American Journal of Epidemiology
Copyright © 2004 by the Johns Hopkins Bloomberg School of Public Health
All rights reserved
Vol. 160, No. 9
Printed in U.S.A.
DOI: 10.1093/aje/kwh296
Bone Density-related Predictors of Blood Lead Level among Peri- and
Postmenopausal Women in the United States
The Third National Health and Nutrition Examination Survey, 1988–1994
Denis Nash1,2, Laurence S. Magder2, Roger Sherwin3, Robert J. Rubin4, and Ellen K.
Silbergeld2,4
1
Center for Urban Epidemiologic Studies, New York Academy of Medicine, New York, NY.
Department of Epidemiology and Preventive Medicine, School of Medicine, University of Maryland, Baltimore, MD.
3 Department of Epidemiology, School of Public Health, Tulane University, New Orleans, LA.
4 Department of Environmental Health Sciences, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD.
2
Received for publication February 10, 2003; accepted for publication June 1, 2004.
Because of the long half-life of lead stored in bone (years), skeletal lead stores may be a source of endogenous
lead exposure during periods of increased bone demineralization, such as menopause. To test the hypothesis
that postmenopausal bone resorption increases blood lead levels, the authors examined cross-sectional
associations of bone density-related factors with blood lead levels among women aged 40–59 years from the
Third National Health and Nutrition Examination Survey (1988–1994). Factors related to bone turnover were
significant predictors of blood lead level. Bone mineral density was significantly inversely related to blood lead
levels in log-linear multivariate models that adjusted for age, race/ethnicity, smoking, education, household
income, alcohol use, and residence (urban/rural). With menopausal status added to the model, naturally and
surgically menopausal women had adjusted median blood lead levels that were 25% and 30% higher,
respectively, than those of premenopausal women (2.0 µg/dl). Current use of hormone replacement therapy was
associated with significantly lower adjusted median blood lead levels (1.8 µg/dl) than past use (2.6 µg/dl) and
never use (2.2 µg/dl). Lead stored in bone may significantly increase blood lead levels in perimenopausal women
because of postmenopausal bone mineral resorption. Attention to factors that prevent bone loss may lessen or
prevent this endogenous lead exposure.
bone density; bone resorption; hormone replacement therapy; lead; menopause; women
Abbreviations: BMD, bone mineral density; HRT, hormone replacement therapy; NHANES, National Health and Nutrition
Examination Survey.
Despite major reductions in environmental lead exposure
for much of the US population (1), lead poisoning remains a
significant public health problem (2). While young children
and some adults continue to be at risk from exogenous
sources of lead exposure, new evidence suggests that some
persons may be at risk for recurring endogenous exposure to
lead previously accumulated over time in the skeleton (3–6).
Like calcium, lead is a bone-seeking element. Over 90
percent of the total burden of lead in the body resides in the
skeleton, where it has a half-life on the order of years to
decades (7–10). Thus, exposure to lead over the course of a
lifetime results in accumulation of lead in the skeletal
compartment (11–13) such that bone lead levels are generally higher among older persons (14). Of particular concern
is the potential for renewed exposure to lead released from
bone in older persons who were excessively exposed to environmental lead, by current standards, prior to its removal
from gasoline (1978–1990) and paint (1977) (15).
The human skeleton is a dynamic physiologic compartment of mineral metabolism. Mineral absorption and resorption are affected by many factors, including age, diet,
weight-bearing activity, trauma, metabolic disorders,
Correspondence to Dr. Denis Nash, Center for Urban Epidemiologic Studies, New York Academy of Medicine, 1216 Fifth Avenue, New York,
NY 10029 (e-mail: [email protected]).
901
Am J Epidemiol 2004;160:901–911
902 Nash et al.
hormonal status, pregnancy, lactation, and menopause (16–
20). In response to these factors, both lead and calcium are
deposited in the skeleton and may act similarly (4, 21, 22).
Lead competes with calcium for transport and binding sites,
and when calcium is released from bone, lead is also liberated (23–30). Additionally, lead does not accumulate
uniformly throughout the skeleton but is selectively taken up
at different sites and in different types of bone (e.g., trabecular bone more than cortical bone) (8, 26, 31, 32). These
events may be of particular importance to women because of
the changing dynamics of bone mineral metabolism related
to events such as pregnancy, lactation, and menopause (33,
34). Over her lifetime, a woman loses up to 50 percent of
trabecular bone and 30 percent of cortical bone, and 30–50
percent of this bone loss occurs in the early postmenopausal
years (35–39). Estrogen deficiency appears to play a significant role in bone loss (37, 40), and hormone replacement
therapy (HRT) after menopause has been shown to prevent
bone loss (38, 41–46) or even increase bone density (47).
There is mounting observational evidence that lead may be
mobilized from the skeleton during periods of increased
bone demineralization, such as pregnancy and lactation (48–
51), very old age (52), and menopause (3, 6, 53, 54).
Increases in blood lead associated with menopause are of
concern because recent research has linked lead in the blood,
at levels previously thought to be safe, to a number of
adverse health outcomes in adults, including increased blood
pressure (55–58), reduced kidney function (59), decrements
in neurocognitive function (60, 61), and increased risks of
atherosclerosis and cardiovascular disease mortality (62).
Two cross-sectional studies of US women that used data
from the Second National Health and Nutrition Examination
Survey (NHANES II; 1976–1980) and the Hispanic Health
and Nutrition Examination Survey (1982–1984) documented that postmenopausal women have significantly
higher blood lead levels than premenopausal women, after
results are controlled for age and other factors related to
exogenous lead exposure (3, 6). A cross-sectional study
conducted in the early 1990s found an inverse relation
between blood lead levels and bone-related factors,
including menopause (63). Another study also identified
menopausal status as an independent predictor of blood lead
levels in a random sample of Scandinavian women (53).
