INTERPRETATION
AND USE OF BLOOD LEAD DATA
Dr.
T Matte
Introduction
This presentation
reviews the interpretation of blood lead data and their use to help
monitor and control harmful lead exposure. In particular, the interpretation
of blood lead measurements in individuals is considered as well as
the use of aggregate data to guide and monitor prevention efforts.
I will emphasize blood lead measures in children, the most vulnerable
group.
Before considering
how blood lead data should be interpreted and used, it is important
to consider how blood lead monitoring fits into a comprehensive strategy
of monitoring and controlling lead exposure (Figure 1). Lead is one
of the most studied environmental toxicants and thus much is known
about the causal pathways leading from the production and uses of
lead to their ultimate impact on human health. This knowledge allows
us to move monitoring efforts "upstream" to identifying
ongoing uses of lead (e.g. as a gasoline additive) that are known
to lead to widespread excessive exposure. In short, we need not depend
on blood lead data as the sole guide to the need for source reduction.
In fact, blood lead data can sometimes be misleading for reasons discussed
later in this presentation. An overall prevention strategy then, should
include monitoring production, modes of consumption, reservoirs (air,
water, soil and food), blood lead levels and their effects on health.
.
Interpretation
of individual blood lead measurements
Health effects
benchmarks: As the levels of blood lead increase, its effects
become more pronounced. Levels of more than 50 micrograms/dl of lead
manifest clinically with death occurring at levels of 150 micrograms/dl.
At levels < 50 ug/dL, subclinical effects occur. Most important
of these "low-level effects" is impaired cognitive performance
and behavioral changes. IQ decrements are associated with increasing
blood lead levels, a relationship that is clear above 10 micrograms/dl
and probably continues at lower levels. This finding is supported
by many studies in diverse populations.
Clinical and public
health response to elevated blood lead levels in individual children
requires a multi-tier approach. According to CDC guidelines, levels
>= 10 micrograms/dl should trigger some preventive measures at
the community level. Medical evaluation is recommended when blood
lead levels are 20 ug/dL or higher. The need for chelation therapy
is widely acknowledged for blood levels more than approximately 45
micrograms/dl. There is less agreement on the value of chelation for
children with lower, but still elevated, blood lead levels (i.e. 20-45
ug/dL). However, these blood lead guidelines may not be appropriate
in other countries. Protocols for follow-up should be based on sources
of exposure, resources for response, and evidence on the efficacy
of interventions. In no case should a program of screening and individualized
follow-up and treatment (secondary prevention) be a substitute for
efforts to identify and reduce sources of lead exposure (primary prevention).
Factors influencing
blood lead measurements: Interpretation of individual blood lead
levels requires and understanding of factors that affect measurement
accuracy and precision. Studies comparing lead measurements of finger-stick
and venous blood samples taken from the same individuals show that,
with proper collection technique, the average positive bias in a fingerstick
sample is less than 1 m g/dL. This is acceptable for screening
purposes. Without good technique, however, contamination of fingerstick
samples and positive bias will be much greater.
Individual blood
lead levels may also vary with age and season, and awareness of these
trends must be considered when evaluating changes in blood lead levels
over time. In certain populations where there is significant exposure
of children to house dust and soil contaminated with lead, there has
been an increase in blood lead levels with age beginning at infants
< 1 year up to a peak at about 2 years of age. This pattern is
apparently related to age-related changes in mouthing behavior and
mobility. It is not seen in some populations that have different predominant
exposures (e.g. diet). Seasonal variation also affect blood lead levels,
which tend to peak in late summer/early fall and are lowest in late
winter/early spring. The explanation for this observation is not fully
known, but it may reflect seasonal changes in actual exposure.
Lead accumulates
in bone where it comprises a high proportion of total body lead burden,
especially for chronically exposed individuals. In these data from
a study by Hu et al (Figure 2), shows bone lead levels in occupationally
exposed adults to be better predicted by their cumulative past lead
than by current blood lead level. Past lead exposure can influence
current blood lead level through release of lead from bone stores.
As a result, when persons chronically exposed to lead have been removed
from the offending environment, blood lead levels tend to decrease
very slowly over time, due to release of bone lead stores.
