RISK ASSESSMENT FOR
SECTION 403 ANALYSES
Draft Report
VOLUME II
APPENDICES A-E
Prepared by
Battelle
505 King Avenue
Columbus, Ohio 43201
for
U.S. Environmental Protection Agency
EPA Contract No. 68-D5-0008
Task 2-10
Chemical Management Division
Office of Pollution Prevention and Toxics
Office of Prevention, Pesticides, and Toxic Substances
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
APPENDIX A
Glossary
Draft - Do Not Cite or Quote A-1 September 27. 1996
-------
APPENDIX A
GLOSSARY
Abatement: The term "abatement" means any set of measures designed to permanently eliminate lead-
based paint hazards in accordance with standards established by Federal agencies.
Accessible or Chewable Surface: The term "accessible surface" means an interior or exterior surface
painted with lead-based paint that is accessible for a young child to mouth or chew.
Arithmetic Mean; The sum of a set of measurements divided by the number of measurements.
Biokinetics: Processes affecting the movement of molecules from one internal body compartment to
another, including elimination from the body.
Blood-Lead Concentration; Blood-lead concentration measures the mass of lead collected per volume of
whole blood collected and is usually expressed in terms of micrograms of lead collected per deciliter of
blood collected (ng Pb/dL blood).
Blue Nozzle Sampler; The term "blue nozzle sampler" refers to the vacuum sampler used with HUD
National Survey and the Baltimore R&M Pilot study.
BRM Sampler; The term "BRM sampler" refers to a device used for the collection of dust over a
specified area using a modified HVS-3 vacuum sampler. This vacuum was initially developed and utilized
in EPA's Baltimore Repair and Maintenance Study.
Deteriorated Paint: The term "deteriorated paint" means any interior or exterior paint that is peeling,
chipping, chalking or cracking or any paint located on an interior or exterior surface or fixture that is
damaged or deteriorated.
Dripline Soil; The term "dripline soil" means any soil sample collected from the drip line area about the
residence. This is usually approximately 1-3 feet from the side (e.g. foundation) of the house, under the
eaves.
Drv Room: (see Wet Room).
Dust-lead concentration; Dust-lead concentration measures the mass of lead collected per mass of dust
collected and is usually stated in terms of micrograms of lead collected per gram of dust collected (ug Pb/g
dust).
Dust-lead loading; Dust-lead loading measures the mass of lead collected per surface area sampled and is
usually expressed in terms of micrograms of lead collected per square foot sampled (ug Pb/ft2).
DVM Sampler; The term "DVM sampler" refers to a device used to collect dust samples using a vacuum
(personal air sampler) operating at a rate of two liters of air per minute.
Draft - Do Not Cite or Quote A-2 September 27, 1996
-------
Efficacy; Refers to the effectiveness of a method of abatement and is defined as the generalized
evaluation of several key factors including the usability of a method, its hazard abatement effectiveness,
and the the amount of hazardous dust lead generated by a method, measured by air and post-cleanup wipe
samples.
Encapsulation; A method of "abatement" that involves the coating and sealing of surfaces with durable
coatings formulated to be elastic, long-lasting (e.g., at least 20 years), and resistant to cracking, peeling,
algae, and fungus.
Enclosure; The resurfacing or covering of surfaces by sealing or caulking them with mechanically
affixed, durable materials so as to prevent or control chalking, flaking, lead-containing substances from
being part of house dust or accessible to children.
Entrvwav Soil; The term "entryway soil" means any soil sample collected immediately adjacent to the
entryway of the residence.
EPI Model; A statistical regression model developed from data collected by an epidemiological study.
The resulting model which predicts blood-lead concentration as a function of environmental lead levels
may be used to predict a national distribution of children's blood-lead levels.
EPI Study; A targeted epidemiology study which measures both children's blood-lead concentrations and
environmental lead levels as well as other factors (e.g., behavioral, demographic) influencing a child's
blood-lead level.
Exposure; Contact between a chemical, physical, or biological agent (e.g., lead) with the outer boundary
of an organism (e.g., a child's skin). Exposure is quantified as the concentration of the agent in the
medium in contact integrated over the time duration of that contact.
Exposure Pathway; The physical course a chemical or pollutant takes from its source to the organism
exposed.
Exposure Route; The manner by which a chemical or pollutant enters an organism after contact (e.g., by
ingestion, inhalation).
Friction Surface; The term "friction surface" means an interior or exterior surface that is subject to
abrasion or friction, including certain window, floor, and stair surfaces.
Geometric Mean; The n* root product of n values. Also, the anti-log of the "arithmetic mean" of a set of
n natural log-transformed values.
Geometric Standard Deviation fGSDl: The anti-log of the "standard deviation" of a set of n natural log-
transformed values.
HEPA; A High Efficiency Paniculate Accumulator vacuum fitted with a filter capable of filtering out
particles of 0.3 microns or greater from a body of air at 99.97 percent efficiency or greater.
IEUBK Model; EPA's Integrated Exposure Uptake Biokinetic Model for Lead is designed to model
exposure from lead in air, water, soil, dust, diet, and paint and other sources with pharmacokinetic
modeling to predict blood-lead levels in children 6 months to 7 years of age.
Draft - Do Not Cite or Quote A-3 September 27. 1996
-------
Impact Surface: The term "impact surface" means an interior or exterior surface that is subject to damage
by repeated impacts, for example, certain parts of door frames.
Interim Controls; The term "interim controls" means a set of measures designed to temporarily reduce
human exposure or likely human exposure to lead-based paint hazards, including specialized cleaning,
repairs, maintenance, painting, temporary containment, ongoing monitoring of lead-based paint hazards or
potential hazards, and the establishment and operation of management and resident education programs.
Lead-Contaminated Dust; The term "lead-contaminated dust" means surface dust in residential
dwellings that contains an area or mass concentration of lead in excess of levels determined by EPA to
pose a threat of adverse health effects in pregnant women or young children.
Lead-Based Paint; Lead-based paint is dried paint film that has a lead content exceeding 1.0 mg/cm2 or
0.5 percent (5,000 parts per million (ppm)) by weight.
Lead-Based Paint Hazard: The term "lead-based paint hazard" means any condition that causes
exposure to lead from lead-contaminated dust, lead-contaminated soil, lead-contaminated paint that is
deteriorated or present in accessible surfaces, friction surfaces, or impact surfaces that would result in
adverse human health effects as established by EPA.
Lead-Contaminated Soil; The term "lead-contaminated soil" means bare soil on residential real property
that contains lead at or in excess of the levels determined to be hazardous to human health by EPA.
Percentile; The term percentile refers to a particular value in a set or distribution of numbers that has the
property that a specified percentage of the numbers are less than the given value. For instance, the Sth
percentile of a set of blood-lead concentrations is the blood-lead concentration value such that 5% of the
numbers are less than the value and 95% are greater than it. The 50th percentile is also known as the
median.
Perimeter Soil: The term "perimeter soil" usually means any soil sample collected from the perimeter or
remote areas of the residence's yard. (Note: in the Rochester Lead-in-Dust study, this terminology referred
to samples collected adjacent to the foundation).
Pharmacokinetics: The study of the time course of absorption, distribution, metabloism, and excretion of
a foreign substance (e.g., a drug or pollutant) in an organism's body.
PICA; The tendency to mouth or attempt to consume non-food objects.
Plavard Soil; The term "playard soil" refers to any soil sample collected at the site where the child
usually played. In the National Survey, this was frequently a local playground. In other studies, this refers
to an exterior site at the residence.
Probability Samples: Samples selected from a statistical population such that each sample has a known
probability of being selected.
Random Samples; Samples selected from a statistical population such that each sample has an equal
probability of being selected.
Draft - Do Not Cite or Quote A-4 September 27, 1996
-------
Reduction; The term reduction means measures designed to reduce or eliminate human exposure to lead-
based paint hazards through methods including interim contrls and abatement.
Regression Model; A statistical representation of the relationship between a dependent variable such as
blood-lead concentration to one or more independent variables such as environmental lead exposures. For
example, a regression model could indicate that blood-lead concentration is an additive function of
environmental lead levels.
Removal and Replacement; A method of abatement that entails removing substrates such as windows,
doors, trim, or soil that have lead-contaminated surfaces and installing new (and presumably lead-free) or
deleaded components.
Risk: The probability of deleterious health or environmental effects.
Sample; A small part of something designed to show the nature or quality of the whole. Exposure-related
measurements are usually samples of environmental or ambient media, exposures of a small subset of a
population for a short time, or biological samples, all for the purpose of inferring the nature and quality of
parameters important to evaluating exposure.
Soil-Lead Concentration; Soil-lead concentration measures the mass of lead collected per mass of soil
collected and is usually stated in terms of micrograms of lead collected per gram of soil collected (ug Pb/g
soil). These units are also sometimes referred to as parts per million (ppm).
Standard Deviation; A measure of the dispersion of a set of values that is the square root of the
"arithmetic mean" of the squares of the deviation of each value from the "arithmetic mean" of the values.
Target Housing; The term "target housing" means any housing constructed prior to 1978, except for
housing of the elderly or persons with disabilities (unless any child who is less than 6 years of age resides
or is expected to reside in such housing for the elderly or persons with disabilities), or any 0-bedroom
dwelling.
u. Microgram: A microgram is 1/1,000,000 of a gram or 1/1,000 of a milligram.
Uptake; The process by which a substance is absorbed into the body.
Vacuum Sample; The term "vacuum sample" refers to collecting dust over a specified area by
vacuuming the area. The contents of the vacuum bag are then analyzed for the amount of dust and the
amount of lead. Results from vacuum sampling are in the form of "dust-lead loadings" and "dust-lead
concentrations".
XRF: "X-ray fluorescence" is a principle used by instruments to determine the lead concentration in
substances, usually in milligrams per square centimeter (mg/cm2).
Wet Room; An interior room in a house which is either a kitchen, bathroom, laundry, or utility room is
classified as a 'wet room', otherwise the room may be classified as a 'dry' room.
Window Sill; The term "window sill" is defined as the horizontal board inside the window.
Draft - Do Not Cite or Quote A-5 September 27, 1996
-------
Window Trough; The term "window trough" is defined as the surface below the window sash and inside
the screen and/or storm window. This is also sometimes referred to as a window well.
Wipe Sample: The term "wipe sample" refers to collecting dust over a specified area by wiping the area
with a moist cloth. The cloth and the dust on the cloth are then analyzed for the amount of lead. Results
from wipe sampling are in the form of dust-lead loadings.
Draft - Do Not Cite or Quote A-6 September 27. 1936
-------
APPENDIX B
Health Effects Associated with Exposure to
Lead and Internal Lead Doses in Humans
Draft - Do Not Cite or Quote B-1 September 27, 1996
-------
I
\
Table B-1. Health Effects Associated with Exposure to Lead and Internal Lead Doses in Humans
Duration of
Exposure
-< 1 yr (occup)
NS (occup)
•< 3 yr (occup)
NS
2 wk - >• 1 yr
(occup)
>• 1 yr (occup)
>• 1 yr (occup)
NS
(general population)
NS
(general population)
NS
(general population)
NS
(general population)
System
Cardiovascular
Cardiovascular
Cardiovascular
Cardiovascular
Cardiovascular
Cardiovascular
Cardiovascular
Effect
Increase in death due to hypertension,
nephritis, neoplasms
Increase in death due to
cerebrovascular disease, nephritis,
and/or nephrosis
No increase in deaths
Acute encephalopathy resulting in
death in children
Increased blood pressure
No effect on blood pressure
Ischemic electrocardiogram changes
Increased blood pressure
Increased systolic pressure by 1 -2
mmHg and increased diastolic pressure
by 1 .4 mmHg with every doubling in
blood-lead level; effect most
prominent in middle-aged white men
No significant correlation between
blood pressure and blood-lead levels
Degenerative changes in myocardium,
electrocardiogram abnormalities in
children
Blood Lead Levels
at which Effect is
Observed (pg/dL)
63-80
NS
34-58 (means)
125-750
t 30- 120
40 (mean)
51 (mean)
44.9 (mean)
7-38
6- 13 (median)
orNS
6-20
Reference
Cooper et al., 1985, 1988
Fanning 1 988; Malcolm and Barnett
1982; Michaels et al. 1991
Gerhardsson et al. 1986b
NAS 1972
deKort et al. 1 987; Pollock and Ibels
1986; Marino et al. 1989; Weiss et al.
1986, 1988
Parkinson et al. 1 987
Kirkby and Gyntelberg 1985
Kheraetal. 1980b
Coate and Fowles 1989; Harlan 1988;
Harlan et al. 1988; Landis and Flegal
1988; Pirkle et al. 1985; Schwartz 1988
Elwood et al. 1 988; Grandjean et al.
1 989; Neri et al. 1 988; Staessen et al.
1990, 1991
Silver and Rodriguez-Torres 1 968
L?
I
3
f
Nl
•M
O)
-------
Table B-1. Health Effects Associated with Exposure to Lead and Internal Lead Doses in Humans (Continued)
Duration of
Exposure
NS (acute) (occup)
NS (acute)
(general population)
NS (occup)
NS
(general population)
NS (occup)
NS
(general population)
NS
(general population)
NS
(general population)
NS
(general population
NS
(general population)
System
Gastrointestinal
Gastrointestinal
Hematological
Hematological
Hematological
Hematological
Hematological
Hematological
Hematological
Hematological
Effect
Colic (abdominal pain, constipation,
cramps, nausea, vomiting, anorexia,
weight loss)
Colic in children
Increased ALAS and/or decreased
ALAD
Decreased ALAD
Increased urinary or blood ALA
Increased urinary ALA
Increased FEP
Increased EP
Increased ZPP
Increased urinary copmporphyrin
Blood Lead Levels
at which Effect is
Observed (pg/dU
40-200
60-100
87 or NS
(correlated with
blood-lead level)
3-56 (adult)
No threshold
(children
•c 40-50, 87
(mean) or NS
> 35 (adult)
25-75 children
t 25-35
30-40 (males
20-30 (females)
t 1 5 (children)
t. 35 (children)
t 40 (adults)
Reference
Awad et al. 1986; Baker et al. 1979;
Haenninen et al. 1 979; Holness and
Nethercott 1988; Kumar et al 1987;
Marino et al. 1989; Matte et al. 1989;
Muijser et al. 1987; Pagliuca et al. 1990;
Pollock and Ibels 1986; Schneitzer et al.
1990
U.S. EPA 1986; NAS 1972
Alessio et al. 1976; Meredith et al. 1978;
Wadaetal. 1973
Chisholm et al. 1985; Hernberg and
Nikkanen 1970; Lauwerys et al. 1978;
Roels et al. 1976; Roels and Lauwerys
1987; Secchietal. 1974
Lauwerys et al. 1 974; Meredith et al.
1978; Pollock and Ibels 1986; Selander
and Cramer 1 970
NAS 1 972; Roels and Lauwerys 1 987
Grandjean and Lintrup 1 978; Roels et al.
1975
Roels and Lauwerys 1987; Roels et al.
1975, 1976, 1979; Stuick 1974
Hammond et al. 1 985; Piomelli et al.
1 982; Rabinowitz et al. 1 986; Roels and
Lauwerys 1987; Roels et al. 1976
U.S. EPA 1986
9
u
CO
I
NO
N
-------
s
Table B-1. Health Effects Associated with Exposure to Lead and Internal Lead Doses in Humans (Continued)
Duration of
Exposure
NS (occup)
NS
(general population)
NS
(general population)
NS (occup)
NS
(general population)
NS (acute)
(general population)
NS (chronic) (occup)
1 -30 yr (occup)
NS (chronic)
(general population)
NS (acute)
(general population)
0.1-20 yr (chronic)
(occup)
NS (chronic)
(general population)
NS
(general population
System
Hematological
Hematological
Hematological
Hematological
Hematological
Hepatic
Renal
Renal
Renal
Renal
Other
Other
Other
Effect
Decreased hemoglobin with or without
basophilic stippling of erythrocytes
Decreased hemoglobin
Anemia (hematocrit of « 35%)
Decreased Py-5'-N
Decreased Py-5'-N
Decreased mixed function oxidase
activity
Chronic Nephropathy
No effect on renal function
Renal (impairment with gout or
hypertension)
Aminoaciduria; Fancoi syndrome
Decreased thyroxin (T4)
No effect on thyroid function in
children
Negative correlation between blood
lead and serum 1,25-dihydroxyvitamin
D in children
Blood Lead Levels
at which Effect is
Observed (//g/dL)
2:40
t 40 (children)
>• 20 (children)
NS
7-80 (children)
NS (children)
40- > 100
40-61
18-26/yg/dL
>• 80 (children)
t 56
2-77 (levels
measured)
12-120
Reference
Awad et al. 1986; Baker et al. 1979;
Grandjean 1979; Lilis et al. 1978;
Pagliuca et al. 1990; Tola et al. 1973;
Wadaetal. 1973
Adebonojo 1974; Betts et al. 1973;
Pueschel et al. 1972; Rosen et al. 1974
Schwartz et al. 1 990
Buc and Kaplan 1978; Paglia et al. 1975,
1977
Angle and Mclntire 1 978; Angle et al.
1982
Alvares et al. 1975; Saenger et al. 1984
Biagini et al. 1977; Cramer et al. 1974;
Lilis et al. 1968; Maranelli and Apostoli
1987; Ong et al. 1987; Pollock and Ibels
1986; Verschoor et al. 1987; Wedeen et
al. 1979
Buchet et al. 1 980; Huang et al. 1 988a
Batumen et al. 1981, 1983
Chisholm 1962; Pueschel et al. 1972
Tuppurainen et al. 1 988
Siegel et al. 1 989
Mahaffey et al. 1982; Rosen et al. 1980
I
Kj
^*
0>
-------
Table B-1. Health Effects Associated with Exposure to Lead and Internal Lead Doses in Humans (Continued)
Duration of
Exposure
NS (chronic)
(general population)
NS (chronic)
(general population)
NS (chronic)
(general population)
< 1 8 yr (occup)
NS (acute)
NS
(acute and chronic)
(occup)
NS
(occup)
NS
(occup)
NS
(occup)
System
Other
Other
Other
Immunological
Neurological
Neurological
Neurological
Neurological
Neurological
Effect
No effect on vitamin D metabolism in
children
Growth retardation in children
No association between blood-lead
levels and growth in children
Depression of cellular immune
function, but no effect on humoral
immune function
Encephalopathy (adults)
Neurological signs and symptoms in
adults including malaise, forgetfulness,
irritability, lethargy, headache, fatigue,
impotence, decreased libido, dizziness,
weakness, paresthesia
Neurobehavioral function in adults;
disturbances in oculomotor function,
reaction time, visual motor
performance, hand dexterity, IQ test
and cognitive performance,
nervousness, mood, coping ability,
memory
No effect on neurobehavioral function
in adults
Peripheral nerve function in adults;
decreased nerve conduction velocity
Blood Lead Levels
at which Effect is
Observed (/ig/dL)
5-24 (levels
measured)
t 30-60; Tooth
lead >• 18.7/yg/g
10-47 (levels
measured)
21-90
50 - >• 300
40-80
40-80
40-60
(levels measured)
30- fc 70
Reference
Kooetal. 1991
Angle and Kuntzelman 1 989; Lauwers et
al. 1986; Lyngbye et al. 1987
Greene and Ernhart 1 991 ; Sachs and
Moel 1989
Alomran and Shleamoon 1 988; Ewers et
al. 1982
Kehoe 1961; Kumar et al. 1987; Smith et
al. 1938
Awad et al. 1986; Baker et al. 1979;
Campara et al. 1984; Haenninen et al.
1979; Holness and Nethercott 1988;
Marino et al. 1 989; Matte et al. 1 989;
Pagliuca et al. 1990; Parkinson et al.
1986; Pasternak et al. 1989; Pollock and
Ibels 1 986; Schneitzer et al. 1 990;
Zimmerman-Tansella et al. 1 983
Arnvig et al. 1 980; Baker et al. 1 983;
Baloh et al. 1 979; Campara et al. 1 984;
Glickman et al. 1984; Haenninen et al.
1978; Hogstedt et al. 1983; Mantere et
al. 1982; Spivey et al. 1980; Stollery et
al. 1989; Valciukas et al. 1978;
Williamson and Teo 1 986
Milburn et al. 1 976; Ryan et al. 1 987
Araki et al.1980; Muijser et al. 1987;
Rosen et al. 1983; Seppalainen et al.
1983; Triebiget al. 1984
I
5
01
I
0>
-------
Table B-1. Health Effects Associated with Exposure to Lead and Internal Lead Doses in Humans (Continued)
Duration of
Exposure
NS
(occup)
NS
(general population)
NS
(general population)
NS
(general population)
NS
(general population)
NS
(general population)
NS
(general population)
prenatal
(general population)
System
Neurological
Neurological
Neurological
Neurological
Neurological
Neurological
Neurological
Developmental
Effect
No effect on peripheral nerve function
Neurological signs and symptoms in
children and encephalopathy
Neurobehavioral function in children:
lower IQS and other neuropsychologic
deficits
Neurobehavioral function in children:
slightly decreased performance on IQ
tests and other measures of
neuropsychological function
No correlation between blood-lead
levels and permanent effects on
neurobehavioral development in
children
Decrease in hearing acuity in children
Alterations in peripheral nerve function
in children
Decreased growth rate
Blood Lead Levels
at which Effect is
Observed U/g/dL)
60-80
(levels measured)
60-450 (effects
other than
encephalopathy);
> 80-800
(encephalopathy)
40-200
Tooth lead:
6 : > 30 /yg/g
Blood lead: 6-60
10-15
4-60
20-30
7.7
Reference
Spivey et al. 1980
Bradley and Baumgartner 1958; Bradley
et al. 1956; Chisolm 1962, 1965;
Chisolm and Harrison 1956; Gant 1938;
Rummo et al. 1979; Smith et al. 1983
dela Burtde and Choate 1972, 1975;
Ernhart et al. 1981; Kotok 1972; Kotok
et al. 1977; Rummo et al. 1979
Bellinger and Needleman 1 983; Bergomi
et al. 1989; Fulton et al. 1987; Hansen
et al. 1989; Hawk et al. 1986;
Needleman et al. 1979, 1985, 1990;
Schroeder et al. 1 985; Schroeder and
Hawk 1987; Silva et al. 198; Wang et al.
