Vol. 91, Issue 3, 1259-1268, September 2001
Effect of race and resistance training status on the density
of fat-free mass and percent fat estimates
Melinda L.
Millard-Stafford1,
Mitchell A.
Collins2,
Christopher M.
Modlesky3,
Teresa K.
Snow1, and
Linda B.
Rosskopf1
1 Exercise Physiology Laboratory, Department of Health and
Performance Sciences, Georgia Institute of Technology, Atlanta
30332-0356; 2 Department of Health, Physical Education, and
Sport Science, Kennesaw State University, Kennesaw 30144-5591; and
3 Department of Exercise Science and Department of Foods and
Nutrition, University of Georgia, Athens, Georgia 30602
 |
ABSTRACT |
The impact of race and
resistance training status on the assumed density of the fat-free mass
(DFFM) and estimates of body fatness via hydrodensitometry
(%FatD) vs. a four-component model (density, water,
mineral; %FatD,W,M) were determined in 45 men: white
controls (W; n = 15), black controls (B;
n = 15), and resistance-trained blacks (B-RT;
n = 15). Body density by hydrostatic weighing,
body water by deuterium dilution, and bone mineral by dual-energy X-ray absorptiometry were used to estimate %FatD,W,M.
DFFM was not different between B and W (or 1.1 g/ml);
however, DFFM in B-RT was significantly lower (1.091 ± 0.012 g/ml; P < 0.05). Therefore,
%FatD using the Siri equation was not different from
%FatD,W,M in W (17.5 ± 5.0 vs. 18.3 ± 5.4%)
or B (14.9 ± 5.6 vs. 15.7 ± 5.7%) but significantly overestimated %FatD,W,M in B-RT (14.0 ± 5.9 vs.
10.4 ± 6.0%; P < 0.05). The use of a
race-specific equation (assuming DFFM = 1.113 g/ml)
did not improve the agreement between %FatD and
%FatD,W,M, resulting in a significantly greater mean
(±SD) discrepancy for B (1.7 ± 1.8% fat) and B-RT (6.2 ± 4.3% fat). Thus race per se does not affect DFFM or
estimates of %FatD; however, B-RT have a DFFM
lower than 1.1 g/ml, leading to an overestimation of
%FatD.
body water; hydrodensitometry; bone mineral; body composition; Schutte equation; blacks
| |
INTRODUCTION |
BODY COMPOSITION
ASSESSMENT is a valuable tool to establish desirable body weight
ranges and to track changes that occur with training. An indirect
method widely used to assess body composition is densitometry
(16). Estimates of body fatness (%fat) from body density
(Db) (%FatD) are based on a two-component
model in which the densities of the fat mass and fat-free mass (FFM)
are considered known and constant. The commonly used equation for estimating %fat developed by Siri (38) assumes that the
densities of these two components are 0.9 and 1.1 g/ml, respectively.
The density of fat was determined from adipose tissue samples in humans (12). Moreover, the density of the FFM (DFFM)
is based on the assumed relative proportions of water, protein, and
mineral (73.8, 19.4, and 6.8%, respectively) and each of their
respective densities (e.g., 0.9937, 1.34, and 3.038 g/ml at 36°C)
(3). The presumed proportional density and composition of
the FFM were determined in white men (3); thus they may
vary in other subgroups of the population, resulting in considerable
error when %fat is estimated using %FatD.
The development of multicomponent models, in which Db
measurements are combined with water and/or mineral, has allowed
researchers to test the validity of the assumed composition of FFM and
DFFM (15). One group in which the
DFFM appears to be different than the reference value of
1.1 g/ml is weight trainers with high musculoskeletal development.
Recent studies using a four-component model (density, water, mineral)
as a criterion (29, 47) concluded that DFFM is
significantly lower in white weight trainers (<1.1 g/ml), resulting in
an overestimation of %fat when densitometry and the Siri equation (38) (%FatD-Siri) are used. The lower
DFFM in resistance-trained men was based on significantly
higher water (W/FFM) and lower mineral (M/FFM) and protein (P/FFM)
fractions of the FFM compared with male controls (29).
Because of the increased use of weight training in athletic programs
and the rising proportion of black athletes in collegiate and
professional sports, it is important to determine the DFFM in resistance-trained black men so that appropriate equations can be
used to estimate %FatD. Whether resistance-trained black men exhibit a DFFM similar to their white counterparts is
difficult to predict. It has been theorized that blacks have a higher
DFFM than whites, because greater skeletal muscle, bone
mineral mass, and bone density have been reported (7, 9, 32, 37,
45). This circumstantial evidence has led to the
development and recommendation of race-specific equations (6, 19,
20, 36, 44, 45) in favor of traditional equations (3,
38) that convert Db to %fat. The most widely used
equation for %FatD estimation in blacks was developed by
Schutte et al. (36) and postulates that the
DFFM is 1.113 g/ml based on calculations derived from
hydrometry and hydrodensitometry. To our knowledge, this equation has
not been adequately cross-validated with a four-component model. In fact, the few published reports that have assessed DFFM in
blacks with a multicomponent model are at odds. Only two investigations support the theory of a higher DFFM in blacks (32,
45), whereas others suggest that there is no difference in the
DFFM between blacks and whites (9, 43). No
studies have simultaneously evaluated the impact that both race and
resistance training status have on DFFM in men.
Considering that weight training and high degrees of muscularity are
linked to a lower DFFM and average black men may or may not
have an elevated DFFM, predicting their combined effects is difficult. If the DFFM in black men is significantly >1.1
g/ml, the opposing effects of resistance training status and race may counteract each other so that the DFFM in black resistance
trainers would be similar to that of white controls (1.1 g/ml). If that is the case, the Siri equation would provide a more accurate estimate of %FatD than the race-specific equation
(36). However, if DFFM is independent of race
and <1.1 g/ml in black weight trainers, consistent with results
observed in white weight trainers (29, 47), then
%FatD would be overestimated by both equations.
