0021-972X/97/$03.00/0
Journal of Clinical Endocrinology and Metabolism
Copyright © 1997 by The Endocrine Society
Vol. 82, No. 2
Printed in U.S.A.
Racial Differences in Bone Density between Young Adult
Black and White Subjects Persist after Adjustment for
Anthropometric, Lifestyle, and Biochemical
Differences*†
BRUCE ETTINGER, STEPHEN SIDNEY, STEVEN R. CUMMINGS,
CESAR LIBANATI, DANIEL D. BIKLE, IRENE S. TEKAWA, KIMBERLY TOLAN,
AND PETER STEIGER
Division of Research (B.E., S.S., I.S.T., K.T.), Kaiser Permanente Medical Care Program, Oakland
94611; Department of Medicine (S.R.C.), University of California, San Francisco 94143; Division of
Endocrinology (C.L.), Department of Medicine, Veterans Administration Medical Center, Loma Linda
92357; Mineral Metabolism Unit (D.D.B.), Veterans Administration Medical Center, San Francisco,
California 94121; Hologic, Inc. (P.S.), Waltham, Massachusetts 02154
ABSTRACT
This study tested whether racial differences in bone density can be
explained by differences in bone metabolism and lifestyle. A cohort of
402 black and white men and women, ages 25–36 yr, was studied at
the Kaiser Permanente Medical Care Program in Northern California, a prepaid health plan. Body composition (fat, lean, and bone
mineral density) was measured using a Hologic-2000 dual-energy
x-ray densitometer. Muscle strength, blood and urine chemistry values related to calcium metabolism, bone turnover, growth factors, and
level of sex and adrenal hormones were also measured. Medical history, physical activity, and lifestyle were assessed. Statistical anal-
yses using t- and chi-square tests and multiple regression were done
to determine whether racial difference in bone density remained after
adjustment for covariates. Bone density at all skeletal sites was statistically significantly greater in black than in white subjects; on
average, adjustment for covariates reduced the percentage density
differences by 42% for men and 34% for women. Adjusted bone density
at various skeletal sites was 4.5–16.1% higher for black than for white
men and was 1.2–7.3% higher for black than for white women. We
concluded that racial differences in bone mineral density are not
accounted for by clinical or biochemical variables measured in early
adulthood. (J Clin Endocrinol Metab 82: 429 – 434, 1997)
T
HE LOWER lifetime risk of hip fracture among black
persons (1) results at least partly because American
black women (2–5) and men (5, 6) achieve 5%-15% greater
peak bone mass than white persons.
The greater bone density among black persons may be
caused by their higher obesity rate (7), greater frame size (7),
and greater muscle mass (8). However, Luckey and coworkers (3) found that premenopausal black women had statistically significantly greater spinal and radial bone density
than white women, even after adjusting for height, weight,
and body mass index.
Many have hypothesized that genetically determined differences in skeletal metabolism account for racial differences
in bone density; this hypothesis is supported by studies that
show racial differences in calciotropic hormones (9, 10) and
bone turnover markers (11). Compared with white persons,
black persons have lower urinary calcium excretion, higher
1,25-dihydroxyvitamin D (1, 25D) level, and lower 25-hydroxyvitamin D (25D) and osteocalcin level (9). Moreover,
bone biopsies in black persons have shown lower bone turnover (12).
Peak bone density can be influenced by lifestyle factors
such as dietary calcium intake (13), physical activity (13),
smoking, (14) and alcohol intake (14, 15). Menstrual and
reproductive factors, including age of menarche (16), pregnancy early in life (15), breast-feeding (17), and oral contraceptive use (13), have also been found by some to influence
peak bone density.
Gonadal steroids may account for gender differences (18)
and for racial differences in bone density. Regardless of
cause, it is generally agreed that lower estrogen level results
in lower peak bone mass. Positive associations between bone
density and serum estrogen (19) and androgen (20) level
have been reported in young women. A positive association
between androgen level and bone density exists in young
men also (21). Black persons may differ from white persons
in sex hormone level. Two studies have demonstrated statistically significantly higher serum testosterone level in
young adult black men (22) and women (23).
Our study tested whether racial differences in bone density could be explained by differences in bone metabolism
and lifestyle (24). Accordingly, we measured most of the
clinical and biochemical variables believed to be related to
Received May 2, 1996. Revision received September 24, 1996.
