••*?'
U.S. EPA
OSWER #9285.7-37
March 2003
TRW RECOMMENDATIONS FOR
PERFORMING HUMAN HEALTH RISK
ANALYSIS ON SMALL ARMS SHOOTING
RANGES
.5
Office of Solid Waste and Emergency Response
U.S. Environmental Protection Agency
Washington, DC 20460
-------
NOTICE
This document provides guidance to EPA staff. It also provides guidance to the public and to the
regulated community on how EPA intends to exercise its discretion in implementing the National
Contingency Plan. The guidance is designed to implement national policy on tiiese issues: The document
does not, however, substitute for EPA's statutes or regulations, nor is it a regulation itself. Thus, it
cannot impose legally-binding requirements on EPA, States, or the regulated community, and may not
apply to a particular situation based upon Hie circumstances. EPA may change this guidance in the
future, as appropriate.
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
TECHNICAL REVIEW WORKGROUP FOR LEAD
The Technical Review Workgroup for Lead (TRW) is an interoffice workgroup convened by the U.S.
EPA Office of Solid Waste and Emergency Response/Office of Emergency and Remedial Response
(OSWER/OERR).
Region 8
JimLuey
Denver, CO
CO-CHAIRPERSONS
NCEA/Washington
Paul White
MEMBERS
Region 1
Mary Ballew
Boston, MA
Region 2
Mark Maddaloni
New York, NY
Region 3
Linda Watson
Philadelphia, PA
Region 4
Kevin Koporec
Atlanta, GA
Region 5
Patricia VanLeeuwen
Chicago, IL
Region 6
Ghassan Khouiy
Dallas, TX
Region 7
Michael Beringer
Kansas City, KS
Region 10
Marc Stifelman
Seattle, WA
NCEA/Washington
Karen Hogan
NCIA/Cincinnati
Harlal Choudhuiy
NCEA/Research Triangle Park
Robert Elias
OERR Mentor
Larry Zaragoza
Office of Emergency and Remedial Response
Washington, DC
Executive Secretary
Richard Troast
Office of Emergency and Remedial Response
Washington, DC
Associate
Scott Everett
Department of Environmental Quality
Salt Lake City, UT
-------
-------
EXECUTIVE SUMMARY
Past management practices at small arms outdoor shooting ranges (subsequently referred to as
'ranges') allowed the spent ammunition to accumulate on site. Many range operators now recognize •
the risk posed to humans and the environment by spent lead ammunition and have implemented
programs to manage and recycle lead shot and bullets. In accordance with its mission to provide
scientifically sound and consistent guidance on lead risk assessment, the TRW has prepared this
document to provide guidance and recommendations for performing risk assessment on land currently
or formerly used as ranges. This document supplements Region 2's Best Management Practices for
Lead at Outdoor Shooting Ranges (U.S. EPA, 2001 a), which serves as national guidance on the
management of lead at ranges to minimize the environmental impact of spent lead ammunition (U.S.
EPA, 2001b).
As used in this document, the term 'small arms' includes rifles, handguns (pistols), shotguns,
submachine guns, and machine guns (NRA, 1999, p. 1-8). Ranges can be divided into high velocity
shooting ranges, where target shooting with pistols and rifles occurs, and low velocity shooting ranges
where shotguns are used (i.e., skeet, trap, and sporting clay ranges). Lead bullets and fragments at
pistol and rifle ranges are typically contained in a relatively small, well-defined area, or volume, of sand
and/or soil. The shotfall zones at skeet and trap ranges may cover 10-50 acres or more, depending
upon the layout of the range.
This document contains brief discussions of the regulatory background for outdoor shooting ranges and
the toxicology of lead on humans, an operational and physical description of the different types of
outdoor shooting ranges, and the fate of spent lead ammunition in the environment and its bioavaOability.
This document provides recommendations on how the Integrated Exposure Uptake Biokinetic model
and the Adult Lead model can be used to predict the risk to human health from spent lead ammunition
on small arms shooting ranges.
-------
BACKGROUND AND PURPOSE
There are approximately 9,000 outdoor small arms shooting ranges (subsequently referred to as
'ranges') in the United States, excluding those located on military sites (U.S. EPA, 2001a). Millions of
pounds of lead are discharged annually at these ranges (U.S. EPA, 200la). In the past, the common
practice at ranges was to allow the spent ammunition to accumulate on site. Many range operators
now recognize the risk posed to humans and the environment by ihe lead in spent ammunition and have
implemented programs to manage and recycle lead shot and bullets.
Given the large number of ranges in Ihe United States and their potential risk to humans, ihe Technical
Review Workgroup for Lead (TRW) has prepared Ihis document to assist Regional and State risk
assessors and project managers in performing risk assessments at range sites. The proper management
of lead at shooting ranges is addressed in an EPA Region 2 document entitled 'Best Management
Practices for Lead at Outdoor Shooting Ranges" (U.S. EPA, 2001 a). This white paper contains the
TRW's recommendations on how to collect data that will be used to provide site specific information
when using the Integrated Exposure Uptake Biokinetic (JEUBK) Model and the Adult Lead Model
(ALM) (U.S. EPA, 2002c). The focus of this document is on formerly used ranges, although exposure
to active ranges may occur and is also discussed.
INTRODUCTION
As used in this document, the term 'small arms' includes rifles, handguns (pistols), shotguns,
submachine guns, and machine guns (NRA, 1999, p. 1-8). Spent ammunition on ranges is not regulated
as solid/hazardous waste unless it is discarded (abandoned) and left to accumulate for a long period of
time (U.S. EPA, 2001 a). Furthermore, it is not regulated if the spent ammunition is recovered or
reclaimed on a regular basis (U.S. EPA, 2001 a). However, if the range poses an imminent or
substantial danger to health or the environment it can be addressed through Resource Conservation and
Recovery Act (RCRA) (U.S. EPA, 2001a).
The intake of lead has a wide variety of effects on humans. Adults and children exposed to lead are
susceptible to neurotoxic effects (ATSDR, 1999). Lead may also increase blood pressure and cause
anemia; high levels of exposure to lead may damage the brain and kidneys and even cause death
(ATSDR, 1999; U.S. EPA, 2002b). High levels of exposure may cause miscarriages, retard fetal
development, and damage the organs responsible for sperm production (ATSDR, 1999; U.S. EPA,
2002b). Other effects of lead exposure include irritability, poor muscle coordination, muscle and joint
pain, memory and concentration problems, digestive problems, and hearing and vision impairment (U.S.
EPA, 2001 a, 2002b). In children, lead exposure can cause behavioral and learning problems, hearing
problems, impairment of vision and motor skills, hyperactivity, and developmental delays (ATSDR,
1999; U.S. EPA, 2001a, 2002b). Blood lead concentrations of 10 (ig/dL or less have been associated
with adverse heahh effects in children (U.S. EPA, 1986a; CDC, 1991). It is EPA policy to limit
exposure to lead such that 1he probability of a typical (or hypolhetical) child, or group of similarly
exposed children, having or exceeding the 10 ng/dL blood lead concentration is less than 5% (U.S.
EPA, 1994a).
Other chemicals of potential concern at shooting ranges include arsenic and antimony (components of
-------
ammunition), nickel (coating on some lead shot), copper, zinc, strontium, and magnesium (present in
tracer munitions that are used in machine guns), and polycyclic aromatic hydrocarbons (present in
petroleum 'pitch' found in clay targets used at skeet and trap ranges and in 'wadding' from shotgun
shells) (Jorgensen and Willems, 1987; EEA, 1992; EA, 1995; Peddicord and LaKind, 2000). Lead
shot contains primarily lead (97%), antimony (2%), arsenic (0.5%), and sometimes nickel (0.5%)
(Jorgensen and Willems, 1987; Lin et al, 1995). The crust material surrounding lead shot contains
between 0.5 and 2.0% antimony, 0.15% nickel, and trace amounts of arsenic (Jorgensen and Willems,
1987), Lead bullets are composed of 90-99% lead, 1-10.5% antimony, and 0.1% copper (EA,
1996). Baer et al. (1995) determined that clay targets ('pigeons') contain approximately 2/3 dolomitic
sandstone and 1/3 pitch; painted targets also contain approximately 1% fluorescent paint
Due to a ban on the use of lead shot for waterfowl hunting and in waterfowl production areas by the
U.S. Fish and Wildlife Service (U.S. FWS, 1999,2001), other less toxic materials have been
introduced, such as bismuth, steel, tungsten/iron, and tungsten polymers (a list is provided in EPA,
2001 a, Appendix B). However, due to the higher cost of shot that is manufactured with the less toxic
material, lead shot continues to be the most commonly used material on skeet and trap shooting ranges
(U.S. EPA, 2001b).
TYPES OF SHOOTING RANGES
Ranges can be divided into high velocity shooting ranges, where target shooting with pistols and rifles
occurs, and low velocity shooting ranges where shotguns are used (i.e., skeet, trap, and sporting clay
ranges). Appendix A contains a list of on-line sources of additional information for small arms outdoor
shooting ranges.
High Velocity (Pistol and Rifle) Shooting Ranges
High velocity ranges consist of a firing line, targets, backstop (to contain bullets and fragments), side
berms (to contain ricochets), and ground and overhead baffles (to contain 'short' and 'long' shots,
respectively) (Vargas, 1996). Not all ranges, particularly older ranges, will include all these
components. Typical lengths for ranges vary between 25 yards and 200 meters (642 yards) (Vargas,
1996); widths depend on the number of firing stations.
The backstop is required to contain the bullets and bullet fragments. Backstops traditionally consist of
earthen berms, typically between 15 and 25 feet high. More recently, earthen berms have been
replaced with sand traps, steel traps, and rubber traps (U.S. EPA, 2001 a). Other innovations have
been developed for backstops, such as Shock Absorbing Concrete (SACON) that has been used on
some Department of Defense (DOD) ranges since the 1980s (U.S. EPA, 2001a). The purpose of the
newer types of backstops/bullet containment devices is to facilitate the collection of bullets and bullet
fragments, which substantially reduces the amount of contaminated media generated by the range.
Low Velocity (Shotgun) Shooting Ranges
Shotguns are used to shoot clay targets ('pigeons' or 'birds') on skeet, trap, and sporting clay ranges.
In many cases, skeet and trap shooting takes place in one range. A typical trap range consists of five-
shooting positions and one structure, the 'traphouse', from which the targets are thrown by a machine
-------
called a 'trap'. The angle at which the targets are thrown varies within an arc of 45 degrees (in a
horizontal plane). The shooting positions are located 16 yards from the traphouse (Capital Trap Club,
2001a). At skeet ranges, targets are released from two structures, the 'high house' and the 'low
house.' There are eight-shooting positions arranged along an arc between the two houses. At the top
of the arc, the shooter is approximately 30 yards from the line that connects the two houses (Capital
Trap Club, 2001 b). The actual layout of skeet and trap ranges varies widely between sites (e.g., EA,
1995; E&E, 1997; Murray et al., 1997).
