International Journal of Obesity (2000) 24, 1200±1206
ß 2000 Macmillan Publishers Ltd All rights reserved 0307±0565/00 $15.00
www.nature.com/ijo
Estimation of body fat in Caucasian and
African-American girls: total-body electrical
conductivity methodology versus a
four-component model{
WW Wong1*, JE Stuff1, NF Butte1, EO Smith1 and KJ Ellis1
1
USDA=ARS Children's Nutrition Research Center and Texas Children's Hospital, Department of Pediatrics, Baylor College of
Medicine, Houston, TX 77030, USA
BACKGROUND: Obesity has been increasing dramatically in recent years among children, particularly AfricanAmerican girls. Total-body electrical conductivity (TOBEC) is a simple way to measure body fat with minimal risk.
OBJECTIVE: This study compared the agreement between the percentage of fat mass (%FM) predicted using two
TOBEC equations with %FM measured by a four-component model in 73 Caucasian and 41 African-American girls.
DESIGN AND MEASUREMENTS: %FM predicted using the TOBEC equations was compared with %FM from the fourcomponent model based on measurements of body density, body water and bone mineral content.
RESULTS: Analyses by linear regression analysis and by the Bland and Altman methods comparison procedure
showed that the equation using the square root of the TOBEC zero-order Fourier coef®cient and the subject's height
yielded more accurate and more reproducible %FM, regardless of race, than the TOBEC linear equation, which was
based on the zero-, ®rst- and second-order Fourier coef®cients. The Bland and Altman comparison further revealed
that the accuracy and limits of agreement of the TOBEC linear equation were related to body fatness among the
Caucasian girls. The relationship, however, disappeared when prepubescent girls and a girl with low %FM were
excluded from the analysis.
CONCLUSIONS: The TOBEC square root equation with adjustment for body geometry and length is recommended for
use in adolescent girls, as it yielded better agreement with the criterion method. However, further validation of the
TOBEC instrumentation for estimating body fat in prepubescent children and children with low body fat is warranted.
International Journal of Obesity (2000) 24, 1200±1206
Keywords: body fat; girls; ethnicity; TOBEC; underwater weighing; isotope dilution; dual-energy X-ray absorptiometry
Introduction
Total-body electrical conductivity (TOBEC) is a
rapid, noninvasive methodology developed for predicting body fatness in humans. The measurement
requires only minimum training of the operator and
cooperation from the subject. The TOBEC method is
based on the principle that fat-free mass (FFM), which
contains large amounts of body water and electrolytes,
is electrically conductive. When a human subject is
placed inside the electromagnetic ®eld of a TOBEC
instrument, the conductive tissues in the body dissipate some of the magnetic ®eld's energy, which
causes changes in the instrument's impedance level.
The magnitude of the perturbation is related to the
subject's body size and the conductivity of the FFM.
*Correspondence: WW Wong, USDA=ARS Children's Nutrition
Research Center, 1100 Bates Street, Houston, TX 77030, USA.
E-mail: [email protected]
{Disclaimer: The contents of this publication do not necessarily
re¯ect the views or policies of the USDA, nor does mention of
trade names, commercial products, or organizations imply
endorsement by the US Government.
Received 17 September 1999; revised 7 February 2000; accepted
18 April 2000
Once FFM is calculated, body fat or fat mass (FM) is
de®ned as the difference between body weight and
FFM. Percentage of fat mass (%FM) is de®ned as
FM=(body weight) 100.
Data on the agreement between %FM predicted
using the TOBEC instrument and %FM based on a
four-component criterion model in humans, particularly among minority populations, are limited. Two
infant studies compared body composition measurements using TOBEC with measurements using an
anthropometric technique1 or an isotope dilution
method.2 Two adult studies compared FFM measured
by TOBEC with that measured by densitometry
and=or 40K counting.3,4 Anthropometry assumes a
constant relationship between %FM and subcutaneous
skinfold thickness. The densitometric technique
assumes a constant composition of FFM. The isotope
dilution method assumes a constant hydration of FFM,
while the 40K counting method assumes a constant
potassium content of FFM. Because fat patterning and
the density, hydration and potassium content of FFM
change with age5 and differ among ethnic groups,5 ± 7
the use of a constant relationship or a single, ®xed
constant with any one of these methods may be
inappropriate, especially during growth.