Uses of aggregate
blood lead data (Figure 3)
Types of data:
Just as it is important to understand factors influencing blood
lead measures at the individual level, it is also crucial to take
into account the type of population sample (Figure 4) before drawing
inferences from blood lead data.. Each type may be subject to its
own biases and limitations. For example, children who with access
to health care and lead screening services may not include the most
socially disadvantaged children, who are often at highest risk for
lead exposure.
The following
sections present some examples of how blood lead data have been used
to support prevention efforts.
Targeting high-risk
groups: In the NHANES (National Health and Nutrition Examination
Survey) III (1988-94) it was found that blood lead levels were high
in lower socio-economic populations. In addition, children living
in homes built in pre-1946 years, when lead was extensively used in
paints, had higher lead levels than those in the post 1946 period
when less was added to paint (Figure 5). These findings have helped
to target prevention efforts in the U.S.
Monitoring
trends: Over time blood lead levels have also decreased as usage
of leaded gasoline was phased out. This was demonstrated in the NHANES
II study (1976-80), a finding that helped to support the phase out
of Pb in gasoline.
Source identification:
In a study carried out in Jamaica, nearly 80% of children who
stayed in the premises of battery repair shops had blood lead levels
greater than 50 ug/dL, much higher than control children. Levels were
higher in children than adults probably because of contamination of
house dust and soil with lead and hand to mouth behavior of children.
Blood lead data were used in Mexico to show a strong relation of use
of ceramic food and beverage containers to blood lead levels. In many
states, blood lead measurements are reported to the health department
where investigations are triggered by high levels. These investigations
may reveal new or unusual sources of lead exposure (Figure 6), leading
to appropriate preventive measures.
Though blood lead
studies have been useful in assessing sources and pathways of lead
exposure, they have important limitations (Figure 7). Some sources
of lead exposure are difficult to measure precisely (such as lead
in diet), others are ubiquitous - all members of a community may be
exposed. When either or both factors are present, the contribution
of a given source to population lead exposure may be underestimated,
or missed altogether. Certain uses of lead may produce little exposure
initially, but then cause widespread lead exposure after many years
of use. A notable example in the U.S. is leaded paint. When initially
applied and well-maintained it may cause less exposure than after
years of deferred maintenance and deterioration when children become
exposed to resulting paint chips and contaminated house dust. Thus,
in the U.S. most cases of childhood lead poisoning due to leaded paint
occurs in houses more than 50 years old. Paint used before that time
often contained 30-50% lead by dry weight and that same paint is now
deteriorated and causing exposure. Thus, if a developing country were
beginning to use leaded paint in a rapidly growing housing stock,
the consequences may not be evident in a blood lead study done today.
Because of the
limitations of blood lead studies, prudent public health policy would
be to phase out uses of lead, such as in residential paint, gasoline,
water and food cans, that have already demonstrated, after years of
use, to cause widespread lead exposure in other countries.
Establishing
exposure-response relationships: The most important remaining
sources of lead in the U.S. are soil and dust (Figure 8) contaminated
mainly by deteriorated house paint and past use of leaded gas. So
an important question for us is the relation of levels of lead in
contaminated dust to blood lead levels. A recent combined reanalysis
(Figure 9) showed a strong relation down to house dust lead levels
as low as 10 ug/ft2. This data is enabling the use of dust lead measurements
as an alternative to blood lead screening to identify hazardous home
environments.
Conclusion
Although blood
lead measurements have been invaluable for guiding both individual
and community-level prevention efforts, they should only be part of
an overall strategy for preventing lead toxicity. In addition, their
proper use requires a knowledge of factors that can influence individual
levels and population distributions as well as the limitations of
blood lead studies for source apportionment. Conclusion, prior to
undertaking a blood lead study, consider (Figure 10):
- What question
is being asked?
- What other
data or studies might already exist to address the question?
- What factors
(age, season, measurement techniques) could affect the measurements?
.
- What is the
best type of sample or study design to address the question?
.
.
The following
paper was submitted by Dr. Matte as supporting evidence of the above
presentation.
.
LEAD EXPOSURE
FROM CONVENTIONAL AND
COTTAGE LEAD
SMELTING IN JAMAICA
.
Thomas D. Matte*,
J. Peter Figueroa**, Stephanie Ostrowski***, Gregory Burr,
Linnette Jackson-Hunt, and Edward L. Baker
*Office of the
Director. National Institute for Occupational Safety and Health. Centers
for Disease Control, Atlanta. Georgia, USA: **Epidemiology Unit, Ministry
of Health. Jamaica: ***Division of Environmental Hazards and Health
Effects. Center for Environmental Health and Injury Control. Centers
for Disease Control, Division of Surveillance. Hazard
Evaluations and Field Studies. National Institute for Occupational
Safety and Health. Centers for Disease Control. Cincinnati, Ohio.
