Body composition
in adults with cerebral palsy by dualenergy X-ray absorptiometry,
bioelectrical
impedance
analysis, and skinfold anthropometry
compared
with the
18
isotope-dilution
technique13
Heidi
G Hildreth,
Rachel
ABSTRACT
The aim
cerebral
ods used
adults
palsy
the
to estimate
(13 men
of cerebral
palsy.
thickness,
body
and
by
cantly
using
different
the
from
analysis
measured
DXA
with
and O
derived
skinfold
thickness
difference:
There
80
(mean
with
-6.55
difference:
± 13.6%)
and
equation
measures
specific
to
(7.34
X sex)
0.84,
P <
validated
Am
predict
for
percentage
SEE
=
in an independent
4.85).
signifi-
use of ‘O
dilution
(P
agreement
body
method.
sample
+
<
± 12.3%)
fat
This
sample
(mean
was
the best
DXA
using
was
X
anthropometric
as
follows:
biceps
equation
of adults
8.76
skinfold)
needs
with
-
(R2
=
to be crosscerebral
palsy.
WORDS
Body
water, dual-energy
X-ray
ance
analysis,
isotope
cerebral
palsy, adults
body
composition,
body fat, total body
absorptiometry,
bioelectrical
impeddilution,
skinfold
anthropometry,
composition
movements,
ments
difficult
during
measurements
neuromuscular
deficits
fat over
affected
assumed
to mimic
the use of a stable
for
people
leanness
noted
common
the
for
to
ing
of
cerebral
palsy
(CP)
exhibit
a broad
range
of
and fatness (1), suggesting
that the nutrition
problems
in children
with CP (2-6) persist with age. CP is the most
physical
disability
in humans
and
most
persons
with
disease
survive to adulthood
(7, 8). Therefore,
it is essential
clinicians
to monitor
the nutritional
status of adults with CP
provide
appropriate
nutrition
intervention
aimed at preventthe occurrence
of over- and undernutrition.
Measurement
body composition
provides
an overall
assessment
of the
body’s
1436
energy
stores
and,
in turn,
nutritional
status.
Am
Because
J C/in
Nutr
such
whole
contractures
height
the subject
from
thereby
facility
and
practical
highly
for
asymmetri-
measures
rather
than
the
water
a single
area
of the whole
body.
Thus,
body composition
is ideal
it is noninvasive,
trained
routine
other
technicians
clinical
less
measureremaining
causing
dilution
body
because
as joint
accurate
the composition
isotope
to measure
CP
consequently
does
the subject
to remain
still for the measurement,
pendent
of height and body symmetry.
However,
tive cost of the isotopes
and the need for a mass
make
use.
expensive,
the prohibispectrometry
this
It is more
more
not require
and is inde-
practical
method
im-
appropriate
for
techniques
for
the assessment
of body composition
in a research
setting.
Bandini
et al (6) examined
the relation
between
anthropometric
indexes
and body fatness estimated
from TBW through
use
of the
From
Sciences,
the
Supported
tional
Institutes
E-mail:
1997:66:1436-42.
Clinical
from
reprint
University
Nutrition
and
Metab-
University
of Pediatrics,
grant
no.
to RK
of Vermont,
Randolf
Educational
Center
requests
CP
of Agriculture
the Energy
Sciences,
Department
the William
Research
College
Burlington;
with
of
College
of
Burlington.
Research
of Health
Address
Sciences,
of
Palsy
the
of Vermont,
by grants
Cerebral
of
and
females
Sciences,
of Vermont,
Department
University
General
3
Unit,
in adolescent
of Nutritional
at Birmingham;
Medicine,
United
method
University
Research
Alabama
2
dilution
180
the Department
Life
olism
with
with
(12),
Isotope
of the
and conve-
and
of body
composition.
Addimay increase
the deposition
of
extremities
composition.
is not
( 1 1 ). The neuromuscu-
setting
CP,
without
approach
accurate
(TBW)
inhibit
motionless
tionally,
body
the most
make
and
adults
of this
water
with
to obtain
for adults with CP,
been derived
from
able-bodied
offer
body
in a research
involuntary
formulas
far have
accuracy
of water
of total
and
I
Adults
normal,
associated
and
INTRODUCTION
for
lar complications
validating
J Clin Nuir l997;66:1436-42.
KEY
isotopes
measure
compartment
its high
regres-
thus
( 1 , 9, 10). The
nient
cal
or BIA
fat in the sample,
The best-fitting
by
(0.38
between
but not between
‘SO. Although
body
x weight)
+ (0.32
0.001,
this
were
prediction
estimates
developed
Stable
the reference
equation,
± 9.6%),
6.33
measure
for predicting
percentage
cost prohibits
its use as a practical
sion
skinfold
and dual-energy
with
favorable
0.06
equations
disabilities
known.
There
was no significant
fat measured
with DXA and
was
difference:
(mean
of 20
degrees
from
H Contompasis
body-composition
fat from total body
water
by 80
use of BIA and skinfold
thickness,
Jackson-Pollock
those
meth-
consisted
Stephen
the lack of established
in adults
various
(BIA),
and compared
0.001
and P < 0.001,
respectively).
difference
between
percentage
body
that
sample
y with
fat was estimated
impedance
and
of standard
20-55
body
body
from
I Goran,
to determine
The
aged
(DXA),
of percentage
Values
derived
estimated
was
practicality
composition.