The size and shape of the shotfall zone is a function of the layout of the site, and ranges from rectangular
for sites with multiple ranges located next to each other (e.g., E&E, 1997), to semi-circular for sites
with one range (e.g., EA, 1995). The outfall zone from trap shooting will tend to be less than for skeet
shooting due to Ihe angle at which shooting occurs. In skeet shooting, ihe targets are thrown overhead
and the shooting angle is approximately 45 degrees from the horizontal. Targets are released much
closer to the ground in tap shooting; the shooting angle is approximately horizontal. Another factor that
affects the distance the shot will travel is the size of the shot used. When the shooting angle is
approximately horizontal, the maximum distance shot will travel varies from 198 yards for No. 8 shot to
330 yards for No. 2 shot (Baldwin, 1994). Number 6 shot will cover an area between 300 and 700
feet from the shooting position when the shooting angle is level; if released from an angle of 40 degrees
from Ihe horizontal, the shot will drop between 400 and 900 feet from the shooting position (Baldwin,
1994).
FATE OF LEAD AMMUNITION IN THE ENVIRONMENT AND BIOAVAILABUJTY
Lead ammunition oxidizes in Ihe environment, forming a crust around the shot; tins crust contains lead
carbonates and sulfates (Jorgensen and Willems, 1987; Manninen and Tanskanen, 1993; Lin et al.,
1995; EA, 1996; Murray et al., 1997). The predominate lead carbonates that have been found in the
crust material include hydrocerussite (Pt^CO^OH^) and cerussite (PbCO3); the predominate lead
sulfate compound is anglesite (PbSO4). The rate of oxidation depends upon several environmental
factors including: oxidation/reduction potential, ionic strength, pH, oxygen content of the soil and the
presence of compounds (e.g., phosphate) that may inhibit oxidation (Jorgensen and Willems, 1987; EA;
1996). At some point, the presence of the crust material appears to inhibit the further weathering of the
ammunition (Jorgensen and Willems, 1987). While metallic lead is insoluble under typical
environmental conditions, lead is released to the environment through the dissolution of the lead
compounds found in the crust material (Jorgensen and Willems, 1987; Manninen and Tanskanen, 1993;
Lin et al., 1995; EA, 1996; Murray et al., 1997). The solubility of the lead compounds is affected by
pH, eH, tiie presence of carbonate, sulfate, sulfide, phosphate and chloride, and the organic matter
content of 1he soil (Jorgensen and Willems, 1987; Manninen and Tanskanen, 1993; Lin et al., 1995;
EA, 1996; Murray, 1997). .
Site-specific environmental factors that affect weathering rates of lead on shooting ranges include the
amount of precipitation, pH of rain water, slope of the ground surface, amount of organic material
present on the ground surface (e.g., leaves, peat) and soil type (U.S. EPA, 2001a). In general,
weathering of lead ammunition will increase with increasing precipitation amounts, increase with
decreasing pH, increase with increasing chloride concentration, decrease with increasing ground slope
-------
(due to decrease in contact time between precipitation and ammunition) and increase with increasing
organic matter cover (U.S. EPA, 2001 a). Disturbance of the soil (e.g., soil cultivation of agricultural
fields) may increase the decomposition of lead ammunition (Jprgensen and Willems, 1987). Bundy et
al. (1996) found high corrosion rates were negatively correlated with corrosion potential and soil
resistance, two characteristics of the soil environment that can be measured in the field.
The weathering rate of small arms ammunition may be affected by the presence of a copper 'jacket' or
metal casing that surrounds the lead core of some ammunition (Major, 2003). The contact between the
different types of metals may create a galvanic couple that increases the rate of corrosion of the
ammunition if the moisture content of the surrounding soil is sufficiently high. However, on some sites,
jacketed bullets do not appear to corrode at a faster rate than unjacketed bullets (Hall, 2002).
Jacketed bullets are common on small arms ranges that are located on DOD sites because the military
is required to use jacketed bullets; jacketed bullets are much less common on non-military ranges (Hall,
2002). Bullet jackets typically contain (by weight) 89-95% copper, 0.03-0.05 % lead, 0.05% iron,
and 5-10 % zinc (Battelle, 1997).
The bioavailability of lead compounds varies greatly from lead sulfates, which have relatively low
bioavailability (<25% bioavailable), to lead carbonates (>75% bioavailable) (Henningsen et al., 1998).
The overall bioavailability of lead at shooting ranges depends upon the relative amounts of lead
carbonates and lead sulfates that are present in the soil. Equilibrium diagrams (i.e., eH-pH diagrams)
predict lead sulfates to be the dominant form of lead at pHs <5.3, carbonates to be the dominant form
at pHs between 5.3 and 8.5, and lead hydroxides to dominate at pHs >8.5 (EA, 1996). The
speciation found at a particular site will also vary depending upon the amount of carbonate and sulfate
present in the soil (EA, 1996).
SITE CHARACTERIZATION AND RISK ASSESSMENT
An important difference between high velocity ranges and low velocity ranges, with respect to risk
assessment, is the size of the areas impacted by each type of range. Lead bullets and fragments at
pistol and rifle ranges are typically contained in a relatively small, well-defined area, or volume, of sand
and/or soil. Very little contamination may be found at a well-designed range equipped with bullet traps,
although some traps and targets (e.g., steel targets) may generate lead dust and particulates tiiat can be
transported by air and water and contaminate the surrounding area.
Due to the relatively small size of typical pistol and rifle ranges, the risks posed to human and ecological
receptors are typically low. Exceptions to this would include former ranges that are planned for
development, or are currently used for activities that could result in exposure to human and ecological
receptors. Risks to human and ecological receptors from exposure to lead and other contaminants at
skeet and trap ranges may be moderate to severe, due to the size of the shot outfall area In some
cases, the outfall areas are located within or near wetlands and surface water, which tends to increase
the risks to ecological receptors, particularly waterfowl (EA, 1995; E&E, 1997; U.S. EPA, 2001a).
-------
The difference between ihe types of exposures at high velocity and low velocity ranges also warrant
consideration in the exposure assessment. These differences are due to ihe nature of the lead
contamination in soil. At high velocity ranges, in addition to whole bullets and fragments, small particles
of lead will be present due to die partial disintegration of the bullets upon impact with the targets and
soil and rock particles in the berms (EA, 1996). Less disintegration of Hie bullets can be expected
where fully jacketed and partially jacketed bullets are used (e.g., military ranges). At low velocity
ranges, Ihe lead shot will tend to be less fragmented due to the lower velocities. Although lead shot
does break down in the environment, complete decomposition of the lead is a slow process that may
take 30-300 years, depending upon site conditions (Jorgensen and Willems, 1987; EA, 1996).
Exposure Scenarios and Pathways
Land use adjacent to and near shooting ranges should be considered when developing current and
future exposure scenarios for a site. Under the current land use scenario, the potentially exposed
human populations of particular concern at an operating range are residents of adjacent residential
properties, residents and farm workers on adjacent agricultural properties, and workers who are
employed on adjacent commercial properties. Other receptors include trespassers who use the site for
recreational purposes such as fishing., hunting, and hiking (Peddicord and LaKind, 2000; U.S. EPA,
200la), as well as other recreational users when the range is located on or within an area that is used -
for recreational activities other than target shooting (e.g., multi-use parks). Under future land use
scenarios, the potentially exposed population depends upon die intended or actual land use, which may
include residential, agricultural, commercial, or industrial uses.
The potentially exposed populations under both future and current land use scenarios that should be
considered are:
* Residential land use:
• children under the age of 7 years old
• adults
• Agricultural land use (farm family or subsistence farm family):
• children under the age of 7 years old
• adults
• farm workers
• Commercial and industrial land use:
• adults
• trespassers
• maintenance stafl7construction workers who may be exposed during invasive work
(e.g., excavating trenches to install/repair utilities)
The main pathway for human exposure to lead at shooting ranges is through incidental ingestion of
contaminated soil (Peddicord and LaKind, 2000; U.S. EPA, 2001a). The ingestion of site-raised meat
(beef, pork, chicken) and fruits and vegetables contaminated with lead dust may also pose a risk for
local residents and especially for local farm families; these pathways must be evaluated in tie latter
scenario. Grazing farm animals may ingest and bioaccumulate large quantities of lead [Braun et
al.,1997]. In addition, ihe inhalation of dust/soil particles may be a potential pathway depending upon
-------
site conditions, particularly during activities that involve the excavation of soil (e.g., during maintenance
and construction work), or during agricultural activities (e.g., tilling, planting, and plowing) that may
release clouds of dust Hunters (and poachers) may be exposed via consumption of lead-contaminated
wildlife (EA, 1995; E&E, 1997; Peddicord and LaKind, 2000; U.S. EPA, 2001a). Children who
exhibit pica behavior may be at risk from exposure to formerly used shooting range sites; however, the
pica child will not be considered in this document In most cases, the primary receptor of concern will
be children and fanners who are potentially exposed to formerly used ranges and adjacent
contaminated areas. (The subsistence farm family ingestion rate of relevant meat or site-raised crops is
assumed to comprise a large part of the diet; data is currently available in the Exposure Factors
Handbook to evaluate this scenario and the EPA Default Exposure Factors Workgroup is developing
default values as well.)
Ecological receptors of concern at shooting ranges include invertebrates, fish, mammals and birds,
particularly waterfowl. Pathways include the incidental ingestion of soil, intentional ingestion of lead
fragments as grit, and the ingestion of contaminated food items. Risk assessments at shooting ranges
have predicted unacceptable levels of risk from lead for raptors, and small (e.g., mice) and large (e.g.,
fox and deer) terrestrial mammals (EA, 1995; Peddicord and LaKind, 2000). The highest risks from
lead have been predicted for small mammals and birds that ingest lead shot incidentally while feeding, or
intentionally as grit (EA, 1995; Peddicord and LaKind, 2000). The predicted adverse effects of lead
on small mammals and birds, are supported by the literature, which has shown mortality may result from
the ingestion of one lead pellet (e.g., Ma, 1989; Roscoe et al., 1989; Hoffman et al., 2000; Vyas et al.,
2000). Additional information is available from the EPA ECOTOX database (U.S. EPA, 2002a).
Consultation with an ecotoxicologist is recommended when planning an ecological risk assessment
Soil Sampling Strategies and Recommendations
This section is divided into three subsections: 1) General Sampling Strategy, 2) Sample Collection at
High Velocity Ranges, and 3) Sample Collection at Low Velocity Ranges. The overall sampling
strategy and soil recommendations that are common to both types of ranges are discussed in the
General Sampling Strategy subsection. Recommendations that are specific to the two types of ranges
are provided in the second and third subsections.
General Sampling Strategy. Sampling may be required to support several different
management strategies at a shooting range. The initial visual inspection of the site and a review of
operating records may indicate that sufficient spent ammunition exists, making removal and recycling
cost effective. For small ranges, particularly pistol and rifle ranges, it may be cost effective to remove
or remediate contaminated material rather than conduct a risk assessment In this case, an initial
sampling maybe conducted to determine the amount and type of sampling data required to characterize
the material for treatment or disposal, followed by a screening to ensure the remaining soil does not
pose a potential risk. Even when recycling will not be a primary management strategy for the site, it is
recommended to reclaim spent ammunition from the range prior to collection of data to support a risk
assessment in order to avoid duplication of effort The risk assessment may be conducted to determine
the potential for imminent risk, evaluate the risk posed by residual levels of lead in soil, or to develop
further management strategies. A first step in the risk assessment is to define the exposure area(s) for
-------
the site. Exposure areas should be small enough to reflect an area of repeated site use by a
hypothetical individual. Additional exposure areas may need to be defined for agricultural activities,
including crop planting, maintenance and harvesting, and grazing.
