group
Material Administered*
(|jg Pb/kg-day)
1109
1
Control
0
1124
1135
1139
1151
1103
2
PbAc
25
1104
1116
1117
1118
1105
3
PbAc
75
1123
1129
1130
1144
1121
4
PbAc
225
1136
1138
1146
1150
1106
5
Murray Smelter
75
1112
Soil
1133
1142
1149
1102
6
Murray Smelter
225
1122
Soil
1128
1143
1154
1126
7
Murray Smelter
675
1137
Soil
1140
1141
1155
1110
8
NIST Paint
75
1115
1134
1148
1153
1101
9
NIST Paint
225
1108
1111
1132
1152
1113
10
NIST Paint
675
1119
1120
1125
1147
*AII materials administered orally
Appendix B_Tables.xls (11) Page 10 Of 11
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX B - ATTACHMENT 1
EXPERIMENT 12 STUDY DESIGN
Pig Number
Group
Material Administered*
Dose
(|jg Pb/kg-day)
1205
1
Control
0
1228
1236
1208
2
PbAc
25
1213
1215
1217
1248
1227
3
PbAc
75
1240
1243
1244
1255
1222
4
PbAc
225
1225
1226
1241
1249
1201
5
Galena-enriched Soil
75
1233
1250
1251
1253
1203
6
Galena-enriched Soil
225
1209
1214
1231
1247
1218
7
Galena-enriched Soil
675
1229
1235
1237
1254
1207
8
Palmerton
25
1223
1230
Location 2 (reproducibility)
1245
1252
1202
9
Palmerton
75
1210
1212
Location 2 (reproducibility)
1220
1232
1211
10
Palmerton
225
1216
1221
Location 2 (reproducibility)
1239
1246
1204
11
California Gulch
225
1224
1238
Oregon Gulch Tailings
1242
*AII materials administered orally
Appendix B_Tables.xls (12)
Page 11 of 11
-------
DRAFT- Do Not Cite, Quote, or Release
OSWER 9285.7-77
APPENDIX C
DETAILED METHODS OF
SAMPLE COLLECTION AND ANALYSIS
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX C
DETAILED METHOD OF SAMPLE COLLECTION AND ANALYSIS
1.0 COLLECTION OF BIOLOGICAL SAMPLES
Blood
Samples of blood were collected from each animal three or four days before exposure began, on
the first day of exposure (day 0), and on multiple days thereafter (usually days 1, 2, 3, 5, 7, 9, 12,
and 15). All blood samples were collected by vena-puncture of the anterior vena cava, and
samples were immediately placed in purple-top Vacutainer® tubes containing EDTA
(ethylenediaminetetra-acetic acid) as anticoagulant. Blood samples were collected each
sampling day beginning at 8:00 AM, approximately one hour before the first of the two daily
exposures to lead on the sampling day and 17 hours after the last lead exposure the previous day.
This blood collection time was selected because the rate of change in blood lead resulting from
the preceding exposures is expected to be relatively small after this interval (LaVelle et al., 1991;
Weis et al., 1993), so the exact timing of sample collection relative to last dosing is not likely to
be critical.
Liver. Kidney, and Bone
Following collection of the final blood sample at 8:00 AM on day 15, all animals were humanely
euthanized and samples of liver, kidney, and bone (the right femur) were removed and stored in
lead-free plastic bags for lead analysis.
Samples of all biological samples collected were archived in order to allow for reanalysis and
verification of lead levels, if needed, and possibly for future analysis for other metals (e.g.,
arsenic, cadmium). All animals were also subjected to detailed examination at necropsy by a
certified veterinary pathologist in order to assess overall animal health.
C-l
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX C
2.0 PREPARATION OF BIOLOGICAL SAMPLES FOR ANALYSIS
Blood
One mL of whole blood was removed from the purple-top Vacutainer and added to 9.0 mL of
"matrix modifier," a solution recommended by the Centers for Disease Control and Prevention
(CDCP) for analysis of blood samples for lead. The composition of matrix modifier is 0.2%
(v/v) ultrapure nitric acid, 0.5% (v/v) Triton X-100, and 0.2% (w/v) dibasic ammonium
phosphate in deionized and ultrafiltered water. Samples of the matrix modifier were routinely
analyzed for lead to ensure the absence of lead contamination.
Liver and Kidney
One gram of soft tissue (liver or kidney) was placed in a lead-free screw-cap Teflon container
with 2 mL of concentrated (70%) nitric acid and heated in an oven to 90°C overnight. After
cooling, the digestate was transferred to a clean, lead-free 10 mL volumetric flask and diluted to
volume with deionized and ultrafiltered water.
Bone
The right femur of each animal was removed, defleshed, and dried at 100°C overnight. The
dried bones were then broken in half, placed in a muffle furnace and dry-ashed at 450°C for 48
hours. Following dry ashing, the bone was ground to a fine powder using a lead-free mortar and
pestle, and 200 mg was removed and dissolved in 10.0 mL of 1:1 (v:v) concentrated nitric
acid/water. After the powdered bone was dissolved and mixed, 1.0 mL of the acid solution was
removed and diluted to 10.0 mL by addition of 0.1% (w/v) lanthanum oxide (La^) in deionized
and ultrafiltered water.
3.0 LEAD ANALYSIS
Samples of biological tissue (blood, liver, kidney, bone) and other materials (food, water,
reagents and solutions, etc.) were arranged in a random sequence and provided to USEPA's
analytical laboratory in a blind fashion (identified to the laboratory only by a chain of custody
C-2
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX C
tag number). Each sample was analyzed for lead using a Perkin Elmer Model 5100 graphite
furnace atomic absorption spectrophotometer. Internal quality assurance samples were run every
tenth sample, and the instrument was recalibrated every 15th sample. A blank, duplicate, and
spiked sample were run every 20th sample. In addition, a series of quality assurance (QA)
samples were prepared and submitted to the laboratory in bland fashion, including a variety of
duplicates, blanks, and standards.
All results from the analytical laboratory were reported in units of |ig Pb/L of prepared sample.
The quantitation 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 ng/L). The standard deviation was usually
about 0.3 ng/L, so the quantitation limit was usually about 0.9 to 1.0 ng/L (ppb). However,
because different dilution factors were used for different sample types, the detection limit varies
from sample type to sample type. For prepared blood samples (diluted 1/10), this corresponds to
a quantitation limit of 10 |ig/L (1 jxg/dL). For soft tissues (liver and kidney, also diluted 1/10),
this corresponds to a quantitation limit of 10 ng/kg (ppb) wet weight, and for bone (final dilution
of 1/500) the corresponding quantitation limit is 0.5 |ig/g (ppm) ashed weight.
4.0 REFERENCES
LaVelle, J. M., R. H. Poppenga, B. J. Thacker, J. P. Giesy, C. Weis, R. Othoudt, and C.
Vandervoot. 1991. Bioavailability of lead in mining waste: An oral intubation study in young
swine. In: The Proceedings of the International Symposium on the Bioavailability and Dietary
Uptake of Lead. Science and Technology Letters 3:105-111.
Weis, C. P., G. M. Henningsen, R. H. Poppenga, and B. J. Thacker. 1993. Pharmacokinetics of
lead in blood of immature swine following acute oral and intravenous exposure. The
Toxicologist 13(1): 175.
C-3
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
OSWER 9285.7-77
APPENDIX D
DETAILED METHODS FOR DATA REDUCTION
AND STATISTICAL ANALYSIS
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
DETAILED METHODS FOR DATA REDUCTION
AND STATISTICAL ANALYSIS
1.0 INTRODUCTION
The 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 principal that equal absorbed doses of lead will
produce equal biological responses. By definition:
Absorbed dose (ref) = Administered dose (ref) ¦ ABA (ref)
Absorbed dose (test) = Administered dose (test) • ABA (test)
When the responses are equal, then:
Admin, dose (ref) • ABA (ref) = Admin, dose (test) • ABA (test)
Thus:
RBA = ABA(test) / ABA(ref) = Admin. Dose (ref) / Admin. Dose (test)
That is, given the dose-response curve for some particular endpoint (e.g., the concentration of
lead in blood or tissue) for both the reference material and the test material, RBA may be
calculated as the ratio of administered doses that produce equal biological responses.
Note that, in this approach, the mathematical form of the dose-response model must be the same
for both reference material and test material. This is because the shape of the dose-response
curve is a function only of the pharmacokinetic response of the biological organism to an
absorbed dose of lead, and the response per unit dose absorbed dose does not depend on the
whether the absorbed lead was derived from reference material or test material. Another way to
envision this is to recognize that, if the unit of exposure were absorbed dose (rather than
administered dose), the dose-response curves for reference material and test material would be
identical.
D-l
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
Based on this, the general procedure for estimating the value of RBA from measured dose-
response data for reference and test materials is as follows:
1. Plot the biological responses of individual animals exposed to a series of oral
doses of reference material. Select an exposure-response model which can fit
smoothly through the observed data points. The model may be either linear or
non-linear, depending on the response endpoint being used.
2. Plot the biological responses of individual animals exposed to a series of doses of
test material. Fit the same exposure-response model as was used for the reference
material. Note that the intercept term must be the same for both curves, but that
other coefficients may be different.
3. To find the ratio of doses that produce equal responses, set the two exposure
response curves equal to each other and solve for the ratio of doses expressed in
terms of the model parameters.
For example, assume that the increase in lead in femur (PbF) is observed to be a linear function
of administered dose. Assume that the best-fit exposure-response models derived from the
experimental data for animals exposed to reference material and test material are as follows:
PbF(ref) = 2 + 6Dose(ref)
PbF(test) = 2 + 3Dose(test)
Setting the two equations equal yields:
2 + 6Dose(ref) = 2 + 3Dose(test)
Solving yields:
Dose(ref) / Dose(test) = 3/6 = 0.5
That is, the ratio of administered doses that produce equal responses is 0.5, so the RBA is 0.5
(50%).
D-2
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
An important assumption used in this approach is that administration of increasing doses of test
material will cause increased biological responses. However, this may not occur in the case of a
test material in which the form of lead has very low solubility. For example, the solubility of
lead sulfide (galena) in water is less than 1 |ig/L. Thus, if a dose of lead sulfide results in
saturation of the gastric fluid, administration of more lead sulfide will not increase the
concentration of bioavailable lead and, hence, little or no increase in response would be
expected. An example of this is shown in Figure D-l. In this case, RBA cannot be defined as
the ratio of doses that produce equal responses, since many different doses of lead sulfide all
produce the same response. However, this is not a substantial difficulty, since the amount of
lead that becomes bioavailable will be small (and hence the response will be close to control),
and simple inspection of the data will demonstrate that the test material is not likely to be of
health concern.
2.0 MEASUREMENT ENDPOINTS
Four independent measurement endpoints were evaluated in each study, based on the
concentration of lead observed in blood, liver, kidney, and bone (femur). For liver, kidney, and
bone, the measurement endpoint was simply the concentration in the tissue at the time of
sacrifice (day 15). For blood, the measurement endpoint used to quantify response was the area
under the curve (AUC) for blood lead vs. time (days 0-15). The area under the blood lead vs.
time curve for each animal was calculated by finding the area under the curve for each time step
(i.e., the interval between successive blood collection days) using the trapezoidal rule:
AUC(ditod^ = 0.5-(ri-+xJ)-(dl-di)
where:
d = day number, where i and j are successive blood sampling events
r = response (blood lead value) on day i (rf) or day j (rj)
The areas of the trapezoids for each time step were then summed to yield the final AUC for each
animal.
D-3
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
Occasionally blood lead values were obtained that were clearly different than expected. A value
was considered to be an outlier if it was clearly different from other values within the same dose
group on the same day, and/or if the value was clearly different from the time trend established
by preceding and following time points in the same animal. A total of 21 such cases occurred
out of a total of 4,284 blood lead data points (0.5%). These values were excluded in the
calculation of AUC, and the missing value was replaced by a value interpolated from the
preceding and following values from the same animal.
3.0 RESPONSES BELOW QUANTITATION LIMIT
In some cases, most or all of the responses in a group of animals were below the quantitation
limit for the endpoint being measured. For example, this was normally the case for blood lead
values in unexposed animals (both on day -4 and day 0 and in control animals), and also
occurred during the early days in the study for animals given test materials with low
bioavailability. In these cases, all animals which yielded responses below the quantitation limit
were evaluated as if they had responded at one-half the quantitation limit. This approach was
used because an assumed value of one-half the detection limit minimizes the potential bias in the
assumption.
4.0 DERIVATION OF STATISTICAL DOSE-RESPONSE MODELS
The techniques used to derive statistical models of the dose-response data and to estimate RBA
are based on the methods recommended by Finney (1978). All model fitting was performed
using JMP® version 3.2.2, a commercial software package developed by SAS®. Details are
provided below.
4.1 Use of Simultaneous Regression
As noted by Finney (1978), when the data to be analyzed consist of two dose-response curves
(the reference material and the test material), it is obvious that both curves must have the same
intercept, since there is no difference between the curves when the dose is zero. This
requirement is achieved by combining the two dose response equations into one and solving for
D-4
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
the parameters simultaneously. For example, if the dose response model is linear, the approach
is as follows:
Separate Models:
Mr(i) = a + b/x/i)
M-t(i) = a + btxt(i)
Combined Model
M-(i) = a + b/x/i) + btXt(i)
where n(i) indicates the expected mean response of animals exposed at dose x(i), and the
subscripts r and t refer to reference and test material, respectively. The coefficients of this
combined model are derived using multivariate regression, with the understanding that the
combined data set is restricted to cases in which one (or both) of x,. and xt are zero (Finney,
1978). The same approach may be extended for use when there are three data sets (reference
material, test material 1, test material 2) that are all derived from a single study and must
therefore all have the same intercept.
4.2 Use of Weighted Regression
Regression analysis based on ordinary least squares assumes that the variance of the responses is
independent of the dose and/or the response (Draper and Smith, 1998). In these studies, this
assumption is generally not satisfied. Figure D-2 provides two example data sets that show a
clear increase in variability in response as a function of increasing dose. This is referred to as
heteroscedasticity. Most other data sets from this study display a similar tendency toward
increasing variance in response as a function of increasing dose.
One method for dealing with heteroscedasticity is through the use of weighted least squares
regression (Draper and Smith, 1998). In this approach, each observation in a group of animals is
assigned a weight that is inversely proportional to the variance of the response in that group:
1
w¦ = —
'
where:
D-5
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
Wj = weight assigned to all data points in dose group i
a2; = variance of responses in animals in dose group i
When the distributions of responses at each dose level are normal, weighted regression is
equivalent to the maximum likelihood method.
There are several options available for estimating the value of o1;.
Option 1: Utilize the observed variance (s2j) in the responses of animals in dose group i.
Option 2: Establish a variance model of the form a2i = amP, where m is the predicted mean
response for dose group i. Simultaneously fit the data to derive values of a and p
along with the other coefficients of the dose-response model using the data from a
particular study. This approach is identical to the non-constant variance approach
used by USEPA's BMDS (USEPA 1995, 2000a).