Estimation of body fat in girls using TOBEC
WW Wong et al
Two studies of the TOBEC methodology for predicting FFM were carried out, one on 114 adults8 and
the other involving 50 teenagers.9 In both studies,
FFM predicted by TOBEC was found to be similar
to the FFM obtained using a four-component model
based on the combined measurements of body density,
total body water and bone mineral content. FFM
calculated using this four-component model is considered superior to any anthropometric method or any
single measurement of body density, total body water,
or total body 40K because the four-component model
reduces the use of constant values. These two studies,8,9 however, primarily involved Caucasian
subjects, and the agreement between the %FM predicted using the TOBEC methodology and the fourcomponent model was not fully evaluated.
Based on the changes in mean weights and heights
of children and adolescents over the last 30 y, as
recorded in the National Health Examination Surveys,
dramatic increases in the incidence of childhood
obesity have occurred in all ages in the USA.10,11
From 1963 to 1991, a 40% increase in the incidence of
obesity, as de®ned by the body mass index, was
observed in Caucasian children, aged 6 ± 17 y. This
trend was even more exaggerated in African-American girls: a 160% increase in the incidence of obesity
was observed in the 6 ± 11-y-old age group, while an
81% increase was reported in the 12 ± 17-y-old range.
In light of this signi®cant obesity trend and the use of
TOBEC technology in many research protocols, it is
important to determine the level of agreement of the
TOBEC estimate for %FM in girls, with %FM measured by the criterion four-component model.
Although newer TOBEC technology is not being
developed, older instruments are being refurbished
and remain available at many institutions. TOBEC
instruments are simple to operate; the measurements
can be repeated as frequently as needed with minimal
risk; and the short time needed for each measurement
makes TOBEC methodology a viable technique for
studies that involve a large number of subjects.
The aim of this study was to evaluate the level of
agreement between the %FM predicted using the
TOBEC methodology and that based on a four-component criterion model in a population of Caucasian
and African-American girls.
Subjects and methods
Human subjects
A group of 73 Caucasian and 41 African-American
girls (Table 1) was recruited from local schools in the
greater Houston metropolitan area. All subjects were
healthy and nondiabetic at the time of the study. The
Institutional Review Board for Human Research at
Baylor College of Medicine approved the protocol.
All subjects and their parents gave written informed
consent.
Table 1 Age, physical characteristics, anthropometric measurements and body composition of the 114 female adolescents
Caucasian
(n ˆ 73)
Age (y)
Weight (kg)
Height (m)
Body mass index
(kg=m2)
Body composition
Body density
(g=ml)
BMC (kg)
TBW (kg)
FFM4C (kg)b
%FM4C (%)b
%FMTOBEC1 (%)c
%FMTOBEC2 (%)c
1201
African-American
P
(n ˆ 41)
values
12.7 1.9a
48.0 13.0
1.54 0.11
20.1 4.0
13.5 1.7a
56.5 13.2
1.60 7.4
22.1 4.7
1.0375 0.0162
1.0371 0.0155
1.53 0.44
26.0 5.8
36.1 7.8
23.4 7.2
24.0 7.5
28.8 13.9
1.94 0.43
30.4 5.1
42.2 6.8
24.0 7.3
24.4 6.4
35.3 7.8
0.02
<0.01
<0.01
<0.02
0.91
<0.01
<0.01
<0.01
0.65
0.79
<0.01
a
x s.d.