The following recommendations for sampling have been developed to produce data that are adequate
for use with the IEUBK. and/or ALM. Given the primary exposure pathways (ie., incidental ingestion
of soil), the objectives of the sampling effort should include producing precise estimates of the exposure
point concentration (EPC) for lead in soil. Sampling designs for outdoor shooting ranges should be
appropriate for estimating the mean of skewed distributions (Cochran, 1977, p.40; Chen, 1995). Data
from skeet and trap ranges indicate the distribution of lead is typically positively skewed, with
concentrations in sieved samples ranging from <1 to 161,000 ppm and coefficients of skewness ranging
from 1.0 to 8.3 (Dragun Corp., 1992; EA, 1994,1995; E&E, 1997; Murray et al., 1997; TTNUS,
2001). Data from rifle and pistol ranges indicate the range of lead concentration may vary from
background levels to 3.9% lead (39,000 ppm), by weight (Dragun Corp., 1992; EEA, 1992). ,
Sampling depth should be appropriate for the exposure scenano(s) that are to be considered in the risk
assessment Typically, this will dictate that samples be collected from the surficial soils (i.e., 0-1" depth
interval) to assess current exposure scenarios. To assess the risk for future exposure scenarios it may
be appropriate to also collect samples at depth. If a large number of samples from different depth
intervals are planned, it is suggested to evaluate the correlation in concentration/shot density between
depth intervals to determine if samples from the different depth intervals can be combined to reduce
sampling costs (U.S. EPA, 2000).
Site conditions (e.g., wetlands) may restrict access to some areas of the site or may increase sampling
costs. The sampling plan should account for this to avoid introducing bias into the estimate of the EPC.
Samples of other media (e.g., surface water) should be collected as appropriate for the exposure
scenarios considered in the risk assessment Assistance of a statistician to develop the sampling design
is recommended.
Sample Preparation. Sieving of soil samples, to evaluate risks, is recommended for two
reasons: the fine particle size fraction (<250 jxm) is the primary source of soil ingestion (U.S. EPA,
2000) and should be used in predicting risk to humans for the incidental soil ingestion pathway;
secondly, sieving will remove lead shot and large bullet fragments from the sample, which are not likely
to be ingested inadvertently by humans. It is recommended that soil samples be sieved twice, first with
a No. 4 (4.75 mm) or No. 10 (2.00 mm) sieve to remove bulk debris, then with a No. 60 (250 urn)
sieve, or smaller sieve size (U.S. EPA, 2000). The No. 4 sieve size recommendation is based on Ihe
maximum size shot feat is typically used on skeet, trap, and sporting clay ranges (No. Wi), which has a
diameter of 2.41 mm. Table B-l lists the diameters of the pellets for different shot sizes and the size of
the openings for different sieve sizes. The portion of the sample that passes through the No. 4 or No.
10 sieve, but retained on Ihe No. 60 sieve, is the 'coarse fraction'; the portion passing through the No.
60 sieve is the 'fine fraction' (U.S. EPA, 2000). The portion of the sample passing through the first
sieve (i.e., No. 4 or No. 10) may be referred to as the 'total' sample (i.e., coarse + fine fractions). The
'total' soil concentration may be appropriate for predicting risks to ecological receptors and for
-------
predicting risks to humans for future exposure scenarios.
It may be possible to reduce sampling costs by developing a relationship between the coarse and fine
sample fractions and use the concentration in one fraction to predict the concentration in the other
fraction (U.S. EPA, 2000). However, Hie coarse fraction will contain fragments from bullets and shot,
which will tend to dominate the concentrations measured in the sample. Unless the lead fragments and
shot are distributed uniformly across 1he site it is unlikely that the lead concentration between the coarse
and fine fractions will be highly correlated. Increasing tiie volume of the sample or collecting composite
samples may be help to improve the correlation.
Analytical Methods. Samples of the fine fraction should be analyzed for total lead to predict
the risk from incidental ingestion of soil. The total fraction should analyzed for total lead to predict the
risks from exposure that may occur after the bullet fragments and shot have undergone additional
weathering. Solid waste test method 7421 is recommended for measuring the concentration of lead in
soil when using fixed-based laboratories for analysis (EPA. 1986b). The use of field-based devices
(e.g., X-ray Fluorescence [XRF]) to measure the concentration of lead may be cost-efFective and
decrease the time to site cleanup. EPA guidance on the use of field-based methods (U.S. EPA, 2001c)
should be consulted prior to developing a sampling plan for the site.
Sample Collection at High Velocity Ranges. The horizontal boundaries of active or
recently closed pistol and rifle ranges should be fairly obvious; the boundaries of pistol and rifle ranges
that have been abandoned for longer periods of time may not be readily apparent from visual inspection
alone. Soil samples should be collected from the berm and Hie rest of the shooting range(s). Samples
of other media should be collected, as appropriate, given the exposure scenarios considered in the risk
assessment.
Soil sample locations should be determined using random sampling methods that provide adequate
coverage of the site, e.g., using systematic or stratified random sampling methods (Gilbert, 1987).
For pistol and rifle ranges, either of Ihese sampling designs may be implemented by sampling on a
rectangular or triangular grid. Jacketed bullets typically travel deeper into berms than unjacketed bullets
(EA, 1996); Hie sampling plans for military ranges and olher ranges lhat use jacketed bullets should
take Ihis into consideration.
Sample Collection at Low Velocity Ranges. Determining the extent of the area potentially
affected by skeet and trap ranges are usually more difficult than it is for pistol and rifle ranges. This is
particularly true for closed ranges. Whether the range is active or closed, records of site operations,
the location of structures on the range (e.g., traphouse) and information gathered from site inspections
can help to prepare a preliminary site layout upon which an initial sampling design can be based.
For skeet and trap ranges, a radial grid, whn the origin located behind the shooting positions, may
produce a more efficient systematic or stratified random sampling design (e.g., EA, 1995). The highest
concentrations of lead are typically found in the top 6-8 inches of soil (EA, 1994,1995; E&E, 1997;
-------
Murray, 1997); however, elevated concentrations of lead have been detected at 24 inches below grade
in skeet and trap range shotfall zones (EEA, 1992; EA, 1995; Murray, 1997).
Lead Risk Modeling Recommendations
This section provides recommendations specific to ranges. General guidance on lead risk assessment,
lead models and model documentation can be downloaded from 1he TRW web site (U.S. EPA,
2002d). The TRW recommends the use of the ffiUBK and ALM to predict risk to children and adults,
respectively (U.S. EPA, 1994b, 1996). The models are intended to predict risk to humans from
exposure to lead that is continuously distributed in various media, including soil. Spent lead ammunition
in soil poses a potential risk to humans in two forms: 1) as lead adsorbed or absorbed to soil particles,
and as very fine lead particles; and, 2) as lead adsorbed/absorbed to larger soil particles, and as lead
shot and bullet fragments. Small soil particles (and by analogy, small lead particles), particularly those
<250 fim in size, are the primary source of soil ingestion (EPA, 2000). Lead shot, bullets and large
bullet fragments, and large soil particles contaminated with lead, are not likely to be ingested by
humans, but represent a source of lead lhat may be released to the environment through weathering
processes. The first form of spent lead ammunition should be considered under the current exposure
scenario, while both forms of spent lead ammunition should be considered under future use scenarios.
The EEUBK and the ALM require the user to input an estimate for the EPC for soil; the IEUBK. also
provides the user with the option of inputting estimates of the EPC for other media: dust, air, drinking
water, and diet (e.g., consumption of game and fish with elevated lead concentrations due to exposure
to spent lead ammunition). Recommendations are limited here to providing estimates of the EPC for
soil; recommendations for other media and model parameters are provided in the IEUBK User's
Manual and other guidance that is available on the TRW website (U.S. EPA, 2002d) and the reference
list of Ibis paper.
Recommendations for Estimating the Exposure Point Concentration Term. When
interpreting analytical data for a shooting range, it is important to distinguish between high and low
. velocity ranges. As described in beginning of Ihe Site Characterization and Risk Assessment section,
soil on high velocity ranges tend to contain very fine particles of lead in addition to bullets and bullet
fragments. Soil on low velocity ranges will tend to contain whole and partially decomposed lead
pellets. The presence of one bullet, bullet fragment, or lead shot in a soil sample will result in very high
measurement of lead concentration which may not yield an accurate prediction of risk for current
exposure scenarios (but may be appropriate for future exposure scenarios.
The use of geostatistics may be useful for shooting ranges, particularly skeet and trap ranges, where soil
concentrations often exhibit spatial patterns. Geostatistical estimators are capable of exploiting these
spatial patterns (i.e., spatial autocorrelation) to determine the extent of contaminated soil, and to
produce more precise estimates of the EPC. Another advantage of geostatistical estimators is they can
be used with data lhat have been collected by random and/or non-random sampling methods. Finally,
geostatistics can take advantage of the correlation between different types of measurements (e.g., soil
concentration and shot count) to obtain more precise estimates of the EPC.
-------
Recommendations for Adjusting Bioavailability. The TRW does not recommend
changing the default value for bioavailability without Hie collection and TRW review of good site-
specific data to support such a change. Bioavailability has been shown to be related to lead speciation
and soil particle size (U.S. EPA, 1999). While site specific data on lead speciation (e.g., from
decomposition of spent ammunition) and particle size are not considered by EPA to be an adequate
basis for adjusting the bioavailability variable in the DEUBK or ALM (U.S. EPA, 1999), this
information can be used to decide if in vivo bioassays are likely to produce estimates of bioavailability
that differ substantially from the EUBK default value. Guidance on the bioavailability variable is
available from the TRW website (U.S. EPA, 1999). Based on the available literature on lead
speciation in soil on shooting ranges, it appears the default values used in the IEUBK and ALM are
appropriate for assessing risks from exposure to soils located on shooting ranges.
Recommendations for Exposure Scenarios. Land use adjacent and near shooting ranges
should be considered when developing current and future exposure scenarios. For currently operating
sites, the populations of primary concern are residents of adjacent residential properties, residents and
farm workers on adjacent agricultural properties, and workers who are employed on adjacent
commercial properties. For future use scenarios, the populations of primary concern depend upon the
proposed site usage.
BEST MANAGEMENT PRACTICES AND RECYCLING
Guidance for implementing best management practices on small arms outdoor shooting ranges is
available from the EPA (EPA, 2001 a). Implementing best management practices (BMPs) on ranges
decreases the risk of spent ammunition to the environment by recycling spent ammunition; containing
lead shot, bullets, and fragments; preventing migration of lead to surface water and groundwater; and
keeping records of site operations (U.S. EPA, 2001a). Site conditions (e.g., wetlands, mud, steep
slopes, wooded areas) will affect the feasibility of removing spent ammunition from ranges.
