Vol. 81, No. 6
Prmted in U.S.A.
0021-972x/96/$03.00/0
Journal
of Clinical
Endocrinology
and Metabolism
Copyright
0 1996 by The Endocrine
Society
Body Composition
in Normal
Black
The Four-Compartment
Model
JOHN
F. ALOIA,
ASHOK
VASWANI,
The Department
of Medicine,
Winthrop
Laboratories,
Upton, New York 11501
RUIMEI
University
MA,
Hospital,
EDITH
AND
Mineola.
ABSTRACT
Women:
FLASTER
New York Brookhaven
National
difference
only for the mineral
compartment.
Various
models were fit
to the data to adjust
for body size and age. The equation
y =
age+height+weight
fits the data as well as the other models. Equations and graphs were developed
to assess each compartment
using
this linear
model and may be used to assess the body composition
status of healthy
and ill black women.
Although
black women tended
to be heavier
than white women,
after controlling
for differences
in
body weight
(and age) black women
had a greater
mass of protein,
mineral,
and water
and a similar
fat store. These differences,
while
statistically
significant,
were not of great
magnitude.
This was a
cross-sectional
study and suspected
trends
with aging must be confirmed
by a longitudinal
study. (J Clin Endocrinol
Metab 81: 23632369, 1996)
The four-compartment
model of body composition
was studied
in
112 healthy
black women
to develop
normative
data to be used to
assess deviations
in illness. Each compartment
was measured
by an
independent
method:
tritiated
water
dilution,
prompt-gamma
neutron activation
analysis
(for nitrogen),
inelastic
neutron
scattering
(for carbon),
and dual energy x-ray absorptiometry
(for calcium).
The
mean age of the population
was 43.2 years. Race was self-declared.
The mean values for the four compartments
were [kg SE]protein:
9.6
(0.07); mineral
3.0 (0.03); fat 24.2 (0.70); and water 33.1(0.29).
Each
of the compartments
changed
significantly
with age (P < .OOOl).
There
were
declines
in total body water,
mineral,
and protein,
whereas
fat increased
linearly
with age. Linear
regressions
performed
on pre- and postmenopausal
women
showed
a significant
T
HERE HAS BEEN a renewed interest in body composition as aging, nutrition, and sports medicine have
received increasing attention. Much of the attention hasbeen
on young white men and women, although recently it has
become clear that body composition is affected by age, race,
and gender (1, 2).
Ethnic differences in body composition have been noted in
the past and have been under study in recent years as more
accurate techniques for measurement of the various components of the body have been developed (3-8). Anthropomorphic and growth differences between blacks and whites
have been noted in several surveys including the HANES I
survey. A study of adult skeletons found that American
blacks had a denser skeleton than American whites (9). Epidemiological studies using radiographic morphometry of
the hand suggest that black women start out with a greater
bone density than whites and lose skeletal massat a slower
rate (10,ll). Cohn et al. (12), measuring total skeletal massby
in viva neutron-activation analysis and bone density of the
radius by single-photon absorptiometry observed that black
women had a greater skeletal massand bone mineral content
of the radius than white women of similar ages.In that study,
lean body masswas assessedby the whole body counting of
40K, a radioactive isotope of potassium, which provides the
value for total body potassium (TBK). The suggestion was
made that the larger muscle mass in black women was a
Received
May 23, 1995. Revised November
22, 1995. Accepted
January 1, 1996.
Address
correspondence
and requests
for reprints
to: John F.
Aloia, M.D., Winthrop
University
Hospital,
259 First Street, Mineola,
NY 11501, Phone: (516) 663-2381,
Fax: (516) 663-8796.
+ This study was supported
by NIH Grants ROl-AR37520-05
and
POl-DK42618.
2363
major determinant of their increased skeletal mass. More
recently we studied a larger group of premenopausal black
and white women and found that black women were heavier
than white women, because of a greater musculoskeletal
mass and the associated increase in body water (13).