Option 3 A: Establish an "external" variance model based on an analysis of the relationship
between variance and mean response using observations combined from all
studies and dose groups. Use that model to predict the expected variance in dose
group i as a function of the predicted mean response for that dose group.
Option 3B: Establish an "external" variance model based on an analysis of the relationship
between variance and mean response using observations combined from all
studies and dose groups. Use that model to predict the expected variance in dose
group i as a function of the observed mean response level for that dose group.
In this study, all four options were investigated for possible use. The advantages and
disadvantages of each are discussed below.
Option 1 (use of group-specific sample variances) is the simplest approach, and does not
require any assumptions or extrapolations. If the number of animals in each dose group
were large enough to provide reliable estimates of the true variance for the dose group,
this would be the preferred method. However, sample variance in a dose group is a
random variable, and because the sample variance based on only five observations (five
D-6
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
animals per dose group) can vary widely (especially when true variance is large), weights
assigned using this approach may occasionally be substantially higher or lower than the
data actually warrant. For example, this approach yielded poor results in cases where
two adjacent groups (usually the control and the low dose group) had very low variance.
In this situation, the weights for those groups were so high that the model fit was
constrained to pass through them with very little deviation, and other dose groups exerted
very little influence. Figure D-3 shows an example of this. Because this outcome was
judged to be inappropriate, Option 1 was not used.
Option 2 (using a non-constant variance model derived from the within-study data only)
utilizes the entire data set from a single study to estimate expected variance as a function
of dose, and so is less vulnerable to random variations in group-specific sample variances
than Method 1. Despite this advantage, however, this approach requires that two
additional parameters (a and p) be derived along with the other model parameters. This
tends to over-parameterize the model, and when this option was tested (using the solver
feature of Excel®) the fits were often not stable (i.e., different results were obtained with
different starting guesses). On this basis, Option 2 was not employed.
Option 3 (both Option 3A and 3B) requires development of an external variance model
based on the consolidated data from all studies. Figure D-4 shows the log-variance in
response plotted as a function of the log-mean response in the group1. One panel is
presented for each of the four different endpoints. As seen, log-variance increases as an
approximately linear function of log-mean response for all four endpoints:
ln^2) = k\ + kl ¦ ln(j(.)
Values of kl and k2 are derived from the data for each endpoint using ordinary least
squares minimization, and the resulting values are shown in the figures. Note that this
variance model is of the same basic form as used in Option 2:
1 In this analysis, some dose groups were excluded if the estimate of variance and/or mean
response was judged to be unreliable, based on the following two criteria: a) the number of animals in the
dose group was <,2, or b) the fraction of responses below the detection limit was more than 20%. For the
blood lead AUC endpoint (where the raw data consist of multiple blood lead values as a function of time),
this corresponds to an AUC less than about 15 |j.g/dL-days.
D-7
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
sf=exp(kl)-(yi)k2
In Option 3 A, the weights for each response are assigned within the model based on the
predicted mean response at each dose level. For example, assuming a linear model:
fx (0 = « + V *1 (0 + b2 • X2 (0
a? = exp[£l + k2 • ln(X (0)]
In Option 3B, the same approach is used, except that the observed mean response rather
than the predicted mean response is used to estimate a2{:
af = exp[A:l + k2 ¦ In(y x (z))]
In testing both options, it was found that Option 3A and 3B gave similar results in most
cases. However, Option 3A (in which weights are not pre-assigned but are optimized
during the fitting procedure) tended to be very sensitive to starting guesses, often failing
to find solutions even when the starting guesses were good, and sometimes yielding
different results depending on the starting guesses. In addition, this approach uses the
expected mean response rather than the observed mean response to estimate the variance,
which tends to diminish the role of the measured data in defining the best fit curve. In
contrast, Option 3B was less prone to unstable solutions, and is based more directly on
the data.
Based on a consideration of the advantages and disadvantages of each approach, Option 3B was
selected for use in this project. This is mainly because it is has relatively less vulnerability than
Option A to random variations in observed variances in a dose group (which results is
assignment of weights that are either too high or too low), and also because it is could be
implemented with relatively few difficulties. It should be noted, however, that Option 3B is
somewhat vulnerable to poor fits when one particular dose group in a data set lies well below the
expected smooth fit through the other dose groups. In this case, the variance assigned to the
group (based on the observed mean response) is lower than typical for that dose level (and hence
D-8
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
the weights assigned to the data are higher than usual), tending to force the line through that data
set at the expense of the other data sets.
4.3 Choice of Model Forms
As noted above, the main objective of the curve-fitting effort is to find a mathematical model
that fits both the reference and test group dose-response data sets smoothly. Note that there is no
requirement that the model have a mechanistic basis or that the coefficients have a biological
meaning. As discussed by Finney (1978), it is generally not appropriate to choose the form of
the dose-response model based on only one experiment, but to make the choice based on the
weight of observations across many different studies. Because simple inspection of the data
suggest that, over the range of doses tested in these studies, some dose-response curves (mainly
those for liver, kidney, and bone) appear to be approximately linear, while others (mainly those
for blood lead AUC) appear to be nonlinear (tending to plateau as dose increases), the linear
model and three alternative non-linear models were evaluated:
1. Linear: y = a + b/Xj. + bt-xt
RB A = bt / br
2. Exponential: y = a + b(l-exp(-crxT)) + b(l-exp(-ctxt))
RB A = ct / cr
3. Michaelis-Menton: y = a + b ^ / (cr + xr) + b xt / (ct+ xt)
RB A = cr / ct
4. Power: y = a + br ¦ xTc + bt ¦ x,c
RBA = (bt / br)1/o
Appendix E presents the detailed results for every data set fit to each of the four different models
investigated. Goodness-of-fit was assessed using the F test statistic and the adjusted coefficient
of multiple determination (Adj R2), calculated as follows (Draper and Smith, 1998):
F - MSE(fit)/MSE{error)
Adj R2 — 1 - MSE (error)/MSE (total)
D-9
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
where:
MSE(flt) = £ w, ¦ (ft - y*)1 l(p- 1)
MSE {error) = ^ wi • (jut - yt )2 / (n- p)
MSE(total) = ^ ¦ (yt — y*)2 / (n — 1)
and:
y* = Y,(wi -y^/J^w,
p = number of parameters in model
n = number of observations (animals)
F is distributed as an F distribution with (p-1) and (n-p) degrees of freedom. Models with p
values larger than 0.05 were not considered to be acceptable. Of the models that were acceptable
(p < 0 .05), the preferred model was identified based on Akaike's Information Criterion (AIC)
(USEPA, 2000a and 2000b), which is calculated as:
AIC = -2L + 2p
where:
L = Log-likelihood function
p = number of parameters in the model
At the kth dose, the sample log-likelihood function is:
Lk=-(Nk/DlnilxaD-^jriy,. - f(xk)f
(Nelson, 1982). The overall log-likelihood is the sum across all dose groups (g):
L = ±L>
k=1
so that
k=1 k=1 J=i
D-10
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
The detailed results are presented in Appendix E, and the findings are summarized in Table D-l.
Inspection of this table reveals the following main conclusions:
• For liver, kidney, and bone, the linear model generally gave the best fit, although this
varied somewhat by endpoint (7/10 for kidney, 6/10 for bone, 4/10 for liver). In cases
where the linear model was not the best fit, the RBA value given by the linear model was
usually close to that given by whatever other model did provide the best fit, with an
average absolute difference of 12% (6% if one data set [study 9] was excluded). On this
basis, the linear model was selected for application to all dose-response data sets for
liver, kidney, and bone.
• For the blood lead AUC endpoint, the linear model usually gave the worst fit, and on this
basis it was rejected as a candidate for the AUC endpoint. In general, each of the three
nonlinear models (exponential, Michaelis-Menton, and power) all tended to give similar
results in terms of RBA value (the standard deviation in RBA for a particular test
material averaged across the three models was usually less than 3%), and differences in
the AIC were usually small. On this basis, it was concluded that any of these three
models would be acceptable. The power model was not selected because it does not tend
toward a plateau, while data from early blood lead pilot studies (using higher doses than
commonly used in the Phase II studies) suggest that the blood lead endpoint does tend to
do so. Of the remaining two models (exponential and Michaelis-Menton), the
exponential model was selected mainly because it yielded the best fit more often than the
Michaelis-Menton model (4 out of 10 vs. 2 out of 10), and because the exponential model
had been used in previous analyses of the data. Thus, the exponential model was selected
for application to all dose-response data sets for the blood AUC endpoint, except in one
special case noted below in section 4.5.
4.4 Assessment of Outliers
In biological assays, it is not uncommon to note the occurrence of individual measured responses
that appear atypical compared to the responses from other animals in the same dose group. For
the purposes of this program, endpoint responses that yielded standardized weighted residuals
greater than 3.5 or less than -3.5 were considered to be potential outliers (Canavos, 1984). When
such data points were encountered in a data set, the RBA was calculated both with and without
D-ll
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
the potential outlier(s) excluded, and the result with the outlier excluded was used as the
preferred estimate.
4.5 Treatment of Problematic Data Sets
Although the data reduction approach described above works well in most cases, a few data sets
yielded atypical results. In particular, fitting the blood lead data set from Experiment 7 proved
difficult. In this study, the blood lead AUC data set did not yield a solution in JMP for the
exponential model, even though solutions could be obtained in Excel using minimization of
weighted squared errors. However, the solutions tended to be unstable. This difficulty in
modeling the data appears to be due to the fact that the data have relatively less curvature than
most blood lead AUC data sets. Because of this lack of curvature, it is not possible to estimate
the exponential plateau value (b) with confidence, which in turns makes it difficult to estimate
the other parameters of the exponential model.
Several alternative solutions were evaluated, including a) using the model fits from one of the
other nonlinear models, b) using the fit for the linear model, and c) fitting the data to the
exponential model using a defined value for the plateau based on results from other data sets.
The results (i.e., the RBA values based on the blood lead AUC endpoint) were generally similar
for all three of these approaches:
Model
RBA of TM1
RBA of TM2
Power
0.65
0.83
Linear
0.69
0.90
Michaelis-Menton
0.69 ±0.01*
0.90 ±0.01*
Exponential fit
0.70 ± 0.02*
0.93 ±0.04*
Exponential fit (parameter b = 126.4)**
0.75
1.04
Exponential fit (parameter b = 169.1)***
0.74
1.01
~Solution was unstable; values represent the mean and standard deviation of five different fitting results.
**Parameter b set to the mean of the estimates obtained for all other blood AUC data sets using the
exponential model.
***Parameter b set to the maximum of the estimates obtained for all other blood AUC data sets using the
exponential model.
All estimates are based on all data (outlier not excluded).
D-12
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
Based on these results, it was concluded that the results from the linear fit were representative of
the range of values derived by other alternatives, so the JMP fit for the linear model was used for
this data set.
4.6 Characterization of Uncertainty Bounds
Each RBA value is calculated as the ratio of a model coefficient for the reference data set and for
the test data set:
RBA (linear endpoints) = bt / br
RBA(blood AUC) = ct / cr
However, there is uncertainty in the estimates of the model coefficients in both the numerator
and denominator and, hence, there is uncertainty in the ratio. As described by Finney (1978), the
fiduciary limits (uncertainly range) about the ratio R of two model coefficients may be calculated
using Fieller's Theorem:
covar t i—
R-g r±±—4w
var b
LB, UB = £ £
1 ~g
W = vart - 2 • R • covart r + R2 ¦ varr - g
covar
.2 ~\
var -
r,t
V
var
t2
g = TT^r
K
where:
R = ratio (bt / br for linear model, ct / cr for exponential model)
varr = variance in the coefficient for the reference material
covarr t = covariance in the coefficients for the reference and test materials
br = coefficient for the reference material (cr in the case of the exponential model)
t = t statistic for alpha (0.05) and (n-p) degrees of freedom
D-13
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
When g is small (<0.05), the variance of the ratio is approximated as (Finney 1978):
var -2 R-covar , + R2 • var
var(R) = —! —* 1
br
4.7 Combination of RBA Estimates Across Endpoints
As discussed above, each study of RBA utilized four different endpoints to estimate absorption
of lead, including blood AUC, liver, kidney, and bone. Consequently, each study yielded for
independent estimates of RBA for each test material. Thus, the final RBA estimate for a test
material involves combining the four end-point specific RBA values into a single value (point
estimate), and estimating the uncertainty around that point estimate. The methods used to
achieve these goals are described below.
Derivation of the Point Estimate
The basic strategy for deriving a point estimate of RBA for a test material is to calculate a
confidence-weighted average of the four endpoint-specific RBA values. If all four endpoints are
considered to be equally reliable, the weighting factors are all equal (i.e., the point estimate is the
simple average). If reliability is considered to differ from endpoint to endpoint, then weights are
assigned in proportion to the reliability:
RBA(point estimate) = S (RBA; • w;) / 2 (w;)
Because each endpoint-specific RBA value is calculated as the ratio of the parameters of the
dose-response curves fitted to the experimental data for reference material and test material, the
reliability of an endpoint-specific RBA is inherently related to the quality of the data that define
the dose-response curve for that endpoint. For endpoints that tend to have low within-group
variability and generate data that fit the dose-response model well, the uncertainty around the
model parameters will tend to be small and hence the uncertainty around the RBA value will also
tend to be small. Conversely, if the underlying dose-response data for an endpoint are highly
variable and the dose-response model does not fit the data well, there will tend to be high
uncertainty in the model parameters and hence in the RBA estimate. Thus, a good indicator of
D-14
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
relative reliability between the four different endpoints is the relative magnitude of the
uncertainty (standard error) around RBA estimates based on each endpoint.
Figure D-5 plots the standard error in each RBA estimate as a function of the RBA value for
each of the four different endpoints. As seen, uncertainty in RBA increases as a function of the
estimated value of RBA in all four cases. This is expected because of the heteroscedastisity in
the underlying dose-response data. Although RBA values based on blood AUC and femur tend
to yield estimates with slightly lower standard errors than RBA values based on liver or kidney,
the magnitude of the standard errors tends to be generally similar for all four endpoints, and the
difference between the four regression lines is not statistically significant (p = 0.699). Based on
this, each endpoint-specific RBA value was judged to have approximately equal validity, and the
point estimate was calculated as the simple average across all four endpoint-specific RBA
values.
Estimation of Uncertainty Bounds Around the Point Estimate
The uncertainty bounds around each point estimate were estimated using Monte Carlo
simulation. For each test material, values for RBA were drawn from the uncertainty
distributions for each endpoint with equal frequency. Each endpoint-specific uncertainty
distribution was assumed to be normal, with the mean equal to the best estimate of RBA and the
standard deviation estimated from Fieller's Theorem (see Section 4.6 above). The uncertainty in
the point estimate was characterized as the range from the 5th to the 95th percentile of the average
across endpoints.