-------
References
ATSDR. 1999. Toxicological Profile for Lead (Update), U.S. Department of Health and Human
Services, Public Health Service, Centers for Disease Control and Prevention, Agency for Toxic
Substances and Disease Registry. Atlanta, GA July. Available online at
http://www. atsdr. cdc.gov/toxprofiles/tpl 3.html.
Baer, K.N., D.G. Button, RL. Boer, T.J. Ward, and R.G. Stahl. 1995. Toxicity Evaluation of Trap
and Skeet Shooting Targets to Aquatic Test Species. Ecotoxicology. 4:3 85-92.
Baldwin, D. 1994. How Far Will a Shotgun Shoot? Gun Club Advisor. Spring. 1994. Available on-line
^
Battelle. 1997. Implementation Guidance Handbook: Physical Separation and Acid Leaching to
Process Small-Arms Range Soils. Prepared for Naval Facilities Engineering Service Center and U.S.
Army Environmental Center. September.
Bundy, K.J., M. Bricka, and A. Morales. 1996. An Electrochemical Approach for Investigating
Corrosion of Small Arms Munitions in Firing Ranges. Proceedings of 1he HSRC/WERC
Conference on the Environment Albuquerque, NM. May 21-23.
Capital Trap Club. 2001 a American Trapshooting. Ottawa, Canada April., 4. Available on-line at:
Capital Trap Club. 2001b. American Skeet Ottawa, Canada April, 4. Available on-line at:
http://www.firemmlrairiing.ca''capital/skeet.hlml.
CDC. 1991. Preventing Lead Poisoning in Young Children. U.S. Department of Health and Human
Services, Public Health Service, Centers for Disease Control and Prevention. Atlanta, GA Available
on-line at http://wonder.cdc.gov/wonder/prevguid/p0000029/p0000029.asp.
Chen, L. 1995. Testing the Mean of Skewed Distributions. J. Amer. Stat. Assoc. 90(430):767-772.
Cochran, W.G. 1977. Sampling Techniques. 3rd Ed. John Wiley & Sons, New York, NY.
Dragun Corp. 1992. Phase II Environmental Assessment Report Parcels 32/36 Carleton Sportsmen's
Club Sumpter Township, Michigan. October.
EA Engineering, Science and Technology (EA). 1994. Phase I Site Assessment Southern Lakes Trap
and Skeet Club. October 28, 1994.
EA Engineering, Science and Technology (EA). 1995. Phase II Site Investigation Camp Buckner Skeet
and Trap Range. Final. U.S. Military Academy. West Point, NY. August
-------
EA Engineering, Science and Technology (EA). 1996. Lead Mobility at Shooting Ranges. Sporting
Arms and Ammunition Manufacturers' Institute, Inc. Newton, CTt
E&E (Ecology and Environment, Inc.). 1997. Preliminary Assessment/Site Inspection forNahant
Marsh Site, Davenport, Iowa Final. September.
EEA, Inc. 1992. Environmental Assessment and Field Investigation at the Blue Mountain Sportsmen's
Center, Town of Cortlandt, Westchester County, New York. Preliminary Draft. November.
Gilbert, R.0.1987. Statistical Methods for Environmental Pollution Momtorwg.Van Nostrand
Reinhold, New York, NY.
Hall, T., Jr. 2002. Personal communication. July 3.
Henningsen, G., C, Weis, S. Casteel, L. Brown, E. Hoffinan, W. Brattin, J. Drexler, and S.
Christensen, 1998. Differential Bioavailability of Lead Mixtures from Twenty Different Soil Matrices at
Superfund Mining Sites. Abstract. Toxicol. Sci. 42(1-s).
Hoffman, D.J., G.H. Heinz, L. Sileo, D.J. Audet, J.K. Campbell, and L.J. LeCaptain. 2000.
Developmental Toxicity of Lead-Contaminated Sediment to Mallard Ducklings. Arch. Environ.
Contam. Toxicol. 39:221-232.
Jorgensen, S. and M. Willems. 1987. The Fate of Lead in Soils: The Transformation of Lead Pellets in
Shooting-Range Soils. Ambio. 16(1):11-15.
Lin, Z., B. Comet, U. Qvarfort, and R. Herbert. 1995. The Chemical and Mineralogical Behaviour of
Pb in Shooting Range Soils from Central Sweden. Environ. Poll. 89(3):303-09.
Ma, Wei-chun. 1989. Effect of Soil Pollution with Metallic Lead Pellets on Lead Bioaccumulation and
Organ/Body Weight Alterations in Small Mammals. Arch. Environ. Contam. Toxicol. 18:617-622.
Major, M. 2003. Personal communication. March 3.
Manninen, S. and N. Tanskanen. 1993. Transfer of Lead from Shotgun Pellets to Humus and Three
Plant Species in a Finnish Shooting Range. Arch. Environ. Contam. Toxicol. 24:410-414.
Murray, KL, A. Bazzi, C. Carter, A. Ehlert, A. Harris, M. Kopec, J. Richardson, and H. Sokol. 1997.
Distribution and Mobility of Lead in Soils at an Outdoor Shooting Range. J. Soil Contam. 6(l):79-93.
NRA (National Rifle Association). 1999. The Range Source Book. National Rifle Association, Range
Department, Field Operations Division. Fairfax, VA. November.
Peddicord, R.K. and J.S. LaKind. 2000. Ecological and Human Health Risks at an Outdoor Firing
-------
Range. Env. Tox, Chem. 19(10):2602-13.
Roscoe, D.E., L. Widjeskog, and W. Stansley. 1989. Lead Poisoning of Northern Pintail Ducks
Feeding in a Tidal Meadow Contaminated with Shot from a Trap and Skeet Range. Bull, Environ.
Contain. Toxicoi 42:226-233.
TTNUS (Tetra Tedi NUS, Inc.). 2001. Ecologically-Based Remediation Goals for Lead and PAHs in
Soil, Site 15, Blue 10 Ordinance Disposal Area, Naval Air Station Cecil Field, Jacksonville, Florida
U.S. EPA. 1986a Air Quality Criteria for Lead. Vol I. Draft Final. EPA-600/8-83/028aF. June.
U.S. EPA. 1986b. SW-846: Test Methods for Evaluating Solid Wastes Physical/Chemical
Methods, Method 7421, Lead (Atomic Adsorption, Furnace Technique). Revision 0. September.
Available on-line at http://\vww'.epa.gov/epaoswer/hazwaste/test/pdfs/7421.pdf
U.S. EPA. 1994a OSWER Directive: Revised Interim Soil Lead (Pb) Guidance for CERCLA Sites
andRCRA Corrective Action Facilities. EPA/540/F-94/043 PB 94-963282. August. Available on-
line at: bttp.v7v\^w.epagov/superfTmd/prograrns/lead/'products/oswerdir.pdf.
U.S. EPA. 1994b. Guidance Manual for the Integrated Exposure Uptake Biokinetic Model for
Lead in Children. Office of Emergency and Remedial Response. EPA/540/R-93/081 PB 93-
9635100. February. Available on-line at: hltp://ww^v.epagov/superfund/programs/lead/producte.htm.
U.S. EPA. 1996. Recommendations of the Technical Review Workgroup for Lead for an Interim
Approach to Assessing Risks Associated with Adult Exposures to Lead in Soil. December.
Available on-line at: http://\v\vw.epagovv'supei1ui:id/prograrns/lead/products.hlm
U.S. EPA, 1999. IEUBKModel Bioavailability Variable. EPA-540-F-00-006. OSWER 9285.7-
32. October. Available on-line at: http://w\\w.epagov/supertiuid/prograrns/lead/products/sspbbioc.pclt!
U.S. EPA, 2000. TRW Recommendations for Sampling and Analysis of Soil at Lead (Pb) Sites.
EPA-540-F-00-010. OSWER 9285.7-38. April. Available on-line at:
http://vvww^epa.gov/superfund''prograina/lead/producLs/sssiev'.pdf
U.S. EPA, 2001a Best Management Practices for Lead at Outdoor Shooting Ranges. EPA-902-
B-01-001. Region 2. January. Available on-line at: http://www.epagov/region2/was£e/leadsliot/.
U.S. EPA, 2001b. OSW Memorandum to RCRA Senior Policy Advisors. Subject: National
Guidance on Best Management Practices for Lead at Outdoor Shooting Ranges. October 10.
U.S. EPA 2001 c. Integrating Dynamic Field Activities Into the Super fund Response Process: A
Guide for Project Managers. Draft Final. December.
-------
U.S. EPA. 2002a. ECOTOX. Office of Research and Development. U.S. Environmental Protection
Agency. Washington, DC. Available on-line at: http://'www,epagov/ecotox/ecotox_home.htm.
U.S. EPA 2002b. Lead and Human Health. Technical Review Workgroup. Available on-line at:
ht1p://wv^^^^epa.gov/suptTflnld'programs,1ead/leadandhurnaDheal^l.hta
U.S. EPA. 2002c. Lead Workgroup. Products. Region 2. U.S. Environmental Protection Agency.
Washington, DC. Available on-line at http://vAv\v.epagov/superfund/prograrns/lead/products.htm
U.S. EPA. 2002d Technical Review Workgroup for Lead (TRW) Home Page. U.S. Environmental
Protection Agency. Washington, DC. Available on-line at:
http://wwv^^epagov/superfund''pI•ograms/lead/index.hlm
U.S. FWS. 1999. Migratory Bird Hunting. 50 CFR 20.210).
U.S. FWS. 2001. Rejuge-Specific Regulations for Fishing and Hunting. 50 CFR 32.2(k).
Vargas, C. 1996, Design Criteria for Shooting Ranges. Third National Shooting Range Symposium
Available on-line at:
http://wvvw.rangeinfo.org/resoiuw Hbraiy/fa
-------
en
&
c§
PS
M
I
§
£
CO
•a
s
X
1
T3
•fan
S
3
48
•3
C*
§
'•§
i
•»
u
^
S
o>
1
*
^H
«!
i
•a
H
CO
£
•o
i
i
H
S
i
3
•8
^
i
e
|
N
"3
1
$
>»
g
41
ftp
"^
g |
(A +*
83 5rt
— U
1 a
*p ?
v< w
1 1
1 1
' -rt
§ 8i
P S C1 '[j
| '§
f | 1 1
S K S tn
>
a
1
s
B-
s
.8
en
a
i
1
j3
>s
HH
I I
G. oc
§i §
of tg
^ •§
.?•*
1J
J 4t
l£S
js
^
I
ap
^!
II
is
w a
1
"o
I
I
I
i
S.