Field assessmentsof body fat such asanthropometric measurements, body weight, or the body massindex are unsatisfactory if quantitative measurement of body composition is
desired. Two-compartment quantitative models have divided the body into fat and fat-free mass (FFM). FFM has
been measured using TBK (as above), total body nitrogen
(TBN), and total body water (TBW). Fat mass (FM), using
(TBW, TBK, or TBN) may be calculated by subtracting FFM
from body weight. Unfortunately, small systematic errors in
FFM can result in a large error in FM when subtracted from
body weight. Other quantitative methods are also dependent
on physical properties of the body, such as body density,
impedance, or body attenuation of photons or x-rays, and the
coefficient of variation of derived FM will be larger than that
of these measured physical properties of the body.
The neutron-activation facility at Brookhaven National
Laboratory (BNL) has been developed to measure the chemical constituents of fat and fat-free tissue using promptgamma neutron-activation, in conjunction with the TBW
measurement for the measurement of TBN and inelastic neutron scattering for the measurement of total body carbon
(TBC). In the current study, we measured TBW by tritiated
water dilution, and estimated total body calcium (TBCa) by
dual-energy x-ray absorptiometry (DXA). The development
of these techniques has allowed the construction of a fourcompartment model of body composition, consisting of mineral ash, fat, protein, and water, with relatively independent
measurement of each compartment. We have recently pre-
ALOIA
sented data on premenopausal black and white women (13),
and we have also developed comparative normative equations for white adult women (14). In this report we provide
additional data on postmenopausal black women so that we
may develop predictive equations for body composition in
black women from age 20-70 yr.
JCE & M . 1996
Volt31 . No 6
ET AL.
administered
orally, and a blood sample was obtained
3 h later. The
water was removed
from the plasma by rapid vacuum
sublimation
and
was counted using a liquid scintillation
counter to determine
TBW. The
radiation
dose is less than 4 mrem. The reproducibility
of the technique
is less than 1%. The values were corrected
for density and proton
exchange.
Calculation
Materials
and
Methods
Subjects
Healthy
black women age 20-70 yr were recruited
by advertising
in
the local media and through
a direct mail campaign.
Black race was
determined
by self-declaration,
and participants
were asked to complete
a family pedigree
extending
to grandparents.
Four black women had one
European
or American
Indian ancestor,
and four had two European
ancestors.
Exclusion
characteristics
consisted
of any chronic illness including
hypertension,
diabetes, obesity, and any past history
of illness
or medication
known
to affect bone metabolism.
Hysterectomy
or oophorectomy
was an exclusion
factor. The project was approved
by the
institutional
review
boards
of Winthrop-University
Hospital
and
Brookhaven
National
Laboratory,
and written
informed
consent was
obtained
from each participant.
After initial screening,
women
were
further
rejected based on abnormal
blood chemistries
(multi-channel
chemistries
(SMAlZ),
CBC, urinalysis,
free T4, TSH) or abnormal
physical findings.
A body mass index (BMI) of 18-33 was considered
acceptable for inclusion
in the study. Less than 10% were excluded
after
initial screening
using all criteria.
The current report includes
data on
women
between
the ages of 20 and 70 yr. Some of the data (on premenopausal
women)
have been previously
reported
(14). Height
was
measured
using a wall-mounted
Harpenden
Stadiometer
(Holtain
Ltd.,
UK). The participants
were fasting overnight
and were weighed
wearing
scrubs using a balanced
beam scale.
Prompt-gamma
neutron-activation
(PGNA)
The PGNA system at BNL has been described
previously
total skin dose to a patient was 80 mrem (16). The coefficient
(CV) was 2.5%.
Inelastic
neutron
scattering
(15,161. The
of variation
system (INS)
The INS facility was built in 1987 (16). The total skin dose is less than
50 mrem (16). The system is calibrated
daily with an Alderson
phantom.
The CV was 3%.
Dual-energy
x-ray absorptiometry
A whole body DXA scanner (DPX-L)
(Lunar Radiation,
Madison,
software
program
1.3Y) was used to measure bone mineral density
The skin entry dose is 0.01 mrem. The CV is under 1%.
Total
TBW
tritiated
body water
TABLE
1. Clinical
by the standard
Approximately
cm
Wt, kg
BMI
Age, yr
Age at menarche
Age at menopause
Median
income
Median
education
Number
premenopausal
NS, not significant.