1989
Cooney et al. 1989; Harvey et al. 1984,
1988; Lansdown et al. 1986; McBride et
al. 1 982; Ernhart and Greene, 1 990;
Dietrich et al. 1987a; Bellinger et al.
1989; McMichael et al. 1986; Pocock et
al. 1989; Smith et al. 1983; Winneke et
al. 1984
Robinson et al. 1985; Schwartz and Otto
1987
Erenberg et al. 1 974; Landrigan et al.
1976; Schwartz et al. 1988; Seto and
Freeman 1964
Shuklaet al. 1989
0)
I
Ko
O
-------
s
Table B-1. Health Effects Associated with Exposure to Lead and Internal Lead Doses in Humans (Continued)
Duration of
Exposure
prenatal
(general population)
NS
(general population)
NS
(general population)
NS
(general population)
NS
(general population
NS
(general population)
NS (occup)
System
Developmental
Developmental
Developmental
Developmental
Reproductive
Reproductive
Reproductive
Effect
Reduced birth weight and/or reduced
gestational age, and/or increased
incidence of stillbirth and neonatal
death
No association between blood-lead
levels and birth weight, gestational
age, or other neonatal size measures
Impaired mental development in
children
Inverse correlation between blood-lead
levels and ALA and ALAD activity
Increased incidence of miscarriages
and stillbirths in exposed women
No association between blood-lead
levels and the incidence of
spontaneous abortion in exposed
women
Adverse effects on testes
Blood Lead Levels
at which Effect is
Observed (//g/dL)
12-17
3-55
10-15
10-33
(mean)
t 10orNS
2
40-50
Reference
Bornschein et al. 1 989; McMichael et al.
1986; Moore et al. 1982; Ward et al.
1987; Wibberley et al. 1977
Greene and Ernhart 1991; Factor-Litvak
etal. 1991
Baghurst et al. 1987; Bellinger et al.
1984, 1985a, 1985b, 1986a, 1986b,
1987a, 1987b; Bornschein et al. 1989;
Dietrich et al. 1986, 1987a, 1987b;
Ernhart et al. 1985, 1986, 1987;
McMichael et al. 1 988; Rothenberg et al.
1989; Wigg et al. 1988; Winneke et al.
1985a, 1985b; Wolf et al. 1985;
Vimpanietal. 1985, 1989
Haas et al. 1972; Kuhnert et al. 1977;
Lauwerys et al. 1 978
Baghurst et al. 1987; Hu et al. 1991;
McMichael et al. 1986; Nordstrom et al.
1979; Wibberley et al. 1977
Murphy etal. 1990
Assennato et al. 1987; Braunstein et al.
1978; Chowdhury et al. 1986; Cullen et
al. 1984; Lancranjan et al. 1975;
Rodamilans et al. 1 988: Wildt et al. 1 983
9
I
r
to
M
i
ALA = 6-aminolevulinic acid; ALAD = 6-aminolevulinic acid dehydratase; ALAS = 6-aminolevulinic acid synthase; EP = erythrocyte protoporphyrins;
FEP = free erythrocyte protoporphyrins; IQ = intelligence quotient; mmHg = millimeters of mercury; NS = not specified; (occup) = occupational;
Py-5'-N = pyrimidine-5-nucleotidase; wk = week(s); yr = year(s); ZPP = zinc erythrocyte protoporphyrin
-------
APPENDIX C
Supporting Information for Chapter 3
Draft - Do Not Cite or Quote C-1 September 27. 1996
-------
APPENDIX C
SUPPORTING INFORMATION FOR CHAPTER 3
C1.0 CHARACTERIZING BASELINE ENVIRONMENTAL-LEAD
LEVELS IN THE NATION'S HOUSING STOCK
As discussed in Section 3.3.1.1, the §403 risk assessment effort used environmental-lead
data from the National Survey of Lead-Based Paint in Housing ("HUD National Survey") to
characterize baseline environmental-lead levels in the nation's 1997 housing stock. Here, the
term "baseline" refers to conditions prior to implementing interventions in response to §403
rules. Data for 284 privately-owned, occupied housing units included in the HUD National
Survey were considered in the characterization. In total, these units represented the entire U.S.
privately-owned, occupied housing stock built prior to 1980. Due to the complex sampling
design employed, the HUD National Survey assigned sampling weights to each unit, which
equaled the number of privately-owned, occupied housing units in the national housing stock
built prior to 1980 that were represented by the unit.
In order to use the information from the HUD National Survey to represent baseline
environmental-lead levels in the 1997 national housing stock, the following steps were taken:
1. Update the sampling weights assigned in the HUD National Survey to reflect the
1997 housing stock (including publicly-owned units).
2. Determine the total number of children residing in the housing units represented by
each sampling weight.
3. Summarize the environmental-lead levels within each surveyed unit.
Methods for conducting each of these steps, and the results from implementing these methods,
are summarized in the following subsections.
C1.1 UPDATING THE NATIONAL SURVEY SAMPLING WEIGHTS
Characterizing the 1997 national housing stock and its distribution of environmental-lead
levels involved updating the sampling weights assigned in the HUD National Survey to reflect
Draft - Do Not Cite or Quote C-2 September 27, 1996
-------
the 1997 national housing stock. The tasks performed to update these weights were the
following:
1. Identify demographic variables that served to group the housing units by their
potential for differing environmental-lead levels.
2. Use information within the National Survey weights and the 1993 American Housing
Survey to determine total numbers of 1997 housing units within each of these
housing groups.
3. Allocate these 1997 totals among the National Survey units within the housing
groups.
The methods developed for each of these tasks are presented in the following subsections.
C 1.1.1 Identify Significant Factors Associated With Environmental-Lead Levels. In
updating the sampling weights of the 284 National Survey units, the units were classified into
housing groups according to a set of demographic factors found to have a statistically significant
influence on environmental-lead levels hi the units. Then, the number of 1997 housing units in
each group was determined. By grouping the housing units according to these factors, units
within the same group had relatively similar distributions of environmental-lead levels, while
units in different groups had considerably different distributions.
In determining an appropriate housing grouping, a set of candidate factors was identified,
where these factors satisfied three criteria: 1) they would be either important in an economic
analysis for §403 rulemaking, or they were likely to be significantly associated with
environmental-lead levels; 2) their values for National Survey units existed within the National
Survey database; and 3) their values were measured within the 1993 American Housing Survey, a
national survey conducted by the Bureau of the Census and the Department of Housing and
Urban Development (HUD) to characterize the nation's housing stock. Then, a stepwise
regression variable selection analysis selected a subset of these factors which explained the
largest proportions of house-to-house variability in the following four environmental-lead
measurements:
Draft - Do Not Cite or Quote C-3 September 27, 1996
-------
A mass-weighted arithmetic average floor dust-lead concentration1 for the unit (i.e.,
each measurement was weighted by the mass of the sample);
An area-weighted arithmetic average floor dust-lead loading for the unit (i.e., each
measurement was weighted by the square-footage of the sample area);
A weighted arithmetic average soil-lead concentration for the unit, where results for
samples taken from remote locations were weighted twice as much as results for
dripline and entryway samples.
Maximum XRF paint-lead level in the unit (for units containing lead-based paint2).
The set of factors included in this analysis are documented in Table C-l.
Table C-1. Demographic Factors Included in the Step wise Regression Analysis
Factor
Year the Unit Was Built
Race of Youngest Child
Urbanicity Status
Region of Country
Ownership Status
Number of Units in the Bldg.
Annual Income of Residents
How the Factor Categorized Housing Units
for the Stepwise Regression Analysis
Pre-1940; 1940-1959; 1960-1979
White/Non-Hispanic; Other
City; Suburb/non-metro
Northeast; Midwest; South; West (U.S. Census regions)
Owner-occupied; renter-occupied
One unit; more than one unit
< $30,000; $30,000 or more
The analysis was performed twice on each endpoint: on data for National Survey units
containing lead-based paint (LBP) and for units with no LBP. Table C-2 provides the observed
significance levels of each factor considered in the stepwise regression analyses when these
levels were below 0.10. Lower significance levels imply a stronger effect on the measurement.
The columns in Table C-2 correspond to separate regression analyses. Across all analyses, the
1 Prior to calculating the mass-weighted average, dust-lead concentrations were adjusted to reduce bias associated with
underestimated sample weights ("low tap weights") reported in the HUD National Survey for dust samples. The adjustment procedure
is documented in Appendix Z.
2 LBP was considered present in a unit if its predicted maximum XRF value (as determined by statistical modeling
techniques within the HUD National Survey) in either the interior or exterior was at least 1.0 mg/cm2.
Draft - Do Not Cite or Quote
C-4
September 27, 1996
-------
year in which a unit was built (as categorized by pre-1940,1940-1959, and 1960-1979) had the
strongest and most consistent effect on the environmental-lead level (with floor dust-lead
concentration an exception). Statistical significance levels for the effect of year built were
consistently less than 0.01. While similar significance levels were occasionally observed for
other factors in the table, the extent of significance across the environmental-lead measurements
was not as consistent for any other factor. Therefore, the year in which the unit was built was the
only factor considered in grouping National Survey units for purposes of updating their weights
to 1997.
Table C-2. Demographic Factors Included in Stepwise Regression Analyses, and
Significance Levels Associated With These Factors When Less Than 0.101
Demographic Factors2
Year the Unit Was
Built
Race of Youngest
Child
Urbanicity Status
Region of Country
Ownership Status
# Units in the Bldg.
Annual Income of
Residents
Units with predicted maximun
below 1.0 mg/cm2 or missii
Floor
Dust-Lead
Loading
<0.01B
0.03
Floor
Dust-Lead
Cone.3
it XRF value
ig(n=40)
Soil-Lead
Cone.
<0.01
0.04
Units with predicted maximum XRF value
at 1 .0 mg/cm1 or above (n = 221)
Floor
Dust-Lead
Loading
<0.01
0.01
Floor Dust-
Lead Cone.3
0.01
Soil-Lead
Cone.
<0.01
Max.
Observed XRF
Value4
<0.01
1 Column headings for this table identify the environmental-lead measurement being considered in the analysis and the group
of National Survey units whose data are included in the analysis. Each column corresponds to a separate regression
analysis. The demographic factors included in the regression analyses are included as rows of the table. As the
significance level for a demographic factor gets closer to zero, the effect of the factor on the given environmental
measurement is considered more highly statistically significant.
1 See Table C-1 for definitions of these factors.
3 This analysis was performed on unadjusted dust-lead concentrations (i.e., no adjustment was made for bias due to
underestimated sample weights).
4 Regression performed on units where the observed maximum XRF value was at least 1.0 mg/cm2.
5 In the regression analysis of floor dust-lead loading in units without LBP, the effect of the year in which the unit was built
was statistically significant with a p-value of less than 0.01 (i.e., significance can be concluded at the O.01 level).
Draft - Do Not Cite or Quote
C-5
September 27, 1996
-------
The stepwise regression analysis assumed that the predicted maximum XRF value is an
accurate indicator of whether or not a unit contains LBP. Also, those units with no predicted
maximum XRF value were assumed not to contain LBP.
C1.1.2 Estimating Numbers of Housing Units in 1997 Within Year-Built Categories.
In this second task, the number of occupied housing units in 1997, both privately- and publicly-
owned, was estimated for each of four categories denoting when the unit was built: pre-1940,
1940-1959,1960-1979, and post-1979. These categories are hereafter referred to as "year-built
categories." The results of this task are presented in Table 3-3 within Chapter 3 of this
document.
The primary data source for determining the number of units within each year-built
category was the 1993 American Housing Survey (AHS), the most recent survey with available
data on estimated numbers of units in the national housing stock. Data from the 1993 AHS
provided estimates of the number of housing units in each year-built category in 1993. However,
it was of interest to obtain estimates for 1997, not 1993. Therefore, the 1993 estimates were
augmented to reflect additions to and removals from the national housing stock from 1994 to
1997. Once the 1997 estimate of the total within each year-built category was obtained, the total
was distributed among the National Survey units in the group using information within the
National Survey weights. Details on each of these procedures are now provided.
Characterizing the 1993 National Housing Stock
As in the National Survey, each unit in the 1993 AHS was assigned a weight that was
interpreted as the number of units in the national housing stock represented by the given unit.
Therefore, placing the AHS units among the four year-built categories and summing the weights
of the units within each category yielded the estimated number of units in 1993 for each category.
Only occupied housing units in the 1993 AHS (either publicly-owned or privately-owned)
were considered in updating to 1997. The definition of an "occupied" unit was one which was
occupied by at least one resident who was classified as not having his/her usual residence
elsewhere. Data for 40,931 occupied housing units were available from the 1993 AHS.
Draft - Do Not Cite or Quote C-6 September 27, 1996
-------
Updating the 1993 Housing Stock to 1997
Once the number of housing units in 1993 was determined for each of the four year-built
categories, these totals were updated to reflect the 1997 housing stock. Updating the 1993 totals
to 1997 was done in the following way:
1. For the post-1979 category, the total number of housing constructed from 1994 to
1997 and occupied in 1997 was estimated and added to the 1993 total.
2. For all four year-built categories, the total number of housing occupied in 1993 and
lost from the housing stock from 1994 to 1997 was estimated and subtracted from
the 1993 total.
In the first step, numbers of new, privately-owned housing units completed in 1994 and 1995
were obtained from Bureau of the Census and HUD (1996). This publication reported estimates
of 1,346,900 such units completed in 1994 and 1,311,300 units in 1995. For this analysis, the
1995 estimate was also used in estimating totals for both 1996 and 1997. Therefore, the 1993
estimate for the post-1979 housing category was incremented by 1,346,900 + 3*1,311,300 =
5,280,800 units. Note that this approach assumes that new housing units are completed and
occupied within the same year. In addition, no provision was considered for adding new
publicly-owned units.
The second step, subtracting the number of housing units occupied in 1993 and lost from
the housing stock from 1994 to 1997 within each of the four year-built categories, was more
complex. Information on losses was not available by considering only the 1993 AHS. To obtain
such information, the 1989 and 1991 AHS databases were obtained. As the AHS retains the
same units from survey to survey, it was possible to determine those units that were occupied in
one survey and lost from the housing stock by the next. Units were considered lost from the
housing stock in a given survey if they were labeled as a "Type C non-interview" in the survey,
meaning the unit no longer exists and is dropped from consideration for future surveys. Such
losses include demolition, disaster loss, abandoned permit, or the unit was merged with another
unit. While moving a house or mobile home from the site also labels the unit as a Type C
noninterview, such an instance was not labeled as a loss from the housing stock for this effort, as
it is assumed that the unit remains habitable in its new location. Using this definition of a loss,
Draft - Do Not Cite or Quote C-7 September 27, 1996
-------
all occupied units in the 1989 AHS were labeled as whether or not they were considered lost
from the housing stock by the 1991 AHS. Similarly, all occupied units in the 1991 AHS were
labeled as whether or not they were considered lost from the housing stock by the 1993 AHS.
It is recognized that the probability of removal from the housing stock is a function of the
age of the unit (in years). However, each of the year-built categories represent units of varying
ages. For this approach, it was necessary to assign a single age (in years) to each category. For
the 1940-1959, 1960-1979, and 1980-1997 categories, this age corresponded to the age of a unit
built in the middle year of the category (48, 28, and 9 years, respectively). The single age
assigned to all units in the pre-1940 category was equal to the age of a unit built in 1939 (58
years).
Because each AHS was separated by two years, a two-year period was considered when
determining the probability of a unit in a given year-built category being lost from the housing
stock. In this approach, each unit in the 1989 AHS and in the 1991 AHS was assigned an age in
years (as of 1989 and 1991, respectively) according to the year-built category in which they were
classified (different year-built categories from the four considered here were used in the AHS).
This information was combined across the two surveys, and a logistic regression analysis was
used to predict the probability of a loss over a two-year period as a function of age (in years).
This regression analysis was weighted by the sample weights assigned to the units in their
respective surveys (1989 or 1991). The resulting prediction model was
P[loss over a two-year period] =
m- 00094. aee
where "age" is the age of the unit in years. The probability for a one-year period is roughly one-
half of the probability for the two-year period. Table C-3 provides the predicted probabilities of
losses over a one-year period for every five years of age.
Draft - Do Not Cite or Quote C-8 September 27, 1996
-------
Table C-3. Estimated Probability of an Occupied Housing Unit Becoming Lost from the
Housing Stock Over a One-Year Period, Given the Age of Unit
Age of Unit
(yrs)
5
10
15
20
25
30
35
40
Probability of
Loss
0.0013
0.0014
0.0015
0.0016
0.0017
0.0018
0.0020
0.0021
Age of Unit
(yrs)
45
50
55
60
65
70
75
80
Probability
of Loss
0.0023
0.0025
0.0026
0.0028
0.0031
0.0033
0.0036
0.0038
Note: These probabilities were estimates from equation (1) and adjusted to cover a one-year period.
Table C-4 illustrates how reductions were determined from 1993 to 1995. First, an age
(in years) associated with each of the four year-built categories was determined for 1993 and
1995. Then, the probability of loss for both ages was determined from equation (1); these
probabilities are labeled as pig93J95 and p,^^ in Table C-4. The total number of units in the
category in 1993 was then reduced by multiplying the total by the product (l-
(i.e, the last column of Table C-4).
Table C-4. Determining Losses from the Housing Stock from 1993-1997
Year-Built
Category
Pre-1940
1940-
1959
1960-
1979
Post- 197 9
Age of
units in
1993 (yrs.)1
54
44
24
7
Prob. of loss
from 1993-1995
(Pl993.9s)2
0.0052
0.0045
0.0034
0.0026
Age of
units in
1995 (yrs.)1
56
46
26
9
Prob. of loss
from 1995-
1997 (p199B.97)2
0.0054
0.0046
0.0034
0.0027
Proportion of
1993 Total That
Remains in 1997 3
0.989
0.991
0.993
0.995
1 A single age is assigned to all units in a given category according to the approach indicated in the text.
2 Determined from equation (1).
3 Equal to (1-p,nMi>*(1-PiMM7>
Draft - Do Not Cite or Quote
C-9
September 27, 1996
-------
Besides additions and removals, changes in the number of occupied homes in the national
housing stock from 1993 to 1997 are also affected by the number of units that are occupied in
1993 and vacant in 1997, as well as by the number of units that are vacant in 1993 and occupied
in 1997. However, in this approach, it was assumed that the number of occupied units in 1993
that become vacant in 1997 was approximately equal to the number of vacant units in 1993 that
become occupied in 1997, thereby canceling each other out.
C1.1.3 Determining the Number of 1997 Units Represented by Each National
Survey Unit. The procedures outlined in the previous subsection provide a method for
estimating total numbers of housing units in 1997 within each of the four year-built categories.
The results are displayed in Table 3-3 in Chapter 3 of this report. The housing units were grouped
within year-built categories to facilitate the linking of numbers of units with estimated
environmental-lead levels. The linking process consisted of classifying the National Survey units
among the four categories, then distributing the 1997 total among the National Survey units
within each category. This distribution yielded an updated weight for each National Survey unit,
reflecting changes in the numbers of units in the year-built category from the time the National
Survey was conducted to 1997. A unit's updated weight represented the number of units in the
1997 housing stock associated with the National Survey unit (and therefore with its
environmental-lead levels).
The 1997 totals include both privately-owned and publicly-owned housing units, while
the 284 National Survey units were privately-owned. Therefore, the revised 1997 weights for the
National Survey units represent publicly-owned as well as privately-owned units.
Updating the Weights to Reflect the Pre-1980 Housing Stock
To update the sampling weights for the 284 National Survey units to reflect the pre-1980
housing stock, the units were grouped according to the three pre-1980 year-built categories.
(Recall that all National Survey units were built prior to 1980). For these three categories, the
updated 1997 weight for each unit in the category was calculated as follows:
1997 weight = (National Survey weight)*(Updating factor for the category) (2)
Draft - Do Not Cite or Quote C-10 September 27. 1996
-------
where the updating factor was determined as follows:
Updating factor =
# units in the category in 1997
Total National Survey weights in the category
(3)
(The sampling weights assigned in the National Survey were determined according to when the
unit was built, whether the unit existed in a single- or multiple-unit building, the Census region
in which the unit was located, and whether or not a child less than aged seven years resided in the
unit).
Table C-5 contains the updating factors applied to the National Survey units according to
year-built category. As an example, Table C-5 indicates that the updated 1997 weight for each of
the 77 National Survey units in the pre-1940 category equaled the weight assigned in the
National Survey multiplied by 0.936.
Table C-5. Number of National Survey Units in the Pre-1980 Year-Built Categories, and
the Multiplicative Factor Used to Update National Survey Weights to 1997
Year-Built
Category
Pre-1940
1940-1959
1960-1979
# National Survey
Units
77
87
120
Sum of National
Survey Weights
21,020,019
20,472,997
35,686,004
Updating Factor
0.936
0.963
0.980
Updating the Weights to Reflect the Post-1979 Housing Stock
Despite the fact that no HUD National Survey units were built after 1979, it was of
interest to use the HUD National Survey to characterize the entire occupied national housing
stock, including those units built after 1979. Therefore, methods were developed to determine
how to use environmental-lead information from the HUD National Survey to represent the post-
1979 occupied housing stock.