USA; Emory University Master of Public Health Program. Atlanta.
Georgia. USA and Public Health Practice Program Office, Centers
for Disease Control, Atlanta. Georgia. USA
Abstract.
A survey was conducted to determine the distribution and determinants
of environmental and blood lead levels near a conventional and several
cottage lead smelters and to assess the relationship between environmental
and blood lead levels in a tropical, developing-country setting. Fifty-eight
households were studied in the Red Pond community, the site of the
established smelter and several back-yard smelters, and 21 households
were studied in the adjacent, upwind Ebony Vale community in Saint
Catherine Parish, Jamaica. Households were investigated, using questionnaires,
soil and housedust lead measurements, and blood lead (PbB) measurements
from 372 residents. Soil lead levels in Red Pond exceeded 500 parts
per million {ppm) at 24% of households (maximum -18.600 ppm), compared
to 0% in Ebony Vale (maximum 150 ppm). Geometric mean PbB in Red Pond.
where 44% of children <6 years of age had PbB levels ³ 25
micrograms per deciliter (m g/dL), was more than twice that Ebony
Vale in all age groups (p < 0.0005). Within Red Pond, proximity
to backyard smelters and to the conventional smelter were independent
predictors of soil lead (p < 0.05). Soil lead was the strongest
predictor of PbB among Red Pond subjects under 12 years of age. The
blood lead-soil lead relationship in children differed from that reported
in developed countries; blood lead levels were higher than expected
for the household-specific soil lead levels that were observed. These
data indicate that cottage lead smelters, like conventional ones,
are a hazard for nearby residents and that children exposed to lead
contamination in tropical, developing countries may be at higher risk
for developing elevated blood lead levels than similarly-exposed children
in developed countries.
Lead poisoning
associated with conventional lead smelting in developed countries
has been described among both workers (Lilis et al. 1977) and
community residents (Popovac et al. 1982; Landrigan et al.
1975; Brunekreef et al. 1981). While airborne lead fume
from smelters is the main vehicle of environmental contamination,
the most important route of community lead exposure in such settings
appears to be ingestion of lead-contaminated soil and housedust, especially
by children (Roels et al. 1980). Children of smelter workers
may be at particularly high risk from exposure to lead dust brought
home on work clothes (Morton et al. 1982; Baker et al. 1977).
Workers and their
families may also be at risk of lead poisoning when lead-related work
is done at home. Lead-related cottage industries have caused lead
poisoning in both developed (Kawai et al. 1983) and in less
developed countries (Koplan et al. 1977), where cottage industries
are relatively prevalent (Phoon 1982). Cases of lead-poisoning associated
with lead smelting near the home have been reported previously (Dolcourt
et al. 1981). However, such establishments are typically scattered
and we are aware of no systematic study of lead exposure from cottage
lead smelting. The relationship between environmental lead contamination
and blood lead levels in children has been extensively studied (Duggan
and Inskip 1985), and data from such studies have been used to propose
"maximum permissible levels" of lead in soil (Madhavan et al. 1989).
However, nearly all previous studies have been carried out in developed
countries with temperate climates. Because unintentional soil and
dust ingestion may be related to hygiene and to time spent outdoors
(Duggan and Inskip 1985), and because lead absorption is influenced
by nutrition (Mahaffey 1982), one might expect soil lead-blood lead
relationships to differ in tropical, developing countries.
Cottage lead smelters
are scattered throughout Jamaica. The clustering of several of these
so-called "backyard" lead smelters in a community near a conventional,
secondary lead smelter provided the opportunity to assess, during
a cross-sectional survey conducted in October 1987, the amount of
environmental contamination associated with both cottage and conventional
smelting in the same community. In addition, the relationship between
environmental contamination and blood lead levels in a tropical, developing
country was examined.
Methods
Study Site
and Population
The conventional
lead smelter ("established smelter") has operated at the southeast
corner of the Red Pond Road community (estimated population 2500)
since 1963. Reclaiming lead from spent car batteries. The smelter,
which usually operated for about two weeks out of every two month
period, did not operate during the several weeks before or during
the survey. Several crude backyard lead smelters are also known to
operate in this relatively poor community, and some residents have
reportedly used lead oxide-containing drums and dross (slag) from
the established smelter grounds for fencing and landfill material.