However, these studies were carried out at a time when environmental exposures and population blood lead levels were
much higher than they are at present. More recently, among
perimenopausal women from the Nurses’ Health Study who
were not using HRT, investigators found higher blood lead
levels and observed that bone lead was more strongly associated with blood lead in comparison with current HRT users
(64). In a study of women from the Third National Health
and Nutrition Examination Survey (NHANES III), Nash et
al. (58) reported the association between blood lead and diastolic hypertension to be most pronounced in postmenopausal women. A recent study of women in Mexico City,
Mexico, where environmental exposure to lead is high,
found no association between blood lead and bone mineral
density (BMD) (65). However, to our knowledge, the association has never been examined in the context of lower envi-
ronmental lead exposures, where the contribution of
endogenous lead to blood lead may be more relevant.
The goal of this investigation was to test whether BMD
and other factors related to bone status (menopausal status,
time since menopause, and use of HRT) are associated with
blood lead levels in ways consistent with the hypothesis that
postmenopausal bone density loss results in increased
endogenous lead exposure in the context of low exogenous
exposure to lead.
MATERIALS AND METHODS
Study population
The study population included adult females from
NHANES III. NHANES III comprised a cross-sectional
population sample that was obtained through a complex
survey design and was intended to be representative of the
US civilian, noninstitutionalized population. Over a 6-year
period (1988–1994), participants took part in a household
survey interview and an in-depth physical examination with
laboratory tests. Other details on the survey design have been
published by the National Center for Health Statistics (66).
The present investigation focused on the 2,575 women
aged 40–59 years who took part in the NHANES III survey
interview. From this group, women were excluded for the
following reasons: 212 did not undergo physical examination or blood testing; 77 did not have information on blood
lead levels; 181 had bone density measurements that were
considered unreliable by the technicians performing the
dual-energy x-ray absorptiometry; and 70 had an indeterminate menopausal status because of missing information. A
further 121 women of ethnicities other than non-Hispanic
Black, non-Hispanic White, and Mexican-American were
excluded because of low numbers in any single category of
self-reported ethnicity. The remaining 1,914 women constituted the sample used in the present analyses.
Definitions
Blood lead. Blood samples were obtained by venipuncture from all adult participants during the physical examination. Blood lead was measured by graphite furnace atomic
absorption spectrophotometry at the laboratories of the
Center for Environmental Health, Centers for Disease
Control and Prevention (Atlanta, Georgia). The assay detection limit was 1.0 µg/dl. Each sample analysis was
performed in duplicate, and the average of both measurements was used in these analyses. Persons with blood lead
levels less than 1.0 µg/dl (n = 201) were assigned a value of
0.5 µg/dl for consistency with previous analyses of
NHANES III data by other investigators (2).
Covariates. Information on race/ethnicity (non-Hispanic
Black, non-Hispanic White, or Mexican-American), age
(years), smoking (current, former, or never), household
income, and education was obtained from the household
interview. The poverty income ratio, a ratio of household
income to the federal poverty-level income (for a given
family size, adjusted to the poverty threshold for the year of
the interview), was used to create a three-level household
Am J Epidemiol 2004;160:901–911
BMD-related Predictors of Blood Lead Level in Women 903
TABLE 1. Characteristics of women aged 40–59 years participating in the Third National Health and
Nutrition Examination Survey, 1988–1994
Menopausal status
Total
(n = 1,914)
Mean age (years)
Premenopausal
(n = 1,111)
Surgical
menopause
(n = 139)
Natural
menopause
(n = 664)
p value*
<0.0001
48.1 (0.22)†
44.9 (0.21)
48.2 (0.65)
54.2 (0.27)
40–44 years
34.8
51.3
31.2
4.2
45–49 years
24.8
34.1
30.5
6.1
50–54 years
20.9
13.1
21.6
35.7
55–59 years
19.5
1.6
16.7
54.0
Non-Hispanic White
84.6
85.1
78.8
84.9
Non-Hispanic Black
11.2
10.6
17.9
11.1
Mexican-American
4.2
4.4
3.3
3.9
Urban
44.2
44.5
37.5
45.0
Rural
55.8
55.5
62.5
55.0
At or below poverty line
5.7
5.3
9.6
5.9
Above poverty line
85.7
87.6
79.9
83.3
Unknown
8.5
7.1
10.6
10.8
No completion of high school
18.1
14.4
26.2
23.4
Completion of high school
40.5
37.2
47.5
45.3
Some college
19.8
23.0
16.9
14.3
Completion of college or more
21.6
25.4
9.4
17.0
None
56.1
51.6
71.8
61.5
<1 drink/week
15.0
16.4
8.9
13.5
1–<3 drinks/week
17.7
20.9
14.9
12.2
≥3 drinks/week
11.3
11.2
4.4
12.8
Current smoker
25.0
23.1
37.3
26.3
Former smoker
25.2
23.9
21.6
28.4
Never smoker
49.8
53.0
41.1
45.4
Age group (%)
0.0001
Race/ethnicity (%)
0.0136
Residence (%)
0.5663
Household income level (%)
0.2123
Educational level (%)
<0.0001
Alcohol use (%)
0.0002
Smoking history (%)
0.0702
* p values for percentage data (categorical variables) were obtained from χ2 tests; p values for age and body
mass index were obtained using analysis of variance.
† Numbers in parentheses, standard error.
income variable. A participant was assigned a household
income level above the poverty line if the poverty income
ratio was greater than 1 and a household income level at or
below the poverty line if the poverty income ratio was less
than or equal to 1; data were considered missing if the survey
participant did not report household income. A four-level
education variable was created on the basis of the number of
years of education reported by the survey participant: less
than high school (<12 years), completion of high school (12
years), some college (>12–<16 years), and completion of
Am J Epidemiol 2004;160:901–911
college or more (≥16 years). A dichotomous variable based
on area of residence (urban/rural) was also included.
Information on participants’ alcohol use (frequency and
amount consumed per week) was obtained using the examination-associated interview questionnaire. A four-level categorical variable for weekly alcohol intake was created, with
the following levels: none; <1 drink per week; 1–<3 drinks
per week; and ≥3 drinks per week.