As the post-1979 housing stock was built after the Consumer Product Safety
Commission's 1978 ban on the sale of LBP and its use in residences, the post-1979 housing
Draft - Do Not Cite or Quote
C-11
September 27, 1996
-------
stock was assumed to be free of LBP. This same assumption was made in the HUD National
Survey and is the reason for not including post-1979 housing in the survey. Therefore, only
National Survey units not containing LBP were considered in representing post-1979 housing.
To determine whether the entire set of National Survey units without LBP should be
considered in representing post-1979 housing or only a subset of these units, data on dust-lead
and soil-lead concentrations for units having maximum and minimum XRF measurements below
0.7 mg/cm2 were investigated. As the top two plots in Figure C-l illustrate, a noticeable
relationship exists between lead concentrations and the age of the unit, with higher
concentrations associated with older units. In contrast, the bottom two plots in Figure C-l show
less of a relationship between concentration and age of unit when only units built from 1960-
1979 were considered. This finding suggests that older units may be free of LBP, but dust and
soil are' more likely to remain contaminated with lead than for newer units, either due to previous
renovation work on the units or from outside contamination.
As a result of the conclusions made from Figure C-l, only the 28 National Survey units
built between 1960 and 1979 and containing no LBP were selected to represent the post-1979
housing stock. As a result, it was assumed that the environmental-lead levels for these 28 units
represented levels that exist in the post-1979 housing stock. These units also were included
among those representing the 1960-1979 housing stock. Therefore, the total 1997 sampling
weight for these 28 units consisted of two parts: that representing the 1960-1979 housing stock,
and that representing the post-1979 housing stock. That portion representing the post-1979
housing stock was determined by dividing the total number of post-1979 units in 1997 by 28.
C1.2 POPULATING HOUSING UNITS WITH CHILDREN
To characterize health benefits associated with §403 interventions, it was necessary to
estimate numbers of children of specific age groups who reside within the national housing stock.
This section documents the methods for populating the 1997 national housing stock with
children.
Draft - Do Not Cite or Quote C-12 September 27. 1996
-------
10000
1000
Soil Concentration
8
10
1
19
10000
1000
1
Concent
8
w
10
1
19
100000
*
1 • 10000
- f ' ' I . . ;, § 1000
* ) , >* vV M
1 » v V g
; ; 100
*
10
00 1920 1940 1960 1980 19
Year Built
100000
10000
; 1
; * § 1000
'•• ' - : • ^^7 2
; ; 100
+ *
10
60 1965 WO WS 1980 19
.
t
t
* I* '* * *
* ** . < *
* » * :
00 1920 1940 1960 1980
Year Built
•
t
, »
t i» *
'; ' ' l
t
60 1965 WO 1975 1980
Year Built
Year Built
Figure C-1. Plots of Dust- and Soil-Lead Concentration (/sg/g) Versus Age of Unit, for HUD
National Survey Units With Maximum XRF Value Less Than 0.7 mg/cm2
- Do A/or C/fe or Quote
C-13
September 27, 1996
-------
Section Cl .1 presented methods to revising the sampling weights for HUD National
Survey units to reflect the 1997 national housing stock of occupied units. Therefore, each weight
represents a subset of the national housing stock. It was desired to link numbers of children with
each weight. Two age groups of children were of interest:
• Children aged 12 to 35 months (1 to 2 years)
• Children aged 12 to 71 months (1 to 5 years)
The 1-2 year age group was the primary group of interest in the §403 risk assessment effort,
while the 1-5 year age group was considered in the sensitivity analysis.
For a given age group of children, the estimated number of children associated with the
units represented within a 1997 sampling weight was the product of three statistics:
# children = (1997 weight)*(Average # residents per unit)*(# children per person) (4)
As the 1997 weight was determined for each National Survey unit using the methods in Section
Cl .1, it was necessary to obtain estimates for the latter two statistics in equation (4).
The factor "average # residents per unit" hi equation (4) was calculated for the housing
group based on information obtained in the 1993 AHS. The 1993 AHS database provided
information on up to 15 residents within each housing unit in the AHS. Once these units were
placed within the four year-built categories, the average number of people residing in a unit
(regardless of age) was calculated for each group. This average ranged from 2.5 to 2.7 across the
four year-built categories. A common average of 2.7 residents per unit was used for all units in
the national housing stock. While this average was based on 1993 data, it is assumed to also
hold for the 1997 housing stock.
The third factor in equation (4), "# children per person," represented the average number
of children (of the given age group) per person residing in units within the housing group. This
factor was calculated from information presented in Day (1993). This document provided two
types of information necessary to calculate average number of children per person:
Draft - Do Not Cite or Quote C-14 September 27. 1996
-------
1. Predicted numbers of births per 1,000 people in the general population within
selected years from 1993 to 2050
2. Predicted numbers of people in the general population of specific ages for these
selected years.
For 1997, Day (1993) predicted a total of 14.8 births predicted per 1,000 people in the U.S.3
Therefore, it was assumed that in any subset of occupied housing in 1997, the units within this
subset will contain 14.8 children less than one year of age for every 1000 residents.
Day (1993) also provided a predicted number of children of various age groups in the
nation in 1997. A total of 3,907,000 children aged 0-11 months, 7,835,000 children aged 12 to
35 months, and 20,066,000 children aged 12 to 71 months were predicted. By dividing each of
these latter two statistics by 3,907,000, approximately 2.01 children aged 12 to 35 months and
5.14 children aged 12 to 71 months are predicted hi 1997 for every child aged 0-11 months.
Thus, using the birth rate in the previous paragraph, a total of 2.01 x 14.8 = 29.7 children aged 12
to 35 months, and 5.14 x 14.8 = 76.1 children aged 12 to 71 months, are predicted in 1997 per
1000 people in the U.S.
Table C-6 contains estimates of average number of children per unit in the 1997 national
housing stock, according to age group. These numbers are the product of the final two factors in
equation (4). Therefore, these numbers are multiplied by the 1997 sampling weights for each
National Survey unit to obtain an estimated number of children residing in units represented
within the weight. By summing the estimates across National Survey units, the total number of
children aged 12-35 months and 12-71 months residing within the 1997 national housing stock is
obtained by year-built category and for the nation. These results are presented in Table 3-24 in
Chapter 3 of this report.
3 This is a "middle series assumption" birth rate, indicating the level at which assumptions are placed on fertility, life
expectancy, and yearly net immigration.
Draft - Do Not Cite or Quote C-15 September 27, 1996
-------
Table C-6. Estimated Average Number of Children Per Unit in the 1997 National Housing
Stock, by Age of Child
Age Group
12-35 months
12-71 months
Estimated Average Number of
Children Per Unit
2.7'0.0297 = 0.080
2.7*0.0761 = 0.205
C1.3 SUMMARIZING ENVIRONMENTAL-LEAD LEVELS WITHIN THE HUD NATIONAL
SURVEY UNITS
The methods of Sections Cl.l and C1.2 were used to link each of the 284 units in the
HUD National Survey with an estimated number of units in the 1997 national housing stock and
an estimated number of children residing within these units. In this final step, it is necessary to
summarize the environmental-lead levels within each National Survey unit.
The following statistics were calculated for each National Survey unit, summarizing the
unit's dust-lead loadings and dust-lead concentrations from floors and window sills, and soil-lead
concentrations:
A mass-weighted arithmetic average floor dust-lead concentration4 for the unit (i.e.,
each measurement is weighted by the mass of the sample);
An area-weighted arithmetic average floor dust-lead loading for the unit (i.e., each
measurement is weighted by the square-footage of the sample area);
A mass-weighted arithmetic average window sill dust-lead concentration4 for the
unit (i.e., each measurement is weighted by the mass of the sample);
An area-weighted arithmetic average window sill dust-lead loading for the unit (i.e.,
each measurement is weighted by the square-footage of the sample area);
A weighted arithmetic average soil-lead concentration for the unit, where results for
samples taken from remote locations were weighted twice as much as results for
dripline and entryway samples. If a unit has no soil-lead results for a particular
Prior to calculating the mass-weighted average, dust-lead concentrations on floors and window sills were adjusted to
reduce bias associated with underestimated sample weights ("low tap weights") reported in the National Survey for dust samples.
Draft - Do Not Cite or Quote
C-16
September 27. 1996
-------
location, the arithmetic average was unweighted (i.e., results for the remaining
locations were not weighted).
• An unweighted arithmetic average soil-lead concentration, considering only the
dripline and entryway samples for the unit.
• The maximum paint-lead concentration in the interior and the exterior of the unit, as
measured by XRF techniques in selected rooms and on selected components within
these rooms.
• The amount of damaged lead-based paint measured in the interior and the exterior
of the unit.
These summary values were used in the statistical models to represent environmental-lead levels
in the national housing stock, in determining health benefits associated with intervention.
In the HUD National Survey database, some units have unrecorded (or "missing") values
for dust-lead loadings or concentrations, or soil-lead concentrations, preventing values for one or
more of the first six summary statistics above from being calculated. As the values of certain
statistics were used as input to the HUD/IEUBK and EPI models to predict health benefits due to
§403 interventions, it was necessary that every housing unit have values for these statistics, even
if no data existed for a particular unit. Therefore, an imputation scheme was devised to obtain
summary values for units having no data in the National Survey database for the given parameter.
In this approach, if a unit did not have data to allow the value of a summary statistic from being
calculated, the value assigned to the unit equaled the weighted arithmetic average of those values
for units within the same year-built category and having the same indicator for the presence of
LBP, with each value weighted by the 1997 weight for the respective unit. For example, a total
of eight National Survey units were built prior to 1940 and contained no LBP. If one of these
units had no floor dust-lead loadings, then the summary value of floor-dust-lead loading for this
unit would equal the weighted average of the summary values across the other seven units.
Table C-7 contains a listing of National Survey units within the three year-built categories
in which they are classified. Also note that the 28 National Survey units built from 1960-1979
and containing no LBP were listed within a fourth category within Table C-7, representing the
national housing stock built after 1979. The dust-lead concentrations summarized in Table C-7
were initially adjusted for underestimated sample weights; see Appendix Z for methods used to
Draft - Do Not Cite or Quote C-17 September 27, 1996
-------
conduct this data adjustment. Also, dust-lead loadings summarized in Table C-7 were initially
adjusted to reflect loadings that would be obtained if wipe collection techniques were used, rather
than the Blue Nozzle vacuum method employed in the National Survey. The method to
converting from vacuum to wipe loadings is presented in Chapter 4.
Table C-7 also contains the updated 1997 sampling weights for each unit (as calculated in
Section Cl.l) and the estimated numbers of children aged 12-35 months and 12-71 months that
reside within the units (as calculated in Section C1.2). For the 28 units listed in both the 1960-
1979 and post-1979 categories, the sampling weights and numbers of children are only that
portion representing units within the category.
Draft - Do Not Cite or Quote C-18 September 27, 1996
-------
Table C-7. Estimated Environmental Lead Levels in the 1997 Housing Stock, As Determined from National Survey Units
o
'
i
Year
Built
<1940
o
to
*
Floor Floor Window Sill
National Dust-Lead Dust-Lead Dust-Lead Window sill Soil-Lead Dnplme/Entry Obs. Max. Obs. Max.
Survey LBP Loading Cone. Loading Dust-Lead Cone. Soil-Lead Interior XRF Exterior XRF Interior
ID Present' (ug/ft2) (ug/g) (ug/ft2) Cone, (ug/g) (ug/g) Cone, (ug/g) Img/cm2) (mg/cm2)
0.60
0.60
0.60
0.60
2.8
0.60
10.
.60
0320408
0320507
1210806
1921709
1932300
1942606
1953009
2022507
0211102
0221101
0221507
0310102
0310607
0310706
0311100
0320705
0350801
0411207
0520106
0520403
0520700
0520908
0711002
0720300
0720706
0721001
0730606
0820506
0911800
0920900
0941005
0950402
0951004
1010909
1011303
1011501
1011600
1041607
1221902
1250406
1251107
1251404
1352608
1353705
1411909
1531201
1531300
1631209
1631308
1740901
1751304
1820802
1830801
1830900
1840503
1851104
1931906
1951904
1952506
2121507
2240406
2311108
2343002
No
NO
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yea
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
17.7
32.9
46.8
39.1
158.
59.2
3.02
142.
41.2
5.90
44.5
27.2
7.19
8.15
220.
15.2
54.2
397.
10.5
198.
127.
28.4
53.3
27.2
30.2
52.9
74.4
14.5
6.38
10.7
10.8
32.6
17.9
51.3
71.5
32.4
78.1 •
7 14
191.
53.5
136.
22.4
276.
10.9
443.
192.
35.6
30.6
12.7
117.
30.0
9.58
17.8
492.
213.
71.6
33.3
68.1
23.6
53.3
279.
119.
13.4
317.
340.
978.
448.
412.
246.
112.
589.
781.
157.
979.
297.
64.1
212.
1600.
399.
2120.
1810.
87.0
299.
938.
630.
537.
340.
535.
328.
527.
126.
96.9
188.
245.
644.
527.
1240.
1100.
623.
487.
0.09
6320.
2370.
1760.
662.
2070.
491.
4340.
831.
303.
215.
123.
860.
198.
106.
279.
3630.
1970.
316.
193.
627.
328.
281.
781.
1280.
292.
3.26
50.0
29.7
56.6
1440.
2500.
2.84
243.
0.36
1.71
484.
5.12
548.
90.4
5.93
1 86
39.9
25.0
9.24
322.
4970.
156.
14.9
2790.
345.
1780. •
33400.
23.0
141.
2.71
1080.
2200.
126.
38.4
1780. •
65.3
1780. •
2.38
9640.
139.
126.
54.1
1780. •
13.2
12.5
681.
55.3
286.
310.
1780. •
0.05
1.61
6.01
13.9
32.6
532.
908.
394.
0.87
1780 •
21700.
1720.
102.
1740. •
762.
618.
579.
1880.
10900.
103.
881.
6700. •
6700. •
1740.
108.
835.
1250.
423.
6700. •
855.
396.
153.
286.
1050.
319.
242.
8710.
2310.
6700. •
10200.
300.
111.
6700. •
4650.
3460.
1890.
1130.
6700. •
1090.
6700. •
6700. •
56500.
949.
1840.
698.
6700. •
6700. •
738.
1090.
481.
1180.
701.
6700. •
1.44
6700. *
74.3
6700. •
526.
2880.
5840.
2080.
39.9
6700. •
8480.
500.
4720.
36.5
113.
305. •
504.
305. •
413.
305. •
326.
84.2
394.
2020.
138.
1240.
534.
711.
274.
25.9
805.
59.6
102.
258.
17.4
157.
1460.
830. •
80.4
372.
1110.
49.8
162.
1620.
2000.
1170.
851.
717.
4620.
392.
39.5
444.
628.
1030.
569.
679.
109.
586.
251.
105.
830. •
524.
137.
358.
1430.
830.
830.
830.
212.
830.
830.
830.
1170.
335.
830. •
256.
19.7
88.1
432.
504.
432.
413.
432.
544.
66.5
424.
1080.
124.
439.
932.
653.
263.
34.1
843.
59.0
113.
296.
23.1
1100.
1930.
1100.
93.7
262.
1110.
34.7
283.
260.
4000.
2290.
1480.
1150.
8960.
589.
44.4
701.
888.
1660.
535.
968.
116.
1040.
283.
179.
1100.
524.
221.
466.
2330.
1100.
1100.
1100.
212.
1100.
1100.
1100.
1170.
270.
1100.
242.
*
*
•
*
*
*
*
0.
3.4
7.1
5.3
0.70
0.40
0.60
0.70
0.60
0.70
0.20
12.
8.0
3.3
10.
0.70
0.60
0.60
0.80
0.30
0.60
10.
0.80
0.40
0.30
0.30
6.4
6.2
5.0
20.
7.0
13.
0.60
0.90
3.3
1.4
1.2
9.4
2.9
0.60
6.6
4.7
1.2
0.60
4.4
6.0
1.9
1.7
0 60
8.6
2 3
8.7
5.1
6.0
0.60
14.
5.8
27.
0.40
0.40
1.8
2.8
0.60
13.
0.60
5.0
0.60
8.8
3.6
0.80
54.
3.8
0.30
6.5
51.
38.
29.
0.30
11.
4.9
0.00
4 0
10.
1.8
7.9
14.
4.4
1.6
1.6
15.
9.5
4.6
2.7
2.4
7.1
3.5
0.70
5.7
Damaged
Interior
LBP (ft2)
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
9.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
11.5
0.0
0.0
0.0
0.0
0.0
89. a
n 6
0.0
18.7
0.0
0.0
0.0
0.8
0.0
0.0
6.2
0.0
21 9
0.5
Damaged
Exterior
LBP (ft2)
0.0
0.0
--
—
0.0
0.0
0.0
0.0
0.0
0.0
4.8
0.0
__
0.0
57.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
24.6
0.0
28.0
226.8
0.0
0.0
0.0
0.0
457. 3
0.0
—
0.0
182.0
0.0
8.4
0.0
0.0
0.0
141.4
0.0
0.0
585.7
112.0
0.0
0.0
0.0
0.0
0.0
_-
__
—
0 0
0 0
0.0
0.0
25 1
604.8
0.0
1.7
1997
Height
183,864
183,864
121,752
199,528
199,528
199,528
199,528
1,140,935
183,864
95,766
183,864
183,864
183,864
95,766
183,864
183,864
95,766
244,799
244,799
114,632
199,528
114,632
111,365
111,365
111,365
60,761
111,365
111,365
111,365
111,365
773,094
773,094
773,094
244,799
244,799
114,632
244,799
244,799
1,140,935
121,752
121,752
1,140,935
111,365
111,365
95,766
773,094
111,365
199,528
199,528
121,752
121,752
60,761
60,761
199,528
60,161
121,752
199,528
60,761
60,761
199,528
244,799
1.140,935
121,752
H Children
12-35 no.
14,744
14,744
9,763
16,000
16,000
16,000
16,000
91,492
14,744
7,679
14,744
14,744
14,744
7,679
14,744
14,744
7,679
19,630
19,630
9,192
16,000
9,192
8,930
8,930
8,930
4,872
8,930
8,930
8,930
8,930
61,994
61,994
61,994
19,630
19,630
9,192
19,630
19,630
91,492
9,763
9,763
91,492
8,930
8,930
7,679
61,994
8,930
16,000
16,000
9,763
9,763
4,872
4,872
16,000
4,872
9,763
16,000
4,872
4,872
16,000
19,630
91,492
9,763
« Children
12-71 mo.
37,779
37,779
25,016
40,997
40,997
40,997
40,997
234,428
37,779
19,677
37,779
37,779
37,779
19,677
37,779
37,779
19,677
50,299
50,299
23,553
40,997
23,553
22,882
22,882
22,882
12,485
22,882
22,882
22,882
22,882
158,848
158,848
158,848
50,299
50,299
23,553
50,299
50,299
234,428
25,016
25,016
234,428
22,882
22,882
19,677
158,848
22,882
40,997
40,997
25,016
25,016
12,485
12,485
40,997
12,485
25,016
40,997
12,485
12,485
40,997
50,299
234,428
25,016
0)
-------
Table C-7. Estimated Environmental Lead Levels in the 1997 Housing Stock, As Determined from National Survey Units (Continued)
1
CJ
0
-
Q
(ft
Q
•ij
—.
3
NJ
O
V)
t5
^«
(O
(0
o>
National
Year Survey
Built ID
<1940 2410801
(cont.l 2441608
2521300
2541209
2542009
2550309
2551802
2651800
2110101
2121009
2931608
3011103
3011905
3020401
1940-1959 0340406
0341107
1312701
1722206
2230100
2611101
2731503
3040706
0120105
0131102
0131201
0251900
0310201
0320101
0321307
0351205
0410100
0411306
0411603
0520809
0531301
0612002
0651901
0710103
0750406
0821009
0911503
0920801
0921304
1010503
1030204
1051200
1120401
1121300
1130806
1140508
1332402
1333806
1352806
1410406
1440205
1450907
1521400
1521509
1530500
1550102
1550607
1551704
LBP
Present?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Floor
Dust-Lead
Loading
Iug/ft2)
18.5
13.0
28.0
77.4
20.5
3.34
36.5
42.6
17.0
29.5
153.
120.
21.8
17.4
7.73
10.6
2.11
75.6
15.1
2.45
18.9
6.31
25.1
190.
23.3
14.0
17.1
17.9
14.7
58.4
28.1
292.
22.1
134.
67.3
19.4
3.88
13.4
125.
6.36
72.2
5.93
6.02
240.
7.98
3.95
81.0
6.45
26 5
58.9
23.2
52.9
9.34
11.3
43.7
68.7
53.5
45.7
8.98
26.5
37.5
58.4
Floor
Dust-Lead
Cone.
(ug/g)
641.
370.
150.
161.
276.
61.6
142.
399.
265.
316.
814.
1310.
821.
330.
60.6
44.7
62.2
373.
32.2
57.0
137.
196.
116.
812.
144.
278.
120.
342.
162.
706.
18.6
1240.
215.
543.
241.
740.
84.1
243.
667.
101.
259.
80.1
138.
1560.
105.
118.
248.
63.0
258.
775.
276.
322.
95.7
174.
236.
73.5
173.
396.
166.
288.
419.
314.
Window Sill
Dust-Lead
Loading
Iug/ft2)
101.
503.
2.18
329.
549
34.3
24.2
10.0
2.36
1220.