-------
t
•a
•-n
"
I
3
S
•a
I
i
•a
1
.1
."2
£
JS
8
i
•
"2
S-* I
2 <= *'
1
J
|
i
I
£
I
I
I
I
I
u
-------
•o
3
.0
T=
o
£
G
u
"6
O
y
$ -
? 1
gl ~
i S
f£ 'S;
1 a*
^ 'i
ll
1
'i
1
'?
g
"S.
i
I
q
I
•3
-------
Table B-l. Shot Size and EPA-Recommended Sieve Sizes
for Use at Small Arms Outdoor Firing Ranges
Pellet Diameter
Shot Size
Buckshot
No. 000 - No. 2
No. 3
No. 4
Shot
F
T
BBB
A No, 4 Si^VK^J^tjscH/i^MSiCrt
BB
1
2
3
4
5
6
7
7'/Z
8
8'/s
9
' ' jVJto. i&^£V$(&fl&lf«^2<^MJItl
Inches
.36 - .27
.25
.24
.22
.20
.19
"KJtSSfeSiWltiRWs
.18
.16
.15
.14
.13
.12
.11
.10
.095
.09
.085
.08
P*$0!t*G®WtWm
i-]t(t£f\.5f^iVfi Sf4^3V^.J>
Note: Shot size is generally limited to a maximum of no. 7 V4 for trap anc
no. 7 Vi and minimum of no. 9 for skeet shooting.
Millimeters
9.14-6.86
6.35
6.10
5.59
5.08
4.83
iaV£ Tim swat saz&s USTKO ABoVji,
4.57
4.06
3.81
3.56
3.30
3.05
2.79
2.54
2.41
2.29
2.16
2.03
i^^mm**um***nt* -y '-;
SSSRSS^^T ^ - :
sporting clay use, and a maximum of
-------
-------
EPA
OSWER #9285.7-52
March 2002
BLOOD LEAD CONCENTRATIONS OF
U.S. ADULT FEMALES:
SUMMARY STATISTICS FROM PHASES 1 AND 2
OF THE NATIONAL HEALTH AND NUTRITION
EVALUATION SURVEY (NHANES III)
Office of Solid Waste and Emergency Response
U.S. Environmental Protection Agency
Washington, DC 20460
-------
NOTICE
This document provides guidance to EPA staff. It also provides guidance to the public and to the
regulated community on how EPA intends to exercise its discretion in implementing the National
Contingency Plan. The guidance is designed to implement national policy on these issues. The
document does not, however, substitute for EPA's statutes or regulations, nor is it a regulation
itself. Thus, it cannot impose legally-binding requirements on EPA, States, or the regulated
community, and may not apply to a particular situation based upon the circumstances. EPA may
change this guidance in the future, as appropriate.
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
TECHNICAL REVIEW WORKGROUP FOR LEAD
The Technical Review Workgroup for Lead (TRW) is an interoffice workgroup convened by the
U.S. EPA Office of Solid Waste and Emergency Response/Office of Emergency and Remedial
Response (OSWER/OERR).
CO-CHAIRPERSONS
Region 8
Kevin Koporec
Region 4
Region 1
Mary Ballew
Boston, MA
Region 2
Mark Maddaloni
New York, NY
Region 3
Linda Watson
Philadelphia, PA
Region 5
Patricia VanLeeuwen
Chicago, IL
Region 6
Ghassan Khoury
Dallas, TX
Region 7
Michael Bertnger
Kansas City, KS
Region 8
Jim Luey
Denver, CO
NCEA/Washington
Paul White
MEMBERS
Region 10
Marc Stifelman
Seattle, WA
NCEA/Washington
Karen Hogan
NCEA/Cincinnati
Harlal Choudhury
NCEA/Research Triangle Park
Robert Elias
OERR Mentor
Larry Zaragoza
Office of Emergency and Remedial Response
Washington, DC
Executive Secretary
Richard Troast
Office of Emergency and Remedial Response
Washington, DC
Associate
Scott Everett
Department of Environmental Quality
Salt Lake City, UT
-------
Blood Lead Concentrations of
U.S. Adult Females: Summary
Statistics from Phases 1 and 2 of
the National Health and
Nutrition Evaluation Survey
(NHANES III)
Technical Analysis by
William C. Thayer
Gary L. Diamond
Syracuse Research Corporation
6225 Running Ridge Road
No. Syracuse, NY 13212
Contract No. GS-10F-0137K
FEDSIM Order No. DABT63-01-F-0133-00
SRC No. FA332
-------
1.0 Introduction
In 1996 the Technical Review Workgroup for Lead (TRW) provided guidance for assessing lead
risks to adults from exposures to lead in soil. The Adult Lead Methodology (ALM) (U.S. EPA,
1996) includes two parameters that are the subject of this report. The background blood lead
concentration (PbBBdu!t „) represents the typical blood lead concentration (PbB) (ng/dL) in women
of child-bearing age, in the absence of exposures at the site being assessed. The parameter
GSDiadult, is the estimated value of the individual geometric standard deviation (GSD); the GSD
among adults (i.e., women of child-bearing age) that have exposures to similar on-site lead
concentrations. Default values for both PbBadult)0 and GSDi)a(JuU were derived from an analysis of
blood lead data for women 17-45 years of age, from Phase 1 of the Third National Health and
Nutrition Evaluation Survey (NHANES IE, Phase 1) as well as consideration of available site-
specific data on PbB GSDs (U.S. EPA, 1996). Based on those analyses the following default
values ranges were recommended: PbBa(klt0,1.7-2.2 jig/dL and GSDia(lu]t, 1.9-2.1.
Data from Phase 2 of the NHANES HI became available subsequent to the latter analysis. The
NHANES in survey was designed to be completed in two phases; while unbiased estimates of
population parameters may be obtained using data from either phase separately, more precise
estimates are obtained from combining the two phases (CDC, 1996a). Therefore, the availability
of the complete NHANES in data prompted a reexamination of the basis for the default values
for these two parameters, the results of which are provided in this report. The analysis reported
here estimates the geometric mean (GM) and GSD of PbBs of U.S. non-institutionalized women
between the ages 17-45 years based on data collected in Phases 1 and 2 of the NHANES. As
was the approach taken in 1996, estimates were made for the major race/ethnicity categories
represented in the NHANES III survey: non-Hispanic white, non-Hispanic black, Mexican-
American, and Other. Additionally, results of the combined Survey Phases are presented
separately for each of the regional quadrants of the NHANES Survey.
Decreases in estimates of GM PbBs observed between Phases 1 and 2 are offset by increases in
the GSD. The net effect is that the ranges of Preliminary Remediation Goals (PRGs) calculated
using the ALM do not differ appreciably between the two phases.
Technical Approach: Information on age, race/ethnicity, and PbB concentration for adults
17-45 years of age was extracted from the NHANES HI database (CDC, 1997). Data from both
phases of the NHANES in was used in this analysis in accordance with CDC recommendations
(CDC, 1996a). An accurate estimate for the GM from any subset of the PbB concentrations can
be made by using the sample weights included in the NHANES HI database. To obtain an
accurate estimate for the GSD from a subset of the PbB concentrations, however, is more
complicated because the mathematical formula that is used to calculate a GSD is not linear.
When estimating a measure of variability, such as the GSD, the sample weights provided in
NHANES do not fully account for the complex sampling design used in NHANES HI.
Furthermore, the nature and degree of bias in the estimate of a GSD that is calculated using only
the sample weights are unknown. To partially address this source of uncertainty, two approaches
were used to estimate the GSD as described below.
In the first approach, estimates for the GM and GSD were obtained using SAS (release 8.00, SAS
Institute Inc.) and the sample weights recommended by CDC (1996a); this was the same
Page 1 ~ Febiuary 28, 2002
-------
approach used in the analysis of the NHANES Phase 1 data (U.S. EPA, 1996), Standard errors
for the estimates of the GM PbB were estimated using SUDAAN (version 7.5, a program that is
implemented within SAS).
In the second approach, a lognormal probability plot was created using the empirical cumulative
distribution (BCD) (i.e., percentiles) estimated with SUDAAN for each race/ethnicity group
defined in NHANES in. The ECDs were estimated using SUDAAN. SUDAAN is designed to
compute statistics (e.g., means and percentiles) and their standard errors for data derived from
complex sample surveys such as the NHANES IE, (SUDAAN does not calculate estimates of
population variance, such as the GSD.) The analysis utilizes the sample weights and pseudo-
primary sampling units and pseudo-stratums provided in the NHANES in (CDC, 1996a). The
sample weights incorporate the differential probabilities of selection of survey participants and
include adjustments for non-coverage and non-response. The pseudo-primary sampling units and
pseudo-stratums account for the multistage sampling design and are necessary to estimate
accurate standard errors of parameter estimates.
The GM PbB and GSD estimated from the probability plots were compared to those estimated
directly from NHANES III with SUDAAN and SAS as a qualitative check on the curve fitting
procedure. A quantitative check on the curve fitting is provided by the coefficient of
determination (R2) that is reported for each probability plot.
Page 2 ~ Febiuaiy 28, 2002
-------
2.0 Results
Table 1 presents the percentiles of PbB estimated for U.S. women, 17-45 years of age, stratified
by race/ethnicity, along with their standard errors and 95% confidence intervals. Table 2
presents estimates of the GM PbB and GSD, stratified by race/ethnicity. The values of the GM
estimated from the probability plots (Figures 1-5) were close to those estimated directly from
NHANES in using SUDAAN, although they were consistently higher (by an average of
0,03 jig/dL). The values of GSD estimated from the probability plots were close to the those
estimated using SAS, however, the values of GSD estimated from the probability plots are
consistently lower (by an average of 0.10 |ig/dL).
The probability plots and the close agreement between the estimates of the GM and GSD based
on the two approaches is a qualitative indication that the lognormal distribution is a reasonable
model for the PbBs included in this analysis. A more quantitative indication is provided by the
high R2s shown in Figures 1-5.
•The results indicate that the GM PbB for the non-Hispanic black and Mexican-American
race/ethnicity groups are greater than the GMs for the non-Hispanic white group and the
combined groups. The results also indicate greater variability in the PbBs of the Mexican-
American group than the non-Hispanic black, non-Hispanic white, or combined groups. These
outcomes are consistent with the results obtained from the analysis of Phase 1 of the NHANES
(U.S. EPA, 1996). Due to the small sample size and related high uncertainty, the results shown
for the Other race/ethnicity group should be interpreted with caution (CDC, 1996a).
Table 3a shows the GM and GSD by census regions and race/ethnicity. Figure 6 shows the
delineation of the states into the four census regions-(U.S. Census Bureau, 2001). In these
analyses, GSDs were estimated using SAS. The pattern of higher GM in non-Hispanic blacks
and Mexican Americans than in non-Hispanic whites persisted when data were stratified by
geographic quadrant. The GMs for all race/ethnicity categories were higher in the northeast
quadrant than in other quadrants. GSDs for the non-Hispanic whites and Others groups were
relatively consistent across quadrants; the GSDs for non-Hispanic black and Mexican-American
groups varied from 1.9-2.2 and 1.9-2.4, respectively. The lowest GSDs for each race/ethnicity
group occurred in the northeast quadrant, while the highest GSDs are found in the midwest
region. The Mexican-American race/ethnicity group has the largest GSD for each census region,
with the exception of the northeast. The GM and GSD estimated for Mexican-Americans in the
northeast region should be interpreted with caution due to the low sample number (24).
In Tables 3b and 3c the 17-45 year age group was further divided into three age groups: 17-25,
26-35, and 36-45. Table 3b shows the GM and GSD by the three age categories and
race/ethnicity, Table 3c shows the GM and GSD by the three age categories and census region.