Statistics
The following
models were fit to each of the four-compartments
in
order to adjust the data for age and body size (height and weight):
1. y = a + b (height)
+ c (weight)
+ d (age) + e (age’)
2. In y = a + b (height)
+ c (wei ht) + d (age) + e (age’)
3. y = a (heighti’)
(weight’)
(age b:1, where y is protein,
water, mineral,
or fat.
The models were used to evaluate
the effects of the covariates
height,
weight,
and age on body composition.
To minimize
collinearity
between
age and age squared,
age was
entered into both terms in the equation
as age minus its mean. We fit the
model for age and age squared
and used those for Figures la to 4a. We
next fit the residuals
from that model to a linear model of weight and
height and used those for Figures lb to 4b. This has the effect of including
the natural
changes of weight and height with age into the age model
and then fitting the effects of weight and height that were independent
of age. For compartments
where the age squared
term was not statistically significant,
the linear effect of age was used in the first model,
before fitting
weight
and height. In addition,
linear regressions
were
performed
for each compartment
zts. age for premenopausal
and postmenopausal
women
and the slopes tested for statistically
significant
differences.
To compare
the black women
with white women,
previously
collected data on white women
were used (14). Since the black women
weighed
more than the white and were younger,
the data were adjusted
to the midpoints
of age, height,
and weight
of the two races using
separate
multiple
regression
equations
(SAS, SAS Institute,
Cary NC).
The percent
bodyweight
of each compartment
was then calculated.
Between
groups,
differences
for each compartment
were evaluated
by
a Student’s
t-test (two-tailed).
Results
WI;
(17).
Subjects
The clinical characteristics of the women are given in Table
1. They are compared to previously reported data in a larger
population of white women (14). The age of menopausewas
younger in black than in white women (49.1 VS.50.5 yr). There
isotope-dilution
method
using
50 &i
of tritiated
water was
characteristics
Variable
Ht.
Total body fat (TBF) was calculated
from TBC and total body protein
(TBPr):
TBF (kg) = [TBC(kg)-0.55
TBPr (kg)1/0.77
and TB Protein = 6.25 (N/H)x
(TBH), where H is hydrogen.
Mineral
ash
was calculated
as 2.94 TBCa. The TBCa value was obtained
from DXA.
In this model, body weight
= TBW + TPr +TBF + 2.94 TBCa (18).
(TBW)
was measured
water dilution.
of body compartments
Mean
Blacks (n = 112)
163.7 (0.56)
69.8 (1.05)
26.1 (0.39)
43.2 (1.17)
12.4 (0.14)
49.1 (0.38)
$50,000-$74,999
4 yr college
72
(SE)
Whites
(n = 163)
163.6 (0.50)
64.0 (0.70)
23.9 (0.25)
52.2 (1.05)
12.6 (0.10)
50.5 (0.31)
$50,000-$74,999
4 yr college
89
Differences
(P-value)
NS
0.0001
0.0001
0.0001
NS
0.01
2365
BODY COMPOSITION
TABLE
2. Mean
values
for black
and white
Compartment
(kg)
Age 20-39
women
(SE)
Blacks
(n = 45)
Mineral
Fat
Protein
Water
Whites
(n = 39)
3.1(0.05)
21.4 (0.92)
10.0 (0.11)
33.5 (0.44)
Mineral
Fat
Protein
Water
3.0
25.6
9.5
33.0
Compartment
(kg)
Age 60-70
(0.04)
(1.06)
(0.10)
(0.41)
2.7
28.7
9.3
32.6
(0.07)
(1.80)
(0.28)
(1.21)
Models
Each of the compartments changed significantly with age
(P < .OOOl).The three models fit the data similarly. Therefore,
only model 1 is presented. All covariates (height, weight, and
age) were significant (P < .05) in model 1 for all compartments. In Table 2, the mean values for three age groups are
given. Table 3 shows the results of statistically adjusting the
data to the midpoint age, height, and weight. A black woman
hasa significantly higher proportion of mineral, protein, and
water, and a nearly significant lower proportion of fat than
a white woman. The regression equations are given in Table
4. Table 4 shows that with the sameheight and weight, total
body fat increaseswhile the other compartments decreasefor
older age; with the sameweight and age, total fat decreases
while the other compartments increase with height; with
similar height and age, all compartments increase with
weight. Total body protein, mineral, and water decreasewith
age while total body fat increases. Regression analysis carried out on pre- VS.postmenopausal women showed a significant difference in slope (P < .05) for the mineral ash
measurement only.