Menopausal status. A variable was created to categorize
women as premenopausal (ovarian function intact), surgically menopausal (both ovaries removed surgically before
904 Nash et al.
TABLE 2. Distributions of values for bone density-related variables among women aged 40–59 years, Third
National Health and Nutrition Examination Survey, 1988–1994
Menopausal status
Total
(n = 1,914)
Premenopausal
(n = 1,111)
Surgical
menopause
(n = 139)
Natural
menopause
(n = 664)
p value*
Mean bone mineral
density† (g/cm3)
Femoral neck
0.83 (0.004)‡
0.81 (0.005)
0.79 (0.017)
0.74 (0.008)
<0.0001
Trochanter
0.72 (0.003)
0.69 (0.004)
0.67 (0.013)
0.63 (0.007)
<0.0001
Intertrochanter
1.14 (0.005)
1.09 (0.007)
1.08 (0.022)
1.02 (0.01)
<0.0001
Ward’s triangle
0.68 (0.006)
0.66 (0.007)
0.62 (0.017)
0.55 (0.009)
<0.0001
Total
0.96 (0.004)
0.92 (0.006)
0.91 (0.018)
0.86 (0.008)
<0.0001
Current use
15.6
9.2
25.7
25.8
<0.0001
Past use
10.2
5.7
17.2
17.6
Never use
74.2
85.1
57.1
56.7
Current users
6.0 (0.44)
4.4 (0.64)
6.9 (1.8)
6.9 (0.59)
0.9973
Past users
2.9 (0.45)
3.8 (0.88)
1.3 (0.54)
2.6 (0.58)
0.1126
12.7 (0.67)
9.5 (0.41)
<0.0001
Use of HRT§ (%)
Mean duration of HRT
use (years)
Mean time (years) since
menopause¶
10.0 (0.36)
* p values for percentage data (categorical variables) were obtained from χ2 tests; p values for age and body mass
index were obtained using analysis of variance.
† Measured by dual-energy x-ray absorptiometry.
‡ Numbers in parentheses, standard error.
§ HRT, hormone replacement therapy.
¶ Calculated among postmenopausal women only.
cessation of menses), or naturally menopausal (nonsurgical
loss of ovarian function). Participants without a history of
reproductive surgery were classified as premenopausal if
they reported having had a menstrual period during the
previous 12 months and postmenopausal if they did not,
consistent with World Health Organization criteria. Women
who had undergone a hysterectomy (without ovariectomy)
that coincided with the date of the last menstrual period were
assigned a menopausal classification on the basis of age (<51
years, premenopausal; ≥51 years, naturally menopausal).
Women without a history of hysterectomy or ovariectomy
who were current users of HRT were classified in the same
way. Women who had undergone bilateral ovariectomy that
coincided with the date of the last menstrual period were
classified as surgically menopausal. Women who had undergone hysterectomy or ovariectomy after the date of the last
menstrual period were classified as naturally menopausal.
HRT status. Women were classified as current users, past
users, or never users of HRT on the basis of self-reported
data from the examination questionnaire. Duration of HRT
use in years was also ascertained by self-report at the time of
the examination.
Time since menopause. Information on time since
menopause was estimated as the difference in years between
age at the time of the NHANES interview and self-reported
age at the time of the last menstrual period or ovariectomy,
whichever came first. Women aged 51 years or more whose
last menstrual period coincided with hysterectomy were
assumed to be postmenopausal.
Bone mineral density. BMD was measured in five
regions of the femur by dual-energy x-ray absorptiometry in
the NHANES III medical examination: femoral neck, intertrochanter, trochanter, Ward’s triangle, and total. All five
measurements were examined in bivariate analyses of BMD
and blood lead level. For multivariate analyses, the BMD
value from the trochanter region was used.
Statistical methods
The associations of blood lead level with trochanter BMD,
menopause, time since menopause, and HRT were modeled
using multiple linear regression, controlling for other factors
related to blood lead level. Because the distribution of blood
lead values was nonnormal, as has been found by other
investigators using NHANES data and other data sets (2), we
used natural logarithmic transformation of the blood lead
variable in multiple linear regression analyses. Covariates
included age, race/ethnicity, smoking, education, household
income, alcohol use, and residence.
Am J Epidemiol 2004;160:901–911
BMD-related Predictors of Blood Lead Level in Women 905
TABLE 3. Geometric mean blood lead levels according to bone
density-related predictor variables among women aged 40–59
years, Third National Health and Nutrition Examination Survey,
1988–1994
Blood
No. of
lead level
participants
(µg/dl)
95%
CI*
p value
(Wald χ2
test)
<0.0001
Menopausal status
Premenopausal
1,122
1.9
1.7, 2.0
Surgically menopausal
139
2.7
2.4, 3.1
Naturally menopausal
653
2.9
2.5, 3.2
192
1.9
1.6, 2.2
Use of HRT*
Current use
Past use
Never use
255
3.0
2.6, 3.4
1,460
2.2
2.0, 2.3
<0.0001
Duration of HRT use (years)
0
1,460
2.2
2.0, 2.3
<1
131
3.0
2.5, 3.6
1–2
100
2.2
1.8, 2.6
3–4
51
1.7
1.3, 2.3
≥5
164
2.1
1.7, 2.5
1,122
1.9
1.7, 2.0
115
2.0
1.7, 2.4
0.004
Menopausal status, HRT use,
and time since
menopause
Premenopausal
Current HRT user
(postmenopausal)
<0.0001
aged 40–59 years varied significantly by race/ethnicity (p =
0.0136). As compared with White women, non-Hispanic
Blacks appeared to be overrepresented among surgically
menopausal women, whereas non-Hispanic Whites and
Mexican Americans were slightly underrepresented. The
mean age of the sample was 48.1 years, and, as expected,
menopausal status was strongly correlated with age. Postmenopausal women were more likely to abstain from weekly
alcohol use (by self-report) and less likely to have a college
education than premenopausal women.
Distributions of data on BMD and other bone densityrelated factors
Menopausal status was significantly associated with BMD
measurements from all five regions of the femur (p <
0.0001), with premenopausal women having a higher BMD
than surgically and naturally menopausal women (table 2).