Accepted September 30, 1996.
Address correspondence and reprint requests to Bruce Ettinger, MD,
Division of Research, Kaiser Permanente Medical Care Program, 3505
Broadway, Oakland, California 94611-5714.
* This study was supported by the National Institute of Health &
Human Services through Control N01-HC-48050 from the National
Heart, Lung, and Blood Institute, National Institute of Health, and
through Grant 5 R01-AR40430-03 from the National Institute of Arthritis
and Musculoskeletal and Skin Diseases.
† The Medical Editing Department, Kaiser Foundation Research Institute, provided editorial assistance.
429
430
JCE & M • 1997
Vol 82 • No 2
ETTINGER ET AL.
skeletal health in young adult black and white men and
women.
Materials and Methods
is the L-3 bone mineral content divided by L-3 estimated volume (26).
Lean body mass, fat mass, and ratio of trunk to leg fat were also measured using DXA in the enhanced total body array scanning mode. In
vivo precision for fat mass based on repeated scans in 18 volunteers was
5.9%.
Subjects
Eight hundred fifty volunteer subjects took part in this study, all age
25–36 yr and all simultaneously participants in the Cardiovascular Risk
Development in Young Adults (CARDIA) study, a study initiated by the
National Heart, Lung, and Blood Institute (NHLBI) in 1983 to determine
risk factors for coronary heart disease among black and white men and
women and to identify lifestyle characteristics associated with these risk
factors (25). The Kaiser Permanente Medical Care Program (KPMCP), a
prepaid health plan that serves about one quarter of the local population
in Oakland, California, was one of four sites selected for recruitment of
participants in the CARDIA study. The study protocol was approved by
the KPMCP Northern California Region Institutional Review Board.
After exclusions for certain laboratory abnormalities and pregnancyrelated criteria, 799 subjects were used in the analysis: 197 black men,
263 black women, 157 white men, and 182 white women. Because of
financial constraints, blood and urine chemistry studies, planned for
only the first 100 in each sex-race subject group, were obtained for 402
of the study subjects (the core group): 109 black men, 95 black women,
114 white men, and 84 white women. We excluded women who were
currently breast-feeding or who, within the previous year, had been
pregnant or had less than 10 spontaneous menstrual cycles. We also
excluded women from the core group if they reported current usage of
oral contraceptive agents. We also excluded those with renal failure
(serum creatinine ⬎1.5 mg/dL), hypocalcemia (⬍8.5 mg/dL), or hypercalcemia (⬎10.3 mg/dL), a total of 4 subjects.
Muscle strength
Average isokinetic muscle strength of the quadriceps and hamstring
muscles was measured with a Cybex II dynamometer (Lumex,
Ronkonkoma, NY). The best of five repetitions was used for both flexion
and extension.
Biochemical tests
The following markers of bone and mineral metabolism were obtained from blood drawn when the subject was fasting: creatinine, calcium, phosphorus, bone-specific alkaline phosphatase, intact parathyroid hormone, 25D, 1,25D, and osteocalcin (Bikle D.D., unpublished
material). The following growth factors were measured by immunoassay: insulin, insulin-like growth factor-I (IGF-I) (27), and IGF-binding
protein 3 (IGFBP3) (28). Sex and adrenal hormone immunoassays were
done at Endocrine Sciences Laboratory, Calabasas Hills, California;
these included assays for testosterone (total, free, and bioavailable),
sex-hormone-binding globulin, estradiol, and dehydroepiandrosterone
sulfate. Estradiol was measured in women only, and this measurement
was obtained between day 3 and day 11 of the menstrual cycle. The
estradiol index was calculated by dividing the estradiol by the sexhormone-binding globulin. Urine (collected during 24 h) was obtained
for calcium, creatinine, and free total pyridinoline measurement. Free
total pyridinoline cross-links were measured by enzyme-linked immunoassay (29) (Metra Biosystems Inc, Palo Alto, CA).