The predicted values from the regression equations are
Variable
Mineral
Fat
Protein
Water
3. Calculated
values
Blacks
% (SE)
4.37
34.1
13.9
47.6
c.052)
C.97)
(.097)
c.37)
for black
and white
Whites
% (SE)
4.10
35.6
13.4
46.9
C.037)
(.63)
c.078)
c.27)
1.11
1.23
1.09
1.05
0.0001
0.003
0.0001
0.01
Whites
(n = 78)
Ratio
B/W
P-value
(0.04)
(0.66)
(0.08)
(0.34)
1.07
1.12
1.06
1.06
0.0002
0.03
0.0001
0.0003
Whites
(n = 46)
Ratio
B/w
P-value
1.16
1.22
1.12
1.08
0.002
0.01
0.001
0.01
2.32
23.6
8.3
29.6
were 112 black women and 163 white women. The black
women were heavier and younger.
TABLE
(0.04)
(0.91)
(0.10)
(0.42)
2.8
22.8
9.0
31.0
Blacks
(n = 10)
Mineral
Fat
Protein
Water
P-value
2.8
17.4
9.2
31.9
Blacks
(n = 57)
Compartment
(kg)
Age 40-59
Ratio
B/W
(0.04)
(0.76)
(0.12)
(0.48)
shown in Figures la, 2a, 3a, and 4a. Those figures show that
fat increaseslinearly with age, while protein and water decline. The mineral ash was first described by two separate
slopes, with a more accelerated loss at the postmenopausal
ages (P < .05). The test for difference between slopesis more
powerful than the test of the quadratic age term in a compartment VS.age model including pre and postmenopausal
women. In order to show the relationship between mineral
and age as a continuous curve, we have chosen to represent
it using a quadratic function for age.
A comparison of the slopesof black and white women for
each compartment US.age showed no significant differences.
Figures lb, 2b, 3b and 4b provide the values to adjust an
individual’s value for height and weight. Suppose a 60-yrold black woman had her mineral massmeasured. She is 165
cm tall and weighs 76 kg, which is slightly taller and heavier
than the average black woman in our study sample. The
correction on the weight-height adjustment graph is obtained at the intersection of 165 cm and 76 kg. That adjustment is +O.l. So we move to the mineral massby age graph
(Figure 4a) and add 0.1 to the Y-axis. She is in the lowest 5%
of women her age if her mineral massis lessthan 2.3 kg. (2.2
from the graph plus the 0.1 adjustment). She is in the lower
half of women her age if her mineral massis lessthan 2.9 kg.
(2.8 from the graph plus the 0.1 adjustment).
Discussion
women
B/W
P-value”
1.066
0.958
1.037
1.015
.OOOl
.054
.OOOl
.025
Values
given are at Age 47.7, Height
163.6 cm, Weight
66.9 kg.
Values are percent
of total body weight,
obtained
by using separate
equations
for the weight of each compartment
and then computing
the
percent
of the total compartments.
a The P-value
given is the result of Student’s
t-test on the difference
between
the percentages.
Several studies have establishedthat the various compartments of the body are influenced by sex, age, and ethnicity.
In this manuscript we provide data applicable to black
women (and compare it with previous findings in white
women). Since an influence of age was found for each compartment, clearly values obtained should be plotted on the
figures provided to determine an individual’s percentile
value for her age. A more complex issueinvolves adjustment
for height and weight. The Brookhaven group previously
suggested adjustment of both TBCa and TBK for height and
weight (19). The relationship of skeletal mass to stature is
ALOIA ET AL.