Postmenopausal women were more likely than premenopausal women to be either current or past users of HRT (p <
0.0001). Current users of HRT had been undergoing HRT
for a longer amount of time than past users (a mean of 6
years vs. 2.9 years). Current users of HRT who were menopausal had the longest duration of HRT use (6.9 years).
Surgically menopausal women were, on average, further
away in time from menopause than naturally menopausal
women (12.7 years vs. 9.5 years; p < 0.0001).
Non-HRT user
(postmenopausal)
Blood lead levels for bone density-related predictor
variables
Time (years) since
menopause
0–2
68
3.7
3.1, 4.3
3–4
62
3.5
2.9, 4.2
5–7
60
3.5
2.7, 4.4
7–8
52
3.4
2.6, 4.6
9–10
228
2.7
1.8, 4.1
≥11
168
3.0
2.5, 3.5
* CI, confidence interval; HRT, hormone replacement therapy.
The statistical analyses were conducted using SAS (67),
incorporating the examination sampling weights and survey
design of NHANES III (66). Tests for trend for categorical
variables were carried out in regression models by coding
levels as integers (scores) and evaluating tests for the probability that the slope of the regression line was nonzero.
SUDAAN statistical software (68) was used to calculate
standard errors for the estimates, accounting for both the
weights and the complex survey design. All estimates
presented here were calculated incorporating the sampling
weights. Adjusted geometric mean blood lead levels were
obtained using the LSMEANS statement in SUDAAN.
RESULTS
Sample characteristics
A demographic description of the sample is provided in
table 1. The distribution of menopausal status among women
Am J Epidemiol 2004;160:901–911
Premenopausal women had a geometric mean blood lead
level of 1.9 µg/dl, as compared with 2.7 µg/dl and 2.9 µg/dl
for surgically menopausal women and naturally menopausal
women, respectively (table 3). Current users of HRT had a
lower geometric mean blood lead level than women who
reported never having used HRT (1.9 µg/dl vs. 2.2 µg/dl).
Past users of HRT had a geometric mean blood lead level of
3.0 µg/dl. Women who had never used HRT and women
who had used HRT for at least 1 year had lower mean blood
lead levels than women who had been using HRT for less
than 1 year. Relative to premenopausal women, geometric
mean blood lead levels peaked among women who were 0–2
years past menopause and were lowest among postmenopausal women who were 9 or more years beyond menopause. Blood lead levels were significantly related to
menopausal status and BMD. In bivariate analyses, BMD
measured in all five regions of the femur was significantly
inversely associated with blood lead level (table 4).
Age-stratified analyses
Compared with premenopausal women of the same age,
postmenopausal women had significantly higher geometric
mean blood lead levels in every age group (figure 1), with
the greatest differences being observed in the group aged
45–49 years. In addition, for every age category, current
users of HRT had lower geometric mean blood lead levels
than past or never users (figure 2).
906 Nash et al.
TABLE 4. Geometric mean blood lead levels (µg/dl) according to quartile of bone mineral density
among 1,914 women aged 40–59 years, Third National Health and Nutrition Examination Survey,
1988–1994
Site of bone mineral
density measurement
Quartile of bone mineral density*
1
2
3
4
p value
(Wald χ2 test)
Femoral neck
2.5 (2.2–5.2)†
2.2 (1.9–4.6)
2.0 (1.8–4.3)
2.0 (1.7–4.2)
0.0209
Intertrochanter
2.5 (2.3–5.3)
2.2 (1.9–4.7)
2.0 (1.8–4.3)
1.9 (1.7–4.2)
0.0049
Trochanter
2.5 (2.3–5.4)
2.2 (1.9–4.7)
2.0 (1.8–4.3)
1.9 (1.7–4.1)
0.0001
Ward’s triangle
2.6 (2.3–5.4)
2.2 (1.9–4.7)
2.0 (1.8–4.3)
2.0 (1.7–4.2)
0.0001
Total
2.5 (2.3–5.3)
2.1 (1.9–4.5)
2.1 (1.9–4.4)
1.9 (1.7–4.1)
0.0006
* The quartile cutpoints for bone mineral density differed according to site.
† Numbers in parentheses, interquartile range (25th percentile–75th percentile) for the geometric mean.
Multivariable analyses
BMD in the trochanter region was significantly inversely
associated with blood lead level (natural log), even after
adjustment for age, ethnicity, residence, household income,
smoking, and alcohol use (table 5, model with BMD only). A
one-unit change in BMD was associated with a 0.6-µg/dl
lower geometric mean blood lead level. Age was positively
associated with blood lead. Non-Hispanic Blacks had significantly higher blood lead levels than non-Hispanic Whites.
Persons living in urban or suburban areas had higher blood
lead levels than those living in more rural areas. Having a
household income at or below the poverty line was associated with a significantly higher blood lead level than having
an income above the poverty line. Education was not signif-
icantly associated with blood lead. Nonusers of alcohol had
significantly lower blood lead levels than frequent users.
Current and former smokers had significantly higher blood
lead levels than persons who had never smoked.
Menopausal status. When menopausal status was added
to the model, the association of BMD with blood lead (logtransformed) remained significant. Differences in blood lead
level between pre- and postmenopausal women also withstood multivariable adjustment for the established predictors
of blood lead level (table 5, model with BMD and menopausal status). Naturally menopausal women had a significantly higher geometric mean blood lead level than
premenopausal women (2.5 µg/dl vs. 2.0 µg/dl; p = 0.0126)
after adjustment for age, race/ethnicity, urban/rural residence, household income, educational level, alcohol use, and
FIGURE 1. Adjusted geometric mean blood lead levels by age and menopausal status, Third National Health and Nutrition Examination Survey,
1988–1994. (*Statistically significant difference in blood lead levels relative to premenopausal women.)
Am J Epidemiol 2004;160:901–911
BMD-related Predictors of Blood Lead Level in Women 907
FIGURE 2. Adjusted geometric mean blood lead levels by age and use of hormone replacement therapy, Third National Health and Nutrition
Examination Survey, 1988–1994. (*Statistically significant difference in blood lead levels relative to premenopausal women.)
smoking status. Surgically menopausal women also had a
significantly higher adjusted blood lead level than premenopausal women (2.6 µg/dl vs. 2.0 µg/dl; p = 0.0009).