Bone and body composition measurement
Dual energy x-ray absorptiometry (DXA) of total body, hip, and
lumbar spine was done by using a Hologic 2000 densitometer (Hologic,
Inc, Waltham, MA) in the array scanning mode. In vivo precision for bone
mineral density (BMD), based on repeated scans of 20 volunteers done
1– 6 weeks apart and expressed as a coefficient of variation, was .9% for
total body BMD, 1.4% for posteroanterior spine BMD, and 2.2% for
femoral neck BMD. Hologic software calculates the estimated volume
for L-3 vertebra from the product of the projected area of the lateral scan
and the vertebral width. The latter is the quotient of the projected area
of the PA scan divided by the vertebral height. L-3 volumetric density
Medical history and habits
Usual diet was assessed by a diet history interview in which food
models and measuring cups and spoons were used to estimate portion
size (30). Daily nutrient intake, including intake of calcium and vitamin
D, was estimated by translating the precoded CARDIA diet history from
the database. Calcium and vitamin D intake estimates included
supplemental sources.
Physical activity was assessed by an interviewer-administered questionnaire, which asked about level of participation during the last year
in 13 specific activities (or groups of activities that have similar inten-
TABLE 1. Characteristics of the core group of 402 subjects age 25–36 yr, by gender and race
Men
Black
Subjects
Age, yr
Height, m
Elbow width, mm
Weight, kg
BMI, kg/m2
Education, yr
Physical activity score:
total
heavy only
Dietary intake/day:
vitamin D, U
calcium, mg
caffeine, mg
kilocalories
Sun exposure, h/wk
109
Women
White
Black
White
30.7 (3.2)
1.78 (0.07)
71.1 (3.9)
83.6 (14.6)
26.3 (4.0)
14.2 (2.1)a
114
95
mean (standard deviation)
31.3 (3.2)
31.0 (3.1)
1.78 (0.07)
1.64 (0.06)a
70.3 (3.7)
62.5 (4.5)a
81.2 (14.1)
75.8 (18.2)a
25.5 (4.1)
28.3 (6.5)a
15.4 (2.6)
14.0 (2.1)a
84
31.8 (3.1)
1.66 (0.06)
61.1 (3.0)
67.9 (15.7)
24.6 (6.0)
15.7 (2.4)
554 (382)a
383 (300)a
425 (271)
256 (214)
278 (229)a
163 (163)
364 (243)
207 (176)
413 (310)
1538 (1090)
95 (139)a
4552 (2950)a
29.0 (15.9)a
345 (290)
1416 (894)
189 (194)
3263 (1194)
17.0 (12.9)
254 (217)
829 (454)a
110 (181)a
2430 (1355)
23.8 (15.5)a
314 (216)
1247 (716)
202 (214)
2413 (855)
14.7 (8.4)
27.1a
10.5
25.3
15.5
13.8
12.3
4.2
4.8
percentage
Smoking-current
Alcohol equivalent
⬎2 drinks/day
BMI, body mass index.
a
Statistically significant racial difference within same gender, P ⬍ 0.05.
RACIAL DIFFERENCES IN BONE DENSITY
sity). Because type and intensity of different activities were documented
separately, we were able to separately analyze the effect of heavyweight-bearing activity. We computed physical activity scores based on
the sum of time spent in each activity weighted by an estimate of
kilocalories expended for each activity (31). The reliability of this instrument was studied by comparing questionnaire results with results
repeated 2 weeks later (n ⫽ 129). The test-retest correlation of the total
score was .84 (32).
Medical histories and histories of tobacco use and alcohol use
were obtained by a self-administered and interview-administered
questionnaire.
Statistics
The purpose of these analyses was to determine how much racial
differences in bone density would remain after adjustment for covariates. The analysis was done separately for men and women in 3 phases.
First, by using subjects from the entire cohort (n ⫽ 799), variables that
showed possible racial differences (P ⬍ .1) using t- and chi-square tests
were identified as possible predictor variables. These variables were
assigned to 1 of 10 categories: 6 clinical categories: body size, muscle, fat,
physical activity, lifestyle, and reproductive history (women only); and
4 biochemical categories: calcium metabolism, bone turnover, growth
factors, and sex and adrenal hormones.
All later analytic phases were completed for subjects in the core cohort
(n ⫽ 402) who had data on both clinical and biochemical variables. Next,
forward, stepwise regression models were used to identify predictor
variables which showed a relation (P ⬍ .1) with bone density, when
considered simultaneously. Relations were examined for posteroanterior spine, lateral L3 spine, volumetric L3 spine, total femur, femoral
neck, femoral trochanter, Ward’s triangle, and total body bone density.
Regression analyses were first done within each of the 10 categories.