Therefore, the present investigation had two purposes. First, our aim
was to determine whether black men who did not participate in
resistance training (B) had a DFFM different from their
white counterparts (W) and the reference value of 1.1 g/ml. We
hypothesized that the DFFM of B would not be different from
that of W or 1.1 g/ml, resulting in a more accurate estimate of
%fat estimated by the four-component model
(%FatD,W,M) when %FatD-Siri equation is
used compared with the race-specific equation developed by Schutte et
al. (36) (%FatD-Schutte). Our second purpose
was to determine whether resistance-trained black men with high
musculoskeletal development (B-RT) have a DFFM lower than
that of B and than 1.1 g/ml. We hypothesized that B-RT would have
a higher W/FFM and lower M/FFM and P/FFM, leading to a
lower DFFM than that of B. Consequently,
%FatD-Siri and %FatD-Schutte would
overestimate %FatD,W,M in B-RT.
| |
METHODS |
Subjects.
Subjects were recruited from the student population as well as from
local bodybuilding competitions. Forty-five men were assigned to three
groups: W, B, and B-RT. Race was determined via self-report. All
subjects reported that both parents were of the same race except for
two subjects who indicated 50% black heritage (one each in B and
B-RT). The criteria for B-RT selection was involvement in a
weight-training program (3 days/wk minimum) for at least 2 yr before
the study and a mesomorphy (Meso) rating >5.5, which was ~1.5 points
above "average" Meso (5). Eight members of the group
were competitive bodybuilders (one acknowledged anabolic steroid use
but cessation 6 mo before the study) and the other seven were
experienced in resistance training. The mean (minimum to maximum)
quantity of resistance training for B-RT was 4 (3-6) days/wk, 86.8 (30) min/session, and 9 (2-22) yr of resistance training experience. Members
of W and B were physically active but never involved in a resistance
training program. All subjects gave written consent in accordance with
the policies established by the Institutional Review Board for the use
of human subjects.
Data collection protocol.
All testing was completed during a single session with subjects
reporting to the laboratory in a euhydrated condition after a 12-h
fast. A baseline urine specimen was obtained to measure urine-specific
gravity with a handheld gravitometer (indicative of hydration status).
Based on a normal (1) baseline urine-specific gravity
(mean ± SD = 1.021 ± 0.02 g/ml) and the ability to
produce urine specimens during the test period, subjects were
considered to be adequately hydrated. No exercise was performed by the
subjects 12 h before testing.
Musculoskeletal development.
Body mass in air was determined on an electronic digital scale to the
nearest 0.01 kg, and height was obtained via a stadiometer. Meso was
assessed via the Heath-Carter anthropometric somatotype equation
(5), which utilizes upper arm and calf circumferences, corrected for skinfold thickness, and elbow and knee joint widths to
provide an index of musculoskeletal development relative to height. The
following equation was utilized: Meso = 0.858 (humerus biepichondylar width) + 0.601 (femur bichondylar width) + 0.188 (upper arm girth corrected for skinfold thickness) + 0.161 (calf girth corrected for skinfold thickness) 
0.131 (height) + 4.5. The Meso score was used as an index of
musculoskeletal development because it is independent of body
composition measures and based on the muscle and bone dimensions
relative to height. As an additional index, the FFM relative to
height (FFM/ht2) was also calculated according to
VanItallie et al. (42). The anthropometric measurements
were performed by the same experienced technician for all subjects.
Densitometry.
Db was determined via hydrostatic weighing using a
custom-built, stainless steel tank to measure body volume based on
Archimedes' principle (16). Weight under water was
measured at residual lung volume (RV) by using a Chatillon autopsy
scale to the nearest 0.025 kg. RV was measured simultaneously by using
the standard oxygen-rebreathing nitrogen-dilution technique modified
from Goldman and Buskirk (17). Nitrogen was measured using
a Med Science 505 nitralizer. Db was computed using mass
and volume with corrections for water density, RV, and gastrointestinal
tract gas volume (0.1 liter). The equations of Siri (38)
and Schutte et al. (36) were utilized to estimate %fat
(via a two-component densitometric model) from Db
(%FatD-Siri and %FatD-Schutte, respectively).
Our laboratory's previously published test-retest reliability
(n = 16) for assessing Db was
r = 0.99 (40). The technical error of
measurement defined as the within-subject standard deviation was 0.001 g/ml (40).
Dual-energy X-ray absorptiometry.
Total body bone mineral content, bone mineral density
(BMD), and %Fat (%FatDXA) were determined from
whole body scans using Lunar DPX-L dual-energy X-ray
absorptiometry (DXA) (Madison, WI; software version 1.3Z; medium mode,
3,000 µA). To ensure quality control, the DXA unit was calibrated on
a daily basis using the standard calibration block provided by the
manufacturer. The calibration block was made of a thermoplastic acrylic
resin that contained three bone-equivalent chambers filled with
hydroxyapatite. All scans were performed and analyzed by two trained
technicians. Our test-retest reliability of the DXA (n = 7) for assessing %fat was r = 0.99 with a technical
error of measurement of 0.4% (8).
DXA was assumed to measure bone ash, which is the total bone mineral
minus volatile components lost from ashing (water of crystallization
and CO2 from carbonate) (18). Bone mineral ash was multiplied by 1.27 to estimate total body mineral content. The
constant 1.27 assumes that 4% of bone mineral is lost during the
ashing process and that nonosseous mineral mass is 23% of bone mineral
ash (3).
Total body water.
Total body water was measured by using deuterium oxide dilution as
previously described (10, 24). After a baseline blood sample, subjects ingested a dosage of deuterium oxide equivalent to 0.3 g/kg body wt in 100 ml of distilled water. A second blood plasma sample
was obtained after a 3-h equilibration period. After centrifugation,
blood plasma samples were stored at 70°C until purification. To
purify the plasma samples, equal volumes (1.5 ml) of plasma and
deionized water were incubated at 37°C for 48 h in Conway
diffusion dishes (Bel-Air Products, Pequannock, NJ). Absorbance of the
purified water samples was analyzed by using a single-beam infrared
spectrophotometer with a 4-µm fixed filter (Miran 1FF, Foxboro,
Foxboro, MA) at 16°C. Total body water was corrected for the
deuterium lost in urine during the 3-h equilibration period and reduced
by 4% to account for hydrogen ion exchange during equilibration
(35). The within-subjects SD of duplicate measures of body
water obtained within 1 wk in five subjects was 0.75 liter
(29).