5.0 RELATION BETWEEN RBA AND IVBA
Choice of Model Form
As discussed in Section 3.3.2, one of the important objectives of this program was to
characterize the degree to which measures of in vitro bioaccessibility (IVBA) correlate with in
vivo measurements of RBA. This was approached by plotting the point estimate of in vivo RBA
vs. the corresponding IVBA value for each of the 19 different test materials and fitting several
D-15
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
different mathematical models to the data. The results are shown in Figure D-6 (Panels A to D),
and are summarized below:
Model
R2
AIC
Linear (RBA = a + b-IVBA)
0.837
-72.75
Power (RBA = a + bTVBA0)
0.881
-75.35
2-Parameter Exponential (RBA = a + bexp(IVBA))
0.866
-73.16
3-Parameter Exponential (RBA = a + b exp(c-IVBA))
0.883
-75.74
As seen, all of the models fit the data reasonably well, with the non-linear models (power,
exponential) fitting somewhat better than the linear model. However, the improved fit of the
non-linear models is due mainly to the fact that the two data points that occur in the central part
of the x-range (IVBA = 0.38 and 0.47) lie below the best fit linear line, and these two data points
tend to pull the central part of the curve down slightly when a non-linear model is used. If these
two data points were absent, or if a third data point were present that were above the linear fit,
the quality of the fits would be approximately equal for linear and non-linear models. Based on
the judgement that two data points are not sufficient evidence to conclude that a non-linear fit is
preferable to a linear model, the linear model is selected as the interim recommended model. As
more data become available in the future, the relationship between IVBA and RBA will be
reassessed and the model will be revised if needed.
Effect of Measurement Errors in IVBA
The process of fitting a linear model to the data is complicated by the fact that there are random
measurement errors in both the IVBA and the in vivo RBA estimates. The general solution for
the maximum likelihood estimate of the slope (b) and the intercept (a) is:
S - AS + J(S -AS )2 + 4 AS2
i yy xx \ ^ yy xx J xy
b =
2 S
xy
a=Y-bX
D-16
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
where:
Sxx= Z
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
• In the third case, the assumption that X = Syy/S^ was used. The model based on this
assumption is referred to as the geometric mean functional relationship (GMFR). For
this data set, the value of A is 1.38.
The results are shown in Figure D-7 Panels A, B, and C, with Panel D presenting an overlay of
the three different fits. As seen, all three approaches yielded fits that were relatively close to
each other, with residuals that do not show any clear pattern (middle) and which were well
described by normal distributions (bottom). Based on this, the relationship based on simple
linear regression was selected as the interim preferred model:
RBA = 1.03TVBA - 0.06
Prediction Interval for RBA
The prediction interval around y (RBA) based on a specified value of x (IVBA) is (Sachs, 1984):
y ~ 9 + Ws t
y = Distribution of possible y values consistent with x
£ = Expected (average) value of y at x
tn_2 = Random variate from a t-distribution with n-2 degrees of freedom
s y = Standard deviation around f
The value of s^ is given by:
where:
where:
D-18
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
Qyx =Qy-b-Qxy
Qw = Z • y>) - ^ <2 xt )(Z y>-)
n
n
where n is the number of data points and b is the slope of the regression line. Based on these
equations and the best fit linear regression equation described above, the 90% prediction interval
(i.e., ranging from the 5th to the 95th percentile) is as shown in Figure D-8.
D-19
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX D
6.0 REFERENCES
Canavos, C. G. 1984. Applied Probability and Statistical Methods. Little, Brown and Co.,
Boston.
Draper, N. R., and H. Smith. 1998. Applied Regression Analysis (3rd Edition). John Wiley &
Sons, New York.
Finney, D.J. 1978. Statistical Method in Biological Assay (3rd Edition). Charles Griffin and
Co., London.
Nelson, W. 1982. Applied Life Data Analysis. John Wiley & Sons, New York.
Sachs, L. 1984. Applied Statistics: A Handbook of Techniques. Second Edition. Springer-
Verlag, New York, pages 443-444.
USEPA. 1995. The Use of the Benchmark Dose Approach in Health Risk Assessment. U.S.
Environmental Protection Agency, Office of Research and Development, Risk Assessment
Forum. EPA/630/R-94/007. February, 1995.
USEPA. 2000a. Help Manual for Benchmark Dose Software Version 1.20. U.S. Environmental
Protection Agency, Office of Research and Development. EPA 600/R-00/014F. February, 2000.
USEPA. 2000b. Benchmark Dose Technical Guidance Document. External Review Draft.
U.S. Environmental Protection Agency, Risk Assessment Forum. EPA/630/R-00/001. October,
2000.
D-20
-------
DRAFT- Do NofCft^'CRjote, or Release
TABLE D-1. MODEL COMPARISONS
Endpoint
Experiment
LINEAR
Adj R RBA1 RBA2 RBA3
EXPONENTIAL
Adj R
RBA1 RBA2 RBA3
MICHAELIS-MENTON
Adj R
RBA1 RBA2 RBA3
Adj R
RBA1 RBA2 RBA3
Blood AUC
Blood AUC
412.4014
428.5143
c 0.001
c 0.001
0.779
0.818
0.38
0.53
0.31
0.63
393.6549
377.8492
c 0.001
c 0.001
0.827
0.896
0.34
0.65
0.30
0.94
391.8262
376.0574
< 0.001
< 0.001
0.831
0.899
0.33
0.65
0.30
0.94
366.1163
374.4287
< 0.001
< 0.001
0.846
0.902
0.34
0.62
0.30
0.85
Blood AUC
Blood AUC
455.6739
385.03
c 0.001
c 0.001
0.787
0.864
0.34
0.50
0.46
0.55
382.9415
345.1702
c 0.001
c 0.001
0.896
0.933
0.47
0.69
0.84
0.72
379.6654
344.7351
< 0.001
< 0.001
0.901
0.934
0.47
0.68
0.84
0.73
374.2627
344.9323
< 0.001
< 0.001
0.909
0.934
0.40
0.61
0.73
0.68
Blood AUC
Blood AUC
333.5653
394.3537
c 0.001
c 0.001
0.820
0.692
0.28
0.69
0.30
0.90
311.8304
NS
c 0.001
NS
0.21
NS
312.3221
NS
< 0.001
NS
0.21
NS
0.19
NS
316.066
394.2826
< 0.001
< 0.001
0.875
0.689
0.24
0.65
0.23
0.83
Blood AUC
Blood AUC
377.1965
328.7634
c 0.001
c 0.001
0.822
0.862
0.26
0.62
337.9125
312.2196
c 0.001
c 0.001
0.898
0.909
0.26
0.82
336.9394
312.6794
< 0.001
< 0.001
0.900
0.908
0.26
0.80
344.3656
316.1967
< 0.001
< 0.001
0.885
0.899
0.20
0.73
Blood AUC
Blood AUC
11
12
436.4331
375.1354
c 0.001
c 0.001
0.857
0.906
0.49
0.01
0.60
0.76
390.4143
370.3602
c 0.001
c 0.001
0.922
0.910
0.70
0.01
0.86
0.71
391.3314
370.7599
< 0.001
< 0.001
0.921
0.910
0.70
0.01
0.86
0.72
402.3932
374.8385
< 0.001
< 0.001
0.905
0.905
0.66
0.01
0.83
0.74 0.07
Liver
Liver
543.2988
562.2981
c 0.001
c 0.001
0.567
0.762
0.35
0.56
0.25
1.20
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
543.0502
561.4696
< 0.001
< 0.001
0.574
0.786
0.39
0.60
0.26
1.08
Liver
Liver
558.5529
674.4086
< 0.001
0.003
0.564
0.268
0.51
0.93
0.86
1.13
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
555.8161
675.8196
< 0.001
0.007
0.586
0.249
0.39
0.87
0.74
1.02
Liver
Liver
468.3743
503.4616
c 0.001
c 0.001
0.622
0.679
0.13
0.54
0.14
0.71
470.3592
505.44
c 0.001
c 0.001
0.612
0.671
0.13
0.54
0.14
0.72
470.3592
NS
< 0.001
NS
0.612
NS
0.13
NS
0.14
NS
470.2987
505.3976
< 0.001
< 0.001
0.613
0.672
0.13
0.54
0.13
0.72
Liver
Liver
c 0.001
c 0.001
0.452
0.727
0.18
0.60
NS
470.6533
NS
c 0.001
NS
0.777
NS
1.11
NS
471.6336
NS
< 0.001
NS
0.774
NS
1.07
630.6132
475.7101
c 0.001
< 0.001
0.441
0.760
0.18
0.89
Liver
Liver
11
12
561.4436
506.975
c 0.001
c 0.001
0.757
0.716
0.58
0.02
0.73
1.25
561.5909
NS
c 0.001
NS
0.757
NS
0.66
NS
0.71
NS
561.5427
NS
< 0.001
NS
0.757
NS
0.65
NS
0.71
NS
< 0.001
< 0.001
0.759
0.746
0.63
0.02
0.73
0.98
Kidney
Kidney
530.2226
533.5966
c 0.001
c 0.001
0.667
0.834
0.22
0.58
0.27
0.91
NS
534.2703
NS
c 0.001
NS
0.833
NS
0.58
NS
0.97
NS
534.219
NS
< 0.001
NS
0.834
NS
0.58
NS
0.97
532.2104
534.0045
< 0.001
< 0.001
0.679
0.834
0.22
0.56
0.27
0.95
Kidney
Kidney
550.1067
547.8196
c 0.001
c 0.001
0.715
0.529
0.31
0.73
0.70
0.76
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
551.8207
548.0081
< 0.001
< 0.001
0.709
0.527
0.30
0.64
0.68
0.74
Kidney
Kidney
500.2596
501.5953
c 0.001
c 0.001
0.552
0.657
0.12
0.51
0.16
0.86
501.6143
NS
c 0.001
NS
0.543
NS
0.11
NS
0.14
NS
501.6373
NS
< 0.001
NS
0.543
NS
0.11
NS
0.14
NS
501.9909
503.5096
< 0.001
< 0.001
0.541
0.649
0.12
0.51
0.14
0.85
Kidney
Kidney
586.5547
535.8631
c 0.001
c 0.001
0.573
0.579
0.14
0.51
565.9632
511.6407
c 0.001
c 0.001
0.571
0.661
0.14
1.62
585.9527
513.5473
< 0.001
< 0.001
0.571
0.655
0.14
1.63
561.2902
518.7502
< 0.001
< 0.001
0.587
0.636
0.13
1.36
Kidney
Kidney
11
12
c 0.001
c 0.001
0.725
0.329
0.36
0.01
0.55
0.47
578.6496
870.0696
c 0.001
c 0.001
0.718
0.315
0.53
0.01
0.47
0.34
578.7016
870.32
< 0.001
< 0.001
0.717
0.315
0.48
0.01
0.48
0.36
578.2471
864.5181
< 0.001
< 0.001
0.720
0.326
0.39
0.01
0.52
0.73
Femur
Femur
180.5215
187.2204
c 0.001
c 0.001
0.24
0.65
0.26
0.75
NS
166.1916
NS
c 0.001
NS
0.870
NS
0.70
NS
0.81
NS
166.1445
NS
< 0.001
NS
0.870
NS
0.70
NS
0.81
182.5211
186.099
< 0.001
< 0.001
0.859
0.870
0.24
0.68
0.26
0.79
Femur
Femur
196.1178
221.1807
c 0.001
c 0.001
0.866
0.856
0.31
0.67
0.89
0.73
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
195.6032
222.5576
< 0.001
< 0.001
0.888
0.854
0.32
0.65
0.96
0.72
Femur
Femur
227.7994
216.3481
c 0.001
c 0.001
0.465
0.615
0.11
0.53
0.10
0.80
229.7051
216.5913
c 0.001
c 0.001
0.451
0.611
0.11
0.56
0.10
0.95
229.7112
NS
< 0.001
NS
0.451
NS
0.11
NS
0.10
NS
229.612
216.3737
< 0.001
< 0.001
0.451
0.612
0.11
0.56
0.11
0.93
Femur
Femur
193.9091
118.6208
c 0.001
c 0.001
0.830
0.855
0.20
0.47
195.1797
112.175
c 0.001
c 0.001
0.828
0.864
0.20
0.50
195.1037
111.9654
< 0.001
< 0.001
0.828
0.885
0.20
0.50
< 0.001
< 0.001
0.850
0.888
0.18
0.48
Femur
Femur
11
12
198.2084
137.1663
c 0.001
c 0.001
0.871
0.865
0.39
0.01
0.74
0.95
NS
139.1501
NS
c 0.001
NS
0.856
NS
0.01
NS
0.95
NS
139.1506
NS
< 0.001
NS
0.856
NS
0.01
NS
0.95 0.01
200.0236
139.1826
< 0.001
< 0.001
0.869
0.861
0.38
0.01
0.73
0.95 0.01
- = The respective test material does not exist for this study.
NS = No solution; the software could not find a solution, or the solution was unstable and/or had unrealistic parameter estimates.
NA = Not applicable; the preferred model has the best fit, or no solution was found for the preferred model.
Appendix D_Tables & Figures.xls (Tbl D-1_Models)
-------
DRAFT- Do N6PE5tel?l®i!ibte, or Release
FIGURE D-1. DOSE-RESPONSE CURVE FOR GALENA
Experiment 12 Blood AUC: Test Material 1 (Galena-enriched soil)
Dose (|jg Pb/kg-day)
Appendix D_Tables & Figures.xls (Fig D-1_Saturation)
-------
DRAFT- Do N6PEfftel?l®i!ibte, or Release
FIGURE D-2. EXAMPLES OF HETEROSCEDASTICITY
Experiment 9 Blood AUC: Lead Acetate
Experiment 7 Liver: Test Material 2
Appendix D_Tables & Figures.xls (Fig D-2_Heteroscedasticity)
-------
DRAFT- Do N6PE5tel?l®i!ibte, or Release
FIGURE D-3. EXAMPLE OF POOR FIT DUE TO LOW VARIANCE
IN SOME DOSE GROUPS
Option 1, Linear Fit: Experiment 12 Liver, Lead Acetate
Appendix D_Tables & Figures.xls (Fig D-3_Poor Fit)
-------
DRAFT-- Do Ndt^EE&P'SQote, or Release
FIGURE D-4. VARIANCE MODELS
All Phase II Lead Studies. Data Quality Exclusion Rules Enforced.