Percentage
absorptiometry
measure
dilution.
study
and
7 women)
bioelectrical
X-ray
of this
accuracy
Michael
at the
Hearst
Foundation,
Foundation,
University
and
the
in part
of Vermont
by
(Na-
RR-109).
Johnson,
304 Terrill
Department
of Nutritional
Hall, Burlington,
VT 05405.
[email protected].
Received
August
Accepted
for publication
Printed
27,
in USA.
1997.
0
1997.
May
30,
1997
American
Society
for Clinical
Nutrition
Downloaded from www.ajcn.org at Norris Med Lib Serials Sect on July 11, 2008
with
K Johnson,
BODY
COMPOSITION
IN ADULTS
9). Body mass index (BMI),
triceps
skinfold
thickness,
subscapular
skinfold
thickness,
suprailiac
skinfold
thickness,
and circumferences
of the biceps,
waist,
forearm,
and knee
were all significantly
correlated
with percentage
body fat.
These associations
suggested
that anthropometric
indexes
can
be useful markers
of body composition
in persons
with CP. In
addition,
Bandini
et al’s work highlighted
the need for developing specific
regression
equations
that can easily assess body
composition
in persons
with CP.
The principal
aim of this study was to determine
the accu(n
racy
and
practicality
of skinfold
anthropometric
measurements,
CEREBRAL
body
percentage
the estimation
measurements
(14).
Bioelectrical
Whole-body
METHODS
AND
Subjects
The sample
consisted
of 20 adults with CP (13 men and 7
women)
ranging
in age from 20 to 55 y. Subjects
were recruited
from
community
organizations
serving
persons
with CP
and through
newspaper
advertisements.
All of the subjects
or
their guardians
gave their informed
written
consent
before
participating
Committee
in the
study.
The
protocol
was
approved
by
the
on Human
Research
at the University
of Vermont.
A physician
(SHC) specializing
in developmental
disabilities
conducted
a neurodevelopmental
exam
on each subject
to
determine
the extent of CP involvement.
Subjects
were classifled according
to type of motor impairment
(spastic,
athetoid,
or mixed),
degree of paralysis
(diplegia,
hemiplegia,
triplegia,
or quadriplegia),
and ambulation
status
(ambulatory
or nonambulatory).
Oral motor
impairment
was assessed
with a questionnaire
adapted
from Stallings
et al (3) that determined
the
presence
or absence
of feeding
problems
(tongue
thrust, fluid
or food loss during eating, coughing
or gagging
during mealtime, excessive
time required
for meals, or difficulty
in consuming
textured
foods).
calculated
Skinfold-thickness
(triceps,
thigh,
Instruments,
biceps,
and
calf)
measurements
subscapula,
with
Cambridge,
were
axilla,
Lange
MA)
chest,
calipers
according
taken
from
abdomen,
(Cambridge
to Lohman
nine
sites
suprailium,
Scientific
et al (13).
All measurements
were made in triplicate
to the nearest
0.5
mm by a single
investigator
(HGH)
and the mean
of the
measurements
was used for analysis.
If the subject had bilateral
involvement,
skinfold-thickness
measurements
were taken on
the least affected
side. The Jackson-Pollock
(10) sex- and
age-specific
equations
were used to estimate
fat-free
mass and
analyzer
(lOlA;
was
measured
RJL
Systems,
with
a
Detroit)
from
the
equation
of
Kushner
and
Schoeller
(16).
We reported
a precision
of r = 0.94 for the estimation
of
fat-free
mass with use of BIA compared
with 180 in normal,
healthy
children
(17). However,
the inclusion
of weight in the
BIA predictive
equation
may reduce its accuracy
in determining
change
in lean
body
X-ray
mass
(18).
absorptiometry
Body-composition
measurements
were
made
with Lunar
DPX-L
total-body
scan software
(version
1 .3Y; Lunar
Co.
Madison,
WI) while
subjects
were in a supine
position
as
described
by Mazess
et al (19). Scan modes were chosen
from
the measure
of anteroposterior
tissue thickness
as defined
by
the Lunar DPX manual
(20). This method
is based on the same
principle
as
that
of
dual-photon
absorptiometry,
which
has
been described
elsewhere
(21-23),
with two adjustments.
First,
the radioactive
source
is replaced
with an X-ray beam tube
behind
a k-edge
filter that converts
the X-ray beam into two
main energy peaks. Additionally,
the attenuation
of soft tissue
is now
measured,
rather
than
assumed
to be
constant,
which
provides
the simultaneous
measurement
of bone mineral,
fat,
and fat-free tissue (24). We reported
a precision
of r = 0.98 for
the estimation
of fat-free mass with use of DXA compared
with
pig
carcass
thropometric
0900
and
analysis
(25).
indexes
1000.
were
The
measurements
made
after
the DXA
of BIA
and
scan
an-
between
dilution
18
TBW
Schoeller
Woburn,
Body weight
was measured
to the nearest
0.1 kg with a
calibrated
digital
scale (Scale-Tronix,
Wheaton,
IL) with subjects
barefoot
and in light clothing.