A middle-class housing development, known as Ebony Vale (estimated
population 1600), was recently completed just east of the conventional
smelter. Prevailing winds in the area come from the northeast. Lead
poisoning cases have been recognized in Red Pond but not in Ebony
Vale.
Potential study
households were sampled within Red Pond and Ebony Vale by numbering
dwellings on community survey maps and selecting dwellings according
to a generated list of random numbers. Households with children less
than six years of age were over-sampled in Ebony Vale to increase
the precision of the mean blood lead estimate in that age group. Survey
procedures were completed at 49 Red Pond households (86% of those
sampled) and 21 Ebony Vale households (68% of those sampled). Of households
sampled but not surveyed, three refused participation (one in Red
Pond, two in Ebony Vale) and no adults were at home at 15 households.
In addition, all nine households in Red Pond reported by established
smelter employees to be at, or adjacent to, backyard smelter sites
("possible backyard smelter" households) were surveyed.
Blood lead measurement
was offered for all residents of surveyed households who were at least
six months of age. All participating subjects or their adult guardians
provided informed consent. Most non-participants were not at home
when the dwelling was surveyed. The results of household and subject
selection are summarized in Table 1.
Data Collected
At each participating
household, a responsible adult was interviewed. Information collected
included household income, whether there had been smelting or use
of lead oxide drums or lead dross in a yard, and demographic information
on household members. Depending on the age of household members, information
about occupation, smoking, pica, and time unattended by an adult was
also collected. At each household, a one centimeter deep core soil
sample in the approximate center of the yard was collected. Lead (Pb)
levels in such samples will be referred to as "Pb soil-area". Soil
was also sampled near potential sources of lead.
Table 1. Household
and subject selection, community lead survey, Saint Catherine Parish,
Jamaica, October 1987
________________________________________________________________________
Community_____________________________
Red Pond Ebony Vale
Selection criteria Possible
backyard Random Random
smelter sample sample
________________________________________________________________________
Households
No. participating
(% of those sampled) 9 (100) 49 (86) 21 (68)
Mean household size 5.3 7.6 4.4
Mean household income
(US $/week) 32 42 93
Subjects tested for
blood lead
(% of eligible subjects in
participating households)
By Age:
6 months-5 years 8 (100) 57 (93) 16 (84)
6-11 years 9 (90) 53 (88) 14 (82)
³ 12 years 20 (69) 157 (64) 38 (67)
________________________________________________________________________
contamination,
such as oxide drum fencing or lead scrap. The highest soil lead found
in each yard will be referred to as "Pb soil-peak". Dust samples were
collected. using a published procedure (Vostal et al. 1974),
from the center of the floor in the room where children spend the
most time. Area soil samples were lost or omitted for five households
(four in Red Pond and one in Ebony Vale) and dust samples were lost
or omitted at four households (three in Red Pond and one in Ebony
Vale). Scrapings of housepaint were taken at households where peeling
paint was observed. Blood samples were obtained by venipuncture.
Lead levels in
whole blood (Searle et al. 1973) and environmental samples
(National Institute for Occupational Safety and Health 1984) were
measured by published procedures. The limits of quantitation (LOQ)
were 5 micrograms per deciliter (m g/dL) for blood lead (PbB),
5 parts per million (ppm or m g/g) for soil lead, 24 m g
per square meter (m g/m2) for dust lead (PbD), and
0.01% for paint lead. Samples with lead below the LOQ were assigned
a value midway between zero and the LOQ for data analysis.
Data Analysis
Continuous variables
were transformed to correct skewness in distributions and apparent
non-linear relationships between variables. T-tests and ordinary least
squares regression were used to analyze environmental lead data. Blood
lead data were analyzed by statistical programs (Shah 1981; Holt 1982)
that employ a Taylor series approximation to compute standard errors
of estimated means and regression coefficients using households (rather
than individuals) as sampling units. Separate analyses were conducted
for three age groups: six months through five years (pre-school children),
six through 11 years (school-aged children), and 12 years and older
("adult", i.e., beyond the age of compulsory schooling). Multivariate
models were arrived at by backwards selection, eliminating the predictors
that were least statistically significant by partial F tests until
only those significant at the p < 0.05 level remained (Kleinbaum
and Kupper 1978).