Variables that were significant within a category were then tested simultaneously in a regression model that included variables from all
categories. Because variables in the fat category were highly correlated,
we separated them into 2 groups, those that measured overall fat (total
fat by DXA, weight, BMI, and sum of skinfolds) and those that measured
fat distribution (ratio of trunk:leg fat by DXA and ratio of waist:hip
circumference). BMI, as a measure of overall fat, and waist:hip ratio, as
a measure of fat distribution, were the most consistent and strong predictors of BMD. In the third phase, variables that still had a statistically
significant relation (P ⬍ .05) with bone density were used in a final
regression model that included race as a covariate. The difference in
adjusted means by race in the final model is a measure of remaining
racial differences in BMD. As a measure of relative magnitude of difference in bone density, we calculated percentage difference for each
skeletal site; this was defined as the mean for black persons minus mean
for white persons divided by mean for white persons. Absolute differences in BMD with 95% confidence intervals were calculated on the basis
of unadjusted means and adjusted means from the regression models.
Results
Demographic, anthropometric, and lifestyle characteristics of the core group are shown in Table 1. Differences were
noted in a number of these variables; statistically significant
racial differences in the full cohort for clinical and laboratory
variables are summarized in Table 2. Among men, physical
activity scores were higher among black participants; among
women, white participants had higher scores. Calcium and
vitamin D intake were higher in white women than in black
women, but no statistically significant differences were seen
among men. Black participants were heavier than white participants, had higher caloric intake, and reported more hours
of sun exposure per week. The proportion of current smokers
among black participants was about twice as high, while the
amount of caffeine consumption was about half. Black
women were more likely to have earlier menarche, to have
more children, and to give birth before age 21 yr. In contrast,
431
TABLE 2. Summary significance for clinical and biochemical
variables that differed between races when tested in the entire
cohort
Variable
Clinical variables:
Body size
height
elbow width
Muscle:
serum creatinine
leg strength
lean body mass
Fat:
weight
body mass index
sum skinfolds
waist:hip ratio
total body
trunk:leg ratio
Physical activity:
total
heavy only
Lifestyle:
caloric intake
calcium intake
caffeine intake
vitamin D intake
current smoking
sun exposure
Biochemical variables:
Calcium metabolism:
serum phosphate
25-hydroxyvitamin D
1,25-dihydroxyvitamin D
parathyroid hormone
urinary calcium
Bone turnover:
osteocalcin
alkaline phosphatase
urinary pyridinium cross-links
Sex and adrenal hormones:
bioavailable testosterone
total testosterone
free testosterone
dehydroepiandrosterone sulfate
Growth factors:
fasting insulin
insulin-like growth factor
Reproductive:
age at menarche
parity
infant delivery before maternal age 21 yr
breast feeding ⱖ26 wk
Race with higher levela
Men
Women
Black
White
Black
Black
Black
Black
Black
Black
Black
White
White†
White
Black
Black
Black
Black
Black
Black
Black
Black
White
White
Black
Black
White
White
White
Black
Black
White
Black
Black
Black
White
Black
White
Black†
White
Black
Black
White
White
Black
White
Black†
Black†
Black†
White
Black
Black
White
Black
White
White
Black
Black
White
a
All values significantly higher at P ⱕ 0.05 level except where
indicated with †. † indicates significantly higher within range, 0.05 ⬍
P ⬍ 0.10.
breast feeding was more common among white women than
black women. Both black men and black women had lower
25D as well as higher 1,25D levels and markedly lower urinary calcium excretion. We plan to report the details of
calciotropic hormone analyses elsewhere (Bikle D.D., unpublished material). Inconsistent racial differences were observed in markers of bone turnover and growth factors.
Mean hip and spine size, determined by DXA, was smaller
in black than in white women; among men, there were no
statistically significant racial differences in these sites. Be-
432
JCE & M • 1997
Vol 82 • No 2
ETTINGER ET AL.