Nonambulatory
subjects
were weighed
in a digital
sling scale (Scale-Tronix).
Body
height was recorded
to the nearest
0.5 cm with a stadiometer
(Scale-Tronix).
For those subjects
who were unable to stand,
segmental
height was recorded
with a tape measure.
or impedance
resistance
with subjects
in a supine
position
and their limbs away from
their trunk. Four surface
electrodes
were placed
on the right
side of the body
as described
by Lukaski
et al (1 5). TBW was
0.15
Anthropometry
analysis
impedance
Dual-energy
SUBJECTS
fat. We reported
a precision
of r = 0.87 for
of fat-free
mass with use of skinfold-thickness
compared
with DXA in normal,
healthy
children
impedance
bioelectrical
I 437
PALSY
which
was measured
by 180 (H218O) dilution
as described
by
et al (1 1) with each subject receiving
an oral dose of
g H218O/kg
body
wt (Cambridge
Isotope
Laboratories,
MA).
The bottle
was rinsed
once with tap water,
was
also
consumed.
For
those
individuals
with
severe
oral-motor
impairment,
a sipping
cup and lid were used in
place of the normal dosing container
to maximize
consumption
of the isotope.
To account
for any fluid loss, a preweighed
towel was placed
around
the bottle containing
the dose. The
towel was then reweighed
to determine
the amount
of spillage,
if any.
A baseline
urine sample was collected
before administration
of the oral dose. After subjects
had fasted overnight
from 1800,
two additional
urine
samples
were collected
the following
morning
and 10 d later. This time between
the administration
of the dose and the collection
of the urine samples
allowed
for
the isotope to become
equilibrated
with TBW. 180 in the water
was
measured
in triplicate
with
an
isotope-ratio
mass
spectrom-
eter (model
Sira2;
VG, Cheshire,
United
Kingdom)
at the
Biomedical
Mass Spectrometry
Facility
at the University
of
Vermont.
TBW
was calculated
as the 180 dilution
space
di-
Downloaded from www.ajcn.org at Norris Med Lib Serials Sect on July 11, 2008
dual-energy
X-ray
absorptiometry
(DXA),
and bioelectrical
impedance
analysis
(BIA) in determining
body composition
in
adults with CP by comparing
these measures
with body-composition
measurements
obtained
from TBW
as estimated
by
180
dilution.
The second
objective
was to develop
a simple
anthropometric
equation
to predict
body composition
in adults
with CP by using 180 as a reference
method.
WITH
1438
HILDRETH
vided
by
1.01,
nonaqueous
correcting
oxygen
derived
from
73.2%
hydrated
Statistical
exchange
body
with
TBW
label with
mass was
fat-free
mass is
of the
180
( 1 1). Fat-free
solids
the assumption
that
(26).
analysis
Independent
test
for
of
and
Student’s
between
means.
t tests
for differences
paired
The
were
t tests
strength
used
to
of the relation
between
variables
was examined
by using simple
regression.
Agreement
between
the estimates
of body composition
(skinfold thickness,
BIA, and DXA) and TBW by 180 dilution
was
assessed
with
method,
a pair-wise
the method
of Bland
and
Altman
(27).
With
this
ET
AL
strength
of the relation
as determined
by simple
regression
was
fat by 180 and that by DXA
(,2
0.90, P < 0.001),
followed
by skinfold
thickness
(r =
0.75, P < 0.001),
and BIA (r = 0.67, P < 0.001).
The level
of agreement
between
percentage
body fat by 180 and percentage body fat by DXA was favorable
(Figure
2), with a relative
strongest
bias
between
percentage
of 0.06%.
The
and the limits
DXA appeared
fatness
tion
95%
tween
CI for
the bias
was
2.16%,
-2.05%,
of agreement
ranged
from 9.67%
to be negatively
biased at the high
(percentage
was
body
body
supported
the mean
fat
> 50%)
(Figure
by the significant
of percentage
2). This
negative
body
to -9.56%.
end of body
fat by 180
observa-
correlation
beand by DXA and
and by DXA (r =
Scotts
analysis
centage
bias
(mean
difference)
difference
age
± 2 SD
and
of the
the
limits
difference)
to show
of
the relative
agreement
between
the
fat from
180 in the sample.
Multivariate
using
BMDP
statistical
software
by
Angeles).
The
level
of significance
for
(mean
estimates
of
body
Valley,
CA).
Stepwise-multiple-correlation-regression
was used to determine
the best predictors
performed
Los
is used
of percentanalyses
(BMDP
were
Inc.
analyses
was
all
at P < 0.05.
specified
fat
from
skinfold
body
fat
and limits
of
(Figure
4).
thickness,
by
agreement
Multivariate
which
with
180
underestimated
a mean
ranging
difference
from
18.62%
Stepwise-multiple-correlation-regression
Subject
The physical
characteristics
of the study group are summarized in Table
1. The male and female subjects
were comparable in age, weight,
and height. The women
had significantly
higher BMIs (in kg/m2) than did the men (27.9 compared
with
24. 1, respectively).