Fat-free mass (FFM) and percentage fat mass (%FM) by the
four-component model based on measurements of body
density, bone mineral content (BMC), and total body water
(TBW).
c
%FM by TOBEC using the square root equation (%FMTOBEC1) or
the linear equation (%FMTOBEC2).
b
Anthropometric measurements and sexual maturity
determination
Each subject's body weight was measured to the
nearest 0.1 kg with an electronic scale (Scale-Tronix,
Wheaton, IL) and height to the nearest 0.1 cm with a
stadiometer (Holtain Ltd, Crymmych, Pembs, UK).
One investigator trained in making anthropometric
measurements performed all of these measurements.
Breast and pubic hair development was determined
by a physician according to the Tanner stages of
classi®cation.12
TOBEC methodology
A TOBEC instrument (Model HA-2, EM-Scan,
Spring®eld, IL) was used. After the subject changed
into a hospital gown and removed all her jewelry, she
lay in a supine position on the instrument's carrier.
With her hands at her sides but not in contact with her
body, and her feet slightly apart, the subject was
moved through the instrument's open core, which is
surrounded by a solenoid coil. The TOBEC zero-,
®rst- and second-order Fourier coef®cients were calculated based on the total body scan. The TOBEC
measurements were performed three times on each
subject, and the FFM estimates averaged. The two
TOBEC equations for estimating FFM are as follows:
Square root equation (TOBEC1)
p
FFMTOBEC1 …kg† ˆ 0:2772 FC0 H ‡ 1:232
Linear equation (TOBEC2)
FFMTOBEC2 …kg† ˆ 22:9986 ‡ 0:1015 FC0
‡ 0:0622 FC1 ÿ 0:2908 FC2
International Journal of Obesity
Estimation of body fat in girls using TOBEC
WW Wong et al
1202
where H is height in cm and FC0, FC1 and FC2 are
the TOBEC zero-, ®rst- and second-order Fourier
coef®cients, respectively.
The square root equation (TOBEC1) is provided by
the manufacturer and applies to subjects between 5
and 19 y of age. The square root equation was developed mathematically based on a cylindrical body with
adjustment for height. The linear equation (TOBEC2)
is widely used in the literature for predicting FFM in
both adults and teenagers.8,9 The linear equation was
generated by multiple regression analysis based on
adult data with the assumption that the TOBEC Fourier coef®cients are linearly related to FFM. The
TOBEC estimate of FM is the difference between
body weight (W) and FFM. %FM is calculated as
follows:
%FM TOBEC1 or TOBEC2 ˆ
W ÿ FFMTOBEC1 or TOBEC2
100
W
Body composition measurements
The criterion %FM4C was measured using the fourcomponent model as follows:13
2:747
TBW
ÿ 0:727 ‡ 1:146
%FM4C ˆ
D
W
BMC
ÿ 2:0503 100
W
where D is body density in g=ml measured by underwater weighing14 using the `force cube' transducer
method15 with correction for residual lung volume by
nitrogen dilution;16 TBW is total body water in kg and
is assumed to be identical to 18O dilution space; and
BMC is bone mineral content in kg measured by dualenergy X-ray absorptiometry (Hologic QDR-2000W,
Hologic Inc., Waltham, MA, USA). For the TBW
measurement, a baseline plasma sample was collected
by venipuncture before each subject drank 1.25 g of
10% H218O (Isotec Inc., Miamisburg, OH, USA) per
kg body weight. Another plasma sample was collected
3 h later. The plasma samples were prepared for
oxygen isotope ratio measurements by gas-isotoperatio mass spectrometry.17 TBW was calculated as
follows:
d A Ea
TBW …kg† ˆ
a Ed 103
where d is the dose of H218O in grams; A is the
amount of laboratory water in grams used in the dose
dilution; a is the amount of H218O in grams added to
the laboratory water in the dose dilution; Ea is the rise
in 18O abundance in parts per thousand ( =oo) in the
laboratory water after the addition of the isotopic
water; and Ed is the rise in 18O abundance in =oo
in the 3 h postdose plasma sample.