TABLE 3. Body composition variables for the core group by gender and race
Men
Skeletal area, cm2:
spine, posteroanterior
hip, total
arms
legs
whole body
L3 volume, cm3
Lean mass, kg
Fat mass, kg
Trunk:leg fat ratio
Waist:hip ratio
Sum of 3 skinfold measurements, mm
a
Women
Black
White
Black
White
65.3 (6.6)
43.7 (4.3)
474a (41)
907a (74)
2436a (156)
39.6 (5.8)
64.2a (7.6)
16.0a (8.0)
0.92a (0.33)
0.82a (0.04)
52.9 (22.2)
65.7 (5.7)
43.7 (4.0)
441 (41)
861 (73)
2355 (160)
40.5 (5.8)
60.2 (6.7)
18.6 (9.0)
1.24 (0.47)
0.85 (0.06)
53.4 (19.8)
55.2a (5.9)
32.9a (3.2)
388a (42)
792 (67)
2096 (158)
31.6a (5.6)
43.7 (5.6)
27.9a (11.7)
0.92 (0.31)
0.75a (0.07)
77.0a (30.2)
57.6 (5.5)
34.3 (3.3)
356 (34)
770 (67)
2075 (144)
35.0 (5.3)
42.6 (4.9)
23.5 (11.9)
0.86 (0.27)
0.72 (0.04)
58.7 (27.2)
Statistically significant racial differences within same gender P ⬍ 0.05.
TABLE 4. Sex and adrenal hormones in the core group by gender and race
Men
Women
Black
White
Black
White
mean (standard deviation)
Testosterone:
total, ng/dL
free, pg/mL
bioavailable ng/dL
Sex-hormone-binding
globulin, ␮g/dL
Estradiol, ng/dL
Free estradiol index, ng/␮g
Dehydroepiandrosterone sulfate, ␮g/dL
a
595 (201)
152 (60)
321 (104)
549 (177)
138 (58)
298 (88)
0.7 (0.3)
0.7 (0.3)
201 (67)a
222 (89)
31 (13)
4 (2)
10 (5)a
28 (12)
4 (3)
8 (4)
1.4 (0.7)
9.1 (6.9)
8.1 (9.1)
133 (64)
1.5 (0.7)
10.1 (10.5)
7.7 (7.7)
144 (72)
Statistically significant racial differences within gender, P ⬍ 0.05.
TABLE 5. Bone mineral density (BMD) in the core group by gender and race
Men
Spine:
posteroanterior, g/cm2
lateral L3, g/cm2
volumetric L3, g/cm3
Femur:
neck, g/cm2
trochanter, g/cm2
Ward’s triangle, g/cm2
total, g/cm2
Total body, g/cm2
a
Women
Black
White
Black
White
1.148 (0.151)
0.909 (0.109)
0.246 (0.032)
1.030 (0.126)
0.800 (0.091)
0.223 (0.023)
1.130 (0.116)
0.862 (0.089)
0.256a (0.029)
1.045 (0.102)
0.812 (0.088)
0.241 (0.027)
1.068 (0.173)
0.903 (0.145)
0.961 (0.201)
1.187 (0.172)
1.295 (0.121)
0.891 (0.128)
0.783 (0.124)
0.769 (0.144)
1.034 (0.137)
1.177 (0.096)
0.962 (0.141)
0.778a (0.108)
0.894 (0.176)
1.036 (0.126)
1.163 (0.086)
0.862 (0.101)
0.728 (0.097)
0.789 (0.119)
0.955 (0.103)
1.090 (0.075)
P ⬍ 0.01. All other racial differences (within gender) significant, P ⬍ 0.001.
cause of their larger appendicular bone area, black men had
3% greater total skeletal area than their white counterparts;
similar trends were observed between women subjects (Table 3). Lean body mass was greater in black persons than in
white persons; differences observed were 6.6% for men and
2.6% for women. Black women had greater mean body fat
and other mean fat indexes than white women, but black men
were leaner than white men. Measures of fat distribution
showed higher ratio of abdominal to lower extremity fat in
men compared with women; however, white men showed
greater abdominal fat than black men, whereas black women
showed greater abdominal fat than white women.
Black men and women tended to have higher mean total,
free, and bioavailable testosterone levels but lower dehy-
droepiandrosterone sulfate levels than their white counterparts. Among women, no statistically significant racial differences were found in mean estradiol or estradiol index
(Table 4).