TABLE
4. Coefficients of the regression equations slope
Intercept
Ht
(SE)
Wt (kg)
(cm)
Black
Protein
5.8
0.01(0.010)
Water
13.3
0.05 (0.038)
Mineral
0.7
0.01(0.004)
Fat
46.4
-0.33 (0.079)
White
Protein
2.3
0.03 (0.008)
Water
-2.0
0.13 (0.027)
Mineral
-0.3
0.01(0.003)
Fat
31.3
-0.22 (0.052)
a Mean age black = 43.2; mean age white = 52.2.
b Percent variation explained.
Age-Mean
.i
i
j
..- 26
24 _
I.. /,.... i
j ............ ...j .... .............
22 “““. ; .... .i -“‘~““’ “-- ~~-i”“” -‘:
~i
i
I
I
I
I
i
..- 24
20
j... ..‘. ..................... .... ...........
30
40
50
i
60
Agea
0.03 (0.006)
0.17 (0.021)
0.01(0.002)
0.45 (0.044)
-0.03 (0.005)
-0.07 (0.019)
-0.01(0.002)
0.13 (0.040)
0.03 (0.005)
0.17 (0.018)
0.01(0.002)
0.44 (0.035)
-0.03
-0.07
-0.02
0.09
A
26 -....
JCE & M . 1996
Vol81.
No 6
Rz*
-0.0003
0.003
(0.0001)
(0.003)
37
42
41
56
(0.0001)
(0.002)
43
55
60
57
-0.0003
-0.006
5o I,..,...,
.;.. ..jI ..;..........,......,.
;... ;..
o
22
(0.003)
(0.012)
(0.001)
(0.022)
(Age-Mean Age)’ a
70
:
:
.. ...... .j ‘... ‘-j ‘.......’ :“’ “..‘.i”.“....p’.
1
I
20
30
40
Age
;..,,.....;,
.;.. .I 5o
.... : ..‘..‘. ..:.. .‘.. ..j ......-..j
I
1
50
60
0
70
Age
B
50
145
150
155
160
165
Height,
170
175
180
185
cm.
FIG. 1 A, The values for total body water in black women at different
ages, depicted in percentiles. These data are not adjusted for body
size. B, The height and weight adjustments for any value of total body
water in black women. The subject’s height and weight should be
plotted on the graph. The value obtained where the two points intersect should be added to or subtracted from the actual value before
plotting the value on Figure 1A (see text).
145
I
150
I
155
I
160
:
I
165
Height,
I
170
I
175
I
180
50
185
cm.
FIG. 2 A, The values for total body fat at different ages in black
women, depicted in percentiles. These data are not adjusted for body
size. B, The height and weight adjustments for any value of total body
fat in black women. The subject’s height and weight should be plotted
on the graph. The value obtained where the two points intersect
should be added to or subtracted from the actual value before plotting
the value on Figure 2A (see text).
BODY COMPOSITION
A
i
:”
I3
:
12 -
:
:
;
j
;
I3
:
2367
A5
jI:jiiiji
j
5
/
/
/
:
:
:
:
:
- 12
:
4
a
2
i?
.-2
H
3
2
i..
6
i
20
.I
..:.
.,....,... j .. . ...
;
I
30
I
I
40
/
I
50
I
<
I
:
I
60
;
i
:
1
1
6
20
70
30
40
50
60
70
Age
Age
90
B Q”
: /
:‘...-..A:. .:
:
.L.
,.
,
90
80
80 -
B
9
4.c
3 70
g
.L$
3
3
60
50
50
145
150
155
160
165
Height,
FIG. 3
women,
size. B,
protein
plotted
tersect
plotting
170
175
180
145
185
150
155
160
165
Height,
170
175
180
185
cm.
cm.
A, The values for total body protein
at different
ages in black
depicted
in percentiles.
These data are not adjusted
for body
The height and weight adjustments
for any value oftotal
body
in black women.
The subject’s
height
and weight
should be
on the graph.
The value obtained
where the two points inshould be added to or subtracted
from the actual value before
the value on Figure
3A (see text).
obvious. However, reduction of a value of body fat because
of increased body weight may obscure the very abnormality
that is being examined. Thus, depending on the information
sought, a decision would have to be made whether to adjust
an individual’s values using the height and weight tables we
provide before plotting the value on the regressions against
age.
This study provides a large data set for the four-compartment model in black women, using techniques that are not
interdependent. The equations developed from these measurements may be used to assess the extent of abnormality
in diseases that affect body composition.