HRT use. HRT use was significantly associated with log
blood lead level in a model that adjusted for BMD, menopausal status, and other covariates (table 5, model with
BMD, menopausal status, and HRT). The addition of HRT to
the model did not materially alter the association of BMD or
menopausal status with blood lead level. Among postmenopausal women, never users of HRT had a significantly higher
adjusted geometric mean blood lead level than women who
were currently using HRT (2.2 µg/dl vs. 1.8 µg/dl; p =
0.0006). Former users of HRT also had a significantly higher
adjusted blood lead level than current users (2.6 µg/dl vs. 1.8
µg/dl; p = 0.0004).
Time since menopause. Time (years) since menopause
was a significant predictor of blood lead level among women
who were not current users of HRT (table 5, model with time
since menopause). Postmenopausal women who were closer
in time to the last menstrual period tended to have significantly higher blood lead levels than both premenopausal
women and postmenopausal women who were further away
in time (≥9 years) from the menopausal transition.
DISCUSSION
These analyses suggest that menopause and bone status
are predictors of blood lead level among women aged 40–59
years in the US population. BMD was significantly inversely
associated with blood lead level after adjustment for other
factors traditionally associated with blood lead. Menopausal
status and HRT use, both of which are associated with BMD,
were important predictors of blood lead level independently
Am J Epidemiol 2004;160:901–911
of age and other factors. Despite recent declines in environmental lead exposure, the estimated effects of menopause
observed in this study are similar to those reported for
NHANES II (1976–1980) by Silbergeld et al. (3) and later by
other investigators (6, 63, 65). Among women not currently
on HRT, time since menopause, a potential marker of bone
density loss, was significantly associated with blood lead,
with postmenopausal women closer in time to the menopausal transition tending to have higher blood lead levels
than those further away from menopause. These associations
withstood multivariate adjustment for current BMD, race/
ethnicity, age, smoking, alcohol use, education, household
income, and urbanicity. In the absence of HRT, time since
the menopausal transition was associated with bone
demineralization; these observations are consistent with the
hypothesis that bone demineralization increases the blood
lead levels of menopausal women.
To our knowledge, this study is the first to have examined
the relation between BMD measured by dual-energy x-ray
absorptiometry and blood lead level. A statistically significant association between concurrent measurements of BMD
and blood lead level was observed after multivariate adjustment for menopausal status. A low BMD measurement may
reflect either recent decline in BMD or longstanding bone
mineral status. It is possible that lead may affect BMD,
rather than BMD’s affecting lead. Experimental studies in
rats demonstrate that the effects of lead in decreasing bone
growth and density are associated with perinatal exposures
(69, 70). Because of the cross-sectional design of NHANES
III, inferences about the temporal nature of any of the associations reported here cannot be made on the basis of these
data alone.
908 Nash et al.
TABLE 5. Regression coefficients from multiple linear regression analysis of log blood lead levels (µg/dl) on
established predictors of blood lead and menopausal status among 1,905 pre- and postmenopausal women, Third
National Health and Nutrition Examination Survey, 1988–1994
Variables included in model
BMD†
BMD and
menopausal
status
BMD, menopausal
status, and
HRT† use
BMD, menopausal
status, HRT use,
and time since
menopause‡
R2
0.23
0.24
0.25
0.26
Intercept
–0.6 (0.25)§
–0.17 (0.31)
–0.48 (0.35)
–0.33 (0.38)
Bone mineral density (trochanter
region) (g/cm3)
–0.52 (0.21)*
–0.47 (0.21)*
–0.46 (0.22)*
–0.47 (0.23)*
Surgically menopausal
0.24 (0.07)*
0.26 (0.07)*
Naturally menopausal
0.21 (0.08)*
0.22 (0.08)*
Menopausal status
Premenopausal¶
Use of HRT
Current use¶
Past use
0.37 (0.10)*
Never use
0.24 (0.06)*
Menopausal status, HRT use, and time
since menopause
Premenopausal¶
Current HRT user (postmenopausal)
–0.07 (0.09)
Non-HRT user (postmenopausal)
Time (years) since menopause
0–2
0.48 (0.09)*
3–4
0.33 (0.09)*
5–6
0.41 (0.14)*
7–8
0.43 (0.15)*
9–10
0.13 (0.19)*
≥11
0.15 (0.10)
Table continues
Women who were current users of HRT had adjusted blood
lead levels lower than those of past or never users, independently of age. This observation is consistent with the findings
of other investigators (64) and supports the hypothesis that
blood lead levels among postmenopausal women are partly
driven by factors that affect bone loss. However, the possibility cannot be ruled out that differing lead exposure profiles
between these groups of women were responsible for the
observed effect, despite our attempt to control for this through
multivariate adjustment for factors known to be associated
with lead exposure. Women who choose to use HRT may be
different from women who do not with regard to many health
behaviors and exposures, both known and unknown (71).
However, it is unlikely that this is the only explanation, since
no differences between former users and never users of HRT
were observed after multivariate adjustment.
Our results reinforce the possibility that blood lead stores
may represent an endogenous source of lead exposure.
Although the differences in average blood lead levels
reported here are relatively small, increases in blood lead by
these amounts may still pose a risk to women who experience significant bone loss or have higher bone lead levels at
the start of menopause. Moreover, small increases in blood
lead level are associated with decrements in neurocognitive
performance (72) and increased risks of atherosclerosis and
hypertension (57, 58, 73). Significant increases in risk of
peripheral arterial disease were reported in adults with
increases in blood lead within the range found in this study
(74).
The magnitude of postmenopausal changes in blood lead
level may depend on a number of factors, including the
amount and bioavailability of lead stored in bone prior to
menopause and the rate and magnitude of bone mineral loss.
Initial bone lead stores at menopause may differ among
different ethnic groups on the basis of prior exposures (2).