Bone density at all skeletal sites was statistically significantly greater in black persons than in white persons (Table
5); the racial differences were about 1.5–2 times as great in
men as in women. The greatest racial differences were noted
in Ward’s triangle. In the relatively narrow age range of our
subjects, no consistent age-related change in BMD was found
for BMD measurements (data not shown). The BMD of white
women and men in our cohort agreed closely with the ageexpected values provided in the Hologic reference database;
for women, posteroanterior spine density was within ⫹0.2%,
RACIAL DIFFERENCES IN BONE DENSITY
433
TABLE 6. Difference in various BMD measurements between black and white subjects, adjusted (and unadjusted)
Adjusted differences in BMD
Skeletal measurement
gm/cm2
(95% CI)c
Sign of parameter estimate for variables in model
Percenta
(unadjusted)d
positiveb
Men:
Spine:
posteroanterior
0.076 (0.037, 0.115)
7.3 (11.4)
lateral L3
0.061 (0.032, 0.091)
7.5 (13.6)
lean mass
caloric intake
lean mass
volumetric L3
0.013 (0.005, 0.021)
5.6 (10.1)
BMI
0.119 (0.073, 0.165)
13.0 (19.9)
0.062 (0.024, 0.100)
7.7 (15.3)
Ward’s triangle 0.127 (0.073, 0.181)
16.1 (25.1)
Femur:
neck
trochanter
total
0.099 (0.060, 0.139)
9.4 (14.8)
0.054 (0.025, 0.083)
4.5 (10.0)
Women:
Spine:
posteroanterior
0.060 (0.025, 0.094)
5.6 (8.1)
lateral L3
0.039 (0.011, 0.067)
4.8 (6.2)
volumetric L3
0.003 (⫺0.006, 0.012)
1.2 (6.2)
Total body
Femur:
neck
0.064 (0.027, 0.102)
7.3 (11.5)
0.049 (0.014, 0.085)
6.8 (6.8)
Ward’s triangle 0.058 (0.014, 0.102)
7.1 (13.3)
trochanter
total
Total body
0.053 (0.017, 0.089)
5.5 (8.5)
0.059 (0.031, 0.087)
5.4 (6.7)
lean mass
hours in sun
caloric intake
lean mass
caloric intake
lean mass
caloric intake
hours in sun
lean mass
caloric intake
lean mass
caloric intake
BMI
hours in sun
BMI
caloric intake
BMI
negativeb
R2 for covariables
except race
DHEAS
0.242
elbow width
IGF-BP3
waist:hip
IGF-BP3
waist:hip
0.296
elbow width
IGF-BP3
0.402
elbow width
urinary calcium
urinary calcium
0.362
elbow width
0.370
elbow width
pyridinoline cross-links
0.453
0.220
0.266
0.141
0.098
height
urinary calcium
BMI
hours in sun
BMI
waist:hip
lean mass
1,25-dihydroxy-vitamin D
hours in sun
BMI
bioavailable testosterone
infant delivery before age 21 yr
BMI
hours in sun
lean mass
waist:hip
BMI
osteocalcin
caloric intake
dietary calcium
0.217
0.235
0.230
0.216
0.162
0.336
IGF-BP3, IGF-binding protein 3; waist:hip, ratio of waist-to-hip circumferences.
a
Elevation in black BMD over white BMD, divided by white BMD.
b
Sign of parameter estimate in regression model.
c
95% confidence intervals of BMD differences appear in parentheses.
d
Unadjusted percentage differences in BMD appear in parentheses.
total hip density was within 1.6%, and whole body BMD was
within ⫺0.4%; for men, posteroanterior spine density was
⫺5.8%, and total hip density was ⫺3.7%.
Of 37 variables that differed between races, less than half
were significantly (P ⬍ .05) associated with bone density in
any of the final regression models. Anthropometric variables
appeared in all the final models; most frequently included for
men was lean mass, and for women, body mass index. At
least one lifestyle variable appeared in most of the final
models, with caloric intake most consistent. Biochemical tests
did not show a consistent pattern of inclusion in these models. Multivariate models that did not include race as a covariate explained from 22– 45% of the variance in 8 BMD
measurements in men and from 10 –34% of this variance in
women (Table 6). Adjustment for covariates reduced racial
differences in the 8 BMD measurements, on average, 42% for
men (range 35–55%) and 34% for women (range 0 – 81%).
After these adjustments, similar racial differences in total
body BMD in both genders remained, but larger racial differences were found between spine and hip densities in men
than in women.