Four-component model.
The equation for estimating %fat from a four-component model
(%FatD,W,M) was derived from the relation of
Db to its primary chemical constituents according to the
following
where F, W, M, and P represent fat, water, mineral, and protein
fractions of body mass and DF, DW,
DM, and DP represent each component's density,
respectively (38). The assumed densities were 0.9007, 0.9937, 3.038, and 1.34 g/ml for DF, DW,
DM, and DP, respectively.
The criterion method for %fat estimates was %FatD,W,M,
which was estimated from Db, body water, and body mineral
based on the following equation of Lohman (22)
The protein content of the FFM was calculated by the difference
(protein = FFM water mass mineral mass) based on
the FFM calculated from %FatD,W,M and body mass.
DFFM, W/FFM, M/FFM, and P/FFM were calculated using
formulas previously described (11, 29). The within-subject
SD of repeated measures of DFFM in five subjects tested
within a 1-wk time interval was 0.002 g/ml (29).
Statistical analysis.
Statistical analyses were done with SAS for Windows version 6.12 (SAS
Institute, Cary, NC). For comparisons of %fat estimates, the data were
analyzed by using a two-way (group × method) ANOVA with repeated
measures. Post hoc simple contrasts were used to determine differences
between cell means. The Bonferroni adjustment was used for the family
of contrasts performed with an alpha level of 0.05. For the other
between-group comparisons of body mass composition based on a
four-component model and physical characteristics, a one-way ANOVA was
used with Tukey post hoc tests. Relationships between variables were
described using linear regression analysis and Pearson correlation
coefficients. Agreement between methods for estimating %fat was
determined using a Bland-Altman plot (2). An alpha level
of 0.05 was used for all other significance testing. All values
reported are means ± SD unless otherwise noted.
| |
RESULTS |
Mean subject physical characteristics of W, B, and B-RT are
presented in Table 1. W and B were not
different in age, height, or weight; however, BR-T subjects were older
(by 5 and 7 yr) and heavier (by 12.6 and 9.4 kg) than W and B,
respectively. Meso and FFM/ht2 were not significantly
different (P > 0.05) between W and B; however, as
expected, B-RT had significantly higher Meso and FFM/ht2
compared with W and B. The minimum and maximum values for Meso were
2.2-5.4 (W), 2.8-6.1 (B), and 5.7-9.3 (B-RT). The
overlap in Meso for B and B-RT was due to one short B control subject (Meso = 6.1) and one tall muscular body builder (Meso = 5.7); otherwise, all other B and B-RT subjects had Meso <5.5 and >6.3, respectively. B-RT had significantly higher BMD compared with W and B,
and B had greater BMD than did W. Db was not significantly different among the three groups.
Estimates of the body mass composition based on the four-component
model are presented in Table 2. There was
no significant difference in fat mass or protein mass among the groups.
B-RT had greater (P < 0.05) FFM and water mass
compared with B and W. B-RT and B had similar body mineral mass, and
both groups had values significantly higher than those of W.
The relative composition of the FFM derived from the four-component
model is presented in Table 3. There was
no difference (P > 0.05) in DFFM between W
and B, with values similar to the reference value of 1.1 g/ml
(3). However, DFFM for B-RT was significantly
lower (P < 0.05) than for W and B and than 1.1 g/ml. The lower DFFM in B-RT compared with B was due to a higher
W/FFM and lower M/FFM. The lower DFFM in B-RT compared with
W was due to a significantly higher W/FFM and lower P/FFM.
Estimates of %fat compared with the criterion (%FatD,W,M)
are presented in Table 4. Analysis
of variance indicated a significant group × method
interaction. Based on tests for simple main effects (between
group), there was no difference in %FatD-Siri or
%FatDXA among the three groups and no difference
in %FatD-Schutte between B and B-RT. However, B-RT had
significantly lower (P < 0.05) %FatD,W,M than did W and B (by 7.9 and 5.3% fat, respectively) and so were leaner according to the criterion (%FatD,W,M). Tests for
simple main effects (within group) also indicated a
difference among the methods in all groups.
%FatD-Siri was not significantly different (P > 0.05) from %FatD,W,M in W. However, %FatDXA was significantly lower than
%FatD,W,M and %FatD-Siri. In B,
%FatD-Siri was not significantly different from
%FatD,W,M or %FatDXA. However, the %FatD-Schutte was significantly higher in B
(P < 0.001) compared with all other %fat estimates
(1.7 ± 1.8% higher compared with %FatD,W,M). In
B-RT, %FatD-Siri was significantly higher (by 3.6 ± 4.5% fat) compared with %FatD,W,M (P < 0.05) but was not different from %FatDXA. Moreover, the
discrepancy between %FatD-Schutte and
%FatD,W,M was even greater (by 6.2 ± 4.3% fat) in
B-RT (P 
0.0001).
A scatterplot of the relationship between %FatD-Siri and
%FatD,W,M is illustrated in Fig.
1A. The correlation for the
two estimates was significant (P < 0.05) for W
[r = 0.90, y = 0.97x + 1.31, SE of estimate (SEE) = 2.4%], B (r = 0.95, y = 0.96x + 1.31, SEE = 1.8%),
and B-RT (r = 0.70, y = 0.72x + 0.25, SEE = 4.4%). A scatterplot of the
relationship between %FatD-Schutte and
%FatD,W,M is illustrated in Fig. 1B. The
correlation for the two estimates was significant (P < 0.05) for B (r = 0.95, y = 1.09x 3.31, SEE = 1.8%) and B-RT
(r = 0.70, y = 0.82x 3.30, SEE = 4.4%) (equation not used for W). On comparing the two
graphs (Fig. 1, A and B), the Schutte et al.
(36) equation yields a higher %FatD compared
with that of Siri (38), especially at the lower range of
%fat estimates.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 1.
A: relationship between estimates of body
fatness (%fat) from densitometry and the Siri equation
(38) (%FatD-Siri) and criterion method
estimated from body density (Db), body water, and body
mineral (%FatD,W,M) in whites (W), blacks (B), and
resistance-trained blacks (B-RT) (n = 15 per
group). B: relationship between %fat from densitometry and
the Schutte et al. equation (36)
(%FatD-Schutte) and %FatD,W,M. Note that W
are included (without using the race-specific equation), and so the
data points are identical to those in A for this group.