International Journal of Obesity
Statistical analysis
Linear regression analysis and the Bland and Altman
methods comparison procedure18 were used to compare %FMTOBEC1 and %FMTOBEC2 with %FM4C. In
the regression analysis, the %FMTOBEC1 and the
%FMTOBEC2 were plotted against %FM4C. The deviation of the slope from 1.0 and the deviation of the
intercept from zero of the regression line were evaluated using the critical t values with the corresponding
degrees of freedom.19 For simplicity, the reproducibility of the TOBEC method was de®ned by the
standard error of estimate (SEE) of the regression
line. With the Bland and Altman procedure, differences between %FMTOBEC and %FM4C were plotted
against the averages of the two %FM values. Regression analysis was used to test the relationship between
the differences and averages. If the slope was not
signi®cant (P > 0.05), the accuracy (mean difference
between methods) and the 95% limits of agreement
(accuracy 2 standard deviation of the differences)
were computed. If the slope relating the differences
and averages was signi®cant (P<0.05), the accuracy
and limits of agreement would change depending on
%FM. Under this circumstance, the limits of agreement would be calculated as 2 SEE around the
regression line and would be represented by the
lower limit at the lowest %FM and the upper limit
at the highest %FM. All statistical analyses were
performed using SPSS for Windows (version 8,
SPSS Inc., Chicago, IL, USA).
Results
The ages, physical characteristics and body composition of the 114 female adolescents are given in Table
1. The African-American girls were older, heavier,
taller and had higher body mass index than the
Caucasian girls. The African-American girls also
had higher BMC, TBW and FFM than the Caucasian
girls. The mean body fatness measured by the fourcomponent model and that predicted by the square
root equation (%FMTOBEC1) did not differ between
the two ethnic groups. However, the mean %FM
predicted using the linear equation (%FMTOBEC2)
was higher among the African-American girls.
The results of the linear regression analyses
between %FM estimated by the two TOBEC equations and %FM4C of the 73 Caucasian and 41 AfricanAmerican subjects are presented in Figure 1. When
the square root equation (TOBEC1, upper panels in
Figure 1) was used, the %FMTOBEC1 was highly
correlated (Pearson correlation or r ˆ 0.85 ± 0.90,
P<0.01) with %FM4C. Although the SEE of the
regression line was lower among the African-American subjects (2.9%) than that of the Caucasian subjects (3.9%), the slope and the intercept of the
African-American subjects were found to deviate
further from 1.0 and zero, respectively than those of
Estimation of body fat in girls using TOBEC
WW Wong et al
the Caucasian subjects. Among the Caucasian subjects, the two points deviated furthest away from the
regression line represented one subject aged 9.1 y with
%FM4C of 11.1% and another subject aged 11.5 y with
%FM4C of 8.2%. When the linear equation (TOBEC2)
was used, a slight improvement in slope (from 0.78 to
0.85) was observed among the African-American
subjects (Figure 1, lower panels) but with deterioration in the Pearson correlation (from 0.90 to 0.80), the
SEE (from 2.9% to 4.8%), and the intercept (from
5.6% to 14.79%). Although an apparent improvement
was observed in the intercept (from 3.40% to 1.94%)
among the Caucasian subjects, dramatic deterioration
in the slope (from 0.88 to 1.32), the Pearson correlation (from 0.85 to 0.69), and the SEE (from 3.9% to
10.2%) were obtained. Furthermore, three Caucasian
subjects had negative %FMTOBEC2 ranged between
Figure 1 Linear regression analysis between %FM predicted
using the TOBEC instrument and the four-component model for
the 73 Caucasian and 41 African-American girls: TOBEC1
(square-root equation) vs four-component model (top panels)
and TOBEC2 (linear equation) vs four-component model (bottom
panels). The solid line represents the line of identity (slope ˆ 1,
and intercept ˆ 0). The short-dotted line represents the regression line for the Caucasian subjects and the dashed line represents the regression line for the African-American subjects. The
open circle represents the individual %FM value of the Caucasian subjects and the solid upside-down triangle represents the
individual %FM value of the African-American subjects. The
equation in each ®gure represents the least-square best-®t
equation of the linear regression analysis. In each equation,
the coef®cient represents the slope and the constant represents
the intercept. An asterisk above the coef®cient or the constant
indicates the slope is signi®cantly different from 1.0 or the
intercept is signi®cantly different from zero. The r value below
each equation is the Pearson correlation for the regression and
the P value represents the signi®cance of the regression. The
SEE is the standard error of estimate around the regression line.