Discussion
By using DXA area measurements, we confirmed that
young, adult, black men and women have substantially
greater BMD than white persons (2– 6); this finding is true at
all skeletal sites, and we have extended these observations to
434
ETTINGER ET AL.
volumetric estimates of spinal BMD. Despite an extensive
search for clinical and biochemical explanations, we were
unable to account for most of the racial differences in BMD.
The racial differences we observed in BMD do not appear
to be artifacts of the DXA method. Although the greater area
bone density observed after measurement with DXA could
be caused by larger bone size and therefore may not represent a true increase in volumetric density, our measurement
of area and volume of both spine and hip show that black
persons do not have larger vertebrae or upper femurs than
white persons. Further, the magnitude of the BMD differences in spinal volumetric density are similar to those for
spinal area density.
Our large, population-based study had sufficient power to
detect racial differences in clinical and biochemical variables
on the order of 0.5 standard deviations. Clinical and biochemical assessment was extensive and enabled us to study
a comprehensive list of variables that others have considered
relevant to bone mass. After multiple adjustments for these
variables, about half the racial differences in BMD still remained. This finding suggests that other factors than those
controlling muscle mass and body size must act specifically
on the skeleton.
We conclude that racial differences in BMD are established
early in childhood (33) and are not explained by clinical and
biochemical variables measured in young adulthood. Studies
of adolescents might find differences in metabolic or lifestyle
factors that account for a larger share of the racial differences
in bone mass than those that we observed. By midadolescence, black boys and girls have 10 –15% greater bone density
than their white counterparts (33). However, most of the
difference would be expected to remain evident 10 –20 yr
later. Thus, the appearance of such a large racial difference
in young adults cannot be attributed to persistent differences
in metabolic or lifestyle factors and supports the view that
bone density differences result from influences operating
during childhood and adolescence.
References
1. Cummings SR, Black DM, Rubin SM. 1989 Lifetime risks of hip, Colles’ or
vertebral fracture and coronary heart disease among white postmenopausal
women. Arch Intern Med. 149:2445–2448.
2. Liel Y, Edwards J, Shary J, Spicer KM, Gordon L, Bell NH. 1988 The effects
of race and body habitus on bone mineral density of the radius, hip, and spine
in premenopausal women. J Clin Endocrinol Metab. 66:1247–1250.
3. Luckey MM, Meier DE, Mandeli JP, DaCosta MC, Hubbard ML, Goldsmith
SJ. 1989 Radial and vertebral bone density in white and black women: evidence
for racial differences in premenopausal bone homeostasis. J Clin Endocrinol
Metab. 69:762–770.
4. Harris SS, Wood MJ, Dawson-Hughes B. 1995 Bone mineral density of the
total body and forearm in premenopausal black and white women. Bone.
16:311S–315S.
5. Looker AC, Wahner HW, Dunn WL et al. 1995 Proximal femur bone mineral
levels of US adults. Osteoporos Int. 5:389 – 409.
6. Nelson DA, Jacobsen G, Barondess DA, Parfitt AM. 1995 Ethnic differences
in regional bone density, hip axis length, and lifestyle variables among healthy
black and white men. J Bone Miner Res. 10:782–787.
7. Burke GL, Jacobs DR Jr, Sprafka JM, Savage PJ, Sidney S, Wagenknecht LE.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
JCE & M • 1997
Vol 82 • No 2
1990 Obesity and overweight in young adults: the CARDIA study. Prev Med.
19:476 – 488.
Ortiz O, Russell M, Daley TL, et al. 1992 Differences in skeletal muscle and
bone mineral mass between black and white females and their relevance to
estimates of body composition. Am J Clin Nutr. 55:8 –13.
Bell NH, Greene A, Epstein S, Oexmann MJ, Shaw S, Shary J. 1985 Evidence
for alteration of the vitamin D-endocrine system in blacks. J Clin Invest.
76:470 – 473.
Dawson-Hughes B, Harris S, Kramich C, Dallal G, Rasmussen HM. 1993
Calcium retention and hormone levels in black and white women on high- and
low-calcium diets. J Bone Miner Res. 8:779 –787.
Meier DE, Luckey MM, Wallenstein S, Lapinski RH, Catherwood B. 1992
Racial differences in pre- and postmenopausal bone homeostasis: association
with bone density. J Bone Miner Res. 7:1181–1189.
Weinstein RS, Bell NH. 1988 Diminished rates of bone formation in normal
black adults. N Engl J Med. 319:1698 –1701.