9
8 -
7
_ 6 -
fl>
0
c
1 «¦
CL
3
e
o
'c'
~ 3-
2 -
1 -
0 -
• Control
~ Lead Acetate
A Test Material
BLOOD AUC
y= 1.5516x-1.3226
Rz = 0.5046
p-value < 0.01
3.5 4.0 4.5
ln(Group Mean Response)
18
16-
14 -
12 -
10 -
2 8-
O
• Control
~ Lead Acetate
A Test Material
6 -
4 -
2 •
0 -
LIVER
A A A
a&a a
y = 2.0999x-2.6015
R2 = 0.7966
p-value < 0.01
4 5 6
ln(Group Mean Response)
16 -
• Control
~ Lead Acetate
A Test Material
g 10-
^ 8-
S
<3 0
4 -
2
KIDNEY
y= 1.9557x- 1.8499
R2 = 0.7035
p-value < 0.01
4 5 6
ln(Group Mean Response)
9 Control
~ Lead Acetate
A Test Material
A
A
FEMUR
6
5 -
4 -
3 -
2 ¦
1 -
0 ¦
-1 -
-2
-3
-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
ln(Group Mean Response)
y= 1.656x-1.9713
0.7022
p-value < 0.01
Appendix D_Tables & Figures.xls (Fig D-4_Variarice)
-------
DRAFT- Do NdPeSWPQBote, or Release
FIGURE D-5. EVALUATION OF RELATIVE PRECISION OF MEASUREMENT ENDPOINTS
0.35
0.30 -
0.25
<
CD
or
o 0.20 H
k-
£
iXI
"2
¦S 0.15 H
c
3
to
~ Blood AUG
¦ Liver
A Kidney
~ Femur
0.10
0.05
0.00
Kidney
0.0 0.2 0.4 0.6
0.8
RBA
1.0
Blood AUG
Femur
1.2 1.4
1.6
Endpoint
Slope
Intercept
R*
Blood AUG
0.177
-0.002
0.867
Liver
0.227
0.000
0.916
Kidney
0.219
0.006
0.914
Femur
0.162
0.008
0.732
Comparison of
Regression Lines
F
0.638
Fcrit{0.05)
2.227
P
0.699
Appendix D_Tables & Figures.xls (Fig D-5_Precision)
-------
DRAFT- Do N6PeWQ9ote, or Release
FIGURE D-6. FIT OF DIFFERENT MODELS TO IVBA-RBA DATA
Panel A: Linear Model (y = a + b*x)
Model Fit
IVBA
Parameter Estimates
a -0.06
b 1.03
Fit Statistics
R2
AIC
0.837
-72.75
Fig D-6JVBA-RBA model fits.xls (FigD-6A_Lin)
-------
DRAFT- Do Nc9feffiftM?iQ0ote, or Release
FIGURE D-6. FIT OF DIFFERENT MODELS TO IVBA-RBA DATA
Panel B: Power Model (y = a + b*xAc)
Model Fit
IVBA
Parameter Estimates Fit Statistics
a 0.09 R2 0.881
b 1.22 AIC -75.35
c 2.22
Fig D-6JVBA-RBA model fits.xls (FigD-6B_Power)
-------
DRAFT- Do N6PeWQ9ote, or Release
FIGURE D-6. FIT OF DIFFERENT MODELS TO IVBA-RBA DATA
Panel C: 2-Parameter Exponential Model (y = a + b*exp(x))
Model Fit
Residual Plot
"0
0
O
TJ
P
T3
0
£
0
<0
O
15
3
¦g
-------
DRAFT- Do Nc9feffiftM?iQ0ote, or Release
FIGURE D-6. FIT OF DIFFERENT MODELS TO IVBA-RBA DATA
Panel D: 3-Parameter Exponential Model (y = a + b*exp(c*x))
Model Fit
IVBA
Parameter Estimates
a -0.06
b 0.13
c 2.47
Fit Statistics
R2 0.883
AIC -75.74
Fig D-6JVBA-RBA model fits.xls (FigD-6D_Exp3)
-------
DRAFT- Do Ncffmp'Sftote, or Release
FIGURE D-7. EVALUATING THE EFFECT OF MEASUREMENT ERROR IN IVBA
Panel A: Ordinary Linear Regression
£
Ct
0.3
B
o
TD
0
D.
0.2 -
0.1 -
IS "°-1 "
CD
3
¦g
& -°-2
-0.3
Residuals
0.0
..A
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
IVBA
0.3
S 0.2 -
t 0.1 -
*
| 0.0
™ -0.1
* -0.2 -
-0.3
-2.5
Normality Plot
= 0.1194x- 2E-16
= 0.9858
-2.0
-1.5
-1.0
-0.5 0.0
Z-score
0.5
1.0
1.5
2.0
2.5
Fig D-7_IVBA-RBA.xls (FigD-7A_OLR)
-------
DRAFT- Do Ncffmp'Sftote, or Release
FIGURE D-7. EVALUATING THE EFFECT OF MEASUREMENT ERROR IN IVBA
Panel B: A = 6.0
Normality Plot
-2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5
Z-score
Fig D-7_IVBA-RBA.xls (FigD-7B_A=6)
-------
DRAFT- Do Ncffmp'Sftote, or Release
FIGURE D-7. EVALUATING THE EFFECT OF MEASUREMENT ERROR IN IVBA
Panel C: A = Syy/Sxx
Residuals
-o
3
o
TD
0
D.
1
-o
£
CD
W
X)
O
CD
a:
0.3
0.2 -
0.1
0.0
-0.1 -
-0.2
~ ~
X
~
~
~ ~
-0.3 -t 1 1 1 1 1—
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
IVBA
0.3
0.2
£ 0.1
Q_
0.0
ra -0.1
-0.2
-0.3
-2.5
Normality Plot
y = 0.1193x - 2E-16
R2 = 0.9798
-2.0
-1.5
-1.0
-0.5
0.0
Z-score
0.5
1.0
1.5
2.0
2.5
Fig D-7_IVBA-RBA.xls (FigD-7C_A=Syy over Sxx)
-------
DRAFT- Do N6PE5tel?l®i!ibte, or Release
FIGURE D-7. EVALUATING THE EFFECT OF MEASUREMENT ERROR IN IVBA
Panel D: Overlay
Method
Intercept
Slope
R2
Linear regression (A = infinity)
A = 6
A = Syy/Sxx
-0.057
-0.073
-0.108
1.034
1.066
1.130
0.837
0.845
0.855
Fig D-7_IVBA-RBA.xls (FigD-7D_Overlay)
-------
DRAFT- Do N6PE5tel?l®i!ibte, or Release
FIGURE D-8. PREDICTION INTERVAL FOR RBA BASED ON MEASURED IVBA
Fig 3-6, D-8_Prediction lntervals.xls (Fig D-8_Linear)
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
OSWER 9285.7-77
APPENDIX E
DETAILED DOSE-RESPONSE DATA AND
MODEL FITTING RESULTS
-------
draft-
do Not Cite, Quote, or Release
APPENDIX E
EXPERIMENT la
Effects of Food
Test Material 1: Lead Acetate, simultaneous with feeding
Test Material 2: Lead Acetate, 2 hours after feeding
Figure la Blood AUC - Linear Model
Figure lb Blood AUC - Exponential Model
Figure lc Blood AUC - Michaelis-Menton Model
Figure Id Blood AUC - Power Model
Figure 2a Liver - Linear Model (All Data)
Figure 2a Liver - Linear Model (Outlier Excluded)
Figure 2b Liver - Exponential Model
Figure 2c Liver - Michaelis-Menton Model
Figure 2d Liver - Power Model
Figure 3a Kidney - Linear Model
Figure 3b Kidney - Exponential Model
Figure 3 c Kidney - Michaelis-Menton Model
Figure 3d Kidney - Power Model
Figure 4a Femur - Linear Model
Figure 4b Femur - Exponential Model
Figure 4c Femur - Michaelis-Menton Model
Figure 4d Femur - Power Model
-------
draft-
do Not Cite, Quote, or Release
APPENDIX E
EXPERIMENT 2
Test Material 1: Bingham Creek Residential
Test Material 2: Bingham Creek Channel Soil
Figure la Blood AUC - Linear Model
Figure lb Blood AUC - Exponential Model
Figure lc Blood AUC - Michaelis-Menton Model
Figure Id Blood AUC - Power Model
Figure 2a Liver - Linear Model (All Data)
Figure 2a Liver - Linear Model (Outlier Excluded)
Figure 2b Liver - Exponential Model
Figure 2c Liver - Michaelis-Menton Model
Figure 2d Liver - Power Model
Figure 3a Kidney - Linear Model
Figure 3b Kidney - Exponential Model
Figure 3 c Kidney - Michaelis-Menton Model
Figure 3d Kidney - Power Model
Figure 4a Femur - Linear Model
Figure 4b Femur - Exponential Model
Figure 4c Femur - Michaelis-Menton Model
Figure 4d Femur - Power Model
-------
draft-
do Not Cite, Quote, or Release
APPENDIX E
EXPERIMENT 3
Test Material 1: Jasper County High Lead Smelter
Test Material 2: Jasper County Low Lead Yard
Figure la Blood AUC - Linear Model
Figure lb Blood AUC - Exponential Model
Figure lc Blood AUC - Michaelis-Menton Model
Figure Id Blood AUC - Power Model
Figure 2a Liver - Linear Model (All Data)
Figure 2a Liver - Linear Model (Outlier Excluded)
Figure 2b Liver - Exponential Model
Figure 2c Liver - Michaelis-Menton Model
Figure 2d Liver - Power Model
Figure 3 a Kidney - Linear Model
Figure 3b Kidney - Exponential Model
Figure 3 c Kidney - Michaelis-Menton Model
Figure 3d Kidney - Power Model
Figure 4a Femur - Linear Model
Figure 4b Femur - Exponential Model
Figure 4c Femur - Michaelis-Menton Model
Figure 4d Femur - Power Model
-------
draft-
do Not Cite, Quote, or Release
APPENDIX E
EXPERIMENT 4
Test Material 1: Murray Smelter Slag
Test Material 2: Jasper County High Lead Mill
Figure la Blood AUC - Linear Model
Figure lb Blood AUC - Exponential Model
Figure lc Blood AUC - Michaelis-Menton Model
Figure Id Blood AUC - Power Model
Figure 2a Liver - Linear Model
Figure 2b Liver - Exponential Model
Figure 2c Liver - Michaelis-Menton Model
Figure 2d Liver - Power Model
Figure 3 a Kidney - Linear Model
Figure 3b Kidney - Exponential Model
Figure 3c Kidney - Michaelis-Menton Model
Figure 3d Kidney - Power Model
Figure 4a Femur - Linear Model
Figure 4b Femur - Exponential Model
Figure 4c Femur - Michaelis-Menton Model
Figure 4d Femur - Power Model
-------
draft-
do Not Cite, Quote, or Release
APPENDIX E
EXPERIMENT 5
Test Material 1: Aspen Berm
Test Material 2: Aspen Residential
Figure la Blood AUC - Linear Model
Figure lb Blood AUC - Exponential Model
Figure lc Blood AUC - Michaelis-Menton Model
Figure Id Blood AUC - Power Model
Figure 2a Liver - Linear Model (All Data)
Figure 2a Liver - Linear Model (Outlier Excluded)
Figure 2b Liver - Exponential Model
Figure 2c Liver - Michaelis-Menton Model
Figure 2d Liver - Power Model
Figure 3 a Kidney - Linear Model
Figure 3b Kidney - Exponential Model
Figure 3 c Kidney - Michaelis-Menton Model
Figure 3d Kidney - Power Model
Figure 4a Femur - Linear Model
Figure 4b Femur - Exponential Model
Figure 4c Femur - Michaelis-Menton Model
Figure 4d Femur - Power Model
-------
draft-
do Not Cite, Quote, or Release
APPENDIX E
EXPERIMENT 6
Test Material 1: Midvale Slag
Test Material 2: Butte Soil
Figure la Blood AUC - Linear Model
Figure lb Blood AUC - Exponential Model
Figure lc Blood AUC - Michaelis-Menton Model
Figure Id Blood AUC - Power Model
Figure 2a Liver - Linear Model (All Data)
Figure 2a Liver - Linear Model (Outlier Excluded)
Figure 2b Liver - Exponential Model
Figure 2c Liver - Michaelis-Menton Model
Figure 2d Liver - Power Model
Figure 3 a Kidney - Linear Model (All Data)
Figure 3 a Kidney - Linear Model (Outlier Excluded)
Figure 3b Kidney - Exponential Model
Figure 3 c Kidney - Michaelis-Menton Model
Figure 3d Kidney - Power Model
Figure 4a Femur - Linear Model
Figure 4b Femur - Exponential Model
Figure 4c Femur - Michaelis-Menton Model
Figure 4d Femur - Power Model
-------
draft-
do Not Cite, Quote, or Release
APPENDIX E
EXPERIMENT 7
Test Material 1: California Gulch Phase I Residential Soil
Test Material 2: California Gulch Fe/Mn PbO
Figure la Blood AUC - Linear Model (All Data)
Figure la Blood AUC - Linear Model (Outlier Excluded)
Figure lb Blood AUC - Exponential Model
Figure lc Blood AUC - Michaelis-Menton Model
Figure Id Blood AUC - Power Model
Figure 2a Liver - Linear Model (All Data)
Figure 2a Liver - Linear Model (Outlier Excluded)
Figure 2b Liver - Exponential Model
Figure 2c Liver - Michaelis-Menton Model
Figure 2d Liver - Power Model
Figure 3 a Kidney - Linear Model (All Data)
Figure 3 a Kidney - Linear Model (Outlier Excluded)
Figure 3b Kidney - Exponential Model
Figure 3c Kidney - Michaelis-Menton Model
Figure 3d Kidney - Power Model
Figure 4a Femur - Linear Model
Figure 4b Femur - Exponential Model
Figure 4c Femur - Michaelis-Menton Model
Figure 4d Femur - Power Model
-------
draft-
do Not Cite, Quote, or Release
APPENDIX E
EXPERIMENT 8
Test Material 1: California Gulch AV Slag
Test Material 2: Lead Acetate - IV (for ABA determination)
Figure la Blood AUC - Linear Model
Figure lb Blood AUC - Exponential Model
Figure lc Blood AUC - Michaelis-Menton Model
Figure Id Blood AUC - Power Model
Figure 2a Liver - Linear Model (All Data)
Figure 2a Liver - Linear Model (Outlier Excluded)
Figure 2b Liver - Exponential Model
Figure 2c Liver - Michaelis-Menton Model
Figure 2d Liver - Power Model
Figure 3a Kidney - Linear Model (All Data)
Figure 3 a Kidney - Linear Model (Outlier Excluded)
Figure 3b Kidney - Exponential Model
Figure 3c Kidney - Michaelis-Menton Model
Figure 3d Kidney - Power Model
Figure 4a Femur - Linear Model
Figure 4b Femur - Exponential Model
Figure 4c Femur - Michaelis-Menton Model
Figure 4d Femur - Power Model
-------
draft-
do Not Cite, Quote, or Release