Lines extended through axes, line of identity; short lines,
group-specific regression lines.
|
|
Individual differences between the criterion method
(%FatD,W,M) and the other %fat estimates
(%FatD-Siri, %FatD-Schutte,
%FatDXA) are illustrated in Fig.
2, A, B, and
C, respectively. The mean differences between
%FatD-Siri and %FatD,W,M were 0.7, 0.8, and 3.6% fat for W, B, and B-RT, respectively. Individual differences ranged from 3.1 to 4.0% fat in W, 2.5 to 4.1% fat in B, and 10.6 to 3.2% fat in B-RT. The mean differences between
%FatD-Schutte and %FatD,W,M were 1.7 and
6.2% fat for B and B-RT, respectively. The individual differences
ranged from 4.7 to 1.2% fat in B and from 12.6 to 0.5% fat in
B-RT. The mean differences between %FatDXA and
%FatD,W,M were 2.6, 2.2, and 2.2% fat for W, B, and
B-RT, respectively. The individual differences ranged from 2.7 to
6.8% fat in W, from 2.1 to 7.1% fat in B, and from 9.6 to 3.3%
fat in B-RT. There was a weak relationship (r = 0.25, P = 0.1) between the differences in
%FatD-Siri and %FatD,W,M at specific levels of body fatness (Fig. 2A). The correlation for the
difference between methods and level of %fat (Fig. 2, B and
C, respectively) was significant for
%FatD-Schutte (r = 0.39, P = 0.0075) but not for %FatDXA (r = 0.06, P = 0.71).
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 2.
Agreement between %FatD,W,M and
%FatD-Siri (A),
%FatD-Schutte (B), and %fat via
dual-energy X-ray absorptiometry (%FatDXA; C).
Subjects are W, B, and B-RT (n = 15 per group).
|
|
Figure 3 illustrates the relationship
between DFFM and the difference between
%FatD-Siri and %FatD,W,M. The correlation
between the %FatD,W,M %FatD-Siri
difference and DFFM for all subjects was essentially 1.0 (r = 0.996; y = 360x 396, SEE = 0.3%). Consequently, the difference between
%FatD,W,M and %FatD-Siri is
explained by the variability in DFFM. The correlation
between the %FatD,W,M %FatD-Siri
difference and W/FFM (Fig. 4) was also
significant (P < 0.05) (r = 0.96;
y = 1.36x + 100.5, SEE = 1.1%).
The correlation between the %FatD,W,M %FatD-Siri difference and M/FFM (Fig. 5) was not as strong but was significant
(r = 0.74; y = 4.0x 28.0, SEE = 2.5%).
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 3.
Relationship between the density of the fat-free mass
(DFFM) and the difference in %FatD,W,M and
%FatD-Siri in W, B, and B-RT (n = 15 per
group).
|
|
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 4.
Relationship between the water fraction of FFM and the
difference in %FatD,W,M and %FatD-Siri in W,
B, and B-RT (n = 15 per group). DFFM is
show on right axis for reference.
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 5.
Relationship between the mineral fraction of FFM and the
difference in %FatD,W,M and %FatD-Siri in W,
B, and B-RT (n = 15 per group). DFFM is
show on right axis for reference.
|
|
Figure 6 illustrates that Meso was
significantly related (r = 0.46, y = 1.0x + 5.0, SEE = 3.3%) to the difference
between %FatD,W,M and %FatD-Siri. However,
BMD was not significantly (r = 0.08,
P > 0.05) correlated to the difference between
%FatD,W,M and %FatD-Siri (Fig.
7).
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 6.
Relationship between mesomorphy rating and the difference
in %FatD,W,M and %FatD-Siri in W, B, and B-RT
(n = 15 per group).
|
|
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 7.
Relationship between bone mineral density (BMD) and the
difference in %FatD,W,M and %FatD-Siri in W,
B, and B-RT (n = 15 per group).
|
|
| |
DISCUSSION |
The present study addressed two issues. First, despite the
widespread notion that DFFM is elevated in blacks,
DFFM in B was not different from that in W or from 1.1 g/ml. Therefore, %FatD-Siri was not different from
%FatD,W,M but %FatD-Schutte overestimated %FatD,W,M. These data suggest that the Siri equation
(38), rather than a race-specific equation, should be used
to estimate %FatD in black men who do not engage in
resistance training. The second finding was that DFFM in
B-RT was lower than DFFM in W and B and the reference value
of 1.1 g/ml. Consequently, %FatD-Siri overestimated %FatD,W,M by 3.6% fat, and %FatD-Schutte
overestimated %FatD,W,M by an additional 2.6% (6.2%
fat). These data support recent investigations that collectively
suggest that race per se (43) does not necessitate the use
of population-specific equations to accurately estimate %FatD but that population-specific equations for male
resistance trainers may be needed (29, 47).
Effect of race on DFFM.
The findings of the present study parallel those of Visser et al.
(43) that DFFM is independent of race and are
in contrast to previous studies (32, 36, 45) that suggest
that race impacts DFFM (>1.1 g/ml for black men). The
DFFM in our B controls was not significantly different from
that in W or from 1.1 g/ml. In a large-scale (n = 703)
study, Visser et al. (43) observed the DFFM of
black men (aged 20-94 yr) to be 1.099 g/ml and not different from
that of whites (1.098 g/ml). A major strength of this study was the
large number of black men (n = 98) and women (n = 96) studied, thereby reducing potential sampling
error. Similar to our findings, Visser et al. also observed higher
M/FFM (7.0 vs. 6.6% FFM) in young adult black men (20-40 yr)
compared with whites but no difference in the W/FFM (74.2%).