517.
950.
958.
206.
2.89
9.37
4.30
15.4
2.46
9.76
161.
26.0 •
80.8
66.4
117.
21.4
15.1
5.84
0.04
1400.
11400.
4990.
64.3
2310.
35.9
18.5
141.
271. •
18.7
50.4
48.6
0.19
271. •
152.
271. •
8.31
12.5
3.04
3.22
33.8
11.9
12 7
4.55
7.47
83 6
0.57
182.
39.1
262.
4.39
211. •
11.5
Window Sill
Dust-Lead
Cone, (ug/g)
1320.
6250.
11.6
2050.
1660.
408.
61.1
15.7
6700. •
1980.
2060.
9040.
13200.
2710.
47.4
318.
316. •
183.
67.5
337.
1050.
316. •
300.
596.
589.
194.
277.
302.
1360. •
6530.
43800.
6570.
521.
19900.
183.
244.
417.
1360. •
378.
608.
989.
1360. *
1360. •
1920.
1360. •
1.02
132.
61.4
61.1
1230.
276.
195.
48.1
183.
772.
41.6
350.
1360. •
1080.
112.
1360. •
1360. •
Soil -Lead
Cone.
(ug/g)
290.
609.
35.0
28.6
125.
76.4
159.
47.4
613.
110.
1160.
1500.
2750.
1390.
25.2
41.6
36.
39. •
42.
15.
5. 0
43.
34.
60.
109.
198.
214.
209.
146.
81.4
43.2
122.
115.
341.
160.
135.
70.9
60.2
52.4
90.5
21.7
9.26
75.8
7030.
65.7
142.
99.0
144.
81.0
90.0
182.
61.1
71.3
130.
24.9
26.0
145.
132.
264.
209.
145.
136.
Drlpline/Entry
Soil -Lead
Cone, (ug/g)
378.
1010.
24.1
38.0
113.
132.
184.
49.4
1110.
183.
1150.
425.
5340.
702.
31.8
9.55
51.3
45.8 •
80.6
90.0
7.96
61.9
55.8
63.9
57.4
245.
lie.
284.
101.
135.
63.0
126.
152.
609.
145.
199.
81.4
366. •
61.9
131.
31.6
11.6
54.4
14E3.
89.6
118.
15.2
145.
98.4
64.2
309.
87.9
92.9
179.
32.4
31.2
142.
219.
249.
336.
155.
241.
Obs. Max.
Interior XRF
(mg/cm2)
5.9
9.4
0.50
1.5
8.2
0.60
1.3
0.60
2.6
2.9
3.9
12.
10.
0.60
0.60
0.60
0.60
—
--
0.30
0.20
—
1.6
1.5
0.60
0.60
0.60
1.0
--
3.2
2.9
0.50
7.0
0.40
0.40
0.50
0.90
0.60
0.70
1.1
0.40
0.50
0.50
11.
0.50
0.30
0.90
0.70
7.3
1.1
0 60
0.60
1.9
0.60
1.4
0 60
2 4
1.5
1.8
3 5
1 2
2 9
Obs. Hax.
Exterior XRF
Img/cm2)
7.6
3.9
0.60
0.50
0.90
6.6
0.50
0.30
5.0
0.50
7.7
6.9
3.3
5.3
0.60
-_
0.00
—
0.60
0.30
0.20
0.60
3.7
1.8
1.9
0.60
0.60
3.1
8.4
3.3
1.4
7.8
0.60
—
0.60
10.
0.60
0.60
1.4
-.
0.30
0.30
0.00
30.
0.50
0.40
2.6
17.
8.3
4.8
0.60
2.2
0 SO
1 9
2.2
0 50
2 8
13
3.1
2 1
2 0
2 6
Damaged
Interior
LBP Ift2)
238.6
0.0
0.0
7.0
139.9
0.0
0.0
0.0
0.0
0.0
28.8
0.9
6.6
0.0
0.0
0.0
0.0
0.0
--
0.0
0.0
—
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0 0
0.0
0.0
0.0
0.0
0.0
6.3
0.0
0.0
0.0
0.0
0 0
0.0
0.0
0.0
0.0
0.0
0.0
0 0
6.3
0 0
0.0
12.5
73.5
0.0
Damaged
Exterior
LBP (ft2)
77.3
0.0
0.0
0.0
0.0
0.0
0.0
0 0
0.0
0.0
0.0
0.0
0.0
16.5
0.0
0.0
0.0
—
0 0
0.0
0.0
0.0
0.0
0.0
10.3
0.0
0.0
0.0
0.0
33.7
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
5.1
0.0
0.0
6.5
0.0
0.0
0.0
0.0
0 0
0 0
0 0
0 0
0.0
0.0
278 5
56 0
3.0
0 0
0 0
1997
Height
121,752
121.752
244,799
114,632
114,632
244,199
244,799
244,799
244,799
244.799
183.864
111,365
773,094
773,094
19,676,320
258,519
273,941
227,108
108,151
213,598
181,223
181,223
227,108
273,941
273,941
108,151
273,941
273,941
273,941
273,941
258,519
108,151
213,598
213.598
108.151
181.223
213.598
108,151
227,108
227,108
227,108
227,108
227,108
227,108
213,598
213,598
213,598
181.223
213,598
213,598
181,223
291,118
227,108
227,108
273,941
213,941
258,519
221,109
221,108
221,108
221.108
221.108
221,108
1 Children
12-35 mo.
9,163
9,163
19,630
9,192
9,192
19,630
19,630
19,630
19,630
19,630
14,744
8,930
61,994
61,994
1,577,844
20.731
21.967
18.212
8.673
17,128
14,532
14,532
18,212
21.967
21,967
8,673
21.967
21,967
21,967
21,967
20,731
8,673
17,128
17,128
8,673
14,532
17,128
8,673
18,212
18,212
18,212
18,212
18,212
18,212
17,128
17,128
17,128
14,532
17,128
17, 128
14,532
23.345
18.212
18,212
21.967
21,967
20.731
18,212
18,212
18,212
18,212
19.212
18.212
« Children
12-71 mo.
25,016
25,016
50,299
23,553
23,553
50,299
50,299
50,299
50,299
50,299
37,779
22,882
158,848
158,848
4,042,893
53,118
56,281
46,664
22,222
43,888
37,236
37,236
46,664
56,287
56,287
22,222
56,287
56,287
56,287
56,287
53,118
22,222
43,888
43,888
22,222
37,236
43,888
22,222
46,664
46,664
46,664
46,664
46,664
46,664
43,888
43,888
43,888
37.236
43,888
43,888
31,236
59,816
46,664
46,664
56,281
56,281
53,118
46,664
46,664
46,664
46,664
46,664
46.664
-------
Table C-7. Estimated Environmental Lead Levels in the 1997 Housing Stock, As Determined from National Survey Units (Continued)
I
£k
Q
0
^
(D
^
c
1
(ft
fO
— »
Co
^
•>.
(D
"^
to
^"1
(a
(0
0>
National
Year Survey
Built ID
1940-1959 1730407
(cent.) 1730704
1730803
1731603
1750108
1831106
1831304
1840305
1841105
2022705
2030302
2110906
2141505
2142107
2211902
2332005
2343606
2421709
2441509
2451805
2520906
2540102
2540201
2541407
2541902
2610103
2651206
2652303
2711505
2730703
2731800
2812204
2840403
2841203
2841500
2910107
2931202
2940708
3011509
1960-1979 0130708
0131003
0150201
0330308
0350306
0420901
0430108
0440305
0440602
0541201
0940700
0940809
1020205
1020502
1021005
1040500
1323609
1441302
2220507
2230209
2511806
2521201
2551000
LBP
Present9
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
No
No
NO
No
No
No
No
No
No
No
No
No
No
No
No
Floor
Dust-Lead
Loading
(ug/ft2)
32.4
13.8
15.7
2.86
17.3
66.4
33.4
31.7
40.4
29.0
11.7
299.
2.08
15.3
41.5
53.3
11.3
23.8
75.8
9.11
73.6
133.
12.6
147.
3.65
18.7
12.7
12.4
40.4
6.45
49.9
32.0
51.4
90.7
14.8
24.0
39.1
11.3
10.5
10.6
19.4
46.9
8.41
7.98
28.9
13.2
42.1
18.1
5.85
5.63
18.2
7.88
9.89
5.64
11.7
5.69
4.17
18.7
6.10
6.20
39.9
4.28
Floor
Dust-Lead
Cone.
(ug/g)
162.
224.
88.4
17.5
316.
839.
445.
244.
286.
94.4
102.
1680.
32.6
93.7
61.7
755.
144.
169.
1700.
201.
321.
254.
271.
378.
65.4
284.
16.8
275.
210.
87.0
114.
484.
1070.
1270.
118.
232.
316.
232.
358.
88.5
110.
68.8
54.7
116.
20.2
68.8
245.
144.
22.2
48.2
458.
167.
161.
198.
210.
108.
92.6
123.
70.4
52.9
182.
52.8
Window Sill
Dust-Lead
Loading
Iug/(t2)
395.
13.5
94.2
271. •
9.74
240.
602.
232.
258.
25.5
10.2
560.
0.34
271. '
17.1
81.1
3.55
142.
78.0
424.
68.0
315.
41.1
29.6
143.
28.5
60.5
271. '
795.
16.5
345.
25.5
16.1
1250.
1.46
6.70
62.8
12.1
20.6
46.9
13.6
20.4
3.46
7.17
1820.
21.9
5.98
12.0
6.21
99.4 •
2.15
25.6
26.4
8.70
17.1
99.4 •
0.05
99.4 •
7.94
1.81
180.
4.10
Window Sill
Dust-Lead
Cone, (ug/g)
504.
1360. •
938.
1360. •
323.
1760.
2930.
1040.
1130.
338.
1360. •
4480.
1360. •
1360. •
246.
1270.
148.
159.
1510.
3230.
287.
31.6
178.
413.
1140.
341.
342.
1360. •
4000.
149.
510.
345.
145.
2760.
1360. •
1360. •
698.
261.
344.
203.
no.
16.6
48.1
261.
287.
172.
101.
154.
562.
541. •
533.
406.
1250.
541. •
156.
541. •
541. •
541. •
124
76. U
4930.
64.3
Soil-Lead
Cone.
(ug/g)
63.9
77.3
77.3
88.0
53.8
2570.
2570.
322. •
322. •
60.1
33.7
491.
58.9
123.
22.0
322. •
225.
52.4
4320.
34.1
55.8
102.
33.0
485.
116.
43.5
26.3
49.0
218.
119.
12.1
162.
52.1
61.9
41.4
51.8
188.
44.3
346.
29.7
5.35
6.16
61.6
14.2
21.0
21.3
97.4
79.3
17.9
7.23
17.7
49.2
58.3
25.5
24.5
20.4
13.0
14.1
5.58
11.6
73.4
22.6
Dripline/Entry
Soil -Lead
Cone, (ug/g)
73.5
83.7
83.7
88.0
70.6
2570.
2570.
366. •
366. •
75.2
47.7
491.
102.
99.5
33.1
366. •
246.
90.9
1680.
35.9
45.5
116.
44.0
109.
89.7
44.7
43.8
38.4
292.
237.
21.2
242.
62.5
78.7
31.8
78.1
188.
53.8
689.
31.5
7.87
9.49
105.
24.7
27.6
37.3
144.
128.
28.9
7.52
20.3
70.0
90.8
35.0
21.3
35.5
11.0
19.9
5.61
9.94
67.4
12.7
1 Obs. Max.
Interior XRF
Img/cm2l
1.8
2 1
1.8
1.5
1.2
2.0
0.60
20.
1.0
0.70
0.60
0.60
1.7
1.2
0.70
8.0
0.80
0.60
0.60
0.60
0.80
2.7
0.60
0.60
0.70
0.20
0.60
0.50
1.7
0.40
0.40
2.8
6.1
9.6
1.0
1.4
0.80
1.7
0.60
0.60
0.60
0.60
0.60
—
0.40
0.30
0.50
—
0.60
0.30
0.30
0.30
0.30
0.50
0.40
0.60
0.60
0.60
0.60
0.60
0.60
0.60
Obs. Max.
Exterior XRF
Img/cm2)
2.3
1.5
1.5
1.4
1.8
--
—
—
__
l.S
0.60
6.3
1.5
—
0.90
5.0
2.5
1.4
3.9
0.60
0.70
l.S
1.2
0.50
0.50
0.20
0.30
0.30
7.6
0.20
1.0
0.60
8.7
13.
1.8
0.50
0.50
l.S
1.4
0.60
0.00
0.50
0.60
—
0.60
0.60
0.60
—
0.60
0.30
0.30
0.50
--
0.30
--
--
0.50
0.60
0.60
0.50
0 10
0.50
Damaged
Interior
LBP Itt2)
4.8
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
201.9
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0 0
0.0
0.0
0.0
0.0
0.0
o n
(1.0
0.0
0.0
Damaged
Exterior
LBP (ft2)
0.0
0.0
0.0
0.0
0.0
--
—
--
--
0.0
0.0
7.3
2.5
—
0.0
77.1
0.0
0.0
118.3
0.0
0.0
20.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
o.o
0.0
0.0
0 0
0.0
0.0
0.0
0.0
--
0.0
0.0
0.0
0.0
0.0
0.0
1997
Height
108, 151
108.151
108,151
108,151
433,850
111,336
108,151
108,151
111,336
433,850
433,850
108,151
291,118
227,108
213,598
433,850
433,850
433. BSO
411,982
433,850
213,598
213,598
213,598
181,223
213,598
213,598
213,598
213,598
181.223
213,598
213,598
213,598
213,598
213,598
213,598
273,941
108,151
273,941
227,108
19,717,970
658,726
291,351
291,351
291,351
658,726
291,351
116,364
116,364
316,764
116,364
291,351
291,351
316,764
316,764
316,764
116,364
312,998
658,726
316.764
316.164
116,364
316,764
316,764
« Children
12-35 mo.
8,673
8.673
8,673
8,673
34,790
8,928
8,673
8,673
8,928
34,790
34,790
8,673
23,345
18,212
17,128
34,790
34,790
34,790
33,037
34,790
17,128
17,128
17.128
14,532
17,128
17,128
17,128
17.128
14,532
17,128
17,128
17,128
17,128
17,128
17,128
21,967
8,673
21,967
18.212
1.581,184
52,823
23,363
23,363
23,363
52,823
23,363
9,331
9,331
25,401
9,331
23,363
23,363
25,401
25,401
25,401
9,331
25,099
52,823
25,401
25,401
9,331
25.401
25,401
t Children
12-71 mo.
22,222
22,222
22,222
22,222
89,143
22,876
22,222
22,222
22,876
89,143
89,143
22,222
59,816
46,664
43,888
89,143
89,143
89,143
84,650
89,143
43,888
43,888
43,888
37,236
43,888
43.888
43,888
43,888
37,236
43,888
43,888
43,888
43,888
43,888
43,888
56,287
22,222
56,287
46,664
4,051,451
135,348
59,864
59,864
59,864
135,348
59,864
23,909
23,909
65,085
23,909
59,864
59,864
65,085
65,085
65,085
23,909
64,312
135,348
65,085
65,085
21,909
{•i.085
65,085
-------
1?
Table C-7. Estimated Environmental Lead Levels in the 1997 Housing Stock, As Determined from National Survey Units (Continued)
Floor Floor Window Sill
Obs. Max. Obs. Max. Damaged Damaged
Interior XRF Exterior XRF Interior Exterior
(mg/cm2) Img/cm2) LBP (ft2) LBP I(t2)
p-
o
5
•*
5'
^
— ^
c
0
n>
»B
to
KJ
C/J
Cft
^
O*
fl)
**
•S3
^
Co
<0
0)
National
Year Survey
Built ID
1960-1979 2552107
(cont.) 2B22005
2B31006
2831709
3050101
0130906
0150102
0250902
0252404
0311209
0331009
0340505
0340802
0341404
0410605
0421206
0430207
0430306
0430702
0440107
0441105
0441204
0530105
0530600
0531400
0540203
0541300
0621607
0631408
0840702
0911404
0930701
1011109
1020304
1020403
1020100
1020809
1050509
1050608
1051408
1150200
1150105
1241801
1311505
1312800
1322601
1353309
1441005
1510403
151090B
1520204
1530104
1530302
1530801
1531601
1531706
1540202
1540400
1540806
1541200
1741701
1141800
1143103
2040301
2122000
LBP
Present'
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Dust-Lead
Loading
Iug/ft2)
4.53
8.03
3.61
9.88
11. B
16.0
24.1
12.3
28.1
11.6
10.5
28.2
58.4
10.5
20.0
139.
36.6
51.2
54.2
52.6
13.3
55. 1
20.8
13.4
93.6
28.1
23.2
1.26
12.3
9.92
4.97
12.9
10.1
7.05
6.84
2.22
7.98
3.15
4.40
5.05
15.0
11.5
7.29
5.08
57.8
4.08
2.40
5.81
13.5
29.7
18.9
21.1
19.1
34.6
37.0
20.3
7.52
32.4
44.0
10.6
152.
31.9
3.79
16.4
85.5
Dust-Lead
Cone.
(ug/g)
39.6
66.3
34.5
65 1
487.
68.2
188.
207.
225.
330.
27.6
15.8
652.
34.3
61.0
643.
87.3
7.04
318.
319.
75.7
177.
246.
193.
143.
59.9
179.
64.1
221.
81.9
104.
388.
368.
159.
111.
52.4
285.
48.1
134.
135.
242.
183.
116.
82.8
251.
104.
16.7
48.1
246.
188.
228.
204.
327.
239.
290.
159.
149.
159.
175.
187.
141.
143.
29.6
124.
395.
Dust-Lead
Loading
Iug/ft2)
1.18
176.
0.31
11.3
99.4 •
10.1
5.22
1.10
394.
1070.
15.4
1.12
12.4
45.3
753.
B21.
26.8
12.7
227. •
176.
18.8
286.
59. 3
227. •
145.
227. •
177.
1.04
42.6
2.54
3.26
3.13
1.04
227. •
1.78
' 0.47
17.1
227. •
0.30
227. •
91.8
574.
2950.
3.32
0.93
1.60
0.57
1.82
227. •
3.66
13.3
78.9
515.
65.3
5940. •
541.
221. •
28.1
221. •
221. •
221. •
221. •
0.20
36.3
155.
Window Sill
Dust-Lead
Cone, (ug/g)
11.0
895.
541. •
134.
541. •
97.9
87.1
534.
481.
2470.
169.
45.5
241.
138.
520.
331.
102.
0.79
1540. •
296.
191.
230.
492.
1540. •
339.
1540. •
397.
1540. •
273.
1540.
1540.
58.7
1540.
1540.
1540.
1540.
1540.
1540.
1540.
1540.
705.
1650.
18400.
120.
53.4
1540. •
12.3
2B.7
1540. •
70.7
1540. •
170.
855.
404.
51200.
180.
1540. •
87.3
1540.
1540.
1540.
1540
1540.
472
494.
Soil-Lead
Cone.
(ug/g)
27.2
B2.5
21.1
40.8
6.68
39.5
4.79
180.
604.
1B6.
15.3
23.7
31. B
20.0
127.
22. B
35.2
27.2
26.4
34.7
5.22
81.4
50.9
215.
56.
14.
7. 2
39.
85.
30.
29.8
19.7
996.
26.6
25.0
23.8
25. 4
116.
57.5
143.
35.3
81.6
196.
20. B
13.8
33.3
51.6
18.8
4.63
14.8
35.9
78.7
68.4
40.5
105.
23.4
15.9
49.9
30.1
17.1
54.7
95.7
28.6
14.8
355
Dripllne/Ei
Soll-Lea<
Cone. {ug>
35.9
123.
14.5
30.7
10.5
34.3
7.49
335.
910.
337.
15.9
26.6
36.5
21.7
123.
21.6
65.2
49.2
40.0
38.3
5.22
91.7
52.1
303.
32.4
15.6
8.10
53.2
79.3
37.0
31.8
24.3
1710.
36.3
27.0
20.3
29.4
206.
72.3
259.
40.5
105.
207.
19.4
10.1
39.9
18.7
20.3
5.79
23.7
33.6
50.5
84.3
61.9
100.
28.4
14.5
31.7
32.5
26.0
60 8
85.3
23.1
20.6
685.
1997 « Children I Children
Weight 12-35 mo. 12-71 mo.
0.60
0.60
0.60
0.60
0.60
0.60
0.80
0.60
1.0
0.80
0.60
0.60
0.90
1.0
1.4
0.50
0.50
0.40
0.40
0.60
0.50
0.40
1.4
1.0
0.70
0.80
0.90
0.70
0.40
0.80
0.30
0.60
11.
0.80
0.70
0.60
0.40
3.0
0.30
0.30
1.0
1.6
0.60
0.60
0.90
0.60
0.60
1.5
0.50
0.30
0.30
3.3
0.90
0.60
22.
0.00
0.00
0.60
0 70
0.70
1.0
2.5
1.0
0.60
1.4
0.50
0.00
0.60
0.60
0.60
0.60
0.60
0.60
0.60
0.60
0.60
0.60
0.60
0.60
0.00
1.7
0.80
0.70
0.40
0.50
0.60
0.60
0.70
0.60
1.7
0.70
0.60
0.30
1.2
0.30
0.30
0.40
0.70
0.30
0.70
0.60
0.30
1.7
9.1
0.40
1.4
0.00
0.00
0.50
0.60
0.20
10.