The
extent
of
the
manifestations
of
CP
Table
2. Most of the
group had spastic CP (75%), had diplegia
(40%), were ambulatory
(55%),
and had no feeding
difficulties
(75%). The distribution
of percentage
body
fat in the men and women
as
estimated
from
80
is shown
in Figure
1 . On average,
the
women had higher percentages
of body fat as estimated
by 180
within
the
±
study
group
1 1.8%)
Bivariate
than
is
did
presented
the
in
men
(25.2
± 8.2%)
(P
=
0.001).
On average,
percentage
1 1.4%) was significantly
(23.4
±
fat estimated
by BIA (36.3 ±
and that estimated
by skinfold
body
higher
12.2%)
was
significantly
lower
than
not
with
cerebral
34.5
Age (y)
Height
(cm)
Weight
(kg)
163.1
BMI (kg/m2)
‘t
2
±
SD; n
Significantly
=
20.
different
=
±
Women
10.9
± 8.5
64.3
±
24.1
± 4.6
13.7
(n
=
37.4
± 9.7
152.1
± 4.7
64.1
±
18.6
27.9
±
8.62
7)
men,
P < 0.05
(independent
t test).
add
a second
entering
the individual
dent
variables
first
ness
variable
(r
=
was
of percentage
fat from
body
as height,
used
body
BIA,
weight,
sex,
to the
Percentage
180.
fat by
skinfold
age,
degree
explained
in
the
fat
-
(7.34
sex,
weight,
following
X sex)
as 0
4.85).
to the
DXA
as potential
and
indepen-
and height.
The
skinfold
thick0.80),
and sex
equation:
=
skinfold
=
nonresearch
omitting
measures
in
of percent-
impractical
performed
to age,
generating
biceps
applied
be
was
in addition
body
SEE
measurement
skinfold
sex is coded
and
of variation
may
analysis
+ (0.38
where
amount
Because
selected
by this model
was biceps
0.78),
followed
by weight
(R2
=
0.84),
=
+ (0.32
x biceps
female
thickness
When
present
X weight)
skinfold
and
1
is in mm
(1)
thickness)
male,
weight
0.84,
(R2
this
multiple-regression
sample,
it increased
power of skinfold
thicknesses
for estimating
fat (R2 = 0.84) compared
with the original
equation
(,2
0.75).
is in
P
<
equation
the
predictive
percentage
body
Jackson-Pollock
DISCUSSION
This
from
as well
body fat by
fat by DXA
settings,
was
13)
DXA,
8.76
palsy’
Men (n
and
analysis
predictors
Percentage
significantly
0.001,
of adults
strongest
sample.
percentage
age body
kg,
TABLE
1
Characteristics
the
percent-
age body fat estimated
by 180 (29.8
± I 1.2%)
(P < 0.001).
Percentage
body
fat estimated
by DXA (29.7 ± 13.9%)
was
not significantly
different
from that estimated
by 180.
The
-5.97%
of paralysis,
ambulation
status,
and oral motor
impairment
were entered
as independent
variables
in the analysis.
The only
variable
selected
by the model was percentage
body fat from
DXA,
which
explained
90% of the variation
in percentage
body fat estimated
by 180. Forcing
in the other variables
did
(R2
analysis
thickness
the
thickness,
characteristics
(38.1
in
180
to
6.33%
analysis
to determine
RESULTS
perof
techniques
study
evaluated
of measuring
the accuracy
body
and
composition
practicality
in adults
of existing
with
CP.
Downloaded from www.ajcn.org at Norris Med Lib Serials Sect on July 11, 2008
body composition
and the reference
measure
(TBW
by 180
dilution)
by plotting
their mean difference
against the mean of
the two methods.
Bivariate
analyses
were performed
by using
QUATFRO
PRO FOR WINDOWS,
version
6.0 (Borland,
the difference
in percentage
body fat by 180
-0.63,
P = 0.002).
The mean difference
between
percentage
body fat estimated
by 180 and that estimated
by BIA was -6.55%,
indicating
that
BIA was biased
toward overestimating
percentage
fat (Figure
3). The limits
of agreement
ranged
from
7.06%
to -20.15%.
Another
bias can be applied
to the measurement
of percentage
comparison
BODY
TABLE
Number
COMPOSITION
IN ADULTS
WITH
CEREBRAL
PALSY
I439
2
of patients
in the
sample
with
a given
level
of cerebral
Hemiplegia
palsy
involvement’
Diplegia
Triplegia
Quadriplegia
Total
Percent
%
Motor
involvement
Spastic
3
8
2
2
15
75
Athetoid
0
0
0
2
2
10
Mixed
0
0
1
2
3
15
Ambulatory
3
5
2
1
11
55
Nonambulatory
0
3
1
5
9
45
Ambulation
status
n
‘
20.