In large part, the theory that DFFM is elevated in blacks
hinges on the understanding that blacks have greater bone mass compared with whites. Previous data indicated that blacks have a dry fat-free skeleton that is 10-20% heavier than that of whites (28,
37, 41). Of the major chemical constituents of the FFM (water,
mineral, and protein), mineral has the highest density. Thus deviations in bone mineral and M/FFM were hypothesized to have a profound effect
on DFFM. Yet mineral differences of the magnitude
demonstrated between blacks and whites in the present study
(difference = 0.4% for M/FFM) have a minimal effect on
DFFM and %FatD (<1% fat; Ref. 30). Schutte et al. (36) tested black
(n = 15) and white men (n = 19) similar
in age, height, and weight and observed that Db tended to
be higher (P > 0.05) in blacks, despite no difference in body water and skinfold measurements. However, bone mineral was not
measured in that study. Assuming water mass is not different in blacks
and whites, M/FFM must be nearly twice as high in blacks to produce the
increased DFFM (1.113 g/ml) hypothesized by Schutte et al.
These calculations further support the present findings that
DFFM in black men is not significantly different from the DFFM in white men.
It has been hypothesized that a greater proportion of skeletal muscle
mass in blacks may also contribute to an increased DFFM (18, 32, 44). However, this theory is contradicted by the present study as well as that of Visser et al. (43). The
FFM, total protein, and water mass were not different between our two control groups of B and W. Because skeletal muscle is primarily water
(~74%), its density is slightly <1.1 g/ml (~1.066 g/ml for fat-free muscle). Therefore, if the FFM were composed of a higher proportion of skeletal muscle in blacks, then DFFM would,
in fact, be reduced. This rationale is supported by the present
findings in the muscular B-RT. It has been suggested that the skeletal muscle of black men may have a higher concentration of protein (44), resulting in a higher DFFM. If the
protein concentration of muscle is elevated in blacks, P/FFM should be
elevated as well. However, P/FFM was not higher for B compared with W
in the present study, and a moderate effect size (Cohen's delta = 0.58) suggests that it was actually lower. These findings are
consistent with those of Visser et al., who found P/FFM to be lower in
blacks compared with whites and suggest that skeletal muscle is not
more concentrated in black compared with white men. The hypothesis that
skeletal muscle has a higher protein concentration leading to a denser
muscle in blacks requires further investigation with more direct
measures of protein.
Effect of resistance training status on DFFM.
Until recently, the DFFM in resistance trainers was
unknown. It was suggested that the DFFM is reduced in
resistance-trained men because of a proportionately larger contribution
of muscle than bone to the FFM (48). Alternatively,
because the density is inherently greater (2.982 g/ml) and weight
training increases bone mineralization (39), others
suggested that the DFFM in resistance-trained men is >1.1
g/ml (26). Although recent studies confirmed that the
DFFM is <1.1 g/ml in white male resistance trainers
(29, 47), the DFFM in black resistance
trainers had not been determined. Because the DFFM in black
men was controversial (36, 43, 45), it was difficult to
predict DFFM in black resistance trainers. The present
study suggests that male B-RT have a DFFM < 1.1 g/ml.
These findings are similar to observations in white resistance trainers
(29, 46) and other muscular male athletes involved in
resistance training (33). Modlesky et al.
(29) observed a significantly lower DFFM in
white resistance trainers compared with controls (1.089 vs. 1.099 g/ml)
based on higher W/FFM (74.8% FFM) and reduced M/FFM and P/FFM. The
result was %FatD-Siri overestimated %FatD,W,M
by ~4% fat. These findings have been confirmed by Withers et al.
(47), who reported a low DFFM (1.094 g/ml) in
three male bodybuilders (Meso = 8.6), resulting in
%FatD overestimating %FatD,W,M by 2.1% fat.
Similarly, Prior et al. (33) observed a DFFM
of 1.092 g/ml in muscular football players (black and white); however, some other resistance-trained athletes did not display a lower DFFM (e.g., female gymnasts). These findings tend to
support the theory that participation in resistance training results in
a dilutional effect on the FFM and skeletal muscle in men;
consequently, muscular male weight trainers may have a reduced
DFFM, leading to an overestimation of %FatD
when a DFFM of 1.1 g/ml is assumed. Yet, whether the
DFFM is reduced in all resistance-trained athletes of both
genders (participating in various sports) is not entirely clear.
Although the DFFM observed in B-RT (1.091 g/ml) was similar
to the DFFM previously observed in white weight trainers
(29), it was higher by 0.003 g/ml. Consequently, the
deviation from 1.1 g/ml was less and the overestimation of
%FatD-Siri was slightly smaller (3.6 vs. 4.1% fat). These
subtle differences are likely linked to the type of densitometer used
(30, 34) and the assumptions regarding body water
measurement from deuterium oxide. Bone mineral estimates from the
densitometer used in the present study (Lunar DPX-L) are known to be
~11% higher than estimates from the densitometer (Hologic QDR 1,000 W) used in the study of white weight trainers (29).
Furthermore, body water measurements from dilution were reduced by 4%
body mass to account for hydrogen exchange with protein and
carbohydrate, whereas a 2% correction was applied in the study of
white weight trainers. If the same densitometer and corrections
employed by Modlesky et al. (29) were used in the present
study, DFFM in the resistance trainers would equal 1.089 g/ml, which is identical to the results observed in white weight
trainers. Another difference in the present study compared with the
study of whites (29) was a small age discrepancy (7 yr)
between B-RT and B. However, age does not appear to affect DFFM in men (43), particularly as all subjects
were young adults. The observation that the difference in
DFFM between W and B-RT was the same as that observed
between white controls and white weight trainers in the Modlesky et al.
study illustrates the small effect of densitometer type, hydrogen
exchange correction factor, and slight age differential on estimates of
DFFM.
The reason for the lower DFFM in B-RT than assumed is
primarily related to W/FFM and secondarily to lower M/FFM and P/FFM compared with controls. The higher W/FFM and lower M/FFM are plausible considering that muscle is predominantly water and weight training may
induce a proportionately greater accrual of muscle than bone mineral.
Skeletal muscle and cross-sectional areas can increase 15-40%
with training (25, 27), whereas the changes in BMD and
mineral mass are ~1-3% (21). Water also represents
an ~11-fold greater proportion of the FFM compared with mineral
(30). It has been observed that weight training increases
total body water (4). Other potential contributors to an
increased W/FFM associated with training include increased
extracellular fluid volume (46) and increased water stored
with glycogen (31). However, because these measurements
were not obtained, their influence is not known.