APPENDIX E
EXPERIMENT 9
Test Material 1: Palmerton Location 2
Test Material 2: Palmerton Location 4
Figure la Blood AUC - Linear Model
Figure lb Blood AUC - Exponential Model
Figure lc Blood AUC - Michaelis-Menton Model
Figure Id Blood AUC - Power Model
Figure 2a Liver - Linear Model
Figure 2b Liver - Exponential Model
Figure 2c Liver - Michaelis-Menton Model
Figure 2d Liver - Power Model
Figure 3a Kidney - Linear Model
Figure 3b Kidney - Exponential Model
Figure 3 c Kidney - Michaelis-Menton Model
Figure 3d Kidney - Power Model
Figure 4a Femur - Linear Model
Figure 4b Femur - Exponential Model
Figure 4c Femur - Michaelis-Menton Model
Figure 4d Femur - Power Model
-------
draft-
do Not Cite, Quote, or Release
APPENDIX E
EXPERIMENT 11
Test Material 1: Murray Smelter Soil
Test Material 2: NIST Paint
Figure la Blood AUC - Linear Model
Figure lb Blood AUC - Exponential Model
Figure lc Blood AUC - Michaelis-Menton Model
Figure Id Blood AUC - Power Model
Figure 2a Liver - Linear Model
Figure 2b Liver - Exponential Model
Figure 2c Liver - Michaelis-Menton Model
Figure 2d Liver - Power Model
Figure 3 a Kidney - Linear Model
Figure 3b Kidney - Exponential Model
Figure 3c Kidney - Michaelis-Menton Model
Figure 3d Kidney - Power Model
Figure 4a Femur - Linear Model
Figure 4b Femur - Exponential Model
Figure 4c Femur - Michaelis-Menton Model
Figure 4d Femur - Power Model
-------
draft-
do Not Cite, Quote, or Release
APPENDIX E
EXPERIMENT 12
Test Material 1: Galena-enriched Soil
Test Material 2: Palmerton Location 2 (Reproducibility Study)
Test Material 3: California Gulch Oregon Gulch Tailings
Figure la Blood AUC - Linear Model
Figure lb Blood AUC - Exponential Model
Figure lc Blood AUC - Michaelis-Menton Model
Figure Id Blood AUC - Power Model
Figure 2a Liver - Linear Model (All Data)
Figure 2a Liver - Linear Model (Outlier Excluded)
Figure 2b Liver - Exponential Model
Figure 2c Liver - Michaelis-Menton Model
Figure 2d Liver - Power Model
Figure 3a Kidney - Linear Model (All Data)
Figure 3 a Kidney - Linear Model (Outliers Excluded)
Figure 3b Kidney - Exponential Model
Figure 3c Kidney - Michaelis-Menton Model
Figure 3d Kidney - Power Model
Figure 4a Femur - Linear Model
Figure 4b Femur - Exponential Model
Figure 4c Femur - Michaelis-Menton Model
Figure 4d Femur - Power Model
-------
DRAFT- Do Not Cite, Quote, or Release
OSWER 9285.7-77
APPENDIX F
DETAILED LEAD SPECIATION DATA
FOR TEST MATERIALS
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
METAL CONTENT OF TEST MATERIALS
Experiment
Test Material
Concentration (ppm)
Al
As
Au
Ba
Be
Ca
Cd
Co
Cr
Cu
Fe
Hg
K
Mg
Mn
Na
Ni
Pb
Sb
Se
Tl
V
Zn
2
Bingham Creek Residential
10,600
51.2
4.1
143
0.71
13,600
4.2
7.5
16.6
691
16,100
-
4,340
7,020
466
362
15.0
1,590
10 U
<17
<17
20.8
903
Bingham Creek Channel Soil
10,100
149.0
17.2
152
0.73
8,500
8.7
7.9
17.9
1,720
22,500
-
4,150
5,970
376
314
15.1
6,330
18.7
<17
<17
22.0
-
3
Jasper County High Lead Smelter
8,850
25.1
1.3
284
1.70
45,800
33.7
19.3
23.8
94
40,200
0.64
1,490
7,860
784
399
44.8
10,800
4.90
1.0U
1.4U
22.5
10,000
Jasper County Low Lead Yard
4,370
10.7
0.6
94
1.00
81,800
188.0
6.4
15.2
144
18,000
1.30
927
1,390
240
403
30.1
4,050
1.0 U
1.0U
1.80
14.8
50,000
4
Murray Smelter Slag
9,370
710
18.3
2,140
0.86
89,600
30.9
45.4
34.0
2,100
170,000
1.00
2,430
11,200
2,640
836
16.7
11,700
55.7
43.90
12.60
73.6
49,500
Jasper County High Lead Mill
9,380
16.4
18.8
211
1.40
19,900
139.0
34.3
64.6
96
26,600
12.10
1,400
2,280
1,270
339
110.0
6,940
1.0 U
1.0U
1.4U
23.0
17,200
5
Aspen Berm
5,070
66.9
92.3
1,640
1.30
37,200
41.9
17.1
7.7
145
33,700
0.77
1,090
14,300
2,220
249
29.8
14,200
5.20
2.00
1.80
11.5
6,580
Aspen Residential
8,440
16.7
18.9
1,030
0.82
17,300
47.4
11.1
10.4
52
23,000
0.23
2,140
6,890
934
114
21.9
3,870
11.4
0.38
0.27
16.0
4,110
6
Midvale Slag
10,500
619
.11U
637
0.58
93,200
24.5
33.0
142.0
1,330
202,000
0.74
4,250
6,180
1,640
7,910
.31U
8,170
71.9
39.70
8.10
10.1U
33,300
Butte Soil
7,540
226
40.5
134
0.56
15,700
42.2
9.2
6.9
838
48,500
2.20
3,560
2,950
12,800
530
8.0
8,530
10.60
0.27
1.80
27.0
12,100
7
California Gulch Phase I Residential
Soil
8,670
203
43.0
605
0.60
20,100
59.9
2.0
9.1
657
68,120
1.26
1,500
9,521
7,090
6,560
5.6
7,510
1.80
1.90
<0.5
33.7
13,738
California Gulch Fe/Mn PbO
11,900
110
16.7
266
1.00
3,930
38.5
6.9
7.5
165
27,500
4.90
1,770
2,520
1,190
279
7.5
4,320
6.00
0.80
3.70
17.9
2,650
8
California Gulch AV Slag
20,800
1,050
21.2
2,430
1.20
117,000
12.8
53.8
43.1
2,080
207,000
0.11
7,390
6,360
6,910
4,080
7.1
10,600
57.2
61.30
1.80
37.2
67,300
g
Palmerton Location 2
7,750
110
9.5
6,850
1.40
1,160
195.0
18.8
30.3
462
25,900
1.70
515
725
6,320
667
15.0
3,230
6.00
11.80
1.90
53.1
6,500
Palmerton Location 4
7,850
134.0
5.1
1,090
2.00
2,480
319.0
17.4
26.6
350
26,700
1.10
512
684
9,230
2,100
26.8
2,150
7.40
6.90
0.85
49.8
19,100
11
Murray Smelter Soil
6,520
310
11.1
584
0.48b
69,000
23.8
11.5
16.4
856
38,700
0.52
2,040
15,000
863
532.0b
10.4
3,200
20.0
6.80
4.80
28.3
10,400
NIST Paint
5,850
4.8
0.63U
1,320
0.47b
11,800
4.0
8.3
20.8
12
8,890
0.92
1,360
2,900
272
81.9b
5.80b
8,350
8.7 U
0.61 U
0.87U
11.6
1,880
12
Galena-enriched Soil
6,340
4.9
0.63U
112
0.49b
2,650
0.8
3.1
10.2
11
10,000
0.06b
1,460
2,790
293
31.20b
3.80b
11,200
8.70
0.61 U
0.87U
12.60b
107
California Gulch Oregon Gulch
Tailings
248
1,290
41.7
14
2.00
8,290
4.0
10.1
8.0
350
391,000
0.24
451
118
126
34
28.2
1,270
74.4
0.53
0.86
47.7
441
All samples were analyzed by inductively coupled plasma-atomic emission spectrometry (ICP-AES) in accord with USEPA Method 200.7.
Appendix F_TAL.xls (TAL)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 2 - BINGHAM CREEK RESIDENTIAL
Lead Speciation Summary Statistics
Mineral
Counts
Total
Lib
Avq
Particle Size
Min
Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Cerussite
2
2
4
2
5
1.0%
1.0%
0.28%
0.28%
6.6
0.776
1.8%
1.8%
Fe-Pb Oxide
30
30
15
2
75
15.1%
15.1%
17.93%
17.93%
4
0.052
4.6%
4.6%
Fe-Pb Silicate*
14
14
10
8
20
7.0%
7.0%
5.52%
5.52%
3.5
0.052
1.2%
1.2%
Mn-Pb Oxide
21
21
22
2
110
10.6%
10.6%
18.13%
18.13%
5.1
0.159
18.1%
18.1%
Pb-As Oxide
3
3
4
2
8
1.5%
1.5%
0.52%
0.52%
6
0.5
1.9%
1.9%
Pb Phosphate
43
43
13
1
110
21.6%
21.6%
21.70%
21.70%
5.1
0.37
50.4%
50.4%
Fe-Pb Sulfate
86
86
10
1
120
43.2%
43.2%
35.91%
35.91%
3.7
0.134
21.9%
21.9%
TOTAL
199
199
13
100.0%
100.0%
100.00%
100.00%
100.0%
100.0%
"This mineral is now considered to be equivalent to Fe-Pb Oxide.
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
38.2%
38.2%
8.2%
8.2%
5-9
22.1%
22.1%
12.2%
12.2%
10-19
19.1%
19.1%
13.0%
13.0%
20-49
15.6%
15.6%
30.3%
30.3%
50-99
3.5%
3.5%
18.8%
18.8%
100-149
1.5%
1.5%
17.6%
17.6%
150-199
0.0%
0.0%
0.0%
0.0%
200-249
0.0%
0.0%
0.0%
0.0%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL
100%
100%
100%
100%
Appendix F_Tables.xls (2_Res)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 2 - BINGHAM CREEK RESIDENTIAL
Speciation and Particle Size Data
Panel A: Relative Lead Mass
Fe-Pb Sulfate
Lead Phosphate
Pb-As Oxide
Mn-Pb Oxide
Fe-Pb Silicate*
Fe-Pb Oxide
Cerussite
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
100%
90%
80%
70%
>, 60%
o
| 50%
2
40%
30%
20%
10%
0%
This mineral is now considered to be equivalent to Fe-Pb Oxid
Appendix F_Figures.xls (2_Res)
~ Frequency of Occurrence
~ Relative Lead Mass
i
&
Panel B: Particle Size Distribution
10-19 20-49 50-99 100-149 150-199 200-249 >250
Particle Size (urn)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 2 - BINGHAM CREEK CHANNEL SOIL
Lead Speciation Summary Statistics
Mineral
Counts
Total Lib
Avg
Particle Size
Min Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Anglesite
57
56
4
1
30
11.6%
11.4%
6.26%
6.23%
6.3
0.684
28.4%
28.3%
Cerussite
1
1
2
2
2
0.2%
0.2%
0.05%
0.05%
6.6
0.776
0.3%
0.3%
Fe-Pb Oxide
25
25
17
4
60
5.1%
5.1%
10.88%
10.88%
4.0
0.053
2.4%
2.4%
FeSbO
1
1
5
5
5
0.2%
0.2%
0.13%
0.13%
0.0%
0.0%
Fe-Pb Silicate*
4
4
15
10
20
0.8%
0.8%
1.56%
1.56%
3.5
0.057
0.3%
0.3%
Galena
1
1
50
50
50
0.2%
0.2%
1.30%
1.30%
7.5
0.866
8.9%
8.9%
Mn-Pb Oxide
5
5
21
5
50
1.0%
1.0%
2.67%
2.67%
5.1
0.159
2.3%
2.3%
Lead Organic
2
2
105
100
110
0.4%
0.4%
5.45%
5.45%
1.3
0.037
0.3%
0.3%
Pb-As Oxide
3
3
4
1
8
0.6%
0.6%
0.29%
0.29%
6.0
0.500
0.9%
0.9%
Lead Barite
1
1
10
10
10
0.2%
0.2%
0.26%
0.26%
4.5
0.031
0.0%
0.0%
Lead Phosphate
42
42
12
1
100
8.6%
8.6%
13.01%
13.01%
5.1
0.370
25.8%
25.8%
Fe-Pb Sulfate
349
349
6
1
110
71.1%
71.1%
58.15%
58.15%
3.7
0.134
30.4%
30.4%
TOTAL
491
490
8
100.0%
99.8%
100.00%
99.97%
100.0%
99.9%
"This mineral is now considered to be equivalent to Fe-Pb Oxide.
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
66.2%
66.0%
13.9%
13.8%
5-9
13.6%
13.6%
17.5%
17.5%
10-19
9.8%
9.8%
18.4%
18.4%
20-49
6.1%
6.1%
20.0%
20.0%
50-99
3.1%
3.1%
20.5%
20.5%
100-149
1.2%
1.2%
9.6%
9.6%
150-199
0.0%
0.0%
0.0%
0.0%
200-249
0.0%
0.0%
0.0%
0.0%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL
100%
100%
100%
100%
Appendix F_Tables.xls (2_Ch)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 2 - BINGHAM CREEK CHANNEL SOIL
Speciation and Particle Size Data
Panel A: Relative Lead Mass
Fe-Pb Sulfate
Lead Phosphate
Lead Barite
Pb-As Oxide
:ide I
Lead Organic
Mn-Pb Oxide
Galena
Fe-Pb Silicate*
Fe-Pb Oxide
Cerussite
Anglesite
h
P
0%
10%
~ Frequency of Occurrence
~ Relative Lead Mass
20%
30%
40%
50%
70%
80%
90% 100%
100%
90% -
80% --
70%
60%
| 50% -
40%
30%
20% -
10% -
0%
Panel B: Particle Size Distribution
-+-
J
-+-
-+-
-+-
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249 >250
Particle Size (|jm)
*This mineral is now considered to be equivalent to Fe-Pb Oxid
Appendix F_Figures.xls (2_Ch)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 3 - JASPER COUNTY HIGH LEAD SMELTER
Lead Speciation Summary Statistics
Mineral
Counts
Total
Lib
Avg
Particle Size
Min
Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Anglesite
1
1
12
12
12
0.25%
0.25%
0.11%
0.11%
6.3
0.684
0.9%
0.9%
Calcite
2
2
48
35
60
0.50%
0.50%
0.87%
0.87%
2.8
0.050
0.2%
0.2%
Cerussite
12
11
31
8
90
3.0%
2.8%
3.39%
3.26%
6.6
0.776
32.1%
30.7%
Clay
2
2
35
10
60
0.50%
0.50%
0.64%
0.64%
3.1
0.005
0.02%
0.02%
Fe-Pb Oxide
24
24
45
10
150
6.0%
6.0%
10.04%
10.04%
4.0
0.037
2.7%
2.7%
Fe-Pb Silicate*
22
22
83
4
175
5.5%
5.5%
16.80%
16.80%
3.7
0.100
11.5%
11.5%
Mn-Pb Oxide
5
5
47
12
100
1.3%
1.3%
2.18%
2.18%
5.1
0.112
2.3%
2.3%
Native Lead
56
0
2
1
9
14.0%
0.0%
1.07%
0.00%
11.3
1.000
22.2%
0.0%
Lead Oxide
6
1
6
1
10
1.5%
0.3%
0.31%
0.02%
4.0
0.037
0.09%
0.01%
Lead Phosphate
117
117
7
1
90
29.3%
29.3%
7.25%
7.25%
5.1
0.310
21.1%
21.1%
Slag
62
62
94
15
300
15.5%
15.5%
53.58%
53.58%
3.7
0.012
4.3%
4.3%
Fe-Pb Sulfate
90
75
5
1
10
22.6%
18.8%
3.75%
3.20%
3.7
0.100
2.6%
2.2%
TOTAL
399
322
27
100.0%
80.7%
100.00%
97.95%
100.0%
76.0%
"This mineral is now considered to be equivalent to Fe-Pb Oxide.