The
major
findings
percentage
body
were
fat
it unacceptable
making
the
limits
for clinical
prediction
equation
1) BIA
agreement
of
overestimated
were
The sample
population
tope of deuterium
as part
wide,
use; 2) the Jackson-Pollock
underestimated
requirements
body
percentage
fat and the limits
of agreement
were wide,
making
it also
unacceptable
for clinical
use; 3) DXA had no bias and good
agreement
with percentage
body fat from 180
making
it an
accurate
measure
of body composition;
and 4) biceps skinfold,
weight,
and
sex
percentage
body
were
the
best,
easily
measured
predictors
of
fat in this sample.
water
Because
as a reference
as the
reference
because
of
there
is no true
death
occurs,
a criterion
accuracy
of other
easily
measure
of body
of body
method
applied
composition
unless
must be used to determine
measures.
In this sample,
the
for
the reasons
discussed
in the Introduction,
isotope dilution
was
chosen
as the criterion
measure
of body
composition.
One
possible
limitation
of this method within this sample may have
been the presence
of a compromised
hydration
status, which
would interfere
with the ratio of extracellular
water to TBW. If
this
were
from
true,
the
the assumed
reason
to believe
hydration
value
that
of fat-free
of 73.2%
the
hydration
(26).
mass
would
However,
status
deviate
there
of this
is no
population
should differ from a normal,
able-bodied
population
except in
the most extreme
cases of leanness
and obesity,
in which the
ratio of extracellular
water to TBW may be increased.
method
The
by
in the
and
reasoning
exchanges
dilution
space
doubly
present
180
study
180
has
hydrogen
was
corrected
for.
mass
were
greater
values
observed
in Berg
This
for percentage
and
generated
weight,
body
body
Isaksson’s
(30)
to be
a dilution
In our study, calculations
analysis
overestimated
even after nonaqueous
negative
values
study
that
with use of
TBW in the
exchange
of
of fat-free
consequently
fat. This
I%
space
yielding
outcome
was
on children
adolescents
lations
for
creased
with
TBW
cost
of
the accuracy
nor research
CP, Bandini
indicative
impedance
Bioelectrical
underlying
et al (6) obtained
of malnutrition.
reliable
Despite
by
in
calcuthe in-
analysis
length
40
this
,0
the
system
of the
biggest
00
BIA
(ie,
the
conductor
question
population
method,
and
when
the
these
the
body
main
impedance
of a
is dependent
its configuration
considering
because
that
body)
human
use
components
on the
(32),
raises
the
of this
method
in
be altered.
It
cannot
is extremely
difficult
to obtain
an accurate
with CP because
of their muscle
contractures.
30
CP
dilution
compared
with deuteriurn,
of 180 make this method
the supe-
isotope
180
and precision
technique.
geometrical
#5o
also
with
and could indicate
an analytic
error of the type described
Ritz et al (31). Using 180 as a measure
of body composition
principle
0
the
chosen
deuterium
is proposed
deuterium
than
both
was
over
is up to 5% greater
(29).
the data from deuterium
extreme
cases of leanness
that
labeled
with
precision
in determining
TBW
for this superiority
is that less
with nonaqueous
organic
cornfor
whereas
the
measured
techniques.
accuracy
TBW,
using
was
isotope-dilution
its
than
CP
TBW
The inability
of BIA to accurately
predict
percentage
fat in the sample
may be related
to several
factors.
The
60
.,
with
Hence,
(1 1). The proposed
180
than deuterium
pounds.
measure
(28).
and deuterium
180
greater
Isotope
dilution
composition
in adults
technique
was also administered
a stable isoof another
study examining
energy
height
in individuals
Height
(length
of
the conductor)
is an important
variable
in the statistical
relation
of BIA to percentage
body fat. An over- or underestimation
of
:
120
10
height
by 2.5 cm can
result
TBW
(32),
a small error in the estimation
5%).
The second
major problem
centage
assumption
0
FIGURE
isotope
dilution
cerebral
palsy.
1. Distribution
in adult
of percentage
men
(U;
n
=
body
I 3) and
fat as estimated
women
(0;
n
by
=
180
7) with
producing
body
fat (<
of a symmetrical
(32). The body
this assumption
While
released
in a l.0-L
error
configuration
of the
asymmetry
observed
in this
of a constant
configuration
this
study
from
the
was
National
being
conducted,
Institutes
in the estimation
human
population
invalid.
a
of Health
statement
regarding
of
of peris the
body
renders
was
the
Downloaded from www.ajcn.org at Norris Med Lib Serials Sect on July 11, 2008
skinfold
as follows:
and
HILDRETH
1440
ET AL
10
a
8
6
4
U
2
0
.0
-2
‘a
U
-4
U
-6
.
U
U
-8
-10
0
10
20
of percentage
FIGURE 2. Difference between percentage body fat estimated by 180
of the measurements
of percentage body fat by 180 and by DXA (n =
standardization
curate
(32).
and
The
and
control
of numerous
reliable
assessment
impedance
value
variables
of body
for
composition
is believed
to be
the
body
dilution
by
mm
depending
on the
size
of the
individual.
and
and by dual-energy
by
X-ray
60
DXA
absorptiometry
(DXA)
plotted
against
the mean
easy to apply in a clinical
setting.
Hydration
status and postprandial
measurement
were controlled
to the best of our ability
by requiring
the subjects
to remain
in the Clinical
Research
Center
overnight,
thereby
prohibiting
any excessive
activity
and allowing
for the consumption
of only water after administration
of the isotope.