Table 3b shows the GM PbB for all race/ethnicity groups combined increases with age; this
pattern is also observed within each of the race/ethnicity groups. The pattern of higher GM in
non-Hispanic blacks and Mexican Americans persisted when data were stratified by age groups.
Estimates of the GSD across age groups for all race/ethnic groups combined varied by only 0.02.
Variance of the estimated GSDs across age groups increases once the data is stratified by race-
ethnicity.
Page 3 ~ Febniary 28, 2002
-------
Table 3c shows the trend of higher GM PbBs in the northeast persisted after the data were
stratified by age groups. Table 3c also shows the largest increase in GM PbBs across age groups
occurs in the Midwest, followed by the Northeast, West, and South Estimates of the GSD across
age groups and within census region varied by 0.01-0.02.
Page 4 ~ February 28, 2002
-------
3.0 Discussion and Conclusions
3.1 SOURCES OF UNCERTAINTY
Table 4 shows the occurrence of non-detects for each of the two phases of the NHANES IE and
for both phases combined. The percentage of non-detects for the combined race/ethnicity groups
was 21%.and ranged from 17% for the non-Hispanic black group to 28% for the non-Hispanic
white group. The increase in the overall rate of non-detects between Phases 1 and 2 was 7.3%
and was fairly consistent across the different race/ethnicity groups. In this analysis, non-detects
were set equal to 1A the detection limit of 1.0 u,g/dL, which is consistent with other reported
analyses of PbB concentrations from the NHANES ni (Brody et al., 1994). Preliminary analysis
indicated the estimates of GM PbB and GSD are highly sensitive to values assigned to non-
detects. Estimates of the GM/GSD for all of the race/ethnicity groups combined were 1.8/1.7,
1.5/2.1, and 1.3/2.7 when non-detects were set equal to the detection limit of 1.0 ng/dL, Vi the
detection limit, and 1/4 the detection limit, respectively. The sensitivity of the parameter
estimates to the method used to treat non-detects should be considered.in interpreting differences
between parameter values estimated with different approaches or with different subsets of the
NHANES IE data. Furthermore, the impact of the uncertainty related to the treatment of
detection limits will increase if the tend of decreasing PbB continues, unless the detection limits
are lowered.
As previously discussed, the method used to estimate the PbB GSD does not fully account for the
complex sampling design employed in the NHANES HI. Research would be required to
determine how to calculate more accurate estimates of the GSD and its standard error. It is not
clear if such an effort would be of great value, in terms of reducing uncertainty in the GSD
estimate. NHANES HI is a well designed study and relatively large sample sizes were available
for developing the GSD estimates. The more consequential issue for risk assessment is variation
of the GSD between population subgroups as compared with the uncertainty in the estimates of
GSD.
Based on this analysis and the above considerations, the lognormal distribution appears to
provide an adequate model for distribution of PbBs for non-institutionalized U.S. women,
17-45 years of age. The results obtained from the probability plots were similar to those
obtained with the direct computation of the GM and GSD; thus, either approach appears to be
reasonable and adequate for parameter estimation. However, direct computation from the
NHANES III is recommended as the preferred approach, due to its simplicity. Estimates for the
PbB GM (point estimates and confidence intervals) and GSD, based on the direct computation
approach, are discussed in the remainder of this report (confidence intervals for GSD could not
be calculated with the approaches used in this analysis).
Page 5 - February 28, 2002
-------
3.2 COMPARISON OF 1996 DEFAULT VALUES AMD UPDATED RANGES BASED ON NHANES
PHASES 1 AND 2
The purpose of this analysis was to incorporate data from Phase 2 of the NHANES HI survey in
the estimates of the GM and GSD of PbB in the non-institutionalized U.S. women, 11-45 years
of age. This is consistent with the recommendations of the CDC (1996a); incorporation of the
phase 2 data will tend to increase confidence in the estimates of the GM and GSD of the
distribution of PbB in non-institutionalized U.S. women, 17-45 years of age.
Comparisons of the values for the GM PbB and GSD based on the data from the combined
Phases 1 and 2 of the NHANES m with the values estimated from the NHANES IE Phase 1
(U.S. EPA, 1996) and the default values for PbBadult)0 and GSDi)adu]t used in the EPA ALM (U.S.
EPA, 1996) are presented in Table 5. Several observations can be made from these comparisons:
a. Both the EPA ALM default value range for PbBadul, 0 (1.7-2.2) and the range of
GM PbB based on the NHANES IE Phase 1 data (1.7-2.1), lie outside and above
the 95% confidence intervals for the GM PbB estimated from the combined data
from the NHANES IH phases 1 and 2 (1.4-1.9). Thus, the combined data from
Phases 1 and 2 of the NHANES III suggest a lower GM PbB than previously
reported in the EPA ALM documentation (U.S. EPA, 1996).
b. Both the upper end of the range of the EPA ALM default values for GSDi)I10 are: 1) 1.4-1,8 u,g/dL,
the range of the estimated GMs for the three major race/ethnicity groups; and 2) a
more conservative and equally supportable range would be 1.6-1.9, the range of
the 95% upper confidence limits of the GM for the major race/ethnicity groups.
d. Stratifying the data by census regions, reasonable updated ranges for PbB^,,,,, 0 are:
1) 1,4-2.0 |ig/dL, the range of the estimated GMs for the four census regions; and
2) a more conservative and equally supportable range would be 1.5-2.2, the range
of the 95% upper confidence limits of the GM for the four census regions.
e. The results also support use of an updated value ranges for GSDi)(ldult in the EPA
ALM. Stratifying the data by race/ethnicity groups, a reasonable range for
GSDadu]tis2.1-2.3.
Stratifying the data by census regions, a reasonable updated range for GSDj adu))
2.0-2.2.
is
Page 6 ~ Febraary 28, 2002
-------
3.3 IMPACTS OF UPDATED VALUE RANGES FOR PBBADULTi0 AND GSD1>ABULT ON PRGs
CALCULATED WITH THE EPA ALM
Table 3a contains preliminary remediation goals calculated with the EPA ALM, by census region
and race/ethnicity, using the estimated GMs and GSDs for the respective regions and
race/ethnicity groups.
a. The range for the PRGs established in 1996. based on the range of GMs and
GSDs provided in the ALM (Table 6), is 749-1754 \ig Pb/g soil (ppm).
b. Based on the range of values shown for the major race/ethnicity groups in Table
3a (i.e., for "All Regions"), the range of the PRGs decreased considerably to
794-1, 288 ppm.
c. Based on the range of values shown for the census regions in Table 3a, the range
of the PRGs decreased, but is shifted higher, to 1,079-1,366 ppm.
The similarity in the PRG ranges that are calculated, when each of the PbBg,,,,,,^ and GSDj adul,
ranges are assumed, suggests that use of the updated ranges for these parameters, although
reasonably supported by the NHANES in, may not produce a large change in the PRG calculated
at any given site.
3.4 RECOMMENDATIONS
Previous recommendations for the Interim Adult Lead Model were presented, in aggregate as
well as separately, for the racial/ethnic categories used by the NHANES in survey. This revision
retains the previous racial/ethnic categories and also presents the GM and GSD for each of the
four geographic quadrants delineated by NHANES HI. For site applications of the ALM,
estimates of the PbB^,,^ and GSDia(hl1t parameters could be based on either race/ethnicity or
geographic categories determined appropriate based on the specific demographic or geographic
characteristics of the site. Perceived gains in specificity achieved from stratifying on both
demographic and geographic characteristics may be offset by increased uncertainty caused by
using less of the available survey data. This uncertainty is evident in the reduction of sample size
and increased standard errors in the PbB (GM). Unfortunately, corresponding uncertainty in the
estimates of the GSD is not quantifiable by usual methods due to the complex sampling design
used in NHANES m.
Estimates for PbBadu]t 0 (GM) and GSDi)adult (GSD) by census region and race/ethnicity group are
provided for information. However, it is not recommended to base estimates of the PbBadll)t0 and
GSDi>adu|, from the NHANES HI survey that are stratified by both census region and race/ethnicity
group in the ALM to estimate site-specific risks because of the small sample sizes, particularly in
the Northeast and Midwest regions (e.g., n = 157 for Mexican-Americans in the Midwest region).
The small sample sizes are reflected in the large standard errors for the GM in those regions
(relative to the South and West regions), hi addition to race/ethnicity and census region, other
factors that should be considered when selecting an estimate for the PbBadll]l0 and GSDUdu]t
include characteristics of current and anticipated future exposed populations, age of the housing
stock in the area of the site and other potential sources of lead (e.g., industrial discharges).
Page 7 ~ Febmaiy 28, 2002
-------
Based on this analysis, updated ranges for the PbBadult0 and GSDjadult parameters in the EPA
ALM are supported by the data collected in the completed NHANES ni survey (Phases 1 and 2).
Although the use of these updated ranges in the EPA ALM may not appreciably change PRGs
calculated with the methodology, it is recommended that data from both phases of NHANES III
be used in all PbB analyses; this is consistent with the CDC's recommendation (CDC, I996a).
Page 8 ~ Febniaiy 28, 2002
-------
bi(PbB)
All Data
-2
z-icore
In (PbB) Non-Hispanic Whites
3 ,
1 -_
1 .
-1
GM = 1.52
GSD =1.95
R'
?
^0.6656x + 0.4197
= 0,9989
012
z-fcore
la (PbB) N on-Hispanic Blacks
y- 0.7.19lx+ 0.5866
In (PbB) Mexican-Americans
4 -\
2 ~
0,
GM = 1.72
GSD •« 2.22
-i
y
R'
*i
0
_ • Aj.
^j&&rr. . :
•7963X +
0.9958
^
0,5428 -
0 1 2 3
z- score
In (PbB) Others
i _
I „
I
GM = 1,74
GSD = 1.87
!
*'."
'• '-^n
'•y=:0-.<
•^^
pT; .
)241x + (
>>.--
,5552
RJ= 0.9811
-1 -1 0
1
1 2 3
z-score
FIGURES 1-5. Probability plots were prepared from the log-transformed percentiles
estimated with SAS-SUDAAN. The geometric mean (GM) was estimated by exp (intercept)
and the geometric standard deviation (GSD) was estimated by exp (Slope). The GM and
GSD estimated with this method compare favorably with the estimates produced with SAS-
SUDAAN. The mean difference between the GMs estimated by the two methods is
approximately 0.03; the mean difference in the GSDs is approximately 0.10.
Page 9 ~ Febmary 28, 2002
-------
20DO Census Regions
Midwest
Northeast
nniO South
223 West
FIGURE 6. Grouping of States into the Four U.S. Census Regions. Hawaii and Alaska (not
shown) are in the West Region.