To determine
whether an individual value differs substantially from av-
FIG. 4
women,
size. B,
mineral
plotted
tersect
plotting
A, The values for total body mineral
at different
ages in black
depicted
in percentiles.
These data are not adjusted
for body
The height and weight adjustments
for any value of total body
in black women.
The subject’s
height and weight
should be
on the graph.
The value obtained
where
the two points inshould be added to or subtracted
from the actual value before
the value on Figure
4A (see text).
erage, it is necessary to adjust the value for body size (Figures
lb-4b), as well as age (Figures la-4a).
Earlier studies of body composition partitioned the body
into two compartments:
FM and FFM. Mineral mass was
included in the FFM. Most methods for measuring body
composition (hydrodensitometry,
TBW, bioimpedance analysis, TBK, and TBN) estimate the FFM and calculate fat mass
by subtraction of FFM from body weight (1). Small systematic errors in FFM lead to greater errors in FM when the latter
is derived by subtraction of FFM from body weight. Because
none of these measurements directly measure fat, they are
dependent on assumptions from a few cadaver analyses and
involve assumptions of constancy with age, a constant state
of hydration of the FFM, as well as equivalence of the effect
ALOIA
of age on body massand various compartments. Difficulties
with these assumptions have been detected, and more elaborate models of body composition as well as efforts to measure each body compartment more directly have been proposed (1).
The four-compartment model of body composition divides the body mass into mineral, protein, fat, and water.
Direct methods for measurement of these compartments
have been sought. The measurement of TBW through isotope-dilution was developed, and neutron-activation analysis and photon (x-ray) absorptiometry were introduced for
the measurement of bone mineral. The development of
prompt-gamma neutron-activation for the measurement of
TBN permitted the most direct measurement of body protein, since 99% of body nitrogen is in protein. More recently,
Keyere et al. (20) developed the technique of measuring body
fat through the measurement of TBC, because fat can be
directly assessedby adjusting the TBC measurement for
body protein. Thus, through the use of DXA (mineral ash),
prompt-gamma neutron-activation (protein), inelastic neutron scattering (fat), tritiated water dilution (water), the fourcompartment model can be constructed from independent
measurements,and the influence of age, gender, and race can
be assessedrather than assumed.
Two decades ago, we used neutron-activation analysis,
whole body counting of 40K,and single-photon absorptiometry to compare skeletal massin black and white women (21).
A higher skeletal mass (of about 16%) was found in black
women. However, when the data were adjusted for height
and TBK, the difference between the two groups was reduced
to 7%. Whereas the TBK may overestimate FFM in black
women, this correction may not be accurate because the
equations for the correction factors were those derived from
white women. It is of interest that, although the black women
had the sameheight aswhite women, they were 15%heavier.
The same characteristics were found in the current study,
despite exclusion based on a BMI greater than 33.
Using the four-compartment model in the current study
leads us to similar conclusions as our previous comparison
of black and white premenopausal women (13). The two
populations were the sameheight, but the black women were
heavier, becauseof higher values for the compartments other
than fat. Similar regression curves (vs age) were obtained for
black and white women. For both black and white women,
the mineral ashwas curvilinear, while TBW and protein were
linear.
The body compartment that has been studied most extensively in the black population is bone mineral density. As
early as 18 months of age it is discernible that black children
have a higher bone mineral density than white children (10,
11). The early finding of a higher bone density and anthropomorphic differences in the two races suggest that whatever differences exist are probably influenced more by genetics than environment. Thus, black women probably
inherit a slightly denser musculoskeletal system than white
women and are therefore slightly heavier. The ratio of fat to
protein in black and white women was 2.52 and 2.48, respectively, in the current study. These values are similar,
thereby failing to indicate an excess of fat compared with
protein in black women. Admittedly, our study is not pop-
ET AL.
JCE C M . 1996
Vol81.
No 6
ulation based, and obesity was an exclusion factor. Nonetheless, an increase of body fat relative to the other body
compartments is clearly not a necessary characteristic of
body composition in black women.