Any disturbances
in intracellular
hydration
status associated
with the condition
of CP could obviously not be controlled,
nor could changes
in hydration
status
related to menstruation
for the female
subjects.
When controlling for all these variables,
a clinician
can no longer consider
BIA an easily applied,
practical
measure.
acthe
amount
of time that the subject
is recumbent.
The longer the
length of time that the subject
remains
in the supine position,
the higher
the impedance
measure.
This measure
is dramatically increased
within the first 10 mm of a subject assuming
the
supine position
and continues
gradually
thereafter
(32). All of
our subjects
were measured
after a DXA
scan,
which
lasted
20-50
fat by 180
50
20).
by BIA
affected
40
In relation
to body positioning,
some obese subjects
were unable to separate their legs completely,
which may have interfered
with the
conductance
of the electrical
current
through
the body.
The remaining
variables
that affect the reliability
and accuracy of BIA were unknowingly
controlled,
but may not be as
Skinfold
Indirect
skinfold
anthropometry
measures
thicknesses
of body
composition
such
have to rely on the statistical
as BIA and
association
10
5
0
4)
Co
.0
U
a
U
-.-
-
.-
-
.
-10
.00
I :;:
-25
10
20
Mean
FIGURE
3. Difference
measurements
of percentage
between
body
percentage
fat by ‘0
30
of percentage
fat estimated
by
and by BIA (n = 20).
body
180
40
body
50
60
fat by 180 and by BIA
and by bioelectrical
impedance
analysis
(BLA) plotted
against
the mean
of the
Downloaded from www.ajcn.org at Norris Med Lib Serials Sect on July 11, 2008
Mean
30
BODY
COMPOSITION
IN ADULTS
20
WITH
PALSY
1441
U
15
ce-,
CEREBRAL
U
U
0
C.)
U
10
8)
U
U
U
U
U
4)
U
U
U
.,0
U
0
20
U
.-
U
-10
-
10
Mean
4. Difference
FIGURE
plotted
against
the
of the method
ence
between
mean
with
population
of the
than
reason, these methods
that differs substantially
Although
the
of percentage
body
body
fat derived
body
on
of percentage
body fat estimated
percentage
measurements
percentage
rather
20
a biophysical
from
model.
equation
(10),
used
required
for the estimation
of percentage
of both
the accessibility
this
chine
by
fat from
X-ray
absorptiometry
Of all the body-composition
techniques
examined,
DXA was
by far the best method
for predicting
percentage
body fat in
comparison
with TBW
by 180
The high correlation
with
percentage
fat from
180 (r
=
0.90, P < .001),
supported
by
the strong agreement
as determined
by the method
of Bland
and Altman
(27), made this an acceptable
method
to substitute
for 180 isotope
dilution.
However,
there are several limitations
to consider
before this method
is advocated
for routine
use in
adults with CP.
DXA was more likely to be negatively
biased at the high end
of percentage
body fatness
(percentage
body fat > 50%) in this
sample.
from
However,
only
this
three
observation
individuals
and
was
should
based
on
certain
the
burden
position
placed
within
remaining
motionless
feasibility
of this
the
during
technique
on
the
subject
dimensions
the
is also
from
of the
20-50-mn
questionable.
scan,
attaining
scan
the
The
table
clinical
expense
a
or
the
the
scan.
rejected
Finally,
the
Prediction
equations
( lO)J
is a limiting
setting.
its use
radiologic
amount
factor
in
If a DXA
is still
limited
technician
of time
mato ad-
required
for
the
of the results could be anywhere
ofthese
limiting
factors,
DXA was
easily
with
scan
use,
a licensed
as a practical,
in adults
sition
actual
in a clinical
for clinical
of
applied
measure
of body
compo-
CP.
equation
In clinical
practice,
anthropornetric
data are more practical
for the assessment
of body composition.
The ease of administration
and lack of highly
sophisticated
equipment
required
make
nience
anthropometric
and cost.
measurements
ideal
Skinfold
anthropornetry
in terms
has been
of convecriticized
because
of its high susceptibility
to interobserver
error.
However, this can be controlled
by having
a properly
trained
mdividual
perform
the measurements.
A precision
within
5% can
be achieved
measurements
determined
measures
in this manner
(33).
correlated
strongly
by 180 in the Bandini
of skinfold
thicknesses
multiple-regression
model
Because
skinfold-thickness
with percentage
body
fat as
et al (6)
were entered
to determine
study,
into
whether
correlation,
combined
with
of agreement,
was found
by 180 and three simple
sex,
and
weight).
The
for any variation
to sex
body
inclusion
in the body
differences
small
sample
with
in the
several
180
as
in the
sample.
it is vital
its ability
for
whom
body-composition
a reference
and
predictive
by generating
CP. A strong
favorable
limits
percentage
fat as estimated
(biceps
skinfold
thickness,
of sex
size,
CP
of bias
fat estimate
to determine
fat in adults
derived.
In summary,
compared
with
a lack
between
measures
individual
a stepwise-
the
power of skinfold
thicknesses
could be increased
a regression
equation
specific
to adults
with
be cross-validated
sample.
Besides
available
prediction
20).