Implications of using a two-component model.
The major practical implications from the present study are that
densitometry via %FatD-Siri can give a reasonably accurate estimate of %fat in black men (mean difference = 0.8% fat,
SEE = 1.8%) and that the use of a race-specific equation is not
justified. This is consistent with other recent studies in the
literature (9, 43). Cote and Adams (9) found
no difference between %FatD-Siri and
%FatD,W,M in black women. Visser et al.
(43) also reported no significant difference between
mean %FatD-Siri and %FatD,W,M in blacks
(0.2%) or whites (1.0%). We observed that %fat is overestimated by
2.7% when the race-specific equation proposed by Schutte et al.
(36) is employed. In black weight trainers with high
musculoskeletal development (and a lower DFFM), neither the
Siri (38) nor Schutte et al. (36) equation
accurately estimate %FatD, with overestimations of
3.6 ± 4.5 and 6.2 ± 4.3% fat, respectively.
Because the lower DFFM in B-RT subjects is reflected in
their measured total Db, the relation between measured
Db and total body fatness will be different in individuals
who do not weight train and different from values assumed by Siri
(38). The two-component model (on which densitometry is
based) may, therefore, result in inaccuracies that would ultimately
lead to inappropriate recommendations to reduce body weight and fat.
The SEE (4.4%) for %FatD-Schutte and
%FatD-Siri in B-RT compared with the criterion was above
the level suggested by Lohman (23) for acceptance (<3%).
Based on the Bland-Altman plots, the key difference between estimates
from the two equations is the systematic error associated with the %FatD-Schutte estimates (which appears to be greatest at
the lower levels of %fat). This is because B-RT had a greater
discrepancy between %FatD and %FatD,W,M
(because of their lower than normal DFFM) and lower
%FatD,W,M than the control groups. Based on the results of
this study as well as those of Visser et al. (43), the use
of the Schutte et al. (36) equation in the estimation of
%fat from Db is not recommended. The Siri equation
(38) is more suitable for blacks who do not resistance
train (and have average musculoskeletal development) but may not be
valid for other population groups (e.g., Asian).
The use of a race-specific %fat equation [as proposed by Schutte et
al. (36)] is based on the unsubstantiated theory that black men have a higher DFFM. Furthermore, race-specific
equations have been frequently employed in the literature to validate
other measures of body composition in blacks (6, 20).
However, the methods used by Schutte et al. were insufficient to test
the validity of the DFFM in blacks as mentioned previously.
Estimates of %fat from body water do not account for variations in
M/FFM and P/FFM, which can result in substantial error. Because their sample sizes were small and the findings conflict with recent studies,
which have utilized a four-component model, it is difficult to justify
the use of the Schutte et al. equation or other equations specific to blacks.
The concept of employing race-specific equations to estimate
%FatD in black men continues to persist in the literature,
however. A recent investigation by Wagner and Heyward (45)
suggests the DFFM in black men is 1.1057 g/ml, and the
authors recommended a new %fat equation based on this value (44, 45).
The proposed DFFM and %fat equation were based on body
composition estimates from a four-component model in a sample of 30 black men. Their recommendations may have been premature for two
reasons. First, because the study of black men did not include a
control group of white men for reference purposes (13,
14), a systematic error in the measurement of DFFM
cannot be discounted. Second, when all recent studies (9, 32, 43,
45) are considered, the concept that DFFM is higher
in blacks compared with whites (and >1.1 g/ml) is not consistently supported.
Alternatively, it has been hypothesized that an index of
musculoskeletal development may improve estimates of
%FatD. The theory was based on observations of an inverse
relationship between DFFM and musculoskeletal development
(r = 0.64) in muscular weight trainers and controls
with average musculoskeletal development. In the present study,
DFFM was also significantly related (r = 0.44) to Meso ratings in the total subject pool (n = 45).
Using musculoskeletal development as a criteria to improve estimates of
%FatD (and a two-component model) is not without its pitfalls, however. Prior et al. (33) indicated that
DFFM was poorly related to musculoskeletal development in a
large group (n = 172) of male and female athletes and
nonathletes. Football players, who engage heavily in weight training,
had a lower DFFM (1.092 g/ml), and thus %FatD
overestimated %FatD,W,M by 3.5%. However,
DFFM was poorly related to musculoskeletal development (r = 0.14) in a large group of athletes of different
sports and genders (as well as nonathletes), suggesting that the
relationship between Meso and DFFM is not strong across a
heterogenous population. The deviations in the DFFM of
athletes are complex and not simply a reflection of differences in
musculoskeletal development (assessed via an anthropometric
measurement). Deviations in DFFM among different types of
athletes may be linked to their varied training regimens, and thus
changes in the chemical constituents of the FFM due to resistance
training may be different from when they are combined with other
physical activities. Further research is needed to determine whether
different types of training have unique effects on the constituents of
the FFM and, consequently, DFFM.
An alternative equation (%FatD = 521/Db 478) based on the low DFFM
(1.089 g/ml) observed in white resistance trainers has been proposed to
provide better estimates of %fat in resistance trainers
(29). Although systematic error of %FatD in
relation to %FatD,W,M was greatly reduced (mean %fat = 10.3 vs. 10.4% for %FatD,W,M) when the proposed
equation was used compared with traditional equations, large
variability between the two estimates remained (SEE = 4.4% fat).
These results suggest that the %fat in resistance trainers cannot be
accurately estimated from Db alone.
In conclusion, the present study suggests that DFFM in B is
not different from that in W or from 1.1 g/ml. Thus %fat is more accurately estimated when Db is used in conjunction with
the Siri equation (38) rather than a race-specific
equation developed by Schutte et al. (36). Furthermore,
B-RT with high musculoskeletal development, like their white
counterparts, appear to have a DFFM <1.1 g/ml, due
primarily to higher W/FFM and secondarily to lower M/FFM and P/FFM. The
result is an overestimation of %fat from Db and the Siri
equation (38). The %fat overestimation is exacerbated when Db is used in conjunction with the race-specific equation.
| |
FOOTNOTES |
Address for reprint requests and other correspondence: M. L. Millard-Stafford, Dept. of Health and Performance Sciences,
Georgia Institute of Technology, Atlanta, GA 30332-0356 (E-mail:
mindy.millardstafford{at}hps.gatech.edu).