ÿ12.2% and ÿ17.0%. All three subjects were younger
than 10 y of age (9.1 ± 9.8 y). The Caucasian girl who
had a %FM4C of 8.2% also had a %FMTOBEC2
approaching zero (0.9%).
Bland and Altman comparisons of the 73 Caucasian
and 41 African-American subjects (Figure 2, top
panels) show that %FM estimated using the
TOBEC1 equation was accurate to within
0.6 4.0% for the Caucasian girls and 0.3 3.2%
for the African-American girls when compared with
those measured by the four-component model. More
importantly, the accuracy and limits of agreement
were independent of %FM. However, on an individual
basis, %FM can be over- or underestimated by
approximately 8% among the Caucasian girls and by
approximately 7% among the African-American girls.
The results were very different when the TOBEC2
equation was used (Figure 2, bottom panels). The
difference between %FMTOBEC2 and %FM4C among
the Caucasian subjects was signi®cantly (P<0.01)
related to %FM. On an individual basis, %FM could
be underestimated by as much as 25.1% for an
individual with a %FM of 5% or overestimated by
as much as 38.1% for an individual with a %FM of
50%. Although the relationship was not evident
between the difference and %FM among the African-American subjects, the accuracy (11.2 4.9%)
and limits of agreement (1.5% to 20.9%) deteriorated
when the TOBEC2 equation was used.
In order to eliminate the possibility that the TOBEC
methodology might not be appropriate for younger
subjects and for subjects with low %FM, the data were
reanalyzed after excluding prepubescent girls (girls
with Tanner stages of breast and pubic hair development lower than 3) and girls with low %FM4C
(<10%) from the database. As shown in Figures 3
and 4 (upper panels), no dramatic changes were
observed in the results of the linear regression analysis
and Bland and Altman methods comparison procedure
when the square-root TOBEC1 equation was used.
With the TOBEC2 equation (Figures 3 and 4, lower
panels), similar results were obtained among the
African-American subjects after excluding the prepubescent girls in the analyses. However, when the
prepubescent girls and girls with low %FM were
excluded from the analyses among the Caucasian
subjects, the relationship between difference and
%FM disappeared and an accuracy of 11.3 4.4%
was obtained.
1203
Discussion
The linear equation for TOBEC based on zero-, ®rstand second-order Fourier coef®cients is the one that is
commonly used with this technology. However, our
results indicated that the simpler square root equation
is superior for predicting %FM in female children and
adolescents, regardless of ethnicity (Figures 1 ± 4).
As shown in Figure 1 (upper panel), a negative
%FM value (ÿ1%) was obtained in an 11.5-y-old girl
International Journal of Obesity
Estimation of body fat in girls using TOBEC
WW Wong et al
1204
Figure 2 Comparisons of FM predicted between the TOBEC square root equation (TOBEC1) and the four-component model (top
panels) and between the TOBEC linear equation (TOBEC2) and the four-component model (bottom panels) for the 73 Caucasian and 41
African-American girls. The solid line represents the accuracy (%FMTOBEC1 or %FMTOBEC2ÿ%FM4C). The dashed lines represent the
limits of agreement, calculated as accuracy 2 s.d. of the differences if the slope is not signi®cant. Numerical values above the solid
line represent the accuracy. Numerical values above and below the two dashed lines are upper and lower limits of agreement. If the
slope relating the differences and averages is signi®cant, the 95% limits of agreement are estimated as 2 standard error of estimate
(SEE) around the regression line. Numerical values above and below the two dashed lines are upper and lower limits of agreement at
corresponding %FM values of 5% and 50%. When the slope is not zero, the limits of agreement are de®ned by the lower limit at %FM
of 5% and the upper limit at %FM of 50%.