0.10
1.5
2.5
11
1.3
0.70
0.20
0.30
0.10
0 80
0.90
3.3
3 6
0.60
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0 0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
28.8
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
12 5
0.0
0.0
0.0
0.0
0.0
0.0
9.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
o.o
5.9
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
21.5
0.0
0.0
0 0
0.0
0 0
0 0
0.0
0.0
0.0
0.0
316,164
316,164
316,164
116,364
312,998
126,312
658,126
352,318
291.351
352,318
352,318
658,126
352,318
658,126
291,351
291,351
316,164
316,164
116,364
316,764
116,364
316,764
116,364
291.351
116,364
316,764
316,764
126,372
126,372
291,351
173,719
312,998
116,364
316,764
316,764
316.764
316.764
316,764
316,764
116,364
291,351
291,351
451,561
173,719
312,998
312,998
291,351
658,726
173,719
312,998
312,998
312,998
312,998
312,998
173,719
312,998
173,119
312,998
312,998
173,719
243,025
451,561
451,561
451,561
291,351
25,401
25,401
25,401
9,331
25,099
10,134
52,823
28,252
23,363
28,252
28,252
52,823
28,252
52,823
23,363
23,363
25,401
25,401
9,331
25,401
9,331
25,401
9,331
23,363
9,331
25,401
25,401
10,134
10,134
23.363
13,931
25,099
9,331
25.401
25,401
25,401
25,401
25,401
25,401
9,331
23,363
23,363
36,211
13,931
25,099
25,099
23,363
52,823
13,931
25,099
25,099
25,099
25,099
25,099
13,931
25,099
13,931
25,099
25,099
13,931
19,488
36,211
36,211
36,211
23.363
65,085
65,085
65,085
23,909
64,312
25,966
135,348
12,391
59, 864
12,391
12,391
135,348
12,391
135,34B
59,864
59,864
65,085
65,085
23,909
65,085
23,909
65,085
23,909
59,864
23,909
65,085
65,085
25.966
25.966
59,864
35,694
64,312
23,909
65,085
65,085
65,085
65.085
65,085
65,085
23,909
59,864
59,864
92,182
35,694
64,312
64,312
59,864
135,348
35,694
64,312
64,312
64,312
64,312
64,312
35,694
64,312
35,694
64,312
64,312
35,694
49,934
92,182
92,182
92,182
59,864
-------
of
Table C-7. Estimated Environmental Lead Levels in the 1997 Housing Stock. As Determined from National Survey Units (Continued)
Floor Floor Window sill
National Dust-Lead Dust-Lead Dust-Lead Window Sill Soil-Lead Drlplme/Entry Obs. Max. Obs. Max. Damaged Damaged
Interior XRF Exterior XRF Interior Exterior
(mg/cm2) Img/cm2) LBP (ft2) LBP Ift2)
l.S
0.60
1 1
1.2
0.70
0.90
1.2
1.1
0.60
0.90
1.0
0.60
0.60
1.2
4.6
0.40
0.40
0.50
0.60
0.50
0.70
0.30
0.50
1.2
0.60
0.60
0.20
0.80
0.70
0.60
1.2
0.70
^^
o
.
c?
^
r\
fe>
0
*»»
1
N)
CO
Co
5
Year Survey
Built ID
1960-1979 2130706
Icont.) 2131902
2141604
2151207
2211308
2230506
2351500
2352201
2430403
2431807
2452605
2520609
2521102
2531804
2541506
2620508
2621704
2622603
2623007
2650208
2711109
2751402
2810307
2812105
2830602
2832004
2832103
2840106
2840205
2841401
2940401
30S1000
>1979 0130708
0131003
0150201
0330308
0350306
0420901
0430108
0440305
0440602
0541201
0940700
0940809
1020205
1020502
1021005
1040500
1323609
1441302
2220S07
2230209
2511806
LBP
Present'
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Loading
Iug/ft2)
27.2
8.19
29.1
30.3
14.6
8.73
16.4
12.6
16.7
22.1
15.4
28.7
17.0
4.07
124.
6.78
15.5
4.15
3.69
8. 85
24.3
398.
9.82
9.68
19.5
19.0
S.ll
26.6 •
26.6 •
8.94
4.13
5.99
10.6
19.4
46.9
8.41
7.98
28.9
13.2
42.1
18.1
5.85
5.63
18.2
7.88
9.89
5.64
11.7
5.69
4.17
18.7
6.10
6.20
Cone
(ug/g)
127
76.
87.
327.
89.
77.
136.
180.
480.
215.
320.
818.
117.
59.
2200.
142.
191.
2.
141.
94.
128.
54300.
146.
141.
no.
287
94
786.
786.
59.
66.
152.
88.
110.
68.
54.
116.
20.
68.
245.
144.
22.
48.
458.
167.
161.
198.
210.
108.
92.
123.
70.
52.
4
0
6
6
a
11
4
4
*
•
9
4
5
8
7
2
8
2
2
,6
,4
,9
Loading
Iug/ft2)
17.0
77.1
7.12
53.4
603.
19.3
270.
227. •
6.62
11.3
227. •
0.54
208.
0.97
380.
227. •
2.93
5.89
3.53
24.9
51.2
28.8
227. •
227. •
20.0
0.72
227. •
227. •
227. •
234.
2.76
227. •
46.9
13.6
20.4
3.46
7.17
1820.
21.9
5.98
12.0
6.21
101. •
2.15
25.6
26.4
8.70
17.1
101. •
0.05
101. •
7.94
1.81
Dust-Lead
Cone, (ug/g)
149.
291.
1540.
59.4
879.
562.
13000.
1540.
10.7
131.
1540.
1540.
200.
1540.
1300.
1540.
103.
55.5
105.
92.8
171.
89.6
1540.
1540.
519.
1540.
1540.
1540.
1540.
1190.
1540.
1540.
203.
170.
16.6
48.1
261.
287.
172.
101.
1S4.
562.
503.
533.
406.
1250.
503.
156.
503.
503.
503.
124.
26.0
•
*
*
*
•
*
•
»
*
•
•
•
*
•
•
•
•
•
•
Cone.
(ug/g)
13.7
21.1
39.2
17 5
20.4
6.11
115.
42.5
69.7
41.1
121.
15.7
26.8
66.4
45.2
49.5
53.0
54.6
26.0
52.7
32.0
27.3
23.2
91.3
32.1
20.8
75.6
63.1 •
63.1 •
35.6
27.2
31.1
29.7
5.35
6.16
61.6
14.2
21.0
21.3
97.4
79.3
17.9
7.23
17.7
49.2
58.3
25.5
24.5
20.4
13.0
14.1
5.58
11.6
Soil-
Cone.
15
21
58
27
- 33
S
141
35
97
57
119
12
34
82
71
49
83
59
28
83
21
27
29
126
33
18
30
83
83
34
26
48
31
7
9
105
24
27
37
144
128
28
7
20
70
90
35
21
35
11
19
5
9
Lead
(ug/i
.8
.4
.9
.5
.0
.84
.3
.8
.9
!«
.7
.9
.7
.5
.9 •
.8
.2
.9
.2
.3
.1
.
.0
.5
.8
.9 •
.9 •
.1
.6
.1
.5
.87
.49
.
.7
.6
.3
.9
.52
.3
.0
.8
.0
.3
.5
.0
.9
.61
.94
1997 * Children t Children
Height 12-35 mo. 12-71 mo.
0.60
0.60
0.60
0.60
0.40
0.30
0.50
0.60
0.30
0.30
0.30
0.30
0.50
0.40
0.60
0.60
0 60
0.60
0 60
1.
1.
0.
0.
1.
_-
0.
0.
3.
5.
0.
0.
1.
0.
0.
0.
8.
1.
0.
3.
0.
0.
0.
1.
1.
0.
0.
5.
2.
1.
—
0.
3
6
60
SO
3
10
10
4
1
60
SO
0
SO
50
60
a
1
60
0
80
60
00
5
5
60
20
1
0
6
60
57
0
0
0
0
0
1
0
0
0
0
0
0
0
12
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.3
.0
.0
.0
.0
.0
.1
.0
.0
.0
.0
.0
.0
.0
.8
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
0
0
0
0
8
0
0
0
72
0
0
0
0
0
0
0
0
0
0
6
0
0
0
0
0
0
0
20
25
0
_
0
.0
.6
.0
.0
.0
.0
.0
.0
.4
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.7
.7
.0
_
.0
173,
312,
173,
312,
316,
316,
243,
243,
451,
451,
451,
116,
316,
316,
116,
126,
126,
291,
126,
116,
316,
291,
291,
291,
116,
116,
316,
126,
316,
116,
6S8,
312,
719
998
719
998
764
764
025
025
561
561
561
364
764
764
364
372
372
351
372
364
764
351
351
351
364
364
764
372
764
364
726
998
13,
25,
13,
25,
25,
25,
19,
19,
36,
36,
36,
9,
25,
25,
9,
10,
10,
23,
10,
9,
25,
23,
23,
23,
9,
9,
25,
10,
25,
9,
52,
25,
931
099
931
099
401
401
488
488
211
211
211
331
401
401
331
134
134
363
134
331
401
363
363
363
331
331
401
134
401
331
823
099
35,
64,
35,
64,
65,
65,
49,
49,
92,
92,
92,
23,
65,
65,
23,
25,
25,
59,
25,
23.
65,
59,
59,
59,
23,
23,
65,
25,
65,
23,
135,
64,
694
312
694
312
085
085
934
934
782
782
782
909
DBS
085
909
966
966
864
966
909
085
864
864
864
909
909
085
966
085
909
348
312
0.60
0.00
0.50
0.60
0.60
0.60
0.60
0.60
0.30
0.30
0.50
0.30
0.50
0.60
0.60
0.50
34,984,547 2,805,411 7,188,275
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0 0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0 0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
—
0.0
0.0
0 0
0 0
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
889,038
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
71,292
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,671
182,67]
182,671
Kj
(O
$
-------
Table C-7. Estimated Environmental Lead Levels in the 1997 Housing Stock, As Determined from National Survey Units (Continued)
Year
Built
>1919
National
Survey
ID
2521201
2551000
2552101
2822005
2631006
2831109
3050101
Floor Floor window sill
Dust-Lead Dust-Lead Dust-Lead window sill Soil-Lead Dripline/Entry Otu. Max. Obs. Max. Damaged
LBP Loading Cone. Loading Dust-Lead Cone. Soil-Lead Interior XRF Exterior XRF Interior
Present? Iug/tt2) (ug/g) Iug/ft2) Cone, (ug/g) (ug/g) Cone, (ug/g) (mg/cn2) Img/cm2) LBP (Ct2)
Ho
No
No
No
No
Ho
No
39.9
4.28
4.53
8.03
3.67
9.88
11.8
182.
52. B
39.6
66.3
34.5
65.1
487.
180.
4.10
1.18
176.
0.31
11.3
101. •
4930.
64.3
71.0
895.
503.
134.
S03.
73.4
22.6
27.2
82.5
21.1
40.8
6.68
67 4
12.7
35.9
123.
14.5
30
10.5
0.60
0.60
0.60
0.60
0.60
0.60
0.60
0.10
0.50
0.50
0.00
0.60
0.60
0.60
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Damaged
Exterior
LBP Ift2)
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1991
Height
889,036
889,038
889,038
889,038
889,038
889,038
889,038
• Children
12-35 mo.
11,292
11,292
11,292
11,292
11,292
11,292
11,292
II Children
12-11 mo.
182,611
182,671
182,671
182,671
182,671
182,671
182,671
TOTALS ACROSS ALL UNITS:
24,893,064 1,996,175 5,114,778
99,211,901 7,960,614 20,397,397
• As no data for this parameter existed in the National Survey database for the given housing unit, this value is the average of the values across all units in the same
year-built category and having the same value for the LBP indicator that had reported data. The average is weighted using the 1997 weights.
Note: Dust-lead loadings are area-weighted arithmetic averages for the unit, with the loadings converted from Blue Nozzle vacuum loadings to wipe loadings (see Chapter
4). Dust-lead concentrations are mass-weighted arithmetic averages of individual sample concentrations for the unit that have been adjusted for low tap weights (see
Appendix Z). Two summaries of soil-lead concentration are presented: one represented a weighted arithmetic average for the unit, with remote sample results
weighted twice that of entryway and dripline samples; and one an unweighted average of dripline and entryway sample results only.
9
8
-------
APPENDIX D1
Assumptions and Scientific Evidence to Account
for the Effect of Pica for Paint
Draft - Do Not Cite or Quote D1 -1 September 27. 1996
-------
APPENDIX D1
ASSUMPTIONS AND SCIENTIFIC EVIDENCE TO ACCOUNT
FOR THE EFFECT OF PICA FOR PAINT
The scientific evidence on paint chip ingestion is scant and can be contradictory. It is
well known that pica for paint and plaster is associated with lead poisoning [1]. However, survey
data and blood-lead concentrations collected in the Rochester Lead-in-Dust Study [3] indicated
that children whose parents responded that they have a tendency to eat paint chips had blood-lead
levels only slightly more elevated, on average, than those who do not exhibit pica. The scientific
evidence and assumptions required to estimate the percentage of children who exhibit pica for
paint and their blood-lead levels are summarized in this section.
PERCENTAGE OF CHILDREN WHO INGEST PAINT CHIPS
In a study involving 2,402 children attending the Child Development Center of the
University of Virginia, de la Burde and Reames [1] reported that 9% of mothers of children
between eight months and seven years of age responded that their child exhibited pica for paint
or plaster. The children in the University of Virginia study were generally from low income
families and lived in substandard housing, where flaking paint or falling plaster were accessible.
Use of this data in the risk assessment assumes that the homes in the de la Burde study represent
the HUD National Survey homes where damaged lead-based paint is present. For the risk
assessment, the incidence of paint pica is assumed to be 9% for children living in homes with
damaged lead-based paint. Both children with recent paint chip ingestion and those who
ingested paint chips at some time are included in the 9%.
For HUD National Survey homes where no damaged lead-based paint is present, the
IEUBK model and EPI model, with the effect of pica set to zero, predicted values are used to
estimate blood-lead concentrations for all children represented by the home. When damaged
lead-based paint is present, the same predicted values are used to estimate blood-lead
concentrations under each model for 91% of the children, who are assumed not to ingest paint
chips. The modeling approaches differ for the remaining 9% of children, who are assumed to
Draft - Do Not Cite or Quote D1-2 September 27, 1996
-------
ingest paint chips. Because the EPI model incorporates the effect of pica for paint, EPI model
predicted values are used to estimate blood-lead concentrations for children who ingest paint
chips. The IEUBK model does not include a direct mechanism for estimating the effect of pica
for paint. The assumptions utilized in the risk assessment, to account for the effect of paint pica
under the IEUBK model, are described in the sections that follow.
Although the University of Virginia study was used to estimate the percentage of children
who ingest paint chips, children in this study would have been exposed to lead from sources,
such as automobile exhaust, no longer present in the environment. Thus their blood-lead levels,
if available, would not be comparable to those of present-day children. A current estimate of the
effect of pica for paint may be derived from a study conducted by the University of Rochester
[3]. In that study, 20 of 205 children (10%) were reported to exhibit pica for paint. The
geometric mean blood lead for children who were reported to have ingested paint chips was 9.1
ug/dL, while the geometric mean blood lead for children who were reported to have never
ingested paint chips was 6.1 ug/dL. This geometric mean blood-lead concentration for children
who ingested paint chips at some time is assumed to be 3.0 ug/dL greater than the IEUBK model
predicted geometric mean for children who do not ingest paint chips..
BLOOD-LEAD CONCENTRATION FOR CHILDREN WITH RECENT PAINT CHIP INGESTION
(IEUBK MODEL)
When the IEUBK model is used, the blood-lead concentration is set equal to 63 ug/dL for
children who have recently ingested paint chips. The basis underlying this blood-lead
concentration and the percentage of children assumed to have recently ingested paint chips are
discussed in this section.
Because the opportunity for pica arises only when paint chips are available, the effect of
pica for paint will be applied only for HUD National Survey homes where damaged lead-based
paint is present. Forty-one of the 284 homes in the HUD National Survey have damaged lead-
based paint. These homes represent 15.2% of U.S. housing, based on 1997-projected weights
used in the risk assessment.
Draft - Do Not Cite or Quote D1-3 September 27, 1996
-------
Of the 924 children ages 1-2 years in the NHANES III Survey [4], just one child had a
blood-lead level greater than 40 ug/dL. The percentage of children ages 1-2 with blood lead
greater than 40 |ig/dL, adjusted for sampling weights, is 0.03%.
It is assumed that blood-lead levels greater than 40 ug/dL are extremely rare in homes
with no damaged lead-based paint. Thus the entire 0.03% of children nationwide with blood lead
greater than 40 |ig/dL are assumed to reside in the 15.2% of homes with damaged lead-based
paint. Combining these figures, we estimate that 0.20% of children in homes with damaged
lead-based paint have blood-lead levels greater than 40 ug/dL.
A St. Louis study [2] found that 13 of 90 (14.4%) children less than age 3 years with
blood-lead levels greater than 40 ug/dL, or less than seven years with blood lead levels greater
than SO ng/dL, had radiographic evidence of recent paint chip ingestion. This information,
combined with the preceding estimate, leads us to conclude that 0.03% of children in homes with
damaged lead-based paint have blood lead greater than 40 ug/dL due to recent paint chip
ingestion. Table Dl-1 shows step by step the methodology for computing the percentage of
children living in homes with damaged lead-based paint who have blood-lead levels greater than
40 ug/dL and have recently ingested paint chips. The underlying assumptions of this approach
are that 1) blood-lead concentrations are greater than equal to 40 for children who have recently
ingested paint chips containing lead and 2) only children who reside in homes with damaged lead
based paint can ingest paint chips containing lead. The 13 children in the St. Louis study, who
were confirmed to have ingested paint chips, had a mean blood-lead level of 63 ug/dL. The
blood lead levels of children with recent pica (0.03% of children in homes with damaged lead-
based paint) will be mapped to 63 ug/dL.
BLOOD-LEAD CONCENTRATION FOR CHILDREN WHO INGESTED PAINT CHIPS AT SOME
TIME (IEUBK)
For HUD National Survey homes with damaged lead-based paint, 9% of the children
represented by those homes are assumed to ingest paint chips, with 0.03% of children assumed to
have recent paint chip ingestion, as described above. The remaining 8.7% of children are
assumed to have ingested paint chips at some time, but not recently. The geometric mean blood-
Draft - Do Not Cite or Quote D1-4 September 27, 1996
-------
lead concentration for the 8.7% of children in homes with damaged lead-based paint, who have
ingested paint chips at some time, is estimated to be 3 ng/dL greater than the IEUBK predicted
value for children who do not eat paint chips. The basis for this adjustment is presented in this
section.
Table D1-1. Calculation of Percentage of Children Who Have Recently Ingested Paint
Chips.
Variable Name
PC_EAT
-------
APPENDIX D2
Results of Three Published Meta-Analyses on the Relationship
Between IQ Point Loss and Childhood Blood-Lead Levels
Draft - Do Not Cite or Quote D2-1 September 27. 1996
-------
APPENDIX D2
RESULTS OF THREE PUBLISHED META-ANALYSES ON THE RELATIONSHIP
BETWEEN IQ POINT LOSS AND CHILDHOOD BLOOD-LEAD LEVELS
INTRODUCTION
The association between blood-lead levels and low IQ scores has been consistently
reported in the scientific literature. The estimates of the dose-response relationship published in
the literature have been combined via meta-analysis and reported in the three articles listed
below. This appendix provides a summary of each article and a discussion of the key results,
relative to the §403 Risk Assessment. The studies cited in these articles are summarized in
Tables D2-landD2-2.
PRIMARY REFERENCES
Schwartz, J., 1993, Beyond LOEL's, p Values, and Vote Counting: Methods
for Looking at the Shapes and Strengths of Associations, Neuro Toxicology
14(2-3):237-246.
Schwartz, J., 1994, Low-Level Lead Exposure and Children's IQ: A Meta-analysis and
Search for a Threshold, Environmental Research 65:42-55.
Pocock, S. J., Smith, M., and Baghurst, P., 1994, Environmental Lead and Children's
Intelligence: A Systematic Review of the Epidemiological Evidence, BMJ 309:1189-
1197.
SUMMARY OF SCHWARTZ. J.. 1993
This paper uses examples from the lead literature to illustrate statistical methods for
determining the shape of dose-response relationships, including the possible existence of
thresholds, and for assessing the strengths of associations within a study and for the literature as a
whole. Of interest to the §403 Risk Assessment is a meta-analysis of the results from 7
Draft - Do Not Cite or Quote D2-2 September 27, 1996
-------
studies that estimated a slope for the relationship between children's blood-lead levels and IQ
scores. These studies used linear, or log-linear, regression models to fit the relationship between
IQ scores and PbB in children. Up to 17 additional covariates were included in the models. The
weighted mean regression slope over the 7 studies, weighted by the inverse of the estimated
variance, was -0.245 (±0.039). That is, a 1 ug/dL increase in PbB was associated with a 0.245
decrease in IQ score.
SUMMARY OF SCHWARTZ. J.. 1994
This article focuses on the relationship between blood lead and IQ scores, while the
earlier paper by Schwartz used this relationship to illustrate a statistical method. The 1994 paper
presents a meta-analysis of 7 studies, some of which had been cited in the earlier paper, that
estimated a slope for the relationship between children's blood-lead levels and IQ scores. Three
longitudinal and four cross-sectional studies were included in the analysis. The studies used
linear, or log-linear, regression models to fit the relationship between IQ scores and PbB in
children. Additional covariates were included in the models. The weighted mean regression
slope over the 7 studies, weighted by the inverse of the estimated variance, was -0.257 (±0.041).
That is, a 1 ug/dL increase in PbB was associated with a 0.257 decrease in IQ score.