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
44.4%
28.1%
17.2%
3.8%
5-9
18.5%
16.0%
15.0%
5.7%
10-19
8.0%
7.5%
7.8%
6.4%
20-49
8.3%
8.3%
14.0%
14.0%
50-99
9.0%
9.0%
31.9%
31.9%
100-149
8.8%
8.8%
9.6%
9.6%
150-199
2.0%
2.0%
3.8%
3.8%
200-249
0.5%
0.5%
0.3%
0.3%
>250
0.5%
0.5%
0.4%
0.4%
TOTAL
100%
81%
100%
76%
Appendix F_Tables.xls (3_HL)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 3 - JASPER COUNTY HIGH LEAD SMELTER
Speciation and Particle Size Data
Panel A: Relative Lead Mass
Anglesite
Calcite
Cerussite
Clay
Fe-Pb Oxide
Fe-Pb Silicate*
Mn-Pb Oxide
Native Lead
Lead Oxide
Lead Phosphate
Slag
Fe-Pb Sulfate
~ Frequency of Occurrence
¦ Relative Lead Mass
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Panel B: Particle Size Distribution
100%
90%
80% T
70%
60% +
50%
40%
30% +
20%
10% T
0%
4-
+
-! 1 1 I
+
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249
Particle Size (jjm)
>250
"This mineral is now considered to be equivalent to Fe-Pb Oxid
Appendix F_Figures.xls (3_HL)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 3 - JASPER COUNTY LOW LEAD YARD
Lead Speciation Summary Statistics
Mineral
Counts
Total
Lib
Avq
Particle Size
Min Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Anglesite
3
3
3
2
6
1.6%
1.6%
0.31%
0.31%
6.3
0.684
0.48%
0.48%
Cerussite
95
gs
15
1
130
52.2%
52.2%
43.37%
43.37%
6.6
0.776
81.1%
81.1%
Clay
1
1
15
15
15
0.5%
0.5%
0.46%
0.46%
3.1
0.005
0.003%
0.003%
Fe-Pb Oxide
18
18
36
8
100
g.g%
g.g%
ig.53%
ig.53%
4
0.037
1.1%
1.1%
Fe-Pb Silicate*
g
g
33
5
100
4.g%
4.g%
g.11%
g.11%
3.7
0.1
1.2%
1.2%
Galena
2
1
53
25
80
1.1%
0.5%
3.21%
0.76%
7.5
0.866
7.6%
1.8%
Mn-Pb Oxide
10
10
25
8
55
5.5%
5.5%
7.73%
7.73%
5.1
0.112
1.6%
1.6%
Pb-As Oxide
1
1
8
8
8
0.5%
0.5%
0.24%
0.24%
7.1
0.243
0.15%
0.15%
Lead Silicate
2
2
2
1
2
1.1%
1.1%
o.og%
o.og%
8
0.167
0.04%
0.04%
Lead Phosphate
32
32
11
1
80
17.6%
17.6%
10.42%
10.42%
5.1
0.31
6.0%
6.0%
Fe-Pb Sulfate
g
g
20
1
100
4.g%
4.g%
5.53%
5.53%
3.7
0.1
0.75%
0.75%
TOTAL
182
181
18
100.0%
gg.5%
100.00%
g7.56%
100.0%
g4.2%
"This mineral is now considered to be equivalent to Fe-Pb Oxide.
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
28.6%
28.6%
5.0%
5.0%
5-g
20.3%
20.3%
8.5%
8.5%
10-ig
20.g%
20.g%
17.1%
17.1%
20-4g
ig.8%
ig.8%
30.2%
30.2%
50-gg
7.7%
7.1%
23.6%
17.8%
ioo-i4g
2.7%
2.7%
15.6%
15.6%
150-igg
0.0%
0.0%
0.0%
0.0%
200-24g
0.0%
0.0%
0.0%
0.0%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL
100%
gg%
100%
g4%
Appendix F_Tables.xls (3_LL)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 3 - JASPER COUNTY LOW LEAD YARD
Speciation and Particle Size Data
Panel A: Relative Lead Mass
Anglesite
Gerussite
Clay
Fe-Pb Oxide
Fe-Pb Silicate*
Galena
Mri-Pb Oxide
Pb-As Oxide
Lead Silicate
Lead Phosphate
Fe-Pb Sulfate
0%
10%
~ Frequency of Occurrence
~ Relative Lead Mass
20% 30% 40% 50%
60%
70%
80% 90% 100%
100%
90% --
80% --
70%
60%
50%
40%
30%
20% --
10% --
0%
Panel B: Particle Size Distribution
-+-
.
L
+
+
<5
5-9
10-19
20-49 50-99 100-149 150-199 200-249
Particle Size (pm)
>250
This mineral is now considered to be equivalent to Fe-Pb Oxid
Appendix F_Figures.xls (3_LL)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 4 - MURRAY SMELTER SLAG
Lead Speciation Summary Statistics
Mineral
Counts
Total Lib
Avg
Particle Size
Min
Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Anglesite
3
0
12
10
15
0.2%
0.0%
0.04%
0.00%
6.3
0.684
1.0%
0.0%
Cerussite
4
3
8
3
15
0.3%
0.2%
0.04%
0.04%
6.6
0.776
1.1%
1.0%
Fe-Pb Oxide
3
3
18
8
35
0.2%
0.2%
0.07%
0.07%
4
0.031
0.04%
0.04%
Fe-As Oxide
3
3
17
4
35
0.2%
0.2%
0.06%
0.06%
0.0%
0.0%
Fe-Pb Silicate*
g
9
28
8
80
0.7%
0.7%
0.32%
0.32%
4
0.22
1.5%
1.5%
Galena
98
7
2
1
15
7.2%
0.5%
0.27%
0.08%
7.5
0.866
9.2%
2.6%
Mn-Pb Oxide
7
7
31
8
110
0.5%
0.5%
0.28%
0.28%
5.1
0.112
0.8%
0.8%
Native Lead
3
2
3
2
4
0.2%
0.1%
0.01%
0.01%
11.3
1
0.7%
0.5%
Pb-As Oxide
39
31
6
1
60
2.9%
2.3%
0.30%
0.27%
7.1
0.5
5.7%
5.1%
Pb(M)0
8
3
18
2
110
0.6%
0.2%
0.19%
0.16%
8
0.5
3.9%
3.3%
Lead Oxide
143
79
8
1
100
10.5%
5.8%
1.48%
1.18%
9.5
0.93
68.7%
54.6%
Slag
1037
1037
73
5
310
76.1%
76.1%
96.71%
96.71%
3.65
0.0038
7.0%
7.0%
Fe-Pb Sulfate
2
2
55
10
100
0.1%
0.1%
0.14%
0.14%
3.7
0.1
0.3%
0.3%
Zn-Pb Silicate
4
3
16
10
30
0.3%
0.2%
0.08%
0.07%
5.1
0.014
0.03%
0.03%
TOTAL
1363
1189
58
100.0%
87.2%
100.00%
99.38%
100.0%
76.8%
"This mineral is now considered to be equivalent to Fe-Pb Oxide.
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
14.5%
4.1%
15.6%
5.4%
5-9
12.6%
11.2%
13.7%
7.9%
10-19
14.7%
13.9%
22.9%
17.3%
20-49
6.2%
6.2%
17.1%
15.3%
50-99
20.3%
20.3%
16.2%
16.2%
100-149
23.8%
23.8%
12.8%
12.8%
150-199
4.2%
4.2%
0.8%
0.8%
200-249
3.2%
3.2%
0.8%
0.8%
>250
0.4%
0.4%
0.1%
0.1%
TOTAL
100%
87%
100%
77%
Appendix F_Tables.xls (4_Mur)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 4 - MURRAY SMELTER SLAG
Speciation and Particle Size Data
Panel A: Relative Lead Mass
Zn-Pb Silicate
Fe-Pb Sulfate
Slag
Lead Oxide
Pb(M)0
Pb-As Oxide
Native Lead
Mn-Pb Oxide
Galena
Fe-Pb Silicate*
Fe-As Oxide
Fe-Pb Oxide
Gerussite
Anglesite
--
I
I
V
u
I
j
i
I
i
]
~ Frequency of Occurrence
i
~ Relative Lead Mass
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Panel B: Particle Size Distribution
100%
90% -
80%
70%
60% -
50%
40%
30%
20% -
10% -
0%
+
+
-I I ; t
¦_
<5 5-9 10-19 20-49 50-99 100-149
Particle Size (pm)
150-199 200-249
>250
This mineral is now considered to be equivalent to Fe-Pb Oxid
Appendix F_Figures.xls (4_Mur)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 4 - JASPER COUNTY HIGH LEAD MILL
Lead Speciation Summary Statistics
Mineral
Counts
Total
Lib
Avg
Particle Size
Min Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Anglesite
1
1
25
25
25
0.36%
0.36%
0.36%
0.36%
6.3
0.684
1.6%
1.6%
Lead Barite
1
1
3
3
3
0.36%
0.36%
0.04%
0.04%
4.5
0.045
0.01%
0.01%
Calcite
1
1
25
25
25
0.36%
0.36%
0.36%
0.36%
2.8
0.05
0.1%
0.1%
Cerussite
90
90
8
1
70
32.0%
32.0%
10.74%
10.74%
6.6
0.776
57.0%
57.0%
Clay
3
3
24
8
40
1.1%
1.1%
1.04%
1.04%
3.1
0.005
0.02%
0.02%
Fe-Pb Oxide
33
33
22
3
110
11.7%
11.7%
10.44%
10.44%
4
0.037
1.6%
1.6%
Fe-Pb Silicate*
41
41
36
1
210
14.6%
14.6%
21.16%
21.16%
3.7
0.1
8.1%
8.1%
Galena
6
0
6
1
30
2.1%
0.0%
0.51%
0.00%
7.5
0.866
3.4%
0.0%
Mn-Pb Oxide
39
39
27
3
125
13.9%
13.9%
14.77%
14.77%
5.1
0.112
8.7%
8.7%
Native Lead
3
0
4
1
10
1.1%
0.0%
0.18%
0.00%
11.3
1
2.2%
0.0%
Lead Oxide
3
1
17
5
40
1.07%
0.36%
0.71%
0.57%
9.5
0.93
6.5%
5.2%
Lead Silicate
1
1
10
10
10
0.36%
0.36%
0.14%
0.14%
8
0.45
0.53%
0.53%
Lead Phosphate
15
15
21
2
100
5.3%
5.3%
4.53%
4.53%
5.1
0.31
7.4%
7.4%
Slag
24
24
92
15
210
8.5%
8.5%
31.45%
31.45%
3.65
0.012
1.4%
1.4%
Fe-Pb Sulfate
20
20
13
3
60
7.1%
7.1%
3.58%
3.58%
3.7
0.1
1.4%
1.4%
TOTAL
281
270
25
100.0%
96.1%
100.00%
99.16%
100.0%
93.1%
"This mineral is now considered to be equivalent to Fe-Pb Oxide.
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
22.8%
20.3%
8.3%
7.2%
5-9
20.6%
19.9%
12.9%
11.6%
10-19
22.1%
21.7%
24.3%
22.7%
20-49
18.9%
18.5%
33.7%
30.9%
50-99
8.5%
8.5%
12.8%
12.8%
100-149
5.7%
5.7%
6.5%
6.5%
150-199
0.7%
0.7%
0.2%
0.2%
200-249
0.7%
0.7%
1.3%
1.3%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL
100%
96%
100%
93%
Appendix F_Tables.xls (4_HL)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 4 - JASPER COUNTY HIGH LEAD MILL
Speciatiori and Particle Size Data
Anglesite
Lead Barite
Calcite
Cerussite
Clay
Fe-Pb Oxide
Fe-Pb Silicate*
Galena
Mn-Pb Oxide
Native Lead
Lead Oxide
Lead Silicate
Lead Phosphate
Slag
Fe-Pb Sulfate
b
b
0%
Panel A: Relative Lead Mass
~ Frequency of Occurrence
~ Relative Lead Mass
10% 20% 30% 40% 50% 60% 70% 80%
90% 100%
Panel B: Particle Size Distribution
100% j
90% --
80% --
70% --
>, 60% --
o
c
CD
= 50% --
S
40% --
30% --
20% --
10% --
0% -I—I L—I 4- I——I I 4— I
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249 >250
Particle Size (pm)
This mineral is now considered to be equivalent to Fe-Pb Oxid
Appendix F_Figures.xls (4_HL)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 5 - ASPEN BERM
Lead Speciation Summary Statistics
Mineral
Counts
Total
Lib
Avg
Particle Size
Min Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Clay
4
4
55
10
120
1.4%
1.4%
3.30%
3.30%
2.6
0.02
0.1%
0.1%
Anglesite
34
34
5
1
90
12.2%
12.2%
2.63%
2.63%
6.3
0.684
6.6%
6.6%
Lead Barite
3
3
10
2
25
1.1%
1.1%
0.45%
0.45%
4.5
0.05
0.1%
0.1%
Cerussite
71
68
20
1
110
25.4%
24.4%
20.80%
20.11%
6.6
0.776
61.7%
59.6%
Fe-Pb Oxide
80
69
35
2
210
28.7%
24.7%
41.43%
36.09%
4
0.095
9.1%
7.9%
Galena
8
6
27
10
50
2.9%
2.2%
3.23%
2.70%
7.5
0.86
12.0%
10.1%
Mn-Pb Oxide
9
9
56
10
150
3.2%
3.2%
7.58%
7.58%
5.1
0.2
4.5%
4.5%
Lead Organic
2
2
70
40
100
0.7%
0.7%
2.10%
2.10%
1.3
0.018
0.0%
0.0%
Lead Phosphate
7
7
45
10
110
2.5%
2.5%
4.73%
4.73%
5.1
0.09
1.3%
1.3%
Fe-Pb Sulfate
61
39
15
4
90
21.9%
14.0%
13.75%
6.87%
3.7
0.16
4.7%
2.4%
TOTAL
279
241
24
100.0%
86.4%
100.00%
86.57%
100.0%
92.5%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
26.5%
25.4%
2.5%
2.3%
5-9
19.0%
15.8%
5.9%
5.6%
10-19
21.5%
17.6%
14.4%
12.6%
20-49
17.2%
14.3%
29.7%
26.3%
50-99
8.2%
5.7%
25.3%
23.4%
100-149
6.1%
6.1%
19.0%
19.0%
150-199
0.7%
0.7%
1.8%
1.8%
200-249
0.7%
0.7%
1.4%
1.4%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL
100%
86%
100%
92%
Appendix F_Tables.xls (5_Berm)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 5 - ASPEN BERM
Speciation and Particle Size Data
Fe-Pb Sulfate
Lead Phosphate
Lead Organic
Mn-Pb Oxide
Galena
Fe-Pb Oxide
Cerussite
Lead Barite
Anglesite
Clay
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Panel A: Relative Lead Mass
~ Frequency of Occurrence
~ Relative Lead Mass
]
A
D
D
Panel B: Particle Size Distribution
w
E
100%
90% --
80% --
70% --
60% --
50% --
40%
30%
20%
10%
0%
+
+
Particle Size (pm)
Appendix F_Figures.xls (5_Berm)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 5 - ASPEN RESIDENTIAL
Lead Speciation Summary Statistics
Mineral
Total
Counts
Lib
Avg
Particle Size
Min
Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Anglesite
2
2
5
4
5
0.7%
0.7%
0.27%
0.27%
6.3
0.684
0.6%
0.6%
Cerussite
35
35
23
2
125
12.0%
12.0%
24.57%
24.57%
6.6
0.776
64.2%
64.2%
Fe-Pb Oxide
138
138
9
1
100
47.4%
47.4%
38.18%
38.18%
4
0.095
7.4%
7.4%
Galena
7
1
25
5
110
2.4%
0.3%
5.21%
3.31%
7.5
0.86
17.1%
10.9%
Mn-Pb Oxide
14
14
23
5
80
4.8%
4.8%
9.73%
9.73%
5.1
0.2
5.1%
5.1%
Lead Organic
1
1
80
80
80
0.3%
0.3%
2.41%
2.41%
1.3
0.018
0.0%
0.0%
Lead Phosphate
7
7
21
3
60
2.4%
2.4%
4.49%
4.49%
5.1
0.09
1.1%
1.1%
Fe-Pb Sulfate
87
87
6
1
60
29.9%
29.9%
15.15%
15.15%
3.7
0.16
4.6%
4.6%
TOTAL
291
285
11
100.0%
97.9%
100.00%
98.10%
100.0%
93.8%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
38.5%
38.5%
4.5%
4.5%
5-9
35.1%
34.0%
9.3%
7.5%
10-19
12.4%
11.7%
9.2%
7.2%
20-49
8.2%
7.9%
22.7%
20.2%
50-99
3.8%
3.8%
8.6%
8.6%
100-149
2.1%
2.1%
45.7%
45.7%
150-199
0.0%
0.0%
0.0%
0.0%
200-249
0.0%
0.0%
0.0%
0.0%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL
100%
98%
100%
94%
Appendix F_Tables.xls (5_Res)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 5 - ASPEN RESIDENTIAL
Speciation and Particle Size Data
Panel A: Relative Lead Mass
Fe-Pb Sulfate
Lead Phosphate
Lead Organic
Mn-Pb Oxide
Galena
Fe-Pb Oxide
Cerussite
Anglesite
?