Ortiz et al. (22) studied 28 pairs of black and white females
matched for age, weight, height, and menstrual status. They
found that the black women had greater appendicular skeletal muscle, bone mineral, and TBK. Although they found
slightly higher values for bone mineral, there were no differences in percent of fat. Ortiz et al. (22) concluded that the
racial differences in body composition probably involve inherited rather than environmental factors. Our findings by
adjusting the data for body size using the four-compartment
model are similar. The heavier weight in the black women in
our population was explained by mineral, protein, and water, and not fat.
The use of the body mass index or weight for height as
assessmentsof ponderosity may be somewhat misleading if
the samevalues are used for black and white women. Black
women have a slightly heavier musculoskeletal system,
which could be misinterpreted as excessfat, in body weight
measurements in particular. Thus, epidemiological studies
using the same criteria for whites and blacks may slightly
overestimate the prevalence of obesity in black women. Obesity may be defined as a BMI over 27 in white women,
becauseBMI of 27 roughly corresponds to the 75% value for
white women in the NHANES data base (23). Using the
equations in Table 4 to calculate the body composition of a
45-yr-old white woman with a BMI of 27 at 164 cm, we can
determine that she has 36.6% fat. A black woman under the
same conditions will have only 35.2% fat. If we use the
percent fat to define obesity as over 36.6% fat, then a black
woman can have a BMI of 28.7 before being declared obese.
Using our sample as an estimate of the distribution of BMI
in black women, we calculate that 14% will be wrongly
classified as obesewhen using the standards developed for
white women.
A dimensionof body composition that isnot consideredwith
whole body measurementsis regional distribution of the various components. This may have particular relevance in considering ethnic differences becausethere are anthropomorphic
differences between black and white women (24). For instance,
black women have a lessersitting height to stature ratio than
whites. Thus, blacks have longer extremities compared to their
trunk. The anthropomorphic difference between blacks and
whites is that although their upper and lower extremity length
is equal, the bra&al index differs (i.e.the forearm to upper arm
length is greater in blacksthan whites), with a similar difference
in the lower extremities. Becausemineral, protein, water, and
fat values may be affected by the geometry as well as the total
weight of the body, anthropomorphic differences must be considered. Analyses using dual-energy x-ray absorptiometry are
underway to resolve someof theseissues,at least the relative
mineral, lean, and fat tissue in the various body regions. A
critical caveat in interpreting thesedata is that this was a crosssectional study, and the changeswith age must be confirmed
by a longitudinal study becauseof secularchangeswith aging.
In summary, we present normative body composition data
in 112 healthy black women. The four-compartment model
was used with independent methods to assesseach com-
BODY COMPOSITION
partment. Body compartments
all changed with age and
must be adjusted for body size as well. Black women compared with white women were heavier because of an increase
in protein, water, and mineral with similar fat stores. Separate normative data should be used to classify black and
white women, including different values for field measurements such as BMI.
Acknowledgment
The
authors
thank
Nancy
Li,
MS.
for
help
in statistical
analysis.
References
1. Heymsfield
SB, Wang J, Lichtman
S, Kamen
Y, Kehayias
J, Pierson
RN Jr.
1989 Body composition
in elderly
subjects: a critical
appraisal
of clinical
methodology.
Am J Clin Nutr. 50:1167-1175.
2. Van Itallie TB, Segal KR, Yang MU, Funk RC. 1985 Clinical
assessmeni
of body
fat content in adults: potential
role of electrical
impedance
methods.
In: Roche AF,
ed. Body composition
assessments
in youth and adults. Report of 6th Ross Conference on Medical
Research.
Columbus,
OH: Ross Laboratories;
58.
3. Cohn SH, Parr RM. 1985 Nuclear-based
techniques
for the in vim study of
human body composition.
1985 Report of an Advisory
Group of the International Atomic
Energy
Agency.
Clin Phys Physiol
Meas. 6:275-301.
4. Burkinshaw
L, Hill GL, Morgan
DB. 1978 Assessment
of the distribution
of
protein
in the human
body by in vim neutron-activation
analysis.
Nuclear
activation
techniques
in the life sciences.
Proc Symp Vienna.
Vienna:
IAEA;
787-798.