Figure 3 Linear regression analysis between %FM predicted
using the TOBEC instrument and the four-component model for
the pubescent and postpubescent girls (73 Caucasians and 41
African-Americans): TOBEC1 (square-root equation) vs fourcomponent model (top panels) and TOBEC2 (linear equation)
vs four-component model (bottom panels). The solid line represents the line of identity (slope ˆ 1, and intercept ˆ 0). The shortdotted line represents the regression line for the Caucasian
subjects and the dashed line represents the regression line for
the African-American subjects. The open circle represents the
individual %FM value of the Caucasian subjects and the solid
upside-down triangle represents the individual %FM value of the
African-American subjects. The equation in each ®gure represents the least-square best-®t equation of the linear regression
analysis. In each equation, the coef®cient represents the slope
and the constant represents the intercept. An asterisk above the
coef®cient or the constant indicates the slope is signi®cantly
different from 1.0 or the intercept is signi®cantly different from
zero. The r value below each equation is the Pearson correlation
for the regression and the P value represents the signi®cance of
the regression. The SEE is the standard error of estimate around
the regression line.
with a %FM4C of 8.2% when the TOBEC1 equation
was used. When the TOBEC2 equation was used
(Figure 1, lower panel), negative %FM values
(ÿ12.2% to ÿ17%) were obtained in three prepubescent girls (<10-y-old). The 11.5-y-old girl with low
%FM yielded a %FMTOBEC2 value of 0.9%. Because
both the TOBEC1 and TOBEC2 equations were
mathematically derived, erratic results, including
International Journal of Obesity
Estimation of body fat in girls using TOBEC
WW Wong et al
1205
Figure 4 Comparisons of %FM predicted between the TOBEC square-root equation (TOBEC1) and the four-component model (top
panels) and between the TOBEC linear equation (TOBEC2) and the four-component model (bottom panels) for the pubescent and
postpubescent girls (73 Caucasians and 41 African-Americans). The solid line represents the accuracy (%FMTOBEC1 or %FMTOBEC2
ÿ%FM4c). The dashed lines represent the limits of agreement, calculated as accuracy 2 s.d. of the differences if the slope is not
signi®cant. Numerical values above the solid line represent the accuracy. Numerical values above and below the two dashed lines are
upper and lower limits of agreement. If the slope relating the differences and averages is signi®cant, the 95% limits of agreement are
estimated as 2 standard error of estimate (SEE) around the regression line. Numerical values above and below the two dashed lines
are upper and lower limits of agreement at corresponding %FM values of 5% and 50%. When the slope is not zero, the limits of
agreement is de®ned by the lower limit at %FM of 5% and the upper limit at %FM of 50%.
negative values, could be obtained when the equations
were not properly calibrated for the age group or for
low body fat. The results further suggest that the
TOBEC2 equation is more sensitive to age than the
TOBEC1 equation. The observation is supported by
the disappearance of the signi®cant relationship
between the difference and body fatness among the
Caucasian girls after excluding the prepubescent girls
from the Bland and Altman analyses (Figures 2 and 4,
lower panels). Our results also suggested that the
TOBEC methodology might not be appropriate for
prepubescent girls and for girls with low %FM.
According to the method selection criteria of
Lohman,20 the TOBEC1 method (Figure 1, upper
panel) with an SEE of 3.9% for the Caucasian subjects
would be rated as fairly good and an SEE of 2.9% for
the African-American subjects would be rated as very
good. However, the SEE criteria of Lohman were
based on a 76.5 kg man and a 60.0 kg woman with
body fatness of 15% and 25%, respectively.20 As
shown by the same author, much higher SEE is
anticipated when greater variability is observed in
the criterion measurements, ie %FM by the fourcomponent model.21 With %FM4C of our subjects
ranged between 8% and 42%, it is reasonable to
expect the SEE to be much higher when the TOBEC
method is used to predict %FM in these girls. Therefore the SEE criteria outlined by Lohman would not
be applicable in our evaluation or in any evaluations
when a wide range of body fatness is expected.20
However, in spite of the wide range of body fatness
in our study subjects, the small SEE associated with
the %FM estimates when the TOBEC1 equation was
used is rather impressive, suggesting that the TOBEC
methodology deserves further evaluation.