Moreover, there is evidence that BMD and rates of loss after
menopause may differ by ethnicity (75–77). Bone lead
Am J Epidemiol 2004;160:901–911
BMD-related Predictors of Blood Lead Level in Women 909
TABLE 5. Continued
Variables included in model
BMD and
menopausal
status
BMD, menopausal
status, and
HRT use
BMD, menopausal
status, HRT use,
and time since
menopause‡
0.03 (0.0)*
0.02 (0.01)*
0.02 (0.01)*
0.02 (0.01)*
Non-Hispanic Black
0.29 (0.06)*
0.27 (0.06)*
0.26 (0.06)*
0.26 (0.06)*
Mexican-American
0.08 (0.07)
0.09 (0.07)
0.09 (0.07)
0.08 (0.07)
0.12 (0.06)
0.11 (0.06)
0.11 (0.06)
0.11 (0.06)
Unknown
0.11 (0.09)
0.11 (0.09)
0.06 (0.09)
0.09 (0.10)
At or below poverty line
0.01 (0.07)
0.01 (0.07)
0.01 (0.07)
0.01 (0.07)
No completion of high school
0.19 (0.08)*
0.17 (0.08)*
0.19 (0.08)*
0.17 (0.08)*
Completion of high school
–0.01 (0.06)
–0.02 (0.06)
0.01 (0.06)
–0.01 (0.06)
Some college
0.03 (0.08)
0.03 (0.08)
0.05 (0.08)
0.02 (0.08)
<1 drink/week
0.11 (0.05)*
0.12 (0.05)*
0.14 (0.05)*
0.11 (0.05)*
1–<3 drinks/week
0.25 (0.06)*
0.27 (0.06)*
0.28 (0.06)*
0.27 (0.06)*
≥3 drinks/week
0.28 (0.07)*
0.29 (0.07)*
0.29 (0.06)*
0.27 (0.06)*
Current smoker
0.52 (0.05)*
0.5 (0.05)*
0.48 (0.05)*
0.51 (0.05)*
Former smoker
0.08 (0.06)
0.07 (0.06)
0.06 (0.06)
0.06 (0.06)
BMD
Age (years)
Race/ethnicity
Non-Hispanic White¶
Residence
Urban
Rural¶
Household income level
Above poverty line¶
Educational level
Completion of college or more¶
Alcohol use (no. of drinks/week)
None¶
Smoking status
Never smoker¶
* p ≤ 0.05.
† BMD, bone mineral density; HRT, hormone replacement therapy.
‡ Current users of HRT were excluded.
§ Numbers in parentheses, standard error.
¶ Reference category.
measurements taken over time may provide additional
insights into these issues.
There is some evidence that lead mobilized from bone
stores may be preferentially partitioned into plasma as
compared with erythrocytes; this suggests that lead released
from bone may be more toxicologically relevant than lead
entering the bloodstream from environmental sources (78),
though there is also evidence to suggest that this is not the
case (79). If this is the case, increases in the release of endogenous lead may result in transfers of more lead to critical
compartments in soft tissue (10, 78). Although this theory
has not been proven directly, this observation may explain
recent findings that bone lead is more predictive of outcomes
such as hypertension in older persons than is blood lead (57,
80), as well as observations that the relation between blood
Am J Epidemiol 2004;160:901–911
lead and diastolic hypertension appears to be stronger among
postmenopausal women (58). We recommend that investigators studying health outcomes associated with lead exposure
in women during the menopausal transition measure bone
lead levels in addition to blood lead levels.
ACKNOWLEDGMENTS
This research was supported by the Centers for Disease
Control and Prevention through a cooperative agreement
(no. TS 288-14/14) with the Association of Teachers of
Preventive Medicine.
The authors acknowledge the important contributions of
Dr. Paul Stolley (University of Maryland School of Medi-
910 Nash et al.
cine) and Debra Brody (National Center for Health Statistics), both of whom provided guidance and invaluable
comments on this project and this paper.
22.
23.
REFERENCES
1. Pirkle JL, Brody DJ, Gunter EW, et al. The decline in blood
lead levels in the United States: the National Health and Nutrition Examination Surveys (NHANES). JAMA 1994;272:284–
91.
2. Brody DJ, Pirkle JL, Kramer RA, et al. Blood lead levels in the
US population. Phase 1 of the Third National Health and Nutrition Examination Survey (NHANES III, 1988 to 1991). JAMA
1994;272:277–83.
3. Silbergeld EK, Schwartz J, Mahaffey K. Lead and osteoporosis: mobilization of lead from bone in postmenopausal women.
Environ Res 1988;47:79–94.
4. Silbergeld EK. Lead in bone: implications for toxicology during pregnancy and lactation. Environ Health Perspect 1991;91:
63–70.
5. Silbergeld E, Watson L. Exposure to lead during reproduction
and menopause. Fund Appl Toxicol 1995;25:167–8.
6. Symanski E, Hertz PI. Blood lead levels in relation to menopause, smoking, and pregnancy history. Am J Epidemiol 1995;
141:1047–58.
7. Rabinowitz MB, Wetherill GW, Kopple JD. Kinetic analysis of
lead metabolism in healthy humans. J Clin Invest 1976;58:260–
70.
8. Manton WI. Total contribution of airborne lead to blood lead.
Br J Ind Med 1985;42:168–72.
9. Borjesson J, Gerhardsson L, Schutz A, et al. In vivo measurements of lead in fingerbone in active and retired lead smelters.
Int Arch Occup Environ Health 1997;69:97–105.
10. Hu H, Rabinowitz M, Smith D. Bone lead as a biological
marker in epidemiologic studies of chronic toxicity: conceptual
paradigms. Environ Health Perspect 1998;106:1–8.
11. Schroeder HA, Tipton IH. The human body burden of lead.
Arch Environ Health 1968;17:965–78.
12. Barry PS, Mossman DB. Lead concentrations in human tissues.
Br J Ind Med 1970;27:339–51.
13. Saltzman BE, Gross SB, Yeager DW, et al. Total body burdens
and tissue concentrations of lead, cadmium, copper, zinc, and
ash in 55 human cadavers. Environ Res 1990;52:126–45.
14. Kosnett MJ, Becker CE, Osterloh JD, et al. Factors influencing
bone lead concentration in a suburban community assessed by
noninvasive K x-ray fluorescence. JAMA 1994;271:197–203.