~ Frequency of Occurrence
~ Relative Lead Mass
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Panel B: Particle Size Distribution
100% -i
90%
80% -
70% - -
^ 60% -
E?
c
0
§. 50% -
S>
40% -
30% -
20% -
10% -
0% -i—' l—M L+J L+J j I I | i i | 1 1
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249 >250
Particle Size (|jm)
Appendix F_Figures.xls (5_Res)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 6 - MIDVALE SLAG
Lead Speciation Summary Statistics
Mineral
Counts
Total Lib
Avg
Particle Size
Min Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Cerussite
7
7
22
10
45
0.4%
0.4%
0.07%
0.07%
6.6
0.776
3.8%
3.8%
Fe-Pb Oxide
4
4
26
12
45
0.2%
0.2%
0.04%
0.04%
4
0.15
0.3%
0.3%
Galena
2
2
90
80
100
0.1%
0.1%
0.08%
0.08%
7.5
0.866
5.7%
5.7%
Native Lead
67
6
4
1
40
3.4%
0.3%
0.12%
0.04%
11.3
1
15.4%
5.0%
Pb-As Oxide
119
41
16
1
100
6.0%
2.1%
0.82%
0.61%
7.1
0.5
32.6%
24.2%
Lead Oxide
61
29
12
1
55
3.1%
1.5%
0.31%
0.26%
9
0.83
25.9%
21.6%
Slag
1721
1721
131
10
600
86.7%
86.7%
98.52%
98.52%
3.65
0.004
16.0%
16.0%
Sulfosalts
1
1
50
50
50
0.1%
0.1%
0.02%
0.02%
6
0.25
0.4%
0.4%
Fe-Pb Sulfate
2
2
15
15
15
0.1%
0.1%
0.01%
0.01%
3.7
0.14
0.1%
0.1%
TOTAL
1984
1813
115
100.0%
91.4%
100.00%
99.65%
100.0%
77.0%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
6.5%
0.1%
8.4%
0.2%
5-9
1.0%
0.5%
3.5%
2.2%
10-19
3.2%
1.8%
17.7%
8.7%
20-49
4.4%
4.1%
33.7%
29.2%
50-99
20.3%
20.3%
17.7%
17.7%
100-149
28.6%
28.6%
9.4%
9.4%
150-199
18.5%
18.5%
4.0%
4.0%
200-249
12.9%
12.9%
3.8%
3.8%
>250
4.7%
4.7%
1.8%
1.8%
TOTAL
100%
91%
100%
77%
Appendix F_Tables.xls (6_Mid)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 6 - MIDVALE SLAG
Speciation and Particle Size Data
Fe-Pb Sulfate
Sulfosalts
Slag
Lead Oxide
Pb-As Oxide
Native Lead
Galena
Fe-Pb Oxide
Cerussite
Panel A: Relative Lead Mass
~ Frequency of Occurrence
~ Relative Lead Mass
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
100%
90%
80% +
70%
60% +
&•
| 50%
P
40%
30% +
20%
10% +
0%
Panel B: Particle Size Distribution
I 1 . I I
4-
4-
+
J
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249
Particle Size (pm)
>250
Appendix F_Figures.xls (6 Mid)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 6 - BUTTE SOIL
Lead Speciation Summary Statistics
Mineral
Counts
Total Lib
Avq
Particle Size
Min
Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Clay
3
3
58
30
100
0.5%
0.5%
0.82%
0.82%
3.2
0.039
0.1%
0.1%
Anglesite
138
134
12
1
100
21.7%
21.1%
7.51%
7.37%
6.3
0.684
36.2%
35.6%
Cerussite
1
1
10
10
10
0.2%
0.2%
0.05%
0.05%
6.6
0.776
0.3%
0.3%
Fe-Pb Oxide
37
27
61
4
180
5.8%
4.3%
10.48%
8.28%
4
0.15
7.0%
5.6%
Galena
37
35
10
1
55
5.8%
5.5%
1.72%
1.70%
7.5
0.866
12.5%
12.4%
Mn-Pb Oxide
161
150
44
3
200
25.4%
23.6%
32.77%
29.29%
5.1
0.108
20.2%
18.1%
Lead Barite
1
1
5
5
5
0.2%
0.2%
0.02%
0.02%
4.5
0.058
0.0%
0.0%
Lead Phosphate
12
1
54
5
200
1.9%
0.2%
3.03%
0.06%
5.1
0.208
3.6%
0.1%
Fe-Pb Sulfate
245
226
38
2
250
38.6%
35.6%
43.61%
40.55%
3.7
0.111
20.1%
18.6%
TOTAL
635
578
34
100.0%
91.0%
100.00%
88.13%
100.0%
90.7%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
23.0%
22.2%
3.4%
3.3%
5-9
14.8%
13.2%
9.8%
9.5%
10-19
14.0%
12.4%
11.4%
10.7%
20-49
23.0%
21.7%
26.5%
25.8%
50-99
13.7%
11.3%
25.0%
22.1%
100-149
9.0%
8.0%
17.0%
15.1%
150-199
1.6%
1.4%
2.9%
2.9%
200-249
0.8%
0.5%
3.3%
1.5%
>250
0.2%
0.2%
0.6%
0.6%
TOTAL
100%
91%
100%
91%
Appendix F_Tables.xls (6_Butte)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 6 - BUTTE SOIL
Speciation and Particle Size Data
Panel A: Relative Lead Mass
Fe-Pb Sulfate -
Lead Phosphate -
Lead Barite
Mn-Pb Oxide -
Galena -
Fe-Pb Oxide -
Cerussite
Ariglesite -
Clay I
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Panel B: Particle Size Distribution
100% ->
90% --
80% --
70% --
. 60% --
0
1 50%
El
"¦ 40% -
30% --
20% --
10% --
0% -M — — L—I I—J I—
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249 >250
Particle Size (pm)
~ Frequency of Occurrence
~ Relative Lead Mass
Appendix F_Figures.xls (6_Butte)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 7 - CALIFORNIA GULCH PHASE I RESIDENTIAL SOIL
Lead Speciation Summary Statistics
Mineral
Total
Counts
Lib
Avg
Particle Size
Min Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Anglesite
54
28
9
1
45
8.1%
4.2%
2.02%
1.58%
6.3
0.684
10.2%
8.0%
Cerussite
53
33
14
1
125
8.0%
5.0%
3.28%
3.11%
6.6
0.776
19.7%
18.7%
Fe-Pb Sulfate
70
65
31
1
120
10.5%
9.8%
9.59%
9.56%
3.7
0.14
5.8%
5.8%
Mn-Pb Oxide
83
83
43
1
250
12.5%
12.5%
15.77%
15.77%
5
0.24
22.2%
22.2%
Lead Phosphate
150
115
19
1
150
22.6%
17.3%
12.57%
11.96%
5.1
0.4
30.1%
28.6%
Pb-As Oxide
3
0
3
1
5
0.5%
0.0%
0.04%
0.00%
7.1
0.24
0.1%
0.0%
Lead Barite
6
1
18
2
100
0.9%
0.2%
0.48%
0.44%
4.5
0.058
0.1%
0.1%
Fe-Pb Oxide
176
166
52
1
300
26.5%
25.0%
40.45%
40.40%
4
0.031
5.9%
5.9%
PbO-Cerussite
15
0
3
1
10
2.3%
0.0%
0.18%
0.00%
6.6
0.776
1.1%
0.0%
Lead Organic
9
9
78
20
110
1.4%
1.4%
3.08%
3.08%
1.3
0.023
0.1%
0.1%
Galena
19
0
3
1
10
2.9%
0.0%
0.27%
0.00%
7.5
0.866
2.0%
0.0%
Lead Silicate
4
4
30
10
50
0.6%
0.6%
0.53%
0.53%
6
0.5
1.9%
1.9%
Lead Vanidate
1
1
10
10
10
0.2%
0.2%
0.04%
0.04%
6.4
0.32
0.1%
0.1%
Slag
22
22
121
25
250
3.3%
3.3%
11.71%
11.71%
3.65
0.012
0.6%
0.6%
TOTAL
665
527
34
100.0%
79.2%
100.00%
98.18%
100.0%
92.0%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
24.4%
8.3%
5.1%
1.7%
5-9
9.0%
5.0%
5.3%
2.0%
10-19
17.7%
17.3%
11.9%
11.2%
20-49
22.0%
22.0%
22.3%
22.3%
50-99
14.6%
14.4%
22.4%
21.7%
100-149
9.2%
9.2%
27.4%
27.4%
150-199
1.2%
1.2%
3.0%
3.0%
200-249
1.1%
1.1%
0.6%
0.6%
>250
0.9%
0.9%
2.1%
2.1%
TOTAL
100%
79%
100%
92%
Appendix F_Tables.xls (7_Res)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 7 - CALIFORNIA GULCH PHASE I RESIDENTIAL SOIL
Speciation and Particle Size Data
Panel A: Relative Lead Mass
Lead Vanidate
Lead Silicate
Slag
PbO-Cerussite
Fe-Pb Sulfate
Lead Phosphate
Lead Organic
Pb-As Oxide
Mn-Pb Oxide
Galena
Fe-Pb Oxide
Cerussite
Lead Barite
Anglesite
h
?