The paper also presents a sensitivity analysis, summarized as follows:
Revised Analysis: resulting slope (± 1 standard error of the mean)
Study with Largest Effect Size Removed: -0.243 (±0.034)
Study with Most Significant Effect Removed: -0.252 (±0.058)
Add 8 Studies with No Effect (each with average weight of the 7 studies): association still
significant, but slope reduced to about half of original estimate
Longitudinal vs. Cross-sectional:-0.296 (±0.125) vs. -0.269 (±0.051)
Disadvantaged vs. Nondisadvantaged Lifestyle: -0.185 (±0.092) vs. -0.289 (±0.050)
Add 2'Studies that Included Younger Children: -0.239 (±0.031)
The question of whether a threshold exists in the relationship between IQ scores and
PbBs was examined through a meta-analysis that compared studies with different mean blood
lead levels. In studies with mean blood lead levels of 15 ug/dL, the estimated slope was -0.323
Draft - Do Not Cite or Quote D2-3 September 27, 1996
-------
(±0.126) compared to -0.232 (±0.040) for studies with means above 15 ug/dL. Thus, if anything,
a trend toward a higher slope at lower concentrations was observed.
An alternative approach to the threshold issue examined the data from the Boston study
(Bellinger, 1992) more thoroughly. The Boston study was chosen because it had the lowest
mean PbB. For this analysis, separate regression models for IQ score and PbB were fit using the
same set of covariates. A nonparametric smoothed curve (LOESS) was fit to the relationship
between the two sets of residuals. Based on this analysis, the author concludes that the
relationship between blood lead and IQ continues at PbB below 5 ug/dL in this study, i.e. no
threshold was evident.
SUMMARY OF POCOCK. S. J.. SMITH. M.. AND BAGHURST. P.. 1994
This paper presents a systematic review and meta-analysis of 26 epidemiological studies:
5 prospective studies, 14 cross-sectional studies of blood-lead, and 7 cross-sectional studies of
tooth-lead. The three types of studies are considered in separate meta-analyses. The results are
summarized follows:
Analysis: resulting slope (± 1 standard error of the mean)
Prospective Studies, PbB at Birth: 0.018 (±0.062)
Prospective Studies, PbB around 2 Years: -0.185 (±0.051)
Prospective Studies, Postnatal Mean PbB: -0.088 (±0.058)
Cross-Sectional Blood-Lead Studies: -0.253 (±0.041)
Cross-Sectional Blood-Lead Studies, Excluding Shanghai: -0.174 (±0.043)
Cross-Sectional Tooth-Lead Studies: -0.095 (±0.025)
Only the analysis of cross-sectional blood-lead studies had a statistically significant slope.
Draft - Do Not Cite or Quote D2-4 September 27, 1996
-------
DISCUSSION
There was considerable overlap in the studies cited by the three meta-analysis papers.
Two studies, Fulton et al. (1987) and Yule et al. (1981), were cited in all three papers, while
several others were cited in two of the three papers. In addition, some studies cited by Schwartz
(1993) or Pocock were used by Schwartz (1994) in the sensitivity analysis.
The three papers are directly comparable in that a common endpoint was used for all
meta-analyses. For the meta-analysis endpoint, the regression coefficients and standard errors
calculated by the original authors were used to estimate the change in IQ for an increase in
blood-lead from 10 to 20 ug/dL. This was necessary, because some of the original authors
worked with log-transformed data, while others did not transform the data. In most cases, the
regression coefficients were adjusted for other covariates included in the model. The other
covariates varied from study to study. For §403 decision-making, we have converted the
estimated change in IQ back to a slope value for untransformed blood-lead data.
The Schwartz (1993) paper focuses on introducing the statistical methods to a non-
technical audience. The Schwartz (1994) and Pocock papers focus on the relationship between
IQ and blood-lead levels. The Schwartz (1994) paper includes a sensitivity analysis and search
for threshold in the relationship. These topics are not covered in the Schwartz (1993) and
Pocock papers. However, in the meta-analysis of prospective studies, the Pocock paper does
include separate analyses for blood-lead measures at three ages. Also, one of the studies
(Schroeder, 1985) used in the Schwartz (1993) paper included approximately 50 children under
30 months of age. This study and another (Emhart, 1989) with younger children were included
in the sensitivity analysis in Schwartz (1994).
The Pocock paper analyzes prospective and cross-sectional studies separately, while the
Schwartz papers include both types of studies in the same meta-analysis. In the analysis of
prospective studies, Pocock includes an analysis of how PbB at approximately age 2 affects IQ
measured at school age. The slope for this analysis (-0.185) is less than the values
(approximately -0.25) from Schwartz (1993 and 1994) and the Pocock cross-sectional studies
analysis.
Draft - Do Not Cite or Quote D2-5 September 27, 1996
-------
Both Schwartz (1994) and Pocock included "full scale IQ score" in school-age children as
a selection criteria for studies used in the meta-analysis. Most of the studies cited used the
Wechsler Intelligence Scale for Children - Revised (WISC-R) test. The 1993 Schwartz paper
includes one study, Schroeder (1985), that uses the Bayley Scales of Infant Development (BSID),
for children less than 30 months of age. The BSID score is not directly comparable with the IQ
scores, as this test measures developmental endpoints as well as cognitive ability.
Draft - Do Not Cite or Quote D2-6 September 27, 1996
-------
Table D2-1. Design Information for Studies that Investigate the Relationship Between Child's IQ and Blood-Lead Level
Primary
References
That Cite the
Study
Schwartz (1993)
Schwartz (1994)
Pocock
Schwartz ( 1993)
Schwartz (1994)
Pocock
Schwartz (1994)
Pocock
Pocock
Pocock
Schwartz (1993)
Schwartz (1993)
Schwartz (1994)
Schwartz (1994)
Pocock
Study
Hatzakis et al.
(1987)
Hatzakis et al.
(1989)
Bellinger et al.
(1991)
Bellinger et al.
(1992)
Baghurst et al.
(1992)
Ernhart et al.
(1989)
Cooney et al.
(1991)
Schroeder et al.
(1985)
Hawk et al.
(1986)
Dietrich et al.
(1993)
Type of Study
Prospective
Prospective
Prospective
Prospective
Prospective
Prospective
Prospective
Prospective
Replication of
Schroeder Study
Prospective
Year(s) of
Study
1985
Mid- to late-
1980s
1979(Aug.)-
198 11 April)
1979-1982
1983-1990
1977-1978
Location of Study
Participants
Lavrion, Greece (a
lead smelter city;
soil lead levels of
1,300-1 8,000 ppm)
Lavrion, Greece (a
lead smelter city)
Boston, MA
Boston, MA
Port Pirie, Australia
Cleveland, OH
Sidney, Australia
Wake County, NC
Lenoir & New
Hanover counties,
NC
Cincinnati, OH
Age of Study
Participants
Blood Lead
Measure
24 months
0 - 3 yrs
at 2yrs
1 and 2 yrs
0 - 3 yrs
IQ Measure
Primary
school age
6- 12 yrs
Approx. 57
mos
School Age
7 yrs
5 yrs
7 yrs
1 0 mos -
6.5 yrs
(half < 30
mos)
3-7 yrs
Approx.
6.5 yrs
IQ Test
Instrument
WISC-R
WISC-R
GCI
WISC-R
WISC-R
WPPSI
WISC-R
BSIDK30
mos)
SBIS it 30
mos)
SBIS
WISC-R
Sample
Size
509
509
150
147
494
212
175
104
75
231
Other Study Information
Study participants enrolled
in one of four schools in
the town in 1984-85.
Middle and upper-middle
class families, not in inner-
city or housing projects.
Children born at Brigham
and Women's Hospital
from 1979-1981
Middle class, advantaged
Smelter town and rural
surroundings, middle class
families
Inner city, disadvantaged,
50% of mothers alcoholic
Mixed urban
Low income families
Black study participants
from low income and SES
families, at high risk of
exposure to deteriorated
LBP
Inner city, black,
disadvantaged
?
D
NJ
-------
s
Table D2-1. Design Information for Studies that Investigate the Relationship Between Child's IQ and Blood-Lead Level
(Continued)
o
I
Primary
References
That Cite the
Study
Schwartz (1993)
Schwartz (1994)
Pocock
Schwartz (1993)
Pocock
Pocock
Schwartz) 1994)
Pocock
Pocock
Pocock
Pocock
Schwartz (1993)
Schwartz (1994)
Pocock
Study
Yule et al.
(1981)
Lansdown et al.
(1986)
Winneke et al
(1990)
Silva (1988)
Harvey et al
(1988)
Wang et al
(1989)
Winneke et al
(1985)
Fulton et al.
(1987)
Type of Study
Pilot Study
Replication of
Yule Study
Multi-Center,
Cross - Sectional
Study
Cross - Sectional
Cross - Sectional
Cross - Sectional
Cross - Sectional
Cross - Sectional
Year(s) of
Study
Summer 1980
(PbB taken 9-
12 months
earlier)
1972-1973
Late 1979-
earh/1981
1983-1985
Location of Study
Participants
Outer London,
England
Within 1 km of a
factory in London,
England
Bucharest
Budapest
Moden
Sofia
Dusseldorf
Dusseldorf
Dunedin, New
Zealand
Birmingham,
England
Shanghai, China
Nordenham,
Germany
Edinburgh, Scotland
Age of Study
Participants
Blood Lead
Measure
IQ Measure
6-12yrs
6-12yrs
9.2 yrs
(mean age)
8.5 yrs
(mean age)
7.8 yrs
(mean age)
7.3 yrs
(mean age)
6.5 yrs
(mean age)
8.3 yrs
(mean age)
11 yrs
(mean age)
5.5 yrs
(mean age)
6- 14 yrs
7 yrs
6-9 yrs
Idlest
Instrument
WISC-R
WISC-R
WISC-Short
Form
WISC-Short
Form
WISC-Short
Form
WISC-Short
Form
WISC-Short
Form
WISC-Short
Form
WISC-R
WPPSI
WISC-R
WISC-R
BAS
Sample
Size
166
166
301
254
216
142
109
109
579
177
157
122
501
Other Study Information
Results for younger
children are reported
elsewhere.
Mostly middle class
families with homes near a
main road
General population
General population
Industrial city, lead industry
General population
Industrial city, near smelter
Industrial city, near smelter
Mixed urban and rural
Mixed, inner city
Near battery plant, rural
control
Smelter town, rural
surroundings
Study participants enrolled
in one of 1 8 primary
schools
Mixed Urban - previously
hish water lead
o
V
CO
t
*
M
O)
-------
Table D2-2. Summary of Results from Studies that Investigate the Relationship Between Child's IQ and Blood-Lead Level
Primary
References
That Che the
Study
Schwartz
(1993)
Schwartz
(1994)
Pocock
Schwartz
(1993)
Schwartz
(1994)
Pocock
Schwartz
(1994)
Pocock
Pocock
Pocock
Study
Hatzakis
et al.
(1987)
Hatzakis
etal.
(1989)
Bellinger
etal.
(1991)
Bellinger et
al
(1992)
Baghurst
etal
(1992)
Ernhart et
aid 989)
Cooney et
aid 991 )
PbB of Study
Participants111 (pg/dL)
Range
7.4-
63.9
7.4-
63.9
0.0-
23.3
Summary
Statistics
AM = 23.7
STD = 9.2
10%ile = 13.9
50%ile = 21.5
90%ile = 36.0
AM = 23.7
STD = 9. 2
AM = 6.4
STD = 4.1
19% were
>10//g/dL
4% were
>15/ig/dL
AM=6.5
STD=4.9
AM = 20
AM=16.7
STD = 6.45
AM = 14.2
IQ of Study Participants121
Endpoint
Type
WISC-R
GCI
WISC-R
WISC-R
WPPSI
WISC-R
Range/Summary
Statistics
AM = 87.7
STD = 14.8
80-150
AM = 115.5
STD = 14.5
71-147
AM=119.1
STD = 14.8
AM =104.7
AM = 87. 5
STD = 16.6
Measure of Association Between IQ and Blood-Lead Levels111
Measure
-0.270 change in IQ
per unit increase in
PbB
(-0.403, -0.137)
-2.7 change in IQ
for increase from
10-20/;g/dLinPbB
-2. 28 change in IQ
per unit increase in
Log(PbB)
(-6.0. 1.4)
-0.250 change in IQ
per unit increase in
PbB from 5-15
pg/dL PbB
-5.8 change in IQ
for increase from 1 0
to 20 //g/dL in PbB
-3.3 change in IQ
for an increase from
10-20/;g/dLinPbB
-1.1 change in IQ
for an increase fi;j<»
10-20/;g/dLinPbB
0.39 change In IQ
for an increase liun;
10-20 //g/dL in PbB
P-Value
<0.001
<0.001
0.23
0.007
0.04
<0.01
Covariates
1 7 potential
Lunfounders or IQ
correlates141
(called the
'optimal* model)
Up to 24,
including mother's
IQ
1 3 covariates161
HOME mother's
IQ,
8 other
covariates1"
HOME, mother's
IQ, 1 1 others '"
HOME , mothers
IQ, and 1 1
others""
HOME .mothers
IQ. and 4
others "2I
Other Information
Dose-response investigation
showed no PbB effect on IQ
when PbB < 25 //g/dL.
Dose-reponse curve showed
evidence of a threshold at the
level of about 25 //g/dL PbB
Regression diagnostics were
used to check the robustness of
estimates. These results reflect
only PbB data at age 57
months.
Slightly elevated blood lead
levels around the age of 24
months are associated with
intellectual and academic
performance deficits at age 10
years.
Found low-level exposure to
lead during early childhood is
inversely associated with
neuropsychological development
through first seven years of life.
a
V
CO
t
I
ISJ
JO
(o
-------
Table D2-2. Summary of Results from Studies that Investigate the Relationship Between Child's IQ and Blood-Lead
Level (Continued)
O
o
I
I
3
Primary
References
That Cite the
Study
Schwartz
(1993)
Schwartz
(1993)
Schwartz
(1994)
Schwartz
(1994)
Pocock
Schwartz
(1993)
Schwartz
(1994)
Pocock
Schwartz
(1993)
Pocock
Study
Schroeder
et al.
(1985)
Hawk et
al. (1986)
Dietrich et
all 1993)
Yule et al.
(1981)
Lansdown
etal.
(1986)
PbB of Study
Participants111 U/g/dL)
Range
6-58
6.2-
47.4
7-33
7-24
Summary
Statistics
AM = 20.9
STD = 9.7
AM =15.2
STD = 11.3
AM = 13.52
STD = 4.13
80% were
>10//g/dL
4.8% were
>20/yg/dL
AM = 12.75
STD = 3.07
77% were
>10//g/dL
1 .5% were
>20jig/dL
IQ of Study Participants121
Endpoint
Type
BSID
«30
mo.)
S8IS
U30
mo.)
SBIS
WISC-R
WISC-R
WISC-R
WISC-R
Range/Summary
Statistics
45-140
59-118
AM = 86.9
STD = 11. 3
AM — 98.21
STD = 13.44
AM = 105.24
STD = 14.20
Measure of Association Between IQ and Blood-Lead Levels111
Measure
-0.199 change in IQ
per unit increase in
PbB
-0.255 change in IQ
per unit increase in
PbB
(-0.554, 0.043)
1 .3 esimated loss in
IQ for an increase
from 1 0 to 20 //g/dL
in PbB
-8.08 change in IQ
per unit increase in
Log(PbB) (4.63)
-0.560 change in IQ
per unit increase in
PbB from 10-20
j/g/dL
2. 15 change in IQ
per unit increase in
Log(PbB)
0 149 change in IQ per
unit increase in PbB
from 10-20 ng/dL
P-Value
<0.01
<0.05
<0.10
0.084
0.63
Covariates
7 covariates1" plus
interaction with
PbB. Quadratic
and cubic
components of
PbB also
considered.
Gender. HOME
score, maternal IQ
HOME score.
maternal IQ, birth
weight, birth
length, child sex,
cigarette
consumption
during pregnancy
Age, social class
Age, social class
Other Information
Unforced stepwise regression.
SES was only other significant
covariate.
Postnatal PbB concentrations
were inversely associated with
Full Scale IQ.
Social class was considered a
crude measure.
N = 86 for regression analysis.
Social class was also a
significant factor.
O
ro
I
I
M
Cft
-------
Table D2-2. Summary of Results from Studies that Investigate the Relationship Between Child's IQ and Blood-Lead
Level (Continued)
Primary
References
That Cite the
Study
Pocock
Schwartz
(1994) Pocock
Pocock
Study
Winneke
et al
(1990)
Bucharest
Winneke
et al
(1990)
Budapest
Winneke
etal
(1990)
Moden
Winneke
etal
(1990)
Sofia
Winneke
etal
(1990)
Dusseldorf
Winneke
etal
(1990)
Dusseldorf
Silva
(1988)
Harvey et
aid 988)
PbB of Study
Participants'11 (pg/dL)
Range
4-50
j/g/dL
0.2-
1.4
mol/L
Summary
Statistics
GM=18.9
STD = 1.3
GM=18.2
STD = 1 .7
GM=11.0
STD =1.3
GM=18.2
STD = 1.6
GM = 8.3
STD = 1.4
AM = 7.4
STD =1.3
AM = 11.1
STD = 4.91
AM = 12.3
STD = 0.2
IQ of Study Participants121
Endpoint
Type
WISC-
Short
Form
WISC-
Short
Form
WISC-
Short
Form
WISC-
Short
Form
WISC-
Short
Form
WISC-
Short
Form
WISC-R
WPPSI
Range/Summary
Statistics
AM = 116
AM =108.9
STD = 15.12
AM = 105.9
STD = 10.6
Measure of Association Between IQ and Blood-Lead Levels131
Measure
Loss of 1.51 in IQ
for an increase in
PbB of 10-20/yg/dL
P-Value
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Covariates
Gender, age, social
class, mother's
education
Gender, age, social
class
Gender, age, social
class, mother's
education
Gender, age, social
class, mother's
education
Gender, age, social
class, mother's
education
Gender, age, social
class, mother's
education
None
None
Other Information
No significant relationship was
found between overall IQ and
PbB
o
5?'
I
o
M
I
M
(O
-------
L?
Table D2-2. Summary of Results from Studies that Investigate the Relationship Between Child's IQ and Blood-Lead
Level (Continued)
Primary
References
That Cite the
Study
Pocock
Pocock
Schwartz
(1993)
Schwartz
(1994)
Pocock
Study
Wang et al
(1989)
Winneke
etal
(1985)
Fulton et
al. (1987)
PbB of Study
Participants"1 (//g/dL)
Range
4.5-
52.8
pg/dL
4.4-
23.8
//g/dL
3.3-
34
Summary
Statistics
AM = 21.1
STD= 10.11
AM = 8.2
STD = 1 .4
GM = 11.5
1 .2% were
>25jig/dL
IQ of Study Participants"1
Endpoint
Type
wise
WISC-R
BASC
Range/Summary
Statistics
AM = 89
AM =120.2
STD = 10.3
AM = 112
STD = 13.4
Measure of Association Between IQ and Blood-Lead Levels131
Measure
A decrease of IQ of
9 per 10/yg/dL
increase in PbB
-3.70 change in IQ
per unit increase in
Log(PbB) (1.31)
-0.256 change in IQ
per unit increase in
PbB from 1 0-20
Xig/dL
P-Value
<0.1
0.003
Covariates
Mother's
education and 4
others l101
Age, sex and
hereditary
background
1 3 covariates1" +
school attended
("optimal*
regression model)
Other Information
Found a dose - effect
relation between PbB and IQ
even after confounding variables
were controlled for by stepwise
regression analysis
Adjusted R2 = 45.5%
o
ro
ro
I
Kj
»•*
(o
-------
Notes for Table D2-2:
111 "Range" indicates the observed range of PbB levels among the study participants. Among the summary statistics, AM =
arithmetic mean; GM = geometric mean; STD = standard deviation; x%ile = x percentile of observed distribution.
121 "Type" indicates the type of IQ endpoint measured in the study. WISC-R = Wechsler Intelligence Scale for Children -
Revised (full-scale IQ measurement); GCI = McCarthy Scales of Children's Abilities: General Cognitive Index; BSID =
Bayley Scales of Infant Development; SBIS = Stanford-Binet Intelligence Scale; BASC = British Ability Scales: Combined
Score. Among the summary statistics, AM = arithmetic mean; STD = standard deviation.
131 Results are the outcome of a regression analysis to predict IQ endpoint based on PbB level and other covariates.
"Measure" is the estimated slope parameter indicating the change in IQ measurement associated with a unit change in the
(possibly transformed) PbB level. If the PbB level is transformed, the change in IQ measurement over a given range of the
untransformed PbB level is also given. When available, a 95% confidence interval associated with the slope estimate is
given, or a standard error associated with the estimate. "P-value" is for the test that the slope parameter is equal to zero
versus an alternative that it is not zero. "Adjusted covariates" indicates the number of covariates included in the
regression model; these covariates are named if the number is small. "Other information" indicates specifics associated
with the regression fit (e.g., method used, whether a log-transformation was taken on the PbB level prior to analysis,
information on the covariates).
141 Covariates include parental IQ, birth order, family size, father's age, parental education, alcoholic mother, age,
bilmgualism, birth weight, length of child's hospital stay after birth, walking age, history of CNS disease, history of head
trauma, illness affecting sensory function, parent's divorce.
151 Covariates include family social class, material IQ, preschool attendance, HOME total score, # hours per week of "out-of-
home" care, # changes in family residence since birth, medication use in preceding month, # adults in household, gender,
race, birth weight, material marital status, birth order.