Table 2.
Environmental and blood lead levels at survey households_
Red Pond
______________________Ebony Vale
Possible
backyard Random Random
Measurement smelter sample_ ___ sample
Pb soil-area (ppm) GMc 1089*
133'*** 6
% ³ 500 ppm 75a 24b 0A
Range 9-31.000 < 5-18.600 < 5-150
Pb soil-peak (ppm) GM 7691 ** 221 *** 7
% ³ 500 ppm 89 27 0
Range 9-320.000
<5-520.000 <5-150
Pb dust (m g/m2) GM 2790* 690*** 100
%³ 1500 m g/m2 56 24b 0a
Range 100-109.180 50-294.680 20-338
.
Blood Pb (m
g/dl),
By age_____________________________________________________________________________________________________
£ 5 years GM 25ns
21'*** 9
% ³ 25 m g/dL 50 44 0
Range < 5-94 5-98 < 5-19
6-11 years GM 62*** 17'*** 7
%³ 25 m g/dL 100 40 7
Range 35-139 <5-95 <5-25
³ 12 years GM 28* 12'*** 5
%³ 25 m g/dL 60 17 3
Range <5-90 <5-85 <5-27
____________________________________________________________________________________________________________
a Missing for
1 household
b Missing for
3 households
c GM = geometric
mean
P values for differences
in geometric means, possible backyard smelter compared with randomly-selected
Red Pond households and randomly-selected Red Pond households compared
with Ebony Vale households: *p < 0.05. **p < 0.005. ***p <
0.0005. NS = not significant.
.
Results
.
Environmental
and Blood Lead Levels
Geometric mean
Pb soil-area. Pb soil-peak, and Pb dust were 22, 31, and 7 times,
respectively, higher in Red Pond than Ebony Vale (p < 0.0005, Table
2). Furthermore. 24% of randomly selected Red Pond households had
soil lead levels greater than 500 ppm. a threshold above which elevated
blood lead levels in children are said to occur (Centers for Disease
Control 1985). In contrast, the highest soil lead level in Ebony Vale
was 150 ppm, and 14 Ebony Vale households had Pb soil-area levels
below 5 ppm. Geometric mean Pb soil-peak was 30 times higher at possible
backyard smelter households than at other Red Pond households (p <
0.005), and four possible backyard smelter households had Pb soil-peak
levels greater than 50,000 ppm. Peeling housepaint was observed at
one Ebony Vale and 27 Red Pond households. Ten samples exceeded 1%
lead by weight (maximum 6.0%), all from Red Pond.
Among subjects
of randomly-selected households, geometric mean PbB was significantly
higher in Red Pond than Ebony Vale in each age group (p < 0.0005,
Table 2), and decreased with age in both communities. Elevated PbB
(³ 25 m g/dL (Centers for Disease Control 1985) levels were
common in Red Pond (44% of children under six years) but unusual in
Ebony Vale. The prevalence of elevated PbB was higher among subjects
at possible backyard smelter households than at randomly-selected
Red Pond households, and the geometric mean PbB levels were significantly
higher at smelter households among subjects six through 11 years (p
< 0.0005) and subjects 12 and older (p < 0.05).
Determinants
of Environmental and Blood Lead Levels
The soil and dust
levels in Ebony Vale, being low and quite uniform, were not significantly
correlated with either proximity to the conventional smelter (Figure
l) or to blood lead levels. Because of this and because there were
no cottage smelters in Ebony Vale, the analyses reported below were
limited to the Red Pond community. For these analyses, households
were classified as confirmed backyard smelter sites (7 households),
suspected smelter sites (4 households), or non-smelter sites (all
others) according to questionnaire responses and/or field observations,
regardless of why the households were sampled. Two additional smelter
sites were confirmed just outside of the community.
Use
of lead oxide drum fencing or dross landfill in Red Pond yards was also
associated with contamination of whole yards (high area soil lead levels),
while leaded paint did not appear to be an important cause of dust or
soil contamination in the area studied. The strong relationship between
soil and dust lead levels and the high house dust lead levels along the
road running northwest from the smelter suggest that tracking of lead-contaminated
dust into dwellings is an important source of housedust contamination
in Red Pond, as has been noted elsewhere (Bornschein et al. 1985).