5. Kehayias
JJ. 1993 Symposium:
aging and body composition:
technological
advances
and physiological
interrelationship.
Aging and body composition:
possibilities
for future
studies.
Nutr. 123(suppl.):454-458.
6. Kehayias
JJ, Heymsfield
SB, LoMonte
AF, Wang J, Pierson
RN. 1991 In vim
determination
of body fat by measuring
total body carbon.
Am J Clin Nutr.
531339-1344.
7. Vartsky
D, Ellis KJ, Cohn SH. 1978 In viva quantification
of body nitrogen
by
neutron
capture
prompt-gamma-ray
analysis.
J Nucl Med. 20~1158-1165.
8. Dilmanian
FA, Weber
DA, Yasumura
S, et al. 1990 Performance
of the delayedand prompt-gamma
neutron-activation
systems
at Brookhaven
National Laboratory.
In: Yasumura
S, et al, eds. Advances
in in vim body composition
studies.
New York: Plenum
Press; 309315.
2369
9. Trotter
M, Broman
GE, Peterson
RR. 1960 Densities
of bones of white and
Negro skeletons.
J Bone Joint Surg. 4250-58.
10. Gam SM. 1975 Bone loss and aging. In: Goldman
R, ed. Physiology
and
pathology
of human
aging. New York: Academic
Press; 39-57.
11. Garn SM, Clark DC. 1975 Nutrition,
growth,
development
and maturation:
findings
from the ten-state
survey
of 1968. Pediatrics.
56~306-319.
12. Cohn SH, Abesamis
C, Zanzi
I, Aloia JF, Yasumura
S, Ellis KJ. 1977 Body
elemental
composition:
comparison
between
black and white adults.
Am J
Physiol.
232:E419-E422.
13. Aloia JF, Vaswani
A, Ma R, Flaster
E. 1996 Comparison
of body composition
in black and white premenopausal
women.
J Lab Clin Med. In press.
14. Aloia
JF, Vaswani
A, Ma R, Flaster
E. 1996 Aging
in Women
- The four
compartment
model
of body composition.
Metabolism.
45:43-48.
15. Stamatelatos
IE, Dilmanian
FA, Ma R, et al. 1993 Calibration
for measuring
total body nitrogen
with a newly upgraded
prompt-gamma
neutron-activation
facility.
Phys Med Biol. 38:615-626.
16. Ma R, Dilmanian
FA, Rarback
HM, et al. 1993 Recent upgrade
of the in vim
neutron
activation
facility
at Brookhaven
National
Laboratory.
In: Ellis KJ, Eastman JD, eds. Human
body composition.
New York: Plenum
Press; 345-350.
17. Mazess RB, Barden
HS, Bisek JP, Hanson
J. 1990 Dual-energy
x-ray absorptiometry
for total-body
and regional
bone-mineral
and soft-tissue
composition.
Am J Clin Nutr. 51:1106-1112.
18. ICRP. 1975 Report of the iask group on reference
man. (The hydrogen
content
of ash is assumed
to be negligible).
ICRI’ Publication
23. Oxford:
l’ergamon
Press; 289.
19. Cohn SH, Vaswani
A, Zanzi I, Aloia JF, Roginsky
MS, Ellis KJ. 1976 Changes
in body chemical
composition
with age measured
by total-body
neutronactivation.
Metabolism.
25:85-89.
20. Keyere
K, Oldroyd
B, Oxby CB, Burkinshaw
L, Ellis RE, Hill GL. 1982 The
feasibility
of measuring
total body carbon by counting
neutron
inelastic
scatter
gamma rays. Whys Med Biol. 37~805-819.
21. Cohn SH, Abesamis
C, Zanzi I, Aloia JF, Yasumura
S, Ellis KJ. 1977 Comparative
skeletal
mass and radial
bone mineral
content
in black and white
women.
Metabolism.
26:171-178.
22. Ortiz 0, Russell
M, Aley 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. 558-13.
23. Shils ME, Young
VR. 1988 Modern
nutrition
in health and disease,
7th ed.
Philadelphia,
PA: Lea and Febiger.
24. Himes
JH. 1988 Racial variation
in physique
and body composition.
Can J
Sport Sci. 13:2:117-126.