The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement"
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Received 16 January 2001; accepted in final form 21 May 2001.
| |
REFERENCES |
1.
Armstrong, LE,
Maresh CM,
Castellani JW,
Bergeron MF,
Kenefick RW,
LaGasse KE,
and
Riebe D.
Urinary indices of hydration status.
Int J Sport Nutr
4:
265-279,
1994[ISI][Medline].
2.
Bland, JM,
and
Altman DG.
Statistical methods for assessing agreement between two methods of clinical measurement.
Lancet
1:
307-310,
1986[ISI][Medline].
3.
Brozek, J,
Grande F,
Anderson JT,
and
Keys A.
Densitometric analysis of body composition: revision of some quantitative assumptions.
Ann NY Acad Sci
110:
1133-1140,
1963.
4.
Campbell, WC,
Crim MC,
Young VR,
and
Evans WJ.
Increased energy requirements and changes in body composition with resistance training in older adults.
Am J Clin Nutr
60:
167-175,
1994[Abstract/Free Full Text].
5.
Carter, JEL,
and
Heath BH.
Somatotyping-Development and Applications. Cambridge, UK: Cambridge Unversity Press, 1990, p. 374.
6.
Clark, RR,
Kuta JM,
and
Sullivan JC.
Cross validation of methods to predict body fat in African-American and Caucasian collegiate football players.
Res Q Exerc Sport
65:
21-30,
1994[ISI][Medline].
7.
Cohn, SH,
Abesamis C,
Zanzi I,
Aloia JF,
Yasumura S,
and
Ellis KJ.
Body elemental composition: comparison between black and white adults.
Am J Physiol Endocrinol Metab Gastrointest Physiol
232:
E419-E422,
1977[Abstract/Free Full Text].
8.
Collins, MA,
Millard-Stafford ML,
Sparling PB,
Snow TK,
Rosskopf LB,
Webb SA,
and
Omer J.
Evaluation of the BOD POD for assessing body fat in collegiate football players.
Med Sci Sports Exerc
31:
1350-1356,
1999[ISI][Medline].
9.
Cote, KD,
and
Adams WC.
Effect of bone density on body composition estimates in young adult black and white women.
Med Sci Sports Exerc
25:
290-296,
1993[ISI][Medline].
10.
DeLuca, JP,
and
Friedl KE.
Validation of a simplified method of total body water measurement using deuterium oxide and diffusion dish purification.
Med Sci Sports Exerc
24:
S8,
1992.
11.
Evans, EM,
Saunders MJ,
Spano MA,
Arngrimsson SA,
Lewis RD,
and
Cureton KJ.
Effects of diet and exercise on the density and composition of the fat-free mass in obese women.
Med Sci Sports Exerc
31:
1778-1787,
1999[ISI][Medline].
12.
Fidanza, F,
Keys A,
and
Anderson JT.
Density of body fat in man and other mammals.
J Appl Physiol
6:
252-256,
1953[Free Full Text].
13.
Friedl, KE,
DeLuca JP,
Marchitelli LJ,
and
Vogel JA.
Reliability of body-fat estimations from a four-compartment model using density, body water, and bone mineral measurements.
Am J Clin Nutr
55:
764-770,
1992[Abstract/Free Full Text].
14.
Fuller, NJ,
Jebb SA,
Laskey MA,
Coward WA,
and
Elia M.
Four-component model for the assessment of body composition in humans: comparison with alternative methods, and evaluation of the density and hydration of fat-free mass.
Clin Sci (Lond)
82:
687-693,
1992[Medline].
15.
Fuller, NJ,
Sawyer MB,
and
Elia M.
Comparative evaluation of body composition methods and predictions, and calculation of density and hydration fraction of fat-free mass in obese women.
Int J Obes Relat Metab Disord
18:
503-512,
1994[ISI][Medline].
16.
Going, S.
Densitometry.
In: Human Body Composition, edited by Roche AF,
Heymsfield SB,
and Lohman TG.. Champaign, IL: Human Kinetics, 1996, p. 3-23.
17.
Goldman, RF,
and
Buskirk ER.
Body volume measurement by underwater weighing: description of a method.
In: Techniques for Measuring Body Composition, edited by Brozek J,
and Henschel A.. Washington, DC: National Academy of Science, 1961, p. 77-89.
18.
Heymsfield, SB,
Lichtman S,
Baumgartner RN,
Wang J,
Kamen Y,
Aliprantis A,
and
Pierson RN, Jr.
Body composition of humans: comparison of two improved four-compartment models that differ in expense, technical complexity, and radiation exposure.
Am J Clin Nutr
52:
52-58,
1990[Abstract/Free Full Text].
19.
Heyward, VH.
Practical body composition assessment for children, adults, and older adults.
Int J Sport Nutr
8:
285-307,
1998[ISI][Medline].
20.
Hortobagyi, T,
Israel RG,
Houmard JA,
O'Brien KF,
John RA,
and
Wells JM.
Comparison of four methods to assess body composition in Black and White athletes.
Int J Sport Nutr
2:
60-74,
1992[ISI][Medline].
21.
Lewis, RD,
and
Modlesky CM.
Nutrition, physical activity, and bone health in women.
Int J Sport Nutr
8:
250-284,
1998[ISI][Medline].
22.
Lohman, TG.
Applicability of body composition techniques and constants for children and youths.
In: Exercise and Sports Science Reviews, edited by Pandolf K.. New York: Macmillan, 1986, p. 325-357.
23.
Lohman, TG.
Advances in Body Composition Assessment. Champaign, IL: Human Kinetics, 1992.
24.
Lukaski, HC,
and
Johnson PE.
A simple, inexpensive method of determining total body water using a tracer dose of D2O and infrared absorption of biological fluids.
Am J Clin Nutr
41:
363-370,
1985[Abstract/Free Full Text].
25.
MacDougall, JD,
Sale DG,
Elder GCB,
and
Sutton JR.
Muscle ultrastructural characteristics of elite powerlifters and bodybuilders.
Eur J Appl Physiol Occup Physiol
48:
117-126,
1982[ISI][Medline].