Page 10 ~ Febmary 28, 2002
-------
TABLE 1. Estimated Cumulative Distribution Function of Blood Lead Concentration
(Hg/dL) in U.S. Women, 17-45 years of Age
Race/ethnicity
All
(n = 5016)
Percentile
minimum
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
91
92
93
94
95
96
97
98
99
maximum11
CTEa
0.5°
__t
~
-
-
0.77
1.11
1.25
1.38
1.49
1.61
1.75
1.89
2.06
2.22
2.47
2.81
3.22
3.81
3.89
4.05
4.26
4.52
4.84
5.11
5.73
6.50
8.13
29.2
SEb
naf
-
-
-
-
0.10
0.03
0.03
0.02
0.02
0.02
0.03
0.03
0.03
0.03
0.04
0.05
0.07
0.05
0.05
0.07
0.08
0.09
0.12
0.09
0.18
0.16
0.14
na
95'hLCL°
na
-
-
-
-
0.58
1.05
1.19
1.34
1.45
1.56
1.70
1.83
2.00
2.16
2.39
2.71
3.09
3.71
3.80
3,90
4.10
4.33
4.61
4.94
5.37
6.17
7.86
na
95'kUCL<
na
-
-
-
-
0.97
1.16
1.30
1.43
1.53
1.66
1.81
1.95
2.11
2.28
2.54
2.90
3.36
3.90
3.98
4.19
4.42
4.71
5.07
5.28
6.09
6.83
8.41
na
'CTB: central tendency estimate
"SB: standard error of the estimate (balanced repeated replication method)
'95th LCL/UCL: lower/upper 95" % confidence limits for the estimated pereentile
'Minimum value shown is the value assigned to non-detects (i.e., '/> detection limit of 1 ug/dL)
'The value 0.5 is the value assign ed to non-detects; th e limit of detection for blood lead concentration reported by CDC is 1.0 ug/dL (CDC,
1996b).
ha: not applicable
'Indicates the presence of non-detects prevented an estimate of the pereentile and its standard error
'Maximum value shown is the observed values extracted from the NHANES III database; it is not an estimate.
Page 11 - Febiuary 28, 2002
-------
TABLE 1. Estimated Cumulative Distribution Function of Blood Lead Concentration
(jig/dL) in U.S. Women, 17-45 years of Age — Continued
Race/ethnicity
no n- Hispanic .white
(n= 1529)
Percentile
minimumd
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
91
92
93
94
95
96
97
98
99
maximum11
CTE8
0.5C
_g
-
-
-
-
1.03
1.18
1.32
1.44
• 1.54
1.67
1.82
1.97
2.13
2.31
2.64
3.06
3.61
3.74
3.84
3.98
4.23
4.52
4.89
5.20
6.03
7.41
12.4
SEb
nar
-
-
. -
-
-
0.05
0.03
0.03
0.02
0.03
0.04
0.03
0.04
0.03
0.05
0.07
0.08
0.07
0.06
0.05
0.09
0.11
0.09
0.14
0.17
0.20
0.42
na
95'h LCL-
na
-
-
-
-
-
0.92
1.11
1.25
1.39
1.48
1.60
1.75
1.89
2.06
2.21
2.50
2.90
3.47
3.62
3.74
3.81
4.01
4.33
4.61
4.86
5.62
6.57
na
QCth
95 UCLC
na
-
-
-
-
• -
• 1.14
1.25
1.39
1.49
1.59
1.74
1.88
2.04
2.19
2.42
2.78
3.22
3.76
3.87
3.94
4.16
4.44
4.71
5.18
5.53
6.43
8.26
na
'CTE: central tendency estimate
kSE: standard error of the estimate (balanced repeated replication method)
'95th LCL/UCL: lower/upper 95* % confidence limits for the estimated percentile
'Minimum value shown is the value assigned to non-detects (i.e., Va detection limit of 1 pg/dL)
'The value 0.5 is the value assigned to non-detects; the limit of detection for blood lead concentration reported by COC is 1.0 ug/dL (CDC,
1996b).
'na: not applicable
'Indicates the presence of non-detects prevented an estimate of the percentile and its standard error
'Maximum value shown is the observed values extracted from the NHANES in database; it is not an estimate..
Page 12 ~ February 28, 2002
-------
TABLE 1. Estimated Cumulative Distribution Function of Blood Lead Concentration
(Hg/dL) in U.S. Women, 17-45 years of Age — Continued
Race/ethnicity •
non-Hispanic black
(«= 1692)
Perc entile
minimum
5
10
15
20
25 •
30
35
40
45
50
55
60
65
70
75
80
85
90
91
92
93
94
95
96
97
98
99
maximum11
CTEa
0.5C
.»
-
-
0.89
1.13
1.25
1.38
1.52
1.66
1.79
1,95
2.16
2.40
2.69
3.03
3.37
3.87
4.53
4.75
4.91
5.11
5.28
5.76
6.25
6.71
7.84
9.77
20.3
SEb
naf
-
-
-
0.11
0.03
0.03
0.04
0.05
0.04
0.05
0.06
0.07
0.08
0.08
0.08
0.13
0.11
0.16
0.13
0.12
0.12
0.23
0.23
0.21
0.25
0.30
0.77
na
95thLCL<
na
-
-
-
0.67
1.07
1.20
1.30
1.43 ,
1.57
1.69
1.83
2.02
2.23
2.54
2.88
3.11
3.65
4.21
4.48
4.68
4.86
4.81
5.30
5.83
6.21
7.25
8.22
na
95'hUCL°
na
-
-
-
1.10
1.19
1.31
1.46
1.61
1.74
1.89
2.08
2.30
2.57
2.85
3.19
3.63
4.09
4.86
5.01
5.15
5.36
5.75
6.22
6.68
7.21
8.43
11.32
na
"CTE: central tendency estimate
"SE: standard error of the estimate (balanced repeated replication method)
'95th LCL/UCL: lower/upper 95* % confidence limits for the estimated percentile
'Minimum -value shown is the value assigned to non-detects (i.e., !/j detection limit of 1 Hg/dL)
The value 0.5 is the value assigned to non-detects; the limit of detection for blood lead concentration reported by CDC is I Si ug/dL (CDC,
1996b).
*na: not applicable
•Indicates the presence of non-detects prevented an estimate of the percentile and its standard error
'Maximum value shown is the observed values extracted from the NHANES III database; it is not an estimate.
Page 13 ~ February 28, 2002
-------
TABLE 1. Estimated Cumulative Distribution Function of Blood Lead Concentration
(Hg/dL) in U.S. Women, 17-45 years of Age — Continued'
Race/ethnicity
Mexican -
American
(n=I562)
Percentile
minimum11
5
10
15
20
25'
30
35
40
45
50
55
60
65
70
75
80
85
90
91
92
93
94
95
96
97
98
99
maximum11
CTE"
0.5"
_t
~
-
-
0.92
1.14
1.31
1.47
1.63
1.81
1.97
2.16
2.37
2.59
2.90
3.29
3.79
4.51
4.74
5.02
5.34
5.77
6.26
7.11
7.85
9.15
12.29
29.2
SEb
naf
-
-
-
-
0.11
0.03
0.05
0.04
0.04
0.05
0.05
0.05
0.05
0.06
0.07
0.08
0.10
0.13
0.16
0.16
0.20
0.20
0.34
0.35
0.40
0.37
0.97
na
95th LCL°
na
-
_
-
-
0.70
1.07
1.22
1.40
1.54
1.72
1.88
2.07
2.26
2.48
2.75
3.14
3,59
4.24
4.42
4.70
4.94
5.37
5.58
6.41
7.05
8.41
10.34
na
QCth
y5 UCLC
na
-
-
-
-
1.15
1.21
1.40
1.55
1.71
1.90
2.07
2.26
2.48
2.70
3.05
3.44
3.99
4.77
5.06
5.34
5.75
6.18
6.94
7.81
8.66
9.88
14.23
na
'CTE: central tendency estimate
kSE: standard error of the estimate (balanced repeated replication method)
'95th LCL/UCL: lower/upper 95" % confidence limits for the estimated percentile
'Minimum value shown is the; value assigned to non-detects (i.e., '/i detection limit of 1 ug/dL)
'The value 0.5 is the value assigned to non-detects; the limit of detection for blood lead concentration reported by CDC is 1.0 ug/dL (CDC,
1996b).
'na: not applicable
'Indicates the presence of non-detects prevented an estimate of the percentile and its standard error
'Maximum value shown is the observed values extracted from the NHANES HI database; it is not an estimate.
Page 14 ~ February 28,2002
-------
TABLE 1. Estimated Cumulative Distribution Function of Blood Lead Concentration
(jig/dL) in U.S. Women, 17-45 years of Age— Continued
Race/ethnicity
other
racial-ethnic groups
(n = 233)
Percentile
minimum11
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
91
92
93
94
95
96
97
98
99
maximum11
CTEa
0.5°
_g
-
-
1.07
1.24
1.37
' 1.52
1.63
1.72
1.81
1.92
2.06
2.20
2.39
2.52
2.80
3.09
3.88
3.93
3.99
4.11
4.28 ,
4.71
5.03
6.21
7.09
7.86
9.20
SEb
naf
-
-
-
0.18
0.09
0.06
0.07
0.05
0.05
0.05
0.07
0.07
0.09
0.08
0.11
0.13
0.32
0.13
0.12
0.12
0.15
0.25
0.36
0.51
0.95
0.89
0.62
na
95'hLCL°
na
_
-
-
0.71
1.06
1.25
1.38
1.53
1.63
1.71
1.78
1.92
2.02
2.23
2.31
2.53
2.45
3.62
3.69
3.74
3.80
3.78
3.99
4.02
4.30
5.30
6.61
na
95th UCLC
na
-
-
-
1.42
1.42
1.50
1.66
1.72
1.81
1.92
2.06
2.20
2.39
2.55
2.74
3.06
3.73
4.13
4.17
4.24
4.42
4.77
5.42
6.05
8.12
8.87
9.10
na
'CTE: central tendency estimate
hSE: standard error of the estimate (balanced repeated replication method)
'95th LCL/UCL: lower/upper 95* % confidence limits for the estimated peicentile
'Minimum value shown is the value assigned to non-detects (i.e., 'A detection limit of 1 (ig/dL)
'The value 0.5 is the value assigned to non-detects; the limit of detection for blood lead concentration reported by CDC is 1.0 ug/dL(CDC,
1996b). .
Via: not applicable
'Indicates the presence of non-detects prevented an estimate of the peicentile and its standard error
'Maximum value shown is the observed value: extracted from the NHANES III database; it is not an estimate.
Page 15 ~ Febraaiy 28, 2002
-------
TABLE 2. Estimated Geometric Means and Geometric Standard Deviations of Blood Lead
Concentration (p.g/dL) in U.S. Women, 17-45 Years of Age
Race/
ethnicity8
All
non- Hispanic
white
non-Hispanic
black
Mexican-
American
Other
n
5016
1529
1692
1562
233
GMb
(prob
plot)
1.55
1.52
1.80
1.72
1.74
GMC
(SUDAAN)
1.53
1.45
1.78
1.70
1.74
GMSEd
(SUDAAN)
0.05
0.06
0.06
0.06
0.11
GSDC
(prob
plot)
2.01
1.95
2.05
2.22
1.87
GSDf
(SAS)
2.11
2.09
2.16
2.29
1.97
R2
>0.99
>0.99
>0.99
>0.99
>0.98
•Race-Ethnicity categories provided in NHANES HI
*OM: Estimates of the geometric mean PbB estimated from the log probability plots (Figures 1-5).
'CM: Estimates of the geometric mean PbB estimated directly from NHANES ffl using SUDAAN software.