i
0%
10%
~ Frequency of Occurrence
~ Relative Lead Mass
20%
30%
40%
50%
60%
70%
80%
90% 100%
100%
90%
80% -
70%
60% -
| 50%
P
40%
30% +
20%
10% -
0%
Panel B: Particle Size Distribution
4-
-+-
+
-+-
+-
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249 >250
Particle Size (pm)
Appendix F_Flgures.xls (7_Res)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 7 - CALIFORNIA GULCH Fe/Mn PbO
Lead Speciation Summary Statistics
Mineral
Counts
Total
Lib
Avg
Particle Size
Min Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Lead Barite
7
1
5
2
10
1.8%
0.3%
0.40%
0.10%
4.5
0.05
0.1%
0.0%
Clay
1
1
50
50
50
0.3%
0.3%
0.61%
0.61%
3.1
0.005
0.0%
0.0%
Fe-Pb Oxide
186
186
20
0
130
48.4%
48.4%
44.85%
44.85%
4
0.031
8.4%
8.4%
Mn-Pb Oxide
71
71
45
2
125
18.5%
18.5%
39.14%
39.14%
5.1
0.24
72.1%
72.1%
Lead Organic
2
2
103
80
125
0.5%
0.5%
2.49%
2.49%
1.3
0.0232
0.1%
0.1%
Lead Silicate
1
1
15
15
15
0.3%
0.3%
0.18%
0.18%
6
0.5
0.8%
0.8%
Lead Vanidate
2
2
6
3
8
0.5%
0.5%
0.13%
0.13%
6.4
0.32
0.4%
0.4%
Lead Phosphate
66
64
8
1
60
17.2%
16.7%
6.16%
6.09%
5.1
0.31
14.7%
14.5%
Fe-Pb Sulfate
48
48
10
3
100
12.5%
12.5%
6.03%
6.03%
3.7
0.1
3.4%
3.4%
TOTAL
384
376
21
100.0%
97.9%
100.00%
99.62%
100.0%
99.7%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
25.5%
24.0%
4.0%
3.8%
5-9
19.3%
19.0%
4.8%
4.7%
10-19
24.0%
23.7%
10.9%
10.8%
20-49
17.2%
17.2%
23.4%
23.4%
50-99
10.4%
10.4%
41.7%
41.7%
100-149
3.6%
3.6%
15.3%
15.3%
150-199
0.0%
0.0%
0.0%
0.0%
200-249
0.0%
0.0%
0.0%
0.0%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL
100%
98%
100%
99.7%
Appendix F_Tables.xls (7_FeMn)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 7 - CALIFORNIA GULCH Fe/Mn PbO
Speciation and Particle Size Data
Lead Vanidate
Lead Silicate
Clay
Fe-Pb Sulfate
Lead Phosphate
Lead Organic
Mn-Pb Oxide
Fe-Pb Oxide
Lead Barite
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Panel A: Relative Lead Mass
~ Frequency of Occurrence
~ Relative Lead Mass
Panel B: Particle Size Distribution
100% -i
90%
80% -
70% -
>, 60% -
o
| 50% -
0
"" 40%
30% -
20% -
10% -
0% -I—I —I— 1—I—I 1—I—I —I— 1 I 1 1 H 1 1
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249 >250
Particle Size (Mm)
Appendix F_Figures.xls (7_FeMn)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 8 - CALIFORNIA GULCH AV SLAG
Lead Speciation Summary Statistics
Mineral
Counts
Total Lib
Avg
Particle Size
Min Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Anglesite
3
3
37
30
45
0.2%
0.2%
0.07%
0.07%
6.3
0.684
2.4%
2.4%
Cerussite
3
3
11
8
15
0.2%
0.2%
0.02%
0.02%
6.6
0.776
0.9%
0.9%
Galena
6
1
16
1
80
0.4%
0.1%
0.06%
0.05%
7.5
0.866
3.1%
2.7%
Native Lead
4
1
6
2
15
0.2%
0.1%
0.02%
0.01%
11.34
1
1.4%
0.9%
Pb-As Oxide
253
34
8
1
125
15.6%
2.1%
1.30%
0.90%
6
0.5
30.9%
21.4%
Lead Oxide
139
18
8
1
125
8.6%
1.1%
0.73%
0.59%
9.5
0.930
51.0%
41.5%
Slag
1206
1206
126
5
450
74.5%
74.5%
97.68%
97.68%
3.65
0.0035
9.9%
9.9%
Fe-Pb Sulfate
5
1
37
10
55
0.3%
0.1%
0.12%
0.04%
3.7
0.091
0.3%
0.1%
TOTAL
1619
1267
96
100.0%
78.3%
100.00%
99.36%
100.0%
79.6%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
19.1%
0.1%
11.3%
0.1%
5-9
8.5%
6.9%
4.7%
0.6%
10-19
8.2%
7.4%
6.8%
4.4%
20-49
5.0%
4.6%
23.5%
20.9%
50-99
8.6%
8.6%
24.2%
24.2%
100-149
19.2%
19.2%
22.4%
22.4%
150-199
10.1%
10.1%
1.7%
1.7%
200-249
12.8%
12.8%
2.7%
2.7%
>250
8.6%
8.6%
2.7%
2.7%
TOTAL
100%
78%
100%
80%
Appendix F_Tables.xls (8_AV)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 8 - CALIFORNIA GULCH AV SLAG
Speciation and Particle Size Data
Panel A: Relative Lead Mass
Slag
Fe-Pb Sulfate
Native Lead
Pb-As Oxide
Galena
Lead Oxide
Cerussite
Ariglesite
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Panel B: Particle Size Distribution
100%
90%
80% -
70% -
>, 60%
o
| 50% -
CD
"" 40%
30% -
20% -
10%
0%
4-
+
+
4-
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249
Particle Size (Mm)
>250
Appendix F_Figures.xls (8_AV)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 9 - PALMERTON LOCATION 2
Lead Speciation Summary Statistics
Mineral
Counts
Total
Lib
Avq
Particle Size
Min
Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Clay
1
1
10
10
10
0.9%
0.9%
0.6%
0.6%
3.1
0.005
0.0%
0.0%
Anglesite
2
2
4
3
4
1.8%
1.8%
0.4%
0.4%
6.3
0.684
6.0%
6.0%
Lead Barite
11
11
8
1
41
9.6%
9.6%
5.0%
5.0%
4.5
0.018
1.4%
1.4%
Fe-Pb oxide
15
15
8
3
20
13.2%
13.2%
7.4%
7.4%
4
0.015
1.5%
1.5%
Mn-Pb Oxide
68
68
17
2
100
59.6%
59.6%
68.8%
68.8%
5.1
0.055
66.1%
66.1%
Lead Phosphate
16
16
19
1
45
14.0%
14.0%
17.4%
17.4%
5.1
0.08
24.4%
24.4%
Fe-Pb Sulfate
1
1
8
8
8
0.9%
0.9%
0.5%
0.5%
3.7
0.1
0.6%
0.6%
TOTAL
114
114
11
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
26.3%
26.3%
10.8%
10.8%
5-9
22.8%
22.8%
5.4%
5.4%
10-19
25.4%
25.4%
16.7%
16.7%
20-49
18.4%
18.4%
27.6%
27.6%
50-99
6.1%
6.1%
32.4%
32.4%
100-149
0.9%
0.9%
7.1%
7.1%
150-199
0.0%
0.0%
0.0%
0.0%
200-249
0.0%
0.0%
0.0%
0.0%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL
100%
100%
100%
100%
Appendix F_Tables.xls (9_P2)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 9 - PALMERTON LOCATION 2
Speciation and Particle Size Data
Panel A: Relative Lead Mass
Fe-Pb Sulfate
Lead Phosphate
Mn-Pb Oxide
Fe-Pb Oxide
Lead Barite
Anglesite
Clay
~ Frequency of Occurrence
~ Relative Lead Mass
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
cr
s
100%
90% --
80% --
70%
60%
50% --
40% --
30%
20%
10% +
0%
Panel B: Particle Size Distribution
-h
4-
+
+
+
-4-
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249 >250
Particle Size (|am)
Appendix F_Figures.xls (9_P2)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 9 - PALMERTON LOCATION 4
Lead Speciation Summary Statistics
Mineral
Counts
Total
Lib
Avq
Particle Size
Min
Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Clay
3
3
24
8
45
2.6%
2.6%
2.90%
2.90%
3.1
0.005
0.1%
0.1%
Anglesite
8
0
1
1
1
6.8%
0.0%
0.32%
0.00%
6.3
0.684
4.0%
0.0%
Lead Barite
1
1
12
12
12
0.9%
0.9%
0.48%
0.48%
4.5
0.018
0.1%
0.1%
Fe-Pb Oxide
14
14
16
8
40
12.0%
12.0%
9.02%
9.02%
4
0.015
1.6%
1.6%
Mn-Pb Oxide
65
65
31
4
110
55.6%
55.6%
80.82%
80.82%
5.1
0.055
65.8%
65.8%
Pb-As Oxide
17
0
1
1
1
14.5%
0.0%
0.68%
0.00%
7.1
0.5
7.0%
0.0%
Lead Silicate
1
1
4
4
4
0.9%
0.9%
0.16%
0.16%
6
0.5
1.4%
1.4%
Lead Vanidate
5
5
15
5
35
4.3%
4.3%
2.98%
2.98%
6.4
0.32
17.7%
17.7%
Lead Phosphate
1
1
15
15
15
0.9%
0.9%
0.60%
0.60%
5.1
0.08
0.7%
0.7%
Zn-Pb Silicate
2
2
26
12
40
1.7%
1.7%
2.07%
2.07%
5.5
0.05
1.6%
1.6%
TOTAL
117
92
15
100.0%
78.6%
100.0%
99.0%
100.0%
89.1%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
24.8%
3.4%
12.7%
1.8%
5-9
14.5%
14.5%
5.0%
5.0%
10-19
21.4%
21.4%
8.8%
8.8%
20-49
24.8%
24.8%
34.4%
34.4%
50-99
12.8%
12.8%
32.3%
32.3%
100-149
1.7%
1.7%
6.8%
6.8%
150-199
0.0%
0.0%
0.0%
0.0%
200-249
0.0%
0.0%
0.0%
0.0%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL
100%
79%
100%
89%
Appendix F_Tables.xls (9_P4)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 9 - PALMERTON LOCATION 4
Speciation and Particle Size Data
Panel A: Relative Lead Mass
Zri-Pb Silicate
Lead Phosphate
Lead Vanidate
Lead Silicate
Pb-As Oxide
Mn-Pb Oxide
Fe-Pb Oxide
Lead Barite
Anglesite
Clay
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
100%
90%
80%
70%
, 60%
ET
c
0
= 50%
a?
^ 40%
30%
20%
10%
0%
Appendix F_Figures.xls (9_P4)
]
r
3
]
~ Frequency of Occurrence
~ Relative Lead Mass
Panel B: Particle Size Distribution
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249 >250
Particle Size (urn)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 11 - MURRAY SMELTER SOIL
Lead Speciation Summary Statistics
Mineral
Total
Counts
Lib
Avg
Particle Size
Min Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
As(M)0
1
1
3
3
3
0.2%
0.2%
0.02%
0.02%
6.5
0.005
0.0%
0.0%
Cerussite
7
6
14
5
40
1.6%
1.4%
0.66%
0.38%
6.3
0.684
14.0%
8.2%
Fe-Pb Oxide
4
4
8
8
8
0.9%
0.9%
0.22%
0.22%
4
0.031
0.1%
0.1%
Galena
55
1
2
1
30
12.9%
0.2%
0.62%
0.21%
7.5
0.866
20.0%
6.6%
Pb-As Oxide
44
16
5
1
55
10.3%
3.7%
1.59%
1.22%
7.1
0.527
29.4%
22.4%
Pb(M)0
6
4
7
2
15
1.4%
0.9%
0.27%
0.18%
7
0.3
2.8%
1.8%
Lead Oxide
10
8
9
2
25
2.3%
1.9%
0.61%
0.56%
9.5
0.93
26.6%
24.2%
Slag
299
299
47
5
310
70.0%
70.0%
95.76%
95.76%
3.65
0.0037
6.4%
6.4%
Fe-Pb Sulfate
1
1
35
35
35
0.2%
0.2%
0.24%
0.24%
3.7
0.14
0.6%
0.6%
TOTAL
427
340
34
100.0%
79.6%
100.00%
98.78%
100.0%
70.4%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
22.7%
3.3%
26.5%
5.2%
5-9
10.3%
9.8%
10.6%
8.8%
10-19
29.3%
29.0%
17.6%
16.9%
20-49
17.1%
16.9%
33.4%
27.5%
50-99
5.9%
5.9%
7.8%
7.8%
100-149
8.4%
8.4%
1.8%
1.8%
150-199
2.8%
2.8%
0.8%
0.8%
200-249
2.6%
2.6%
1.0%
1.0%
>250
0.9%
0.9%
0.5%
0.5%
TOTAL
100%
80%
100%
70%
Appendix F_Tables.xls (11_Mur)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 11 - MURRAY SMELTER SOIL
Speciation and Particle Size Data
Fe-Pb Sulfate
Slag
Lead Oxide
Pb(M)0
Pb-As Oxide
Galena
Fe-Pb Oxide
Cerussite
As(M)0
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Panel A: Relative Lead Mass
Panel B: Particle Size Distribution
100%
90% -
80%
70% -
60%
50% -
40% -
30%
20% -
10%
0%
<5
+
4-
X
=L-bJ
X
5-9
10-19
20-49 50-99 100-149 150-199 200-249
Particle Size ((jm)
>250
'This mineral is now considered to be equivalent to Fe-Pb Oxid
Appendix F_Figures.xls (11_Mur)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 11 - NIST PAINT
Lead Speciation Summary Statistics
Mineral
Counts
Total
Lib
Avq
Particle Size
Min
Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Anglesite
3
3
7
4
12
1.1%
1.1%
0.87%
0.87%
6.3
0.684
0.6%
0.6%
Cerussite
183
183
9
1
110
66.8%
66.8%
67.80%
67.80%
6.6
0.776
55.3%
55.3%
Lead Oxide
88
88
9
1
80
32.1%
32.1%
31.32%
31.32%
9.5
0.93
44.1%
44.1%
TOTAL
274
274
9
100.0%
100.0%
100.00%
100.00%
100.0%
100.0%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
75.5%
75.5%
15.0%
15.0%
5-9
4.4%
4.4%
3.1%
3.1%
10-19
5.8%
5.8%
6.4%
6.4%
20-49
8.0%
8.0%
27.8%
27.8%
50-99
5.8%
5.8%
43.9%
43.9%
100-149
0.4%
0.4%
3.7%
3.7%
150-199
0.0%
0.0%
0.0%
0.0%
200-249
0.0%
0.0%
0.0%
0.0%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL
100%
100%
100%
100%
Appendix F_Tables.xls (11_NIST)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 11 - NIST PAINT
Speciation and Particle Size Data
Panel A: Relative Lead Mass
~ Frequency of Occurrence
~ Relative Lead Mass
Lead Oxide
Cerussite
Anglesite
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Panel B: Particle Size Distribution
100% -i
90% -
80% -
70% -
^ 60% -
e
0
= 50% -
2
^ 40% -
30% -
20% -
10% -
0% -I—' I I L-M —I 1 1 1
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249 >250
Particle Size (|jm)
Appendix F_Figures.xls (11_NIST)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 12 - GALENA-ENRICHED SOIL
Lead Speciation Summary Statistics
Mineral
Counts
Total Lib
Avq
Particle Size
Min
Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Galena
224 224
17
1
80
100.0% 100.0%
100.00% 100.00%
7.5
0.866
100.0% 100.0%
TOTAL 224 224 17 100.0% 100.0% 100.00% 100.00% 100.0% 100.0%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
47.8%
47.8%
4.9%
4.9%
5-9
2.2%
2.2%
0.7%
0.7%
10-19
4.5%
4.5%
3.3%
3.3%
20-49
41.1%
41.1%
75.9%
75.9%
50-99
4.5%
4.5%
15.3%
15.3%
100-149
0.0%
0.0%
0.0%
0.0%
150-199
0.0%
0.0%
0.0%
0.0%
200-249
0.0%
0.0%
0.0%
0.0%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL
100%
100%
100%
100%
Appendix F_Tables.xls (12_Gal)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 12 - GALENA-ENRICHED SOIL
Speciation and Particle Size Data
Galena
0%
Panel A: Relative Lead Mass
~ Frequency of Occurrence
a Relative Lead Mass
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
100%
90% +
80%
70%
60% +
&•
| 50%
P
40% +
30%
20%
10% +
0%
Panel B: Particle Size Distribution
i i
jZZl
m
+
+
+
<5 5-9 10-19 20-49 50-99 100-149 150-199 200-249
Particle Size (pm)
>250
Appendix F_Figures.xls (12_Gal)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 12 - CALIFORNIA GULCH OREGON GULCH TAILINGS
Lead Speciation Summary Statistics
Mineral
Counts
Total Lib
Particle Size
Avg Min Max
Count Freq (%)
Total Lib
LW Freq (%)
Total Lib
Density
Lead
Fraction
Relative Lead Mass (%)
Total Lib
Galena
217 4
2 1 25
100.0% 1.8%
100.00% 5.14%
7.5
0.866
100.0% 5.1%
TOTAL 217 4 2 100.0% 1.8% 100.00% 5.14% 100.0% 5.1%
Particle Size Distribution
Size
Total Freq
Lib Freq
Total RLM
Lib RLM
<5
85.3%
0.9%
46.8%
1.2%
5-9
8.3%
0.0%
21.5%
0.0%
10-19
6.0%
0.9%
26.7%
4.0%
20-49
0.5%
0.0%
4.9%
0.0%
50-99
0.0%
0.0%
0.0%
0.0%
100-149
0.0%
0.0%
0.0%
0.0%
150-199
0.0%
0.0%
0.0%
0.0%
200-249
0.0%
0.0%
0.0%
0.0%
>250
0.0%
0.0%
0.0%
0.0%
TOTAL 100% 2% 100% 5%
Appendix F_Tables.xls (12_Or)
-------
DRAFT- Do Not Cite, Quote, or Release
APPENDIX F
EXPERIMENT 12 - CALIFORNIA GULCH OREGON GULCH TAILINGS
Speciation and Particle Size Data
Panel A: Relative Lead Mass
~ Frequency of Occurrence
~ Relative Lead Mass
Galena
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Panel B: Particle Size Distribution
100%
90% --
80% --
70%
>, 60% +
o
| 50%
JD
"" 40%
30% +
20%
10%
0%
4-
+
4-
4-
-+-
<5
5-9
10-19
20-49 50-99 100-149 150-199 200-249
Particle Size (pm)
>250
Appendix F_Figures.xls (12_Or)
-------
DRAFT- Do Not Cite, Quote, or Release
This page intentionally left blank to facilitate double-sided printing.
------- |