15. Silbergeld E, Nash D. Lead and human health: is this mine
exhausted? Prog Environ Sci 2000;2:53–68.
16. Robinson CJ, Hall J, Beshir SO. Hormonal modulation of mineral metabolism in reproduction. Proc Nutr Soc 1983;42:169–
80.
17. O’Flaherty EJ. Physiologically based models for bone-seeking
elements. I. Rat skeletal and bone growth. Toxicol Appl Pharmacol 1991;111:299–312.
18. O’Flaherty EJ. Physiologically based models for bone-seeking
elements. II. Kinetics of lead disposition in rats. Toxicol Appl
Pharmacol 1991;111:313–31.
19. O’Flaherty EJ. Physiologically based models for bone-seeking
elements. III. Human skeletal and bone growth. Toxicol Appl
Pharmacol 1991;111:332–41.
20. Garel JM. Hormonal control of calcium metabolism during the
reproductive cycle in mammals. Physiol Rev 1987;67:1–66.
21. Rosen JF, Pounds JG. The cellular metabolism of lead and cal-
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
cium: a kinetic analysis in cultured osteoclastic bone cells.
Contrib Nephrol 1988;64:74–82.
Pounds JG, Long GJ, Rosen JF. Cellular and molecular toxicity
of lead in bone. Environ Health Perspect 1991;91:17–32.
Rosen JF, Wexler EE. Studies of lead transport in bone organ
culture. Biochem Pharmacol 1977;26:650–2.
Pounds JG. Effect of lead intoxication on calcium homeostasis
and calcium-mediated cell function: a review. Neurotoxicology
1984;5:295–331.
Pounds JG, Rosen JF. Cellular metabolism of lead: a kinetic
analysis in cultured osteoclastic bone cells. Toxicol Appl Pharmacol 1986;83:531–45.
Bronner F. Bone and calcium homeostasis. Neurotoxicology
1992;13:775–82.
O’Flaherty EJ. Modeling bone mineral metabolism, with special reference to calcium and lead. Neurotoxicology 1992;13:
789–97.
Simons TJ. Lead-calcium interactions in cellular lead toxicity.
Neurotoxicology 1993;14:77–85.
Zuscik MJ, Pateder DB, Puzas JE, et al. Lead alters parathyroid
hormone-related peptide and transforming growth factor-β1
effects and AP-1 and NF-κB signaling in chondrocytes. J
Orthop Res 2002;20:811–18.
Berglund M, Akesson A, Bjellerup P, et al. Metal-bone interactions. Toxicol Lett 2000;112-113:219–25.
Wittmers LJ, Aufderheide AC, Wallgren J, et al. Lead in bone.
IV. Distribution of lead in the human skeleton. Arch Environ
Health 1988;43:381–91.
Inskip MJ, Franklin CA, Subramanian KS, et al. Sampling of
cortical and trabecular bone for lead analysis: method development in a study of lead mobilization during pregnancy.
Neurotoxicology 1992;13:825–34.
Gulson BL, Mizon KJ, Korsch MJ, et al. Mobilization of lead
from human bone tissue during pregnancy and lactation—a
summary of long-term research. Sci Total Environ 2003;303:
79–104.
Vahter M, Berglund M, Akesson A, et al. Metals and women’s
health. Environ Res 2002;88:145–55.
Lindquist O, Bengtsson C, Hansson T, et al. Bone mineral content in relation to age and menopause in middle-aged women: a
study of bone density in lumbar vertebrae by dual photon
absorptiometry in a population sample of women. Scand J Clin
Lab Invest 1981;41:215–23.
Krolner B, Pors NS. Bone mineral content of the lumbar spine
in normal and osteoporotic women: cross-sectional and longitudinal studies. Clin Sci 1982;62:329–36.
Heaney RP, Recker RR, Saville PD. Menopausal changes in
bone remodeling. J Lab Clin Med 1978;92:964–70.
Lindsay R, Hart DM, Aitken JM, et al. Long-term prevention of
postmenopausal osteoporosis by oestrogen: evidence for an
increased bone mass after delayed onset of oestrogen treatment.
Lancet 1976;1:1038–41.
Riggs BL, Melton LD. Involutional osteoporosis. N Engl J Med
1986;314:1676–86.
Nordin B. Osteoporosis with particular reference to the menopause. In: Alvioli L, ed. The osteoporotic syndrome: detection,
prevention, and treatment. New York, NY: Grune and Stratton,
1983.
Christiansen C, Christensen MS, McNair P, et al. Prevention of
early postmenopausal bone loss: controlled 2-year study in 315
normal females. Eur J Clin Invest 1980;10:273–9.
Gotfredsen A, Nilas L, Riis BJ, et al. Bone changes occurring
spontaneously and caused by estrogen in early postmenopausal
women: a local or generalised phenomenon? BMJ (Clin Res
Ed) 1986;292:1098–100.
Ballard JE, McKeown BC, Graham HM, et al. The effect of
Am J Epidemiol 2004;160:901–911
BMD-related Predictors of Blood Lead Level in Women 911
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
high level physical activity (8.5 METs or greater) and estrogen
replacement therapy upon bone mass in postmenopausal
females, aged 50–68 years. Int J Sports Med 1990;11:208–14.
Riggs BL, Melton LD. The prevention and treatment of
osteoporosis. N Engl J Med 1992;327:620–7.
Barrett-Connor E, Wehren LE, Siris ES, et al. Recency and
duration of postmenopausal hormone therapy: effects on bone
mineral density and fracture risk in the National Osteoporosis
Risk Assessment (NORA) Study. Menopause 2003;10:412–19.
Bagger YZ, Tanko LB, Alexandersen P, et al. Two to three
years of hormone replacement treatment in healthy women
have long-term preventive effects on bone mass and
osteoporotic fractures: The PERF Study. Bone 2004;34:728–
35.
The Writing Group for the PEPI. Effects of hormone therapy on
bone mineral density: results from the Postmenopausal Estrogen/Progestin Interventions (PEPI) Trial. JAMA 1996;276:
1389–96.