161 HOME score, maternal IQ, child's age, child's sex, SES of parents, type of IQ test, presence of father in home, number of
siblings.
171 Parent's vocabulary and matrices tests, child's interest score, age, father's qualifications, length of gestation, parental
involvement with school score, class year, # days absent from school, sex, standardized height, car/telephone ownership,
employment status of father.
181 Child stress, maternal age, race, SES, sex, birth order, martial status, number of residence changes prior to age 57
months
191 Sex, parents' level of education, maternal age at delivery, parents' smoking status, socio-economic status, quality of the
home environment, birth weight, birth order, feeding method (breast feeding, bottle, or both), duration of breast - feeding,
and whether the child's natural parents were living together
1101 Age, sex, father's education, father's occupation, father's daily smoking quantity
1111 Sex, race, birth weight, birth order, gestational age at birth, parental education, maternal variables like PPVT-R, API,
MAST SCORE, AA/day in pregnancy, cigarettes per day, and use of marijuana and other drugs in pregnancy, medical
problems and psychosocial problems.
1121 Gestational age, education of the mother, education and occupational status of the father.
Draft - Do Not Cite or Quote D2-13 September 27, 1996
-------
APPENDIX E1
Generating Distribution of Blood-lead Concentrations
Based on Model-predicted Geometric Mean and Geometric
Standard Deviation
Draft - Do Not Cite or Quote E1-1 September 27, 1996
-------
APPENDIX E1
GENERATING DISTRIBUTION OF BLOOD-LEAD CONCENTRATIONS
BASED ON MODEL-PREDICTED GEOMETRIC MEAN AND GEOMETRIC
STANDARD DEVIATION
Geometric mean blood-lead levels predicted for specific conditions are not sufficient to
characterize the national distribution of children's blood-lead levels, nor are they sufficient to
estimate the arithmetic average blood-lead level in the nation. This section discusses how the
geometric mean blood-lead concentrations predicted at each housing condition were combined to
characterize the national distribution of children's blood-lead levels for children aged 1-2.
First, as described above, either the EPI or the IEUBK model was used to predict the
geometric mean blood-lead concentration for specific housing conditions defined by the
demographic and environmental-lead variables. If blood-lead concentrations have a log-normal
distribution, then the geometric mean represents the predicted median blood-lead concentration
for the distribution of childhood blood-lead concentrations. Rather than lumping the population
of children associated with each specific housing condition to a single blood-lead concentration
represented by the predicted geometric mean they were allocated to seven blood-lead
concentrations distributed about the predicted geometric mean. This approach allowed us to
account for the variability in blood-lead concentrations that is expected to take place for children
with similar environmental exposures.
We now present how the distribution of childhood blood-lead concentrations associated
with each specific housing condition were allocated to the seven blood-lead concentrations.
Blood-lead concentrations are usually assumed to have a log-normal distribution. Under this
assumption, the distribution of blood-lead concentrations associated with each specific housing
condition may be characterized by two numbers: the geometric mean and the geometric standard
deviation. The geometric mean was taken to be that predicted by either the EPI or IEUBK model
and the geometric standard deviation was set equal to 1.6. The default geometric standard
deviation of blood-lead concentrations for children at similar environmental-lead levels for the
IEUBK model is 1.6 (EPA, 1994c). Note that for the EPI model fitted to the Rochester data, the
geometric standard deviation of observed blood-lead levels from the model was 1.63.
Draft - Do Not Cite or Quote El-2 September 27, 1996
-------
If blood-lead concentrations have a log-normal distribution with a geometric mean GM
and geometric standard deviation 1.6, then the logarithms of the blood-lead concentrations have a
normal distribution with mean M = log(GM) and standard deviation S = log(l .6) =0.47. The
distribution of log blood-lead concentrations was partitioned into the seven intervals about the
log of the geometric mean shown in Table El-1. Figure El-1 graphically illustrates the
partitioning of the log blood-lead concentrations. The first row of the table represents the lower
tail of the distribution, log blood-lead concentrations more than 2.5 standard deviations below the
mean. The percentage of the distribution assigned to this interval is based on the area under a
standard normal curve for x-values less than -2.5,0.62%. The assigned log blood-lead
concentration for this interval is the expected value of a standard normal random deviate lying in
the interval -°° to -2.5. The assigned blood-lead concentration for this interval was obtained by
exponentiation:
eM-282.S = gM^-2.82.8 = GM/GSD282.
If N children were associated with the specific housing condition then 0.62 percent of the
N children were assigned a blood-lead concentration of GM/GSD282. The remaining 99.38
percent of the N children were assigned to the other six blood-lead concentrations presented in
Table El-1 using the percentages given in second column of the table. A similar procedure was
used to determine the log blood-lead concentration and relative frequency for each of the six
remaining intervals. In this manner, the distribution of blood concentration of the N children
-were allocated to a distribution of blood-lead concentrations centered around the GM predicted
by the EPI model with a GSD of 1.6.
Draft - Do Not Cite or Quote El-3 September 27. 1996
-------
Table E1-1. Allocation of Blood-Lead Distribution to Seven Blood-Lead Concentrations
Log Blood-Lead Concentrations
Interval for Log
Blood Lead1
[-«, M-2.5"S]
[ M-2.5*S, M-1.5*S]
[M-1.5*S, M-0.5"S]
[ M-0.5*S, M + 0.5*S]
[ M+0.5*S. M + 1.5*S]
[M + 1.5'5, M + 2.5*S]
I M + 2.5*S, ~l
Percentage of
Distribution in
Interval
0.0062
0.0606
0.2417
0.3830
0.2417
0.0606
0.0062
Assigned Log
Blood Lead for
Interval
M - 2.82*Sb
M- 1.85*Se
M - 0.92*S"
M'
M + 0.92 *S'
M + 1.85*S"
M + 2.82'Sh
Assigned Blood Lead
Concentration for Interval
GM/GSD282
GM/GSD1-85
GM/GSD092
GM
GM'GSD092
GM«GSD185
GM'GSD282
• Blood-lead concentrations were assumed to have a log-normal distribution with the geometric mean (GM) predicted by the
EPI model and a geometric standard deviation (GSD) of 1.6. The default geometric standard deviation for the IEUBK model
is 1.6. The distribution of log blood-lead concentrations was assumed to be normal; with mean M given by the log of GM
and standard deviation S = log(GSD) = log(1.6).
• The expected value of a normal random deviate known to lie in the interval I—, -2.5] is -2.82.
c The expected value of a normal random deviate known to lie in the interval 1-2.5,-1.5] is -1.85.
° The expected value of a normal random deviate known to lie in the interval 1-1.5,-0.5] is -0.92.
• The expected value of a normal random deviate known to lie in the interval (-0.5, 0.51 is 0.00.
1 The expected value of a normal random deviate known to lie in the interval [ 0.5, 1.5] is 0.92.
9 The expected value of a normal random deviate known to lie in the interval [ 1.5, 2.5] is 1.85.
" The expected value of a normal random deviate known to lie in the interval [ 2.5, - ] is 2.82.
The predicted distributions at each housing condition were then combined to generate a
distribution of childhood blood-levels over all of the housing conditions present in the HUD
National Survey.
Draft - Do Not Cite or Quote
El-4
September 27, 1996
-------
:
-j
i
i
j
•
i
i
1
i
1
1
H
1
1
: ,
M - 2*2 S
-- "
• M - 148 S
a
M
1
g
/
/
/
/
/
/
/
/
/
i '
• U - O82 S
*
^
i
2
/
1 1
D M
§
1
S
\
\
\
\
\
\
\
\
\
K\_
\
1
» M + 0828
5
+
s
' '
\ ' !
° M + 1.85 S m M -f 2JS2 S
ID in
r- «
+ +
5 5
Distribution of Log(Blood-lead levels)
M=LOQ(QEOMETRIC MEAN). S=LOQ(QEOMETRIC STANDARD DEVIATION)
Figure E1-1. Distribution of Blood-Lead Levels About Geometric Mean on Log Scale.
Draft - Do Not Cite or Quote
E1-5
September 27, 1996
-------
APPENDIX E2
Methodology for Estimating Health Benefits Associated with
Interventions Resulting from Proposed §403 Rules
Draft - Do Not Cite or Quote E2-1 September 27, 1996
-------
APPENDIX E2
METHODOLOGY FOR ESTIMATING HEALTH BENEFITS ASSOCIATED WITH
INTERVENTIONS RESULTING FROM PROPOSED §403 RULES
This appendix details the procedures used to compute the health risks associated with
lead exposure to children. The approach to conducting risk characterization in the §403
rulemaking effort is focused around characterizing the national distribution of blood-lead (PbB)
concentrations in children aged 1-2 years in 1997 under two scenarios:
• Immediately prior to initiating any intervention strategies under the proposed §403
rules.
• Immediately after performing the relevant intervention strategies on the nation's
housing stock under the proposed §403 rules.
Once these two blood-lead concentration distributions are characterized, predicted health benefits
associated with various options for the §403 standards are obtained by calculating a series of
health endpoints (e.g., PfPbB > 10 ug/dL], P[IQ < 70]) and comparing the values of these
endpoints between the two distributions.
Outline of the Methodology
This methodology characterizes the pre-§403 blood-lead distribution for children aged 1-
2 years using reported information from NHANES III. A model-based procedure (either the EPI
or IEUBK model) is used to characterize the distribution of blood-lead concentrations at both
pre-§403 and post-§403, and the observed differences between the two distributions are
identified. Then, a post-§403 distribution that is comparable to the pre-§403 NHANES III
distribution is derived based on the differences between the two model-based estimates and the
pre-§403 NHANES III distribution.
The methodology consists of the following five steps:
Draft - Do Not Cite or Quote E2-2 September 27. 1996
-------
#1. Use blood-lead concentration data reported in the NHANES III to estimate the
geometric mean (GM), the geometric standard deviation (GSD), and the 10th
percentile associated with the baseline (i.e., pre-§403^> distribution of blood-lead
concentration for children aged 1-2 years (i.e., 12-35 months).
#2. Use the environmental-lead levels for HUD National Survey units as input to either
the IEUBK or EPI model to estimate the geometric mean, the geometric standard
deviation (GSD), and the 10th percentile associated with the baseline (i.e., pre-§403')
distribution of blood-lead concentration for children aged 1-2 years (i.e., 12-35
months).
#3. Use adjusted (post-§403) environmental-lead levels for HUD National Survey units as
input to the model used in Step #2 to estimate the geometric mean, the geometric
standard deviation (GSD), and the 10th percentile associated with the post-§403
distribution of blood-lead concentration for children aged 1-2 years (i.e., 12-35
months).
#4. Estimate the geometric mean, 10th percentile, and GSD of a post-§403 blood-lead
distribution that is derived from pre-§403 NHANES III distribution determined in
Step #1 and the changes in the blood-lead distributions estimated in Steps #2 and #3.
#5. Calculate values for the health endpoints of interest using information from the final
post-§403 distribution in Step #4.
Details of the Revised Methodology
A key assumption in this methodology is that blood-lead concentrations are assumed to
be lognormallv distributed, regardless of whether they represent pre- or post-§403 concentrations
or whether the distribution is based on NHANES III data or is model-based. With this
assumption and by estimating the geometric mean and GSD of the distribution, the entire
distribution is characterized.
All five steps of the methodology are now discussed in detail. Following this discussion,
applications of the revised methodology are illustrated for two intervention strategies.
#1. Use NHANES III to characterize the pre-§403 distribution.
The NHANES III database includes blood-lead concentration data for 924 children aged
12-35 months at the time of their survey interview. Each child in the survey was assigned a
Draft - Do Not Cite or Quote E2-3 September 27, 1996
-------
sampling weight corresponding to the number of children in the country being represented by the
child. In Step #1, a weighted geometric mean and weighted geometric standard deviation of the
blood-lead concentration data are calculated across these children, where the weights correspond
to the sampling weights assigned at the time that the child was examined in the survey (variable
WTPEXMH1 in the NHANES III database). Call these variables GM, and GSD,, respectively.
These values are calculated as follows:
GM, = 4.046 ug/dL; GSD, = 2.057 ug/dL .
Thus, the pre-§403 blood-lead distribution characterized using NHANES III data is assumed to
be lognormally-distributed with geometric mean = 4.046 ug/dL and geometric standard deviation
= 2.057 ug/dL. The 10th percentile of this distribution is
P10, = GM, (GSD,)-1282 = 4.04*(2.057)-'282 = 1.60 ug/dL ,
where -1.282 is the 10th percentile of a standard normal distribution.
#2, #3. Obtain pre- and post-§403 distributions that are model-based.
Because interventions under §403 have not yet occurred, thereby precluding post-§403
blood-lead concentrations from being directly measured, the blood-lead distribution resulting
from the proposed §403 rules must be estimated. For this reason, this methodology characterizes
pre- and post-§403 blood-lead distributions that are model-based (i.e., predicted blood-lead
concentrations as a function of environmental-lead levels are obtained using either the IEUBK or
EPI model).
This approach uses the information on environmental-lead levels in the HUD National
Survey database as input to the model. The result of the modeling is a series of blood-lead
concentrations and the number of 1997 children aged 1-2 years associated with each
concentration. A weighted geometric mean and weighted geometric standard deviation of these
concentrations are calculated, where the weights correspond to the numbers of children
associated with each concentration. Call these variables GM2 and GSD2, respectively. Thus, the
model-based pre-§403 blood-lead distribution is assumed to be lognormally-distributed with
geometric mean GM2 and geometric standard deviation GSD2. The 10th percentile of this
distribution is
Draft - Do Not Cite or Quote E2-4 September 27, 1996
-------
P102 = GM,(GSD2)-1282
where -1.282 is the 10th percentile of a standard normal distribution.
The same method used in Step #2 is also used to characterize a model-based post-§403
distribution (Step #3). Step #3 differs from Step #2 in that the environmental-lead levels from
the HUD National Survey are initially adjusted to reflect the effects of intervention. This
adjustment is documented in Table 5-3. Let GM3 and GSD3 be the weighted geometric mean and
geometric standard deviation, respectively, of the predicted post-§403 blood-lead concentrations.
Thus, the model-based post-§403 blood-lead distribution is assumed to be lognormally-
distributed with geometric mean GM3 and geometric standard deviation GSD3. The 10th
percentile of this distribution is
P103 = GM3 (GSD3)-'282
where -1.282 is the 10th percentile of a standard normal distribution.
#4. Derive a post-§403 distribution from NHANES III and Steps #2 and #3.
The three distributions calculated in Steps #1 through #3 are used to characterize a post-
§403 blood-lead distribution that is directly comparable with the pre-§403 distribution
determined in Step #1. This distribution is assumed to be lognormal with geometric mean GM4,
10th percentile P104, and geometric standard deviation GSD4 calculated by the formulas
provided in Table E2-1. Note that while GSD4 is expressed in (3) as a function of the 10th
percentile P104, it could have been calculated from any of the distribution's percentiles, provided
that the denominator in (3) is replaced by the same percentile for the standard normal
distribution, multiplied by -1.
Draft - Do Not Cite or Quote E2-5 September 27, 1996
-------
Table E2-1. Formulas to Calculate the Geometric Mean, 10th Percentile, and
Geometric Standard Deviation for the Post-§403 Blood-Lead
Distribution in Step #4
Statistic
GM4
(geometric mean)
P104
(10th percentile)
GSD4
(geometric S.D.)
Formula
GM4 = GM,*(GM3/GM2) (1)
P104 = P10, *(P103/P10,) (2)
GSD4 = exp[(ln(GM4) - in(P104))/1.282] (3)
#5. Calculate post-§403 health and blood-lead effects.
The procedure for calculating health and blood-lead endpoints for the nation's children
aged 1-2 years, based on the post-§403 blood-lead distribution obtained in Step #4, is as follows:
a. P[PbB > X]. where X=10 ug/dL or 25 ug/dL
Because it is assumed that the post-§403 blood-lead concentration distribution is
lognormally distributed, the probability of observing a blood-lead concentration greater
than X is expressed as
P[PbB > X] = 1 -
ln(X) - ln(GM4)
ln(GSD4)
(4)
where $(z) is the probability of observing a value less than z under the standard normal
distribution. Therefore, setting X=10 and X=25 in equation (4) will provide estimates of
the probability of observing a post-§403 blood-lead level exceeding 10 ug/dL and 25
Hg/dL, respectively. Note that these probabilities under the pre-§403 distribution can be
calculated by replacing GM4 and GSD4 in equation (4) with GM, and GSD,, respectively.
Draft - Do Not Cite or Quote
E2-6
September 27, 1996
-------
b. P[IO < 70]
As indicated in Table E2-2, the estimated probability that a child will have an IQ
score less than 70 given the child's blood-lead concentration (PbB) is expressed as a
piecewise linear function of PbB. To estimate the probability that a child in the national
population has an IQ score less than 70 following §403 interventions, the post-§403
blood-lead distribution derived in Step #4 is used with the information in Table E2-2.
Table E2-2. Formulas for Estimating the Probability of Observing IQ Score Less Than 70.
Given a Child's Blood-Lead Concentration
Interval #
(i)
1
2
3
4
5
6
7
8
9
Range of PbB
(XM < PbB s x,J
0 < PbB s 5.0
5.0 < PbB * 7.5
7.5 < -bB s. 10.0
10.0 < PbB* 12.5
12.5 < PbBs 15.0
15.0 < PbB s 17.5
17.5 < PbBs 22.5
22.5 < PbB s 25.0
25.0 < PbB < ~
P[IQ < 70 | PbB]
(= a, + p,*PbB)
0.00360 + 0.000204* PbB
0.00218 + 0.000488* PbB
-0.00217 + 0.001 068* PbB
-0.00193 + 0.001 044*PbB
-0.00108 + 0.000976*PbB
-0.00534 + 0.001 260* PbB
-0.00653 + 0.001 328*PbB
-0.01112 + 0.001 532*PbB
-0.00942 + 0.001 464*PbB
Source: Wallsten, T.S., and Whitfield, R.G. "Assessing the Risks to Young Children of Three
Effects Associated with Elevated Blood-lead Levels." Report by Argonne National Laboratory.
Report No. ANL/AA-32. Sponsored by the U.S. EPA Office of Air Quality Planning and
Standards. 1986.
Using the notation xi5 ai5 and PJ (i=l,...,9) introduced in the column headings in Table
E2-2, and letting LGM4 = InCGM,) and LGSD4 = ln(GSD4), the expected value of the
probability of observing an IQ score less than 70, given the post-§403 blood-lead
distribution derived in Step #4, is
Draft - Do Not Cite or Quote
E2-7
September 27, 1996
-------
P[PbB<70] =
1=1
4
f \ f
ln(x) - LGM4 ln(x
! - O
LGSD4 1
w \ \ •
-LGM4-(LGSD4)
LGSD4
LGSD4
ln(x|.,)-LGM4-(LGSD4)'
LGSD4
(5)
where K = exp(LGM4 + (LGSD4)2 IT) and O(z) is the probability of observing a value less
than z under the standard normal distribution. In calculating (5) use the following
conventions: In (0)=-~, ln(°°)=«, $(-«)=0, and $(~)=1. Equation (5) is equivalent to
(6)
£ f(ai+PiX)(t)(x)dx
i=l J
Vi
where 4>(x) is the probability density function of the lognormal distribution with
parameters LGM4 and LGSD4. Note that the expected probability under the pre-§403
distribution can be calculated by replacing LGM4 and LGSD4 in equation (S) with LGM,
= ln(GM,) and LGSD, = ln(GSD,), respectively.
c. P[IO decrement > x] for x=l. 2. 3
In this risk characterization, it is assumed that each ug of lead per dL of blood
corresponds to a 0.257 decline in IQ score (see Chapter 4 of the §403 Risk Assessment
report). Therefore, an IQ decrement exceeding 1 is associated with blood-lead
concentrations exceeding 1/0.257 = 3.9 iig/dL. Similarly, blood-lead concentrations
exceeding 2/0.257 = 7.8 ug/dL are associated with an IQ decrement exceeding 2, and
concentrations exceeding 3/0.257 = 11.7 ug/dL are associated with an IQ decrement
exceeding 3. Therefore,
Draft - Do Not Cite or Quote
E2-8
September 27, 1996
-------
P[IQ decrement > 1] = P[PbB > 3.9 ug/dL]
P[IQ decrement > 2] = P[PbB > 7.8 ug/dL]
PPQ decrement > 3] = P[PbB > 11.7 ug/dL]
where the right-hand side of each of these equations is calculated using equation (4) with
X=3.9,7.8,orll.7.
d. Average IP points lost (and associated standard deviation")
The (arithmetic) average IQ points lost in the population of children aged 1-2 years is
calculated using the properties of the lognormal distribution. If X corresponds to a
child's blood-lead concentration and Y is the associated decline in IQ for the child due to
the presence of the blood-lead, then it is assumed in this risk assessment that Y =
0.257*X. As X is assumed to be lognormally distributed, it can be shown that Y is also
lognormally distributed. Furthermore, under the post-§403 blood-lead distribution in
Step #4, estimates of the expected value of Y (average # IQ points lost) and the standard
deviation of Y (S.D. of # IQ points lost) are as follows:
Avg.#IQ points lost = 0.257 *GM4 *exp(ln(GSD4)2/2) (7)
S.D. of #IQ points lost = 0.257 *GM42*^exp(2*ln(GSD4)2) - exp(ln(GSD4)2) (8)
Note that if 0.257 is excluded from the formulas in equations (7) and (8), the result
would be the arithmetic average and standard deviation associated with the distribution of
blood-lead concentrations.