*GM SE: Standard error of the geometric mean estimated with SUDAAN (Taylor series method).
'GSD: Geometric standard deviation estimated from the log probability plots (Figures 1-5).
'GSD: Geometric standard deviation estimated directly from NHANES Ul using SAS and the WTPFEX6 sample weight.
*R:: Coefficient of variation from the probability plots shown in Figures 1-5.
Page 16 - February 28, 2002
-------
TABLE 3a. Estimated Geometric Means and Geometric Standard Deviations of
Blood Lead Concentration (ng/dL) in
U.S. Women, 17 - 45 Years of Age,
By Census Region and Race/Ethnicity
All Regions
Race/Ethnicity"
All
non-Hispanic white
non-Hispanic black
Mexican-American
Other
n
5016
1529
1692
1562
233
GM"
1.53
1.45
1.78
1.70
1.74
GMSE'
0.05
0.06
0.06
0.06
0.11
GSD"
2.11
2.09
2.16
2.29
1.97
PRO0
1,197
1,288
938
794
1,321
Northeast Region
Race/Ethnicity"
All
non-H ispanic white
non-Hispanic black
Mexican- American
Other .,
«
629
240
273
24
92
GMb
1.98
1.93
2.55
3.32
1.83
GMSEC
0.16
0.18
0.24
0.60
0.16
GSDd
2.00
2.01
1.94
1.89
1.94
PRG"
1,092
1,107
823
NRf
NR
Midwest Region
Race/Ethnicity8
All
non-Hispanic white
non-Hispanic black
Mexican-American
Other
n
945
428
347
157
13
GMb
1.53
1.42
2.11
1.88
2.83
GMSEC
0.12
0.14
0.12
0.25
0.52
GSD"
2.18
2.11
2.24
2.39
2.07
PRO0
1,079
1,273
582
535
NR
Page 17 ~ Febniary 28, 2002
-------
TABLE 3a. Estimated Geometric Means and Geometric Standard Deviations of
Blood Lead Concentration (ng/dL) in
U.S. Women, 17 - 45 Years of Age,
By Census Region and Race/Ethnicity — Continued
South Region
Race/Ethnicity*
All
non-Hispanic white
non-Hispanic black
Mexican-American
Other
n
2159
595
947
560
57
GMb
1.39
1.30
1.51
1.82
1.76
GMSEC
0.04
0.05
0.07
0.16
0.20
GSDd
2.07
2.04
2.11
2.16
1.85
PRO"
1,366
1,485
1,211
910
NR
West Region
Race/Ethnicity
All
non-Hispanic white
non-Hispanic black
Mexican-American
Other
«
1283
266
125
821
71
GMb
1.40
1.30
1.87
1.59
1.48
GMSEC
0.09
0.08
0.13
0.05
0.20
GSDd
2.11
2.08
2.04
2.31
1.92
PRO0
1,287
1,410
1,089
842
NR
•Race-Ethnicity categories provided in NHANESIH
"CM: Estimates of the geometric mean PbB estimated using SUDAAN software.
"CM S£: Standard error of the geometric mean estimated with SUDAAN (Taylor series method).
'GSD: geometric standard deviation estimated using SASand the WTPFEX6 sample weight.
'PRO: Preliminary Remediation Goal; determined with the EPA Adult lead Model using the indicated GMs and GSDs and
with the other AIM parameters set to default values.
'NR: Not Reported; PRGs arc not reported when the number of observations (n) is less than 100.
Page 18 ~ Febiuary 28, 2002
-------
TABLE 3b. Estimated Geometric Means and Geometric Standard
Deviations of Blood Lead Concentration (ng/dL) in
U.S. Women, By Age and Race/Ethnicity
Age Group: 17-25
Race/Ethnicity*
All
non-Hispanic white
non-Hispanic black
Mexican- American
Other
n
1625
417
547
577
84
GMb
1.23
1.12
1.50
1.55
1.39
GMSE'
0.05
0.06
0.07
0.08
0.14
GSD"
2.08
2.02
2.07
2.35
2.00
Age Group: 26-35
Race/Ethnicity*
All
non-Hispanic white
non-Hispanic black
Mexican-American
Other
n
1789
S68
599
555
67
GMb
1.55
1.47
1.72
1.74
1.85
GMSEC
0.06
0.07
0.08
' 0.08
0.16
Age Group: 36-45
Race/Ethnicity1
All
non-Hispanic white
non-Hispanic black
Mexican- American
Other
n
1602
544
546
430
82
GMb
1.80
. 1.71
2.20
1.86
2.01
GMSEC
0.07
0.07
0.11
0.09
0.19
GSDd
2.07
2.05
2.23
2.27
1.78
GSDd
2.09
2.09
2.06
2.21
2.00
•Race-Ethnicity categories provided in NHANES ffl
hGM: Estimates of the geometric mean PbB estimated using SUDAAN software.
°GM SE: Standard enor of the geometric mean estimated with SUDAAN (Taylor series method).
'GSD: geometric standard deviation estimated using SAS and the WTPFEX6 sample weight.
Page 19 ~ Febmary 28, 2002
-------
TABLE 3c, Estimated Geometric Means and Geometric Standard
Deviations of Blood Lead Concentration (jig/dL) in
U.S. Women, By Age and Census Region
Age Group: 17-25
Census
Region'
AH
Northeast
Midwest
South
West
n
1625
211
267
727
420
GMb
1.23
1.67
1.10
1.16
1.07
GMSEC
0.05
0.15
0.11
0.05
0.08
. GSDd
2.08
2.01
2.00
2.05
2.09
Age Group: 26-35
Census
Region8
All
Northeast
Midwest
South
West
n
1789
214
370
744
461
GMb
1.55
2.00
1.54
1.40
1.44
GMSEC
0.06
0.26
0.10
0.04
0.11
GSDd
2.07
1.94
2.19
2.05
1.98
Age Group: 36-45
Census
Region"
All
Northeast
Midwest
South
West ,
n
1602
204
308
688
402
GMb
1.80
2.30
1.89
1.62
1.63
GMSEC
0.07
0.14
0.19
0.06
0.14
GSDd
2.09
1.99
2.12
2.02
2.16
•Census regions provided in NHANES ID
"GM: Estimates of the geometric mean PbB estimated using SUDAAN software.
'CM SE: Standard error of the geometric mean estimated with SUDAAN (Taylor series method).
"GSD: geometric standard deviation estimated using SAS and the WTPFEX6 sample weight.
Page 20 ~ February 28, 2002
-------
TABLE 4. Comparison of the Rate of Non-Detects in Blood Lead Concentrations Between Phases
1 and 2 of the NHANES III for U.S. Women, 17 - 45 Years of Age
Ethnicity"
All
non-
Hispanic
white
non-
Hispanic
black
Mexican-
American
Other
Phases 1 and 2
«
5016
1529
1692
1562
233
non-
detects
1070
434
285
312
39
%of
sample
21.3
28.4
16.8
20.0
16.7
Phase 2
»
2769
788
1035
800
146
non-
detects
681
259
202
191
29
%of
sample
24.6
32.9
19.5
23.9
19.9
Phase 1
n
2247
741
657
762
87
non-
detects
389
175
83
121
10
%of
sample
17.3
23.6
12.6
15.9
11.5
Page 21 ~ Febniary 28, 2002
-------
TABLE 5. Comparison of Blood Lead Concentration Estimates
of U.S. Women, 17-45 Years of Age,
with Default Values Used in the EPA Adult Lead Methodology
Ethnicity9
All
non-Hispanic
white
non-Hispanic
black
Mexican-
American
NHANES Phases 1 and 2
n
5016
1529
1692
1562
GM
1.5
(1.4-1.6)
1.4
(1.3-1.6)
1.8
(1.7-1.9)
1.7
(1.6-1.8)
GSD
2.1
2.1
2.2
2.3
NHANES Phase 1 (U.S. EPA, 1996)
«
2250
742
658
763
GM
1.8
1.7
2.1
2.0
GSD
1.9
1.9
2.0
2.1
U.S. EPA ALM( 1996)
-
1.7-2.2
1.8-2.1
-
1.7-2.2
1.8-2.1
Page 22 ~ Febraary 28, 2002
-------
TABLE 6. Comparison of PRGs Calculated with the EPA ALM Using Default Value Ranges
or Updated Ranges for the PbB^,,,,,, and GSDUdult Parameters
GSDi)8dult
1.8
2.1
PbBaduh>0
(default)1
1.7
1754
1096
2.2
1406
749
GSDi>8du]t
1.9
2.3
PbBadlll, ,0
(GM range)*
1.4
1712
988
1.8
1434
710
GSDijadult
1.9
2.3
PbBadult>0
(95% UCL range)0
1.6
1573
849
1.9
1365
641
•EPA ALM (U.S. EPA, 1996)
"Race/ethnicity range of the GM PbBs
'Race ethnicity range of the 95% upper conftdoice limit on the GM PbBs
m
Page 23 ~ Febiuaiy 28, 2002
-------
4.0 References
Brody, D.J., J.L. Pirkle, R.A. Kramer, K.M. Flegal, T.D. Matte, E.W. Gunter, and D.C. Paschal.
1994. Blood Lead Levels in the U.S. Population. Phase 1 of the Third National Health and
Nutrition Examination Survey (NHANES III, 1988 TO 1991). JAMA. 272(4):277-283.
CDC. 1996a. National Health and Nutrition Examination Survey III, Weighting and Estimation
Methodology. Executive Summary. Prepared for National Center for Health Statistics, Hyattsville,
MD. Westat, Inc. Rockville, MD. February.
CDC. 1996b. NHANES in Reference Manuals and Reports. CD-ROM. October.
CDC. 1997. National Health and Nutrition Examination Survey, HI 1988-1994. CD-ROM Series
11, No. I.July.
Palisade Corporation. 1997. "@ Risk" Risk Analysis and Simulation Add-In for Microsoft Excel
or Lotus 1-2-3. Newfield, NY: Palisade Corporation.
SAS Institute Inc. 1989. SAS Language and Procedures: Usage, Version 6, First Edition, Gary,
NC: SAS Institute Inc.
Shah, B.V., B.C. Barnwell, and G.S. Bieler. 1997. SUDAAN User's Manual, Release 7.5.
Research Triangle Park, NC: Research Triangle Institute.
Thayer, B. and G. Diamond. 2000. Memorandum to Mark Follansbee, ISSI, Inc. Subject: ISSI
Subcontract No. 0038-96-02, EPA Contract No. 68-W6-0038. Task 5: Hotline Support. April 11.
U.S. Census Bureau. Census Regions Cartographic Boundary Files Descriptions and Metadata.
Geography Division, Cartograhic Operations Branch. Searched November, 26, 2001. Available
from: http://www.census.gov/geo/www/cob/rg_metadata.html
U.S. Environmental Protection Agency. Technical Review Workgroup for Lead. 1996.
Recommendations of the Technical Review Workgroup for Lead for an Interim Approach to
Assessing Risks Associated with Adult Exposures to Lead in Soil. Available from:
hrtrj://www.epa.arjv/superrund/prograrns/lead/prods.htm
Page 24 ~ Febiuary 28, 2002
,1
------- |