Lagerkvist BJ, Soderberg HA, Nordberg GF, et al. Biological
monitoring of arsenic, lead and cadmium in occupationally and
environmentally exposed pregnant women. Scand J Work
Environ Health 1993;1:50–3.
Knight EM, Spurlock BG, Edwards CH, et al. Biochemical profile of African American women during three trimesters of
pregnancy and at delivery. J Nutr 1994;124(suppl):943S–53S.
West WL, Knight EM, Edwards CH, et al. Maternal low level
lead and pregnancy outcomes. J Nutr 1994;124(suppl):981S–
6S.
Gulson BL, Jameson CW, Mahaffey KR, et al. Pregnancy
increases mobilization of lead from maternal skeleton. J Lab
Clin Med 1997;130:51–62.
Webber CE, Chettle DR, Bowins RJ, et al. Hormone replacement therapy may reduce the return of endogenous lead from
bone to the circulation. Environ Health Perspect 1995;103:
1150–3.
Grandjean P, Nielsen GD, Jorgensen PJ, et al. Reference intervals for trace elements in blood: significance of risk factors.
Scand J Clin Lab Invest 1992;52:321–37.
Muldoon SB, Salamone LM, Cauley JA, et al. Effect of bone
mineral density changes on blood lead levels in peri-menopausal women. (Abstract). Am J Epidemiol 1997;145(suppl):
S21.
Sharp DS, Becker CE, Smith AH. Chronic low-level lead exposure: its role in the pathogenesis of hypertension. Med Toxicol
1987;2:210–32.
Hertz-Picciotto I, Croft J. Review of the relation between blood
lead and blood pressure. Epidemiol Rev 1993;15:352–73.
Hu H, Aro A, Payton M, et al. The relationship of bone and
blood lead to hypertension: The Normative Aging Study.
JAMA 1996;275:1171–6.
Nash D, Magder L, Lustberg M, et al. Blood lead, blood pressure, and hypertension in perimenopausal and postmenopausal
women. JAMA 2003;289:1523–32.
Kim R, Rotnitsky A, Sparrow D, et al. A longitudinal study of
low-level lead exposure and impairment of renal function: The
Normative Aging Study. JAMA 1996;275:1177–81.
Muldoon SB, Cauley JA, Kuller LH, et al. Effects of blood lead
levels on cognitive function of older women. Neuroepidemiology
1996;15:62–72.
Am J Epidemiol 2004;160:901–911
61. Payton M, Riggs KM, Spiro A 3rd, et al. Relations of bone and
blood lead to cognitive function: The VA Normative Aging
Study. Neurotoxicol Teratol 1998;20:19–27.
62. Lustberg M, Silbergeld E. Blood lead levels and mortality.
Arch Intern Med 2002;162:2443–9.
63. Muldoon SB, Cauley JA, Kuller LH, et al. Lifestyle and sociodemographic factors as determinants of blood lead levels in elderly women. Am J Epidemiol 1994;139:599–608.
64. Korrick SA, Schwartz J, Tsaih SW, et al. Correlates of bone and
blood lead levels among middle-aged and elderly women. Am
J Epidemiol 2002;156:335–43.
65. Garrido Latorre F, Hernandez-Avila M, Tamayo Orozco J, et
al. Relationship of blood and bone lead to menopause and bone
mineral density among middle-age women in Mexico City.
Environ Health Perspect 2003;111:631–6.
66. Ezzati T, Massey J, Waksburg J, et al. Sample design: Third
National Health and Nutrition Evaluation Survey. Hyattsville,
MD: National Center for Health Statistics, 1992.
67. SAS Institute, Inc. SAS/STAT user’s guide, version 6. Cary,
NC: SAS Institute, Inc, 1990.
68. Shah BV, Barnwell BG, Bieler GS. SUDAAN user’s manual,
release 7.0. Research Triangle Park, NC: Research Triangle
Institute, 1996.
69. Hamilton JD, O’Flaherty EJ. Influence of lead on mineralization during bone growth. Fundam Appl Toxicol 1995;26:265–
71.
70. Ronis MJ, Aronson J, Gao GG, et al. Skeletal effects of developmental lead exposure in rats. Toxicol Sci 2001;62:321–9.
71. Barrett-Connor E. Postmenopausal estrogen and prevention
bias. Ann Intern Med 1991;115:455–6.
72. Needleman HL, Leviton A, Bellinger D. Lead-associated intellectual deficit. (Letter). N Engl J Med 1982;306:367.
73. Harlan WR. The relationship of blood lead levels to blood pressure in the U.S. population. Environ Health Perspect 1988;78:
9–13.
74. Navas-Acien A, Selvin E, Sharrett AR, et al. Lead, cadmium,
smoking, and increased risk of peripheral arterial disease.
Circulation 2004;109:3196–201.
75. Finkelstein JS, Lee ML, Sowers M, et al. Ethnic variation in
bone density in premenopausal and early perimenopausal
women: effects of anthropometric and lifestyle factors. J Clin
Endocrinol Metab 2002;87:3057–67.
76. Finkelstein JS, Sowers M, Greendale GA, et al. Ethnic variation
in bone turnover in pre- and early perimenopausal women:
effects of anthropometric and lifestyle factors. J Clin Endocrinol Metab 2002;87:3051–6.
77. Melton LJ 3rd, Achenbach SJ, O’Fallon WM, et al. Secondary
osteoporosis and the risk of distal forearm fractures in men and
women. Bone 2002;31:119–25.
78. Cake KM, Bowins RJ, Vaillancourt C, et al. Partition of circulating lead between serum and red cells is different for internal
and external sources of lead. Am J Ind Med 1996;29:440–5.
79. Gulson BL, Mizon KJ, Palmer JM, et al. Urinary lead isotopes
during pregnancy and postpartum indicate no preferential partitioning of endogenous lead into plasma. J Lab Clin Med 2000;
136:236–42.
80. Korrick SA, Hunter DJ, Rotnitzky A, et al. Lead and hypertension in a sample of middle-aged women. Am J Public Health
1999;89:330–5.