Example of Applying the Methodology
The values of the geometric mean, geometric standard deviation, and 10th percentile for
the pre-§403 blood-lead distributions determined from the NHANES III data (Step #1) and the
Draft - Do Not Cite or Quote E2-9 September 27, 1996
-------
IEUBK model (Step #2) are provided in Table E2-3. Table E2-4 documents three §403
intervention strategies upon which the above methodology is applied. These strategies cover a
wide range of expected benefits, from minimal benefit (low) to substantial benefit (severe).
With these strategies are the values of GM;, GSD,, and PlOj for i=3 and 4 (i.e., the post-§403
distributions generated in Steps #3 and #4).
The health endpoints calculated in Step #5 under the NHANES III pre-§403 distribution
(Step #1) and the final post-§403 distributions associated with the three intervention strategies in
Table 3 (Step #4) are presented in Table E2-5. These numbers are calculated from the entries in
Tables E2-3 and E2-4, using the equations found in Step #5.
Draft - Do Not Cite or Quote E2-10 September 27, 1996
-------
Table E2-3. Values of Geometric Mean, Geometric Standard Deviation, and 10th Percentile (/ig/dL) for the Pre-§403
Blood Lead Distributions in Steps #1 and #2
O
§
NHANES III
Pre-§403 Distribution
(Step #1 )
GM,
4.05
GSD,
2.06
P10,
1.60
HUD/IEUBK
Pre-S403 Distribution
(Step #2}
GM2
3.94
GSD2
2.23
P102
1.41
Table E2-4. Intervention Strategies and Values of Geometric Mean, Geometric Standard Deviation, and 10th Percentile
(fjgldl) for the Pre-§403 Blood Lead Distributions in Steps #3 and
Intervention Strategy
Strategy
Name
Low
Mid
Severe
Triggers for Dust-
Cleaning
Floors
--
100/ig/ft2
25 /ig/ft2
Window
Sills
-
500//g/ft2
25 //g/ft2
Trigger for
Soil Cover
-
400 iiglg
50 j/g/g
Trigger for
Soil
Abatement
--
3000 /ig/g
1 500 /
-------
Table E2-5. Values of Health Endpoints Under the NHANES III Pre-§403
Distribution and Under the Post-§403 Distributions of Three
Intervention Strategies
Health Endpoint
P[PbB > 10/yg/dL]
P[PbB > 25 /vg/dL]
P[IQ < 70]
P[IQ decrement > 1]
P[IQ decrement > 2]
P[IQ decrement > 3]
Avg. IQ Point Loss
S.D. of IQ Point Loss
NHANES III Pre-
§403
0.1048
0.0058
0.0057
0.5216
0.1822
0.0709
1.35
1.11
Post-S4031
Low
Intervention
Strategy
0.0965
0.0047
0.0055
0.5142
0.1719
0.0641
1.32
1.06
Mid Intervention
Strategy
0.0336
0.00026
0.0048
0.4444
0.0836
0.0175
1.08
0.65
Severe
Intervention
Strategy
'0.0141
0.00003
0.0045
0.3756
0.0449
0.0062
0.968
0.519
See Table 4 for definitions of the three intervention strategies.
Draft - Do Not Cite or Quote
E2-12
September 27. 1996
-------
APPENDIX E3
Estimation of Primary Prevention Efficacy
Using Model of Bone-Lead Mobilization
Draft - Do Not Cite or Quote E3-1 September 27. 1996
-------
APPENDIX E-3
ESTIMATION OF PRIMARY PREVENTION EFFICACY
USING MODEL OF BONE-LEAD MOBILIZATION
Though the scientific literature documents the effectiveness of a range of behavioral and
environmental intervention strategies on their ability to reduce childhood lead exposure, efficacy
is measured only among already exposed children [I]5. Specifically, declines in children's blood-
lead concentration on the order of 25% as measured 6 to 12 months following a variety of
intervention strategies were reported [2-7]. This secondary prevention intervention effectiveness
is likely not representative of the effectiveness being sought from the promulgation of §403. The
§403 standards for lead in dust, soil, and paint are mostly intended to prevent childhood lead
exposure before it occurs and, therefore, their effectiveness will be assessed by measures of
primary prevention efficacy.
Secondary prevention efficacy results are not necessarily representative of those expected
from primary prevention because lead present in blood is a combination of current environmental
exposure and internal sources of lead. A significant internal source of lead is bone tissue. After
prolonged exposure to lead, bone tissue retains much more lead than the other body tissues [8-
12]. Nordberg et al. [13] suggest that bone can become an internal source of lead during periods
of reduced external exposure to lead; see also [10,14-16]. The reported declines in blood-lead
concentration, therefore, may underestimate the primary prevention effectiveness of the
associated intervention strategy.
Unfortunately, there is limited empirical evidence regarding the extent to which bone-lead
stores are able to keep blood-lead levels elevated following an intervention, especially
concerning children. One study [17] measured bone-lead levels in children before and after an
intervention, but found no significant decline in the levels over a period of six weeks. Despite the
lack of studies concerning children, Nordberg et al. [13] claim that "skeletal turnover is highest
among children under 10 years of age." Several studies have been conducted to study bone-lead
mobilization in adults [14-15, 18-29]. For example, Gulson et al. [29] show that 45% to 70% of
5 The references for this appendix in this draft report are provided at the end of this appendix. The format
will be revised to agree with the rest of the document in the next revision.
Draft - Do Not Cite or Quote E3-2 September 27, 1996
-------
lead in the blood of adult women comes from long-term tissue stores, primarily the bone tissue.
A similar result was observed in another study on five adult subjects undergoing knee and hip
replacement [30].
If the contribution of mobilized bone-lead stores can be characterized, however, it would
be possible to translate the documented secondary prevention results into estimated primary
prevention results. An approach is presented here for estimating the efficacy of a primary
prevention intervention given an observed effectiveness for a secondary prevention intervention.
The approach is based on a bone-lead mobilization model developed to estimate the degree to
which bone-lead stores could mask the full effectiveness of an intervention by mobilizing into
the child's blood. This model is extensively discussed and its basis documented elsewhere [31],
though a summary is provided below.
A Model for Bone-Lead Mobilization
To evaluate the potential for continuing elevated blood-lead levels due to bone-lead
mobilization, a two-compartment model (see Figure E3-1) was adopted for the transfer of lead
between the blood and bone tissues within the body and elimination of lead from the body.
BONE
t
Uptake
BLOOD
Elimination
Figure E3-1. Two Compartment Model of Bone-Lead Mobilization.
In this model, lead is taken into the body (from the gastrointestinal tract and lungs) via the blood,
transfers between the blood and bone tissue, and is eliminated from the body via the blood. It is
Draft - Do Not Cite or Quote
E3-3
September 27, 1996
-------
assumed that the transfer of lead between the blood and bone tissues, and elimination of lead
from the blood follows a first-order kinetic relationship.
While the adopted model is most certainly an oversimplification, model results will
approximate those of other more complicated models involving additional tissue compartments
for two reasons:
While lead does mobilize from non-bone tissues following a decrease in
environmental lead uptake, the effects are believed to be limited to a period of days or
weeks due to the lower concentrations of lead amassed in these tissues, and
While all lead elimination from the body does not occur via a direct pathway from the
blood, the kinetic parameters used in the model properly include these other pathways
(endogenous fecal and via other soft tissues) as if they were directly from the blood.
Based on the model illustrated in Figure E3-1, blood-lead concentrations (PbB) after
intervention would follow the relationship illustrated in Figure E3-2. More specifically,
immediately after intervention there would be an initial drop from the pre-§403 PbB level
(PbBPre) to achieve an immediate post-§403 PbB level (FoB,™^,). PbB,,,,,,,^ represents the
blood-lead concentration that can be supported by the amount of lead being transferred from the
bone. After this initial drop, blood-lead concentrations would follow an exponential decline
toward the long-term post-§403 PbB level (PhS^^). FoB^,,^,,,, is the blood-lead level that
can be supported by the post-§403 exposure level, with no additional lead from the bone. At any
a particular length of time following the intervention, illustrated by the symbol "T" on the
horizontal axis in Figure E3-2, a target post-§403 PbB level (PbB0bseived) will be observed. The
original analysis using this model [31] estimated the maximum length of time (T) the bone-lead
stores would be capable of keeping the blood-lead concentration above the targeted observed
level (PbB0bserved) for a given value of PhE^,,^,,,,. For the purposes of the sensitivity analysis for
§403, the maximum long-term effectiveness is estimated instead. As the long-term percent
decline reflects the post-§403 PbB that can be support by the post-§403 exposure level, it is
assumed this decline is equal to the primary prevention effectiveness of the intervention.
Draft - Do Not Cite or Quote E3-4 September 27, 1996
-------
Pre-lntervention
PbB Level
Immediate
Post-Intervention
PbB Level
Target
Post-Intervention
PbB Level
Long-Term
Post-Intervention
PbB Level
T Time (days)
Figure E3-2. Blood-Lead Concentration Versus Time Following a Reduction
in Lead Uptake
The child's blood-lead concentration at t days post-§403 is given by the equation
PbB = PbB + (PbB - PbB^gTeJ-exp(-t-KBONEBLNet)
(D
where KBONEBL,^, is the net rate of lead flow from bone to blood to elimination. This rate is a
function of the blood-lead level following the initial drop (PbBimmPost) as well as other kinetic
parameters (e.g., the lead mass ratio of bone to blood and the elimination rate of lead from the
blood) which can be estimated from existing scientific literature [31]. As portrayed in Figure 2,
the blood-lead concentration follows an exponential decline toward PhS^,,^,,,,. Setting PbB in
Equation (1) equal to PbEob^^ and solving for the long-term percent decline in blood-lead
concentration (RungTerJ results in the following equation:
R
LongTerm
LongTerm
PbB
Pre
Observed " RpmnPost ^ "*' KBONEBLNet)
1 - exp(-t-KBONEBLNet)
(2)
Draft - Do Not Cite or Quote
E3-5
September 27, 1996
-------
where
R _ serve g R
"Observed pbfi ImmPost
Pre rre
The maximum efficacy of an intervention, then, may be calculated given two parameters:
1. the observed percent decline (Reserved)m an exposed child's blood-lead concentration
following an intervention (i.e., the observed secondary prevention efficacy); and
2. the length of time (t) following the intervention when the decline was observed.
Note that this process estimates the maximum value of R^g,.^ that might have yielded the
inputed values of PbB0bseived and t based on Equation (1). The specific value may lie between
Reserved and RLonBTenir The estimated primary prevention efficacy is a maximum in that RImmPost,
and therefore KBONEBL^,, cannot be estimated from available data [31J. It is necessary to
estimate the maximum efficacy over a range of possible values for RimmPost.
Results of Modeling Bone-Lead Mobilization
To illustrate the efficacy of primary prevention, values of 25%, 50%, and 75% are
considered for the observed secondary prevention efficacy and values of 6,12, 18, and 24 months
are considered for the lengths of time. Table E3-1 presents the maximum primary prevention
efficacy for these scenarios for children 1 to 7 years of age. The standard error of the estimated
efficacy—calculated by propagating, through the model, the standard errors of the underlying
model parameters—is enclosed in parentheses.
As an example of the results in Table E3-1, note that if the observed effectiveness of a
secondary intervention is assumed to be 25% (i.e., PbB decline to 75% percent of the pre-§403
level) at 6 months post-§403 for a 2 year old, then the implied effectiveness of primary
prevention will be at most 47%. The scientific literature reports secondary prevention efficacy of
approximately 25% declines in blood-lead concentration 12 months following dust abatements,
lead-based paint abatements, elevated soil lead abatements, and intensive educational efforts [1].
Draft - Do Not Cite or Quote E3-6 September 27, 1996
-------
Depending upon the age of the child benefitting from the intervention, the results in Table E3-1
would suggest these interventions would prompt primary prevention efficacy of between 30%
and 59% (column: "Length of Time, 12 Months"; row: "Observed Efficacy of Secondary
Prevention, 25%").
Empty cells in Table E3-1 indicate that those scenarios cannot possibly occur based on
Equation (1). For example, for a 7 year old, the impact of mobilized bone-lead stores would
result in less than a 25% decline in blood-lead concentration at 6 months, even for a 100%
effective intervention. Estimates of primary prevention efficacy under these "impossible"
scenarios are not meaningful and are therefore not shown.
Consistent with the limited data available on bone-lead mobilization, the standard errors
in Table E3-1 are quite large. By incorporating the 95% upper confidence bounds on the
maximum primary prevention efficacy, the resulting bounded estimates are 1.2 to 1.9 times larger
than the mean estimates reported in the table.
As described above, this analysis estimates the maximum efficacy of primary prevention
interventions. Consideration was also given to obtaining the minimum efficacy. It was
determined that the present model can provide a meaningful solution for the maximum case only,
and that additional empirical data and extensive model enhancement are required to solve the
minimum case. Only the maximum efficacy, therefore, is reported.
Draft - Do Not Cite or Quote E3-7 September 27. 1996
-------
Table E3-1. Maximum Efficacy of Primary Prevention For Blood-Lead Levels (PbB)
Observed at 25%, 50%, and 75% of Pre-§403 Levels at 6, 12, 18, and 24
Months
Observed
Efficacy of
Secondary
Prevention1
25%
50%
75%
Child's Age
(years)
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
Length of Time2
(months)
6
0.39 (0.16)
0.47 (0.18)
0.56(0.21)
0.67 (0.25)
0.79 (0.27)
0.91 (0.32)
0.78 (0.32)
0.94 (0.36)
12
0.30 (0.05)
0.33 (0.08)
0.36 (0.14)
0.41 (0.19)
0.47 (0.19)
0.53(0.21)
0.59 (0.22)
0.60 (0.09)
0.65 (0.16)
0.73 (0.27)
0.83 (0.37)
0.93 (0.38)
0.90(0.14)
0.98 (0.25)
18
0.28 (0.03)
0.30 (0.04)
0.31 (0.07)
0.34(0.10)
0.37 (0.14)
0.40(0.19)
0.44(0.19)
0.56 (0.05)
0.59 (0.08)
0.63 (0.13)
0.68 (0.21)
0.73 (0.29)
0.81 (0.37)
0.89 (0.37)
0.84 (0.08)
0.89(0.13)
0.94 (0.20)
24
0.27 (0.02)
0.28 (0.03)
0.29 (0.04)
0.31 (0.06)
0.33 (0.08)
0.35 (0.12)
0.37 (0.15)
0.55 (0.04)
0.56 (0.06)
0.59 (0.08)
0.62 (0.13)
0.66(0.17)
0.70 (0.24)
0.75 (0.31)
0.82 (0.05)
0.85 (0.09)
0.88 (0.13)
0.93 (0.19)
0.98 (0.25)
Note: An empty cell means that the scenario is not possible according to model predictions.
1 This is equivalent to the observed percent decline in an exposed child's blood lead levels at a specified time
point following the intervention.
2 This is equivalent to the length of time following the intervention when the decline was observed.
Draft - Do Not Cite or Quote
E3-8
September 27. 1996
-------
APPENDIC E3 REFERENCES
[ 1 ] Environmental Protection Agency. Review of studies addressing lead abatement
effectiveness. Office of Pollution Prevention and Toxics. EPA Report No. 747-R-95-006
(July 1995).
[2] Amitai, Y., Brown, M.J., Graef, J.W., Cosgrove, E. Residential deleading: Effects on
the blood lead levels of lead-poisoned children. Pediatrics. 88(5):893-897 (1991).
[3] Copley, C. G. The effect of lead hazard source abatement and clinic appointment
compliance on the mean decrease of blood lead and zinc protoporphyrin levels.
Mimeo. City of St. Louis, Department of Health and Hospitals, Division of Health,
Office of the Health Commissioner, St. Louis, MO. (1983).
[4] Staes, C., Matte, T., Copley, G., Flanders, D., and Binder, S. Retrospective study of the
impact of lead-based paint hazard remediation on children's blood lead levels in St. Louis,
Missouri. American Journal of Epidemiology. 139(10): 1016-1026 (1994).
[5] Weitzman, M., Aschengrau, A., Bellinger, D., Jones, R., Hamlin, J.S., and Beiser, A.
Lead-contaminated soil abatement and urban children's blood lead levels. Journal of
the American Medical Association. 269(13):1647-1654 (1993).
[6] Chamey, E., Kessler, B., Farfel, M., and Jackson, D. Childhood lead poisoning: A
controlled trial of the effect of dust-control measures on blood lead levels. New
England Journal of Medicine. 309:1089-1093 (1983).
[7] Kimbrough, R.D., LeVois, M., and Webb, D.R. Management of children with slightly
elevated blood-lead levels. Pediatrics. 93(2):188-191 (1994).
[8] Barry, P.S.I, and Mossman, D.B. Lead concentration in human tissues. British Journal
of Industrial Medicine. 27:339-351 (1970).
[9] Barry, P.S.I. A comparison of concentrations of lead in human tissues. British Journal
of Industrial Medicine. 32:119-139 (1975).
[ 10] Barry, P.S.I. Concentrations of lead in the tissues of children. British Journal of
Industrial Medicine. 38:61-71 (1981).
[11] Schroeder, H. A. and Tipton, I.H. The human body burden of lead. Archives of
Environmental Health. 17:965-977 (1968).
[ 12] Leggett, R.W., Eckerman, K.F., and Williams, L.R. Strontium-90 in Bone: A case
study in age-dependent dosimetric modeling. Health Physics. 43(3):307-322 (1982).
Draft - Do Not Cite or Quote E3-9 , September 27, 1996
-------
[13] Nordberg, G.R., Mahaffey, K.R., and Fowler, B.A. Introduction and summary.
International workshop on lead in bone: Implications for dosimetry and toxicology.
Environmental Health Perspectives. 91:3-7 (1991).
[14] Hyrhorczuk, D., Rabinowitz, M, Hessel, S., Hoffinan, D., Hogan, M., Mallin, K.,
Finch, H., Orris, P., and Herman, E. Elimination kinetics of blood lead in workers with
chronic lead intoxication. American Journal of Industrial Medicine. 8:33-42 (1985).
[15] Rabinowitz, M.B., Wetherill, G.W., and Kopple, J.D. Kinetic analysis of lead
metabolism in healthy humans. Journal of Clinical Investigation. 58:260-270 (1976).
[16] Rabinowitz, M.B. Toxicokinetics of bone lead. Environmental Health Perspectives.
91:33-37(1991).
[ 17] Markowitz, M.E., Bijur, P.E., Ruff, H.A., and Rosen, J.F. Effects of calcium disodium
versenate (CaNajEDTA) chelation in moderate childhood lead poisoning. Pediatrics.
92(2):265-271 (1993).
[18] Wrenn, M.E., Cohen, M., Rosen, J.C., Eisenbud, M., Blanchard, R.L. In-vivo
measurements of lead-210 in man. Assessment of radioactive contamination in man.
IAEA. pp. 129-146 (1972).
[ 19] Cohen, N., Jaakkola, T., Wrenn, M.E. Lead-210 concentrations hi the bone, blood, and
excreta of a former uranium miner. Health Physics. 24: 601-609 (1973).
[20] Rabinowitz, M.B., Wetherill, G.W., Kopple, J.D. Lead metabolism in the normal
human: Stable isotope studies. Science. 182: 725-727 (1973).
[21 ] Batschelet, E., Brand, L., Steiner, A. On the kinetics of lead in the human body. J.
Math Biology. 8: 15-23 (1979).
[22] Heard, M.J. and Chamberlain, A.C. Uptake of Pb by human skeleton and comparative
metabolism of Pb and alkaline earth elements. Health Physics. 47: 857-865 (1984).
[23] Marcus, A.H. Multicompartment kinetics models for lead. I. Bone diffusion models for
long-term retention. Environmental Research. 36: 441-458 (1985).
[24] Christoffersson, J.O., Ahlgren, L., Schutz, A., Skerfving, S. and Mattsson, S. Decrease
of skeletal lead levels in man after end of occupational exposure. Archives of
Environmental Health. 41:5: 312-318 (1986).
[25] Cristy, M., Leggett, R.W., Dunning, D.E., Jr. and Eckerman, K.F. Relative age-specific
radiation dose commitment factors for major radionuclides released from nuclear fuel
facilities. NUREG/CR-4628 ORNI/TM-9890 (1986).
Draft - Do Not Cite or Quote E3-10 September 27, 1996
-------
[26] Schutz, A., Skerfving, S., Ranstam, J. and Christoffersson, J. Kinetics of lead in blood
after the end of occupational exposure. Scand. J. Work Environ. Health. 13: 221-231
(1987).
[27] Bert, J.L., van Dusen, L.J. and Grace, J.R. A generalized model for the prediction of
lead body burdens. Environmental Research. 48: 117-127 (1989).
[28] Nilsson, U., Attewell, R., Christoffersson, J.O., Schutz, A., Ahlgren, S., Skerfving, S.
and Mattson, S. Kinetics of lead in bone and blood after end of occupational exposure.
Pharmacology and Toxicology. 69: 477-484 (1991).
[29] Gulson, B.L., Mahaffey, K.R., Mizon, K.J., Korsch, M.J., Cameron, M.A., and
Vimpani, G. Contribution of tissue lead to blood lead in adult female subjects based on
stable lead isotope methods. Journal of Lab and Clinical Medicine. 125: 703 (1995).
[30] Smith, D.R., Osterloh, J.D., Flegal, A.R. Use of Endogenous Stable Lead Isotopes to
Determine Release of Lead from the Skeleton. Environmental Health Perspectives.
104(l):60-66 (1996).
[31 ] Rust, S. W., Kumar, P., Burgoon, D. A., and Schultz, B. Influence of bone-lead stores on
the observed effectiveness of lead hazard intervention. In press. (1996).
Draft - Do Not Cite or Quote E3-11 September 27, 1996
------- |