The linear equation has been shown to yield similar
FFM in 50 teenagers to that obtained using the fourcomponent model.9 However, the Bland and Altman
procedure was not performed in that study, so it is
dif®cult to evaluate the agreement in FFM or %FM
between the two methods. Furthermore, only one
Hispanic, three African-Americans, and four Asians
were among the 50 teenagers who participated in that
study.9 Therefore, an evaluation of the TOBEC methodology for predicting %FM in African-American
adolescents or other minority groups cannot be performed using that data.9 Our results (Figure 4, upper
panels) clearly demonstrated that the accuracy and
limits of agreement of the TOBEC methodology for
estimation of body fatness are similar between the
International Journal of Obesity
Estimation of body fat in girls using TOBEC
WW Wong et al
1206
Caucasian and African-American pubescent and postpubescent girls when the square-root equation with
adjustment for body geometry and length is used.
It has been demonstrated that changes in body
geometry affect the accuracy and precision of the
TOBEC measurement.22 In an earlier study, the
square-root of the TOBEC score with adjustment for
the subject's height improved the accuracy and precision of FFM predicted using the TOBEC methodology.3 Our results (Figures 1 ± 4) provide further
evidence that the use of the square-root of the
TOBEC zero-order Fourier coef®cient and the subject's height yielded estimates of body fatness which
are in better agreement with that measured by the
four-component model than the linear equation.
In conclusion, our results show that when TOBEC
methodology is used to assess body composition in
female adolescents, the equation based on the squareroot of the TOBEC zero-order Fourier coef®cient and
the subject's height is recommended. The negative
%FM values obtained in the prepubescent girls and in
the girl with low %FM indicate that further validation
of the TOBEC methodology for estimation of body
fatness in prepubescent children and in children and
adolescents with low body fatness is warranted.
Acknowledgements
The authors are indebted to the volunteers; to the staff
of the Metabolic Research Unit at the Children's
Nutrition Research Center, for meeting the needs of
the subjects during the study; to Drs A Hergenroeder
and R Hill for performing the physical examination to
determine Tanner stages of sexual maturity; to Dr J
Hoyle in the Pediatrics Department of Kelsey-Seybold
West Clinic, Dr M desVignes-Kendrick, director of
the City of Houston Health and Human Services
Department, Ms X Earlie, director of sciences of the
Aldine Independent School District, Ms S Wooten,
principal of the Teague Middle School, Dr B Shargey,
dean of instruction, Ms CC Collins, principal at the
High School for Health Professions and Ms K Wallace
for subject recruitment; Mr M Puyau and Mr FA
Vohra, for the underwater weighing measurements;
Mr R Shypailo and Ms J Joe for the dual-energy X-ray
absorptiometry measurements; Mrs L Clarke, Mr S
Zhang and Ms K Usuki for the isotope ratio mass
spectrometric measurements; and Ms L Loddeke, for
editorial assistance in the preparation of the manuscript. This work was funded in part with federal
funds from the US Department of Agriculture
(USDA)=Agricultural Research Service under Cooperative Agreement no 58-6250-6-001.
References
1 de Bruin NC, van Velthoven KAM, Stijnen T, Juttmann RE,
Degenhart HJ, Visser HKA. Quantitative assessment of infant
body fat by anthropometry and total-body electrical conductivity. Am J Clin Nutr 1995; 61: 279 ± 286.
International Journal of Obesity
2 de Bruin NC, Westerterp KR, Degenhart HJ, Visser HKA.
Measurement of fat-free mass in infants. Pediatr Res 1995; 38:
411 ± 417.