26.
Martin, AD,
and
Drinkwater DT.
Variability in the measures of body fat: assumptions or technique.
Sports Med
11:
277-288,
1991[ISI][Medline].
27.
McDonagh, MJN,
and
Davies CTM
Adaptive response of mammalian skeletal muscle to exercise with high loads.
Eur J Appl Physiol Occup Physiol
52:
139-155,
1984[ISI][Medline].
28.
Merz, AL,
Trotter M,
and
Peterson RR.
Estimation of skeleton weight in the living.
Am J Phys Anthropol
14:
589-609,
1956.
29.
Modlesky, CM,
Cureton KJ,
Lewis RD,
Prior BM,
Sloniger MA,
and
Rowe DA.
Density of the fat-free mass and estimates of body composition in male weight trainers.
J Appl Physiol
80:
2085-2096,
1996[Abstract/Free Full Text].
30.
Modlesky, CM,
Evans EM,
Millard-Stafford ML,
Collins MA,
Lewis RD,
and
Cureton KJ.
Impact of bone mineral estimates on percent fat estimates from a four-component model.
Med Sci Sports Exerc
31:
1861-1868,
1999[ISI][Medline].
31.
Olsson, K-E,
and
Saltin B.
Variation in total body water with muscle glycogen changes in man.
Acta Physiol Scand
80:
11-18,
1970[ISI][Medline].
32.
Ortiz, O,
Russell M,
Daley TL,
Baumgartner RN,
Waki M,
Lichtman S,
Wang Pierson J,
Pierson RN,
and
Heymsfield SB.
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,
1992[Abstract/Free Full Text].
33.
Prior, B,
Modlesky CM,
Evans EM,
Sloniger MA,
Saunders MJ,
Lewis RD,
and
Cureton KJ.
Muscularity and the density of the fat-free mass in athletes.
J Appl Physiol
90:
1523-1531,
2001[Abstract/Free Full Text].
34.
Roubenoff, R,
Kehayias JJ,
Dawson-Hughes B,
and
Heymsfield SB.
Use of dual-energy x-ray absorptiometry in body-composition studies: not yet a "gold standard."
Am J Clin Nutr
58:
589-591,
1993[Free Full Text].
35.
Schoeller, DA.
Hydrometry.
In: Human Body Composition, edited by Roche AF,
Heymsfield SB,
and Lohman TG.. Champaign, IL: Human Kinetics, 1996, p. 25-43.
36.
Schutte, JE,
Townsend EJ,
Hugg J,
Shoup RF,
Malina RM,
and
Blomqvist CG.
Density of lean body mass is greater in blacks than in whites.
J Appl Physiol
56:
1647-1649,
1984[Abstract/Free Full Text].
37.
Seale, RU.
The weight of the dry fat-free skeleton of American whites and negroes.
Am J Phys Anthropol
17:
37-48,
1959.
38.
Siri, WE.
Body composition from fluid spaces and density: analysis of methods.
In: Techniques for Measuring Body Composition, edited by Brozek J,
and Henschel A.. Washington, DC: National Academy of Science, 1961, p. 223-244.
39.
Snow-Harter, C,
Bouxsein ML,
Lewis BT,
Carter DR,
and
Marcus R.
Effects of resistance and endurance exercise on bone mineral status of young women: a randomized exercise intervention trial.
J Bone Miner Res
7:
761-769,
1992[ISI][Medline].
40.
Sparling, PB,
Millard-Stafford M,
Rosskopf LB,
Dicarlo LJ,
and
Hinson BT.
Body composition by bioelectrical impedence and densitometry in black women.
Am J Human Biol
5:
111-117,
1993.
41.
Trotter, M,
and
Hixon BB.
Sequential changes in weight, density, and percentage ash weight of human skeletons from an early fetal period through old age.
Anat Rec
179:
1-18,
1974[Medline].
42.
VanItallie, TB,
Yang M,
Heymsfield SB,
Funk RC,
and
Boileau RA.
Height-normalized indices of the body's fat-free mass and fat mass: potentially useful indicators of nutritional status.
Am J Clin Nutr
52:
953-959,
1990[Abstract/Free Full Text].
43.
Visser, M,
Gallagher D,
Deurenberg P,
Wang J,
Person RN, Jr,
and
Heymsfield SB.
Density of fat-free body mass: relationship with race, age, and level of body fatness.
Am J Physiol Endocrinol Metab
272:
E781-E787,
1997[Abstract/Free Full Text].
44.
Wagner, DR,
and
Heyward VH.
Measures of body composition in blacks and whites: a comparative review.
Am J Clin Nutr
71:
1392-1402,
2000[Abstract/Free Full Text].
45.
Wagner, DR,
and
Heyward VH.
Validity of two-component models for estimating body fat of black men.
J Appl Physiol
90:
649-656,
2001[Abstract/Free Full Text].
46.
Waki, M,
Kral JG,
Mazariegos M,
Wang J,
Pierson RN,
and
Heymsfield SB.
Relative expansion of extracellular fluid in obese vs. nonobese women.
Am J Physiol Endocrinol Metab
261:
E199-E203,
1991[Abstract/Free Full Text].
47.
Withers, RT,
Noell CJ,
Whittingham NO,
Chatterton BE,
Schultz CG,
and
Keeves JP.
Body composition changes in elite male bodybuilders during preparation for competition.
Aust J Sci Med Sport
29:
11-16,
1997[Medline].
48.
Womersley, J,
Durnin JVGA,
Boddy K,
and
Mahaffy M.
Influence of muscular development, obesity, and age on the fat-free mass of adults.
J Appl Physiol
41:
223-229,
1976[Abstract/Free Full Text].
J APPL PHYSIOL 91(3):1259-1268
8750-7587/01 $5.00
Copyright © 2001 the American Physiological Society
TOP
ABSTRACT
INTRODUCTION
![%Fat<SUB>D,W,M</SUB><IT>=</IT>[(2.747<IT>/</IT>D<SUB>b</SUB>)<IT>−</IT>0.714W<IT>+</IT>1.146M<IT>−</IT>2.0503]<IT>×</IT>100](/content/vol91/issue3/fulltext/1259/img002.gif)