=
and
requirement
the study’s
in a larger
(n
DXA scan and interpretation
from 0.5 to 2 h. On the basis
utable
measures
be replicated
use of Jackson-Pollock
of DXA
is readily
the
50
thickness
the instrument
minister
the Jackson-Pollock
(10) equation,
were difficult
to obtain in
the wheelchair-bound
subjects.
This could have been a potential source
of error in estimating
percentage
body
fat in this
sample
with use of the Jackson-Pollock
equation.
Dual-energy
by skinfold
a refer-
to predict
body
fat by 180 and
For
percentage
fat from skinfold
thicknesses,
was validated
with
the well-established
method
of hydrostatic
weighing,
it was
derived
from a healthy
adult population.
The suppressed
correlation
coefficient
(r = 0.75)
and poor
agreement
with the
reference
measure
of percentage
body fat from 180 confirm
the
inability
of the Jackson-Pollock
equation
to accurately
predict
percentage
body fat in this sample.
In this population,
suprailiac,
abdomen,
midthigh,
and subscapular
skinfold
thicknesses,
40
by ‘o and by skinfold
thickness
[with
fat by ISO and by skinfold
thickness
are usually not effective
for a population
from the original
reference
population.
Jackson-Pollock
30
standard
equation
adjusted
that would
However,
be attribbecause
that this new
to predict
the
percentage
formula
was
techniques
in adults
of
equation
not
were
with
CP.
Downloaded from www.ajcn.org at Norris Med Lib Serials Sect on July 11, 2008
0
1442
HILDRETH
Neither
BIA nor predictive
equations
for skinfold
thickness
generated
from normal,
able-bodied
adults accurately
determined percentage
body fat in the sample.
DXA was an accurate
and precise
method
but is not practical
for routine
clinical
use.
A new prediction
equation
including
biceps skinfold
thickness,
sex, and weight
was developed
that explained
84% of the
individual
variation
in percentage
body fat, thus providing
some
direction
for the development
of new equations
for
predicting
body composition
in adults with CP. With an accurate
indicator
of body composition,
clinicians
can monitor
changes
in nutritional
status and evaluate
the effectiveness
of
nutrition
with CP.
intervention
to improve
quality
of
life
for
14.
15.
16.
17.
the nursing
staff
of the Clinical
Research
Center,
John
manuscript.
their families,
in this research
enthusiasm
and
Most
their
importantly
caregivers,
we
who
20.
Hall
HC,
Bolonchuk
position.
J AppI
Kushner
RF,
electrical
impedence
Goran
MI,
Forbes
l986;60:
1327-32.
DA.
analysis.
MC,
22.
REFERENCES
Am
WH,
composition
WA.
composition.
Am
J Clin
Lunar
Corporation.
DPX
bone
WI: Lunar
A, Borg
by
of
com-
body
water
body
Poehlman
ET,
in young
by bio-
l993;75:
Ravussin
E,
from
children
1776-80.
Is bioimpedance
predic-
a good
Am J Clin Nutr 1992;56:4-6.
Nutr
1990;51:1l06-12.
densitometer.
Operators
Corporation,
J, Christiansen
A, Borg
in vivo
Validation
1986;44:4l7-24.
dual
C, Mazess
J, Christiansen
photon
manual
version
1990.
RB.
in vivo by dual photon absorptiometry.
Clin Physiol 1984;4:343-55.
Gotfredsen
X-ray
Barden
HS, Bisek JP, Hanson
J. Dual-energy
x-ray
for total-body
and regional
bone-mineral
and soft-
tissue
Gotfredsen
JM.
change?
l996;63:299-305.
Nutr
body
Cross-
Nutr
of total
Carpenter
W, Amatruda
G.
dual-energy
Siders
J Clin
J Appl Physiol
resistance.
Simon
Hunger
to assess human
Estimation
AM. Estimating
GB,
CB,
method
Physiol
TR,
reference
against
J Clin
impedance
Schoeller
Kaskoun
Am
WW,
bioelectrical
mineral
cedures.
project.
Nagy
Lukaski
8/90. Madison,
21.
with
RK,
techniques
19. Mazess
RB,
absorptiometry
the
participated
Driscoll
calibration
of body-composition
absortiometry
in young
children.
tor of body-composition
Hiser
thank
P, Johnson
MI,
standardization
1988:55-80.
C, Mazess
absorptiometry.
Total
body
bone
I. Measurement
RB.
II.
pro-
body mineral
Total
Accuracy.
Clin
Physiol
1984;4:357-62.
1. Ferrang
TM,
Johnson
assessment
of
MS. Dietary
RK, Ferrara
adults
with
cerebral
palsy.
J Am
and anthropometric
Diet Assoc
1992;
23.
body
92:1083-6.
2.
Krick
J, Van
Duyn
MAS.
The
relationship
between
VA,
Charney
EB,
Davies
JC,
Cronk
growth failure of children with quadriplegic
Child Neurol 1993;35:l26-38.
4.
Thommessen
oral-motor
M, Kase
BF,
Riis
intakes
difference?
6. Bandini
of disabled
LG, Schoeller
Body composition
7.
palsy
Evans
and energy
plan for survival.
8.
Granet
Nutrition-related
Dev Med
palsy.