3 Presta E, Segal KR, Gutin B, Harrison GG, van Itallie TB.
Comparison in man of total body electrical conductivity and
lean body mass derived from body density: validation of a new
body composition method. Metabolism 1983; 32: 524 ± 527.
4 Horswill CA, Geeseman R, Boileau RA, Williams BT,
Layman DK, Massey BH. Total-body electrical conductivity
(TOBEC): relationship to estimates of muscle mass, fat-free
mass, and lean body mass. Am J Clin Nutr 1989; 49: 593 ± 598.
5 Lohman TG. Applicability of body composition techniques
and constants for children and youths. In Pandolf KE (ed.).
Exercise and sports sciences reviews. Macmillan: New York,
1986, pp 325 ± 357.
6 Schutte JE, Townsend EJ, Hugg J, Shoup RF, Malina RM,
Blomqvist CG. Density of lean body mass is greater in Blacks
than in Whites. J Appl Physiol 1984; 56: 1647 ± 1649.
7 Ortiz O, Russell M, Daley TL, Baumgartner RN, Waki M,
Lichtman S, Wang J, Pierson RN Jr, Heyms®eld 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 1992; 55: 8 ± 13.
8 van Loan MD, Koehler LS. Use of total-body electrical
conductivity for the assessment of body composition in
middle-aged and elderly individuals. Am J Clin Nutr 1990;
51: 548 ± 552.
9 van Loan MD. Assessment of fat-free mass in teen-agers: use
of TOBEC methodology. Am J Clin Nutr 1990; 52: 586 ± 590.
10 Gortmaker SL, Dietz WH, Sobol AM, Wehler CA. Increasing
pediatric obesity in the United States. Am J Dis Child 1987;
141: 535 ± 540.
11 Troiano RP, Flegal KM, Kuczmarski RJ, Campbell SM, Johnson
CL. Overweight prevalence and trends for children and adolescents. Arch Pediatr Adolesc Med 1995; 149: 1085 ± 1091.
12 Tanner JM, Whitehouse RH. Variations of growth and development at puberty. In Atlas of children's growth, normal
variation and growth disorders. Academic Press: New
York, 1982, pp 122 ± 127.
13 Boileau RA, Lohman TG, Slaughter MH. Exercise and body
composition of children and youth. Scand J Sports Sci 1985; 7:
17 ± 27.
14 Siri WE. Body composition from ¯uid spaces and density:
analysis of methods. In Brozek J, Hanschel A (eds). Techniques for measuring body composition. National Academy of
Science: Washington, DC, 1961, pp 223 ± 244.
15 Akers R, Buskirk ER. An underwater weighing system
utilizing `force cube' transducers. J Appl Physiol 1969; 26:
649 ± 652.
16 Wilmore JH. A simpli®ed method for determination of residual lung volumes. J Appl Physiol 1969; 27: 96 ± 100.
17 Wong WW, Lee LS, Klein PD. Deuterium and oxygen-18
measurements on microliter samples of urine, plasma, saliva,
and human milk. Am J Clin Nutr 1987; 45: 905 ± 913.
18 Bland JM, Altman DG. Statistical methods for assessing
agreement between two methods of clinical measurement.
Lancet 1986; i: 307 ± 310.
19 Glantz SA. The special case of two groups: the t-test. In
Laufer RS, Geno TK (eds). Primer of biostatistics. McGrawHill: New York, 1981, pp 63 ± 93.
20 Lohman TG. Basic concepts in body composition assessment.
In Advances in body composition assessment.
Human
Kinetics: Champaign, IL, 1992, pp 1 ± 5.
21 Lohman TG. Prediction equations and skinfolds, bioelectric
impedance, and body mass index. In Advances in body
composition assessment. Human Kinetics: Champaign, IL,
1992, pp 37 ± 56.
22 Fiorotto ML, Cochran WJ, Funk RC, Sheng H-P, Klish WJ.
Total body electrical conductivity measurements: effects of
body composition and geometry. Am J Physiol 1987; 21:
R794 ± R800.