Kanders B, Dempster DW, Lindsay R. 1988 Interaction of calcium nutrition
and physical activity on bone mass in young women. J Bone Miner Res.
3:145–149.
Stevenson JC, Lees B, Devenport M, Cust MP, Ganger KF. 1989 Determinants
of bone density in normal women: risk factors for future osteoporosis? Br
Med J. 298:924 –928.
Sowers M, Wallace RB, Lemke JH. 1985 Correlates of forearm bone mass
among women during maximal bone mineralization. Prev Med. 14:585–596.
Rosenthal DI, Mayo-Smith W, Hayes CW, et al. 1989 Age and bone mass in
premenopausal women. J Bone Miner Res. 4:533–538.
Sowers M, Corton G, Shapiro B, et al. 1993 Changes in bone density with
lactation. JAMA 269:3130 –3135.
Kelly PJ, Twomey L, Sambrook PN, Eisman JA. 1990 Sex differences in peak
adult bone mineral density. J Bone Miner Res 5:1169 –75.
Dhuper S, Warren MP, Brooks-Gunn J, Fox R. 1990 Effects of hormonal status
on bone density in adolescent girls. J Clin Endocrinol Metab 71:1083– 8.
Buchanan JR, Myers C, Lloyd T, Leuenberger P, Demers LM. 1988 Determinants of peak trabecular bone density in women: the role of androgens,
estrogen, and exercise. J Bone Miner Res 3:673– 80.
Kelly PJ, Pocock NA, Sambrook PN, Eisman JA. 1990 Dietary calcium, sex
hormones, and bone mineral density in men. Br Med J. 300:1361– 4.
Ross R, Bernstein L, Judd H, Hanisch R, Pike M, Henderson B. 1986 Serum
testosterone levels in healthy young black and white men. J Natl Cancer Inst
76:45– 48.
Henderson BE, Bernstein L, Ross RK, Depue RH, Judd HL. 1988 The early
in utero oestrogen and testosterone environment of blacks and whites: potential
effects on male offspring. Br J Cancer 57:216 –218.
Kelly PJ, Eisman JA, Sambrook PN. 1990 Interaction of genetic and environmental influences on peak bone density. Osteoporos Int 1:56 – 60.
Friedman GD, Cutter GR, Donahue RP, et al. 1988 CARDIA: study design,
recruitment, and some characteristics of the examined subjects. J Clin Epidemiol. 41:1105–1116.
Jergas M, Breitenseher M, Glüer C-C, Yu W, Genant HK. 1995 Estimates of
volumetric bone density from projectional measurements improve the discriminatory capability of dual X-ray absorptiometry. J Bone Miner Res.
10:1101–1110.
Mohan S, Bautista CM, Herring SJ, Linkhart TA, Baylink DJ. 1990 Development of valid methods to measure insulin-like growth factors-I and -II in
bone cell-conditioned medium. Endocrinology 126:2534 – 42.
Blum WF, Ranke MB, Kietzmann K, Gauggel E, Zeisel HJ, Bierich JR. 1990
A specific radioimmunoassay for the growth hormone (GH)-dependent somatomedin-binding protein: its use for diagnosis of GH deficiency. J Clin
Endocrinol Metab. 70:1292–1298.
Seyedin SM, Kung VT, Daniloff YN,et al. 1993 Immunoassay for urinary
pyridinoline: the new marker of bone resorption. J Bone Miner Res. 8:635– 641.
McDonald A, Van Horn L, Slattery M, 1991 The CARDIA dietary history:
development, implementation, and evaluation. J Am Diet Assoc. 91:1104 –1112.
Sidney S, Jacobs DR Jr, Haskell WL, et al. 1991 Comparison of two methods
of assessing physical activity in the Coronary Artery Risk Development in
Young Adults (CARDIA) Study. Am J Epidemiol. 133:1231–1245.
Jacobs DR Jr, Hahn LP, Kaskell WL, Pirie P, Sidney S. 1989 Validity and
reliability of a short physical activity history: CARDIA and the Minneapolis
Heart Health Program. J Cardiopulmon Rehab. 9:448 – 459.
Bell NH, Shary J, Stevens J, Garza M, Gordon L, Edwards J. 1991 Demonstration that bone mass is greater in black than in white children. J Bone Miner
Res. 6:719 –723.