A. The
KM.
Cerebral
Jackson
body
AS,
impact
palsy.
and
N EngI
body
make
Predicting
26.
a
with cerebral
28.
;29:70-7.
J Med
1994;330:
with cerebral
equations.
E, Peterson
DW,
we must
29.
1760
(letter).
ET. Athetosis
Schoeller
DA,
PD. Total
labeled
12.
Lee
13. Lohman
body
water.
MMC.
in chronic
C. Measurement
lean
van
of lean
absorptiometry.
Thickening
TG, Roche
Pace
N,
Bland
Dicta
W, Jaspan
30.
31.
of the subcutaneous
Hum
Biol
AF, Martorell
l959;31
tissues
with
180
in paralyzed
and
R. Skinfold
32.
2H
limbs
thicknesses
and mea-
Nutr
1993;58:589-9l.
CM,
Goran
by dual-energy
in the
pediatric
EN.
33.
Altman
Metab
Johnson
RK,
body
Studies
on
Hildreth
HG,
of
not yet
MI.
Cross-calibration
of fat
absorptiometry
to pig
weight
body
Statistical
range.
Am
composition.
content
methods
of clinical
SB. Use
studies:
X-ray
combined
nitrogen
l945;l58:685-91.
DO.
two methods
for
measurement.
Contompasis
Nutr
body
to fat
111 The
in relation
assessing
Lancet
SH,
J Clin
agreement
l986;l
Goran
MI.
Total
:307-10.
energy
expenditure
in adults with cerebral palsy as assessed by doubly labeled
water. J Am Diet Assoc l997;97:966-70.
Goran
MI, Poehiman
ET, Nair KS, Danforth E Jr. Effect of gender,
body composition,
and equilibrium
time on the 2H-to-’8O
dilution
ratio.
Am
Berg
K, Isaksson
with
cerebral
Ritz
P, Johnson
water:
J Physiol
1992;263:E1119-24.
B. Body
palsy.
Acta
PG.
composition
Paediatr
Coward
analytical
WA.
considerations
Statement.
surement.
Bioelectrical
impedance
Bethesda,
MD: National
Plenum
nutrition
of school
children
l970;204(suppl):4l-52.
Measurements
and
of 2H and
physiological
in
180
implications.
1994;72:3-12.
Institutes
Cameron
and
Scand
National
Tanner
:187-93.
J Clin
Duthie
between
body
Exerc
J, Klein
Am
Rathburn
JM,
B, Heymsfleld
in body-composition
TR,
measurements
Br J Nuts
Santen
water measurements
in humans
Am J Clin Nuts l980;33:2686.-93.
hemiplegia.
SJ, Nagy
lean
JJ, Dawson-Hughes
absorptiometry
standard.”
Pintauro
space
in-
palsy. J Am Diet
of body density,
Med Sci Sports
accuracy
volume
why
x-ray
water and chemically
content. J Biol Chem
1977;9: 197-202.
11.
J, Christiansen
fat using dual photon
body
R, Kehayias
carcass analysis
1996;63:293-8.
27.
MS, Poehlman
Roubenoff
and
Li, Dicta WH.
in adolescents
1991
25.
of feeding
problems
NK, Wykes
rate in adults
ML.
total
feeding
E. Cerebral
palsy:
1990;65:1329-33.
MI, Ferrara
Pollock
weight,
Res
Albennan
Dis Child
Arch
RK, Goran
do
expenditure
Pediatr
creases resting metabolic
Assoc 1996;96:l45-8.
10.
J, Borg
and total
dual-energy
l991;91:1522-5.
DA, Fukagawa
or myelodysplasia.
PM, Evans
SJW,
9. Johnson
children:
J Am Diet Assoc
24.
with
a “gold
CE.
cerebral
G, Heiberg
in-
population
problems on growth and energy intake in children with cerebral palsy.
Eur J Clin Health l991;45:479-87.
5. Thommessen
M, Riis G, Kase BF, Larsen 5, Heiberg
A. Energy and
nutrient
A, Jensen
mass
Clin Exp 1986;35:88-93.
volvement
and growth: a pilot study in a pediatric
cerebral palsy. J Am Diet Assoc 1984;84:555-9.
3. Stallings
Gotfredsen
N. The
of
methods
JM, eds. Human
Press, 1978:35-90.
Health
Technology
ofanxological
growth.
Assessment
analysis
in body
Institutes
of Health,
anthropometry.
2 Postnatal
Conference
composition
growth.
mea-
1994:12-4.
In: Falkner
New
York:
F,
Downloaded from www.ajcn.org at Norris Med Lib Serials Sect on July 11, 2008
volunteers,
of this
Goran
bioelectrical
Laboratory
of the Sims Obesity
and Nutrition Research Center, Dave Ebenstein
and Bruce O’Rourke
from
the Biomedical Mass Spectrometer
Facility at the University
of Vermont
for their technical expertise, and Diantha Howard for her statistical insight
development
In: Anthropometric
IL: Human Kinetics,
Fontivielle
18.
We thank
surement
technique.
manual. Champaign,
tetrapolar
persons
and Kathleen Olwell from the Physiology
in the
ET AL