Page 1
Bioelectricimpedancephaseangle andbodycomposition13
16
Am JClinNuir
l988;48:l623.Printedin USA.© 1988AmericanSocietyfor ClinicalNutrition
RichardNBaumgartner,PhD;W CameronChumlea,PhD; andAlexFRoche,MD,PhD
ABSTRACF
tion
amount
resistance
leg, and
was determined
(4??)
had significant
andpositive
4’t was associated
weight/stature2
%BF in clinical
Theuse of bioelectric
in 53 males
(X?)
in a conductor
reactance
Weight,stature,
fromdensitometry.
(p
correlations
significantly
in eachsex. Bioelectric
andsurvey
impedance
69 females
relative
were determined
andskinfold
Phase
negative
fatfree
mass
with %BF after
phaseangle for predicting
Thephase
of resistance
body
measured.
(4?),
leg
body
multiple
body
composi
was determined
of reactance
(R) and
trunk.
and962
to the
for the whole
thicknesses
anglesfor the trunk
correlations
(FFM)
controlling
phaseangle
Am J Clin Nutr
y ofage.
amount
angle
describes
Bioelectric
the
(R).
(X?)
and separately
Body
(4?),
fat (%BF)
regression
skinfold
be useful
for arm,
were
composition
whole
in each
analyses,
thickness,
for predicting
and
body
sex,
< 0.05)
with
with
in
percent
males.In
for age, mean
for the trunk
l988;48:1623.
and
may
research.
KEY WORDS
skinfold
Body
composition,bioelectric impedance, fatfreemass, percentbodyfat,
thickness
Introduction
Theprinciples
for
of bioelectric
> 40y (14)
of body
specifically
extracellular
comparatively
phasesensitive
useof bioelectric
tion in thefields
ogy (10),physiology
14). Bioelectric
square root of the
impedance
methods
have
for
bioelectric
(5),
mass
of
interest
body
(8, 9), human
sports
measured
squares of resistance
been es
tablished
components
pedance,
and
butestimating
composition
total
water
recent.
electronics
impedance
ofhuman
(1 1, 12),
impedance
sum
from im
bodywater
fatfree
availability
increased
to estimate
nutrition
and
(Z),
of the
intracellular
(6),
The
and
(7),
are
accurate,
hasin the
composi
biol
(13,
is the
(R)
medicine
in ohms,
andreactance(Xe), or
z2
R2 + ?#{231}2
and
the
ofan
resistive
terfaces
of electric
lag behind
is quantified
of the
phase
impedance,
quency
1 . At very
it is frequency
pureopposition
alternating
effect
and
dependent.
ofa
electric
due to capacitance
membranes.
chargeby
thevoltage,
geometrically
of reactance
(0). The
resistance,
of an electrical
lowfrequencies
Bioclectric
conductor
whereas
produced
Capacitance,
causes
a phase
as the angular
(X?jto
geometrical
reactance,
current
(ft) the
resistance
to the
reactance
by tissue
or the
the
shift.
transformation
(R),
is
biological
current
flow
is the
in
cellstorage
a condenser,
creating
current
This
to
shift
ratio
angle
resistance
relationships
phase
illustrated
capacitive
or the
among
and
in Figure
component
angle,fre
are
of the
tance
is purely
tance
the
critical
critical
portion
very
sentially
is purely
Most
phase
early
angles
bolic
variation
cally,
fixed
tative
in
system
is equal
is effectively
to zero
(R,).
increases
angle to open
frequency
frequency
to resistance
high
frequencies
short
resistive
biologic
angleat the critical
investigations
withphysiological
rate(1, 2), but
in phase
variation
frequency
behavior
cell size,
an
the
open
measured
frequency
circuit so that
impedance
increases,
reac
and
As the
in proportion
until
specific
reactance
with
the
so that
again
systems
(Z)
resistive reac
(Xe)to resistance causing
phase a maximum
to the
begins
increasing
capacitive
the
(R?).
are highly
frequency
onthe
variables,
is known
among
individuals
be dueto differences
tissuesassociated
permeability,
is reached
Beyond
at a
the
(fe)
the
system.
to decrease
frequency
component
measured
in pro
and
is es
at
(fh)
circuitedimpedance
once
conductive
is small
associations
such
about
individuals.
in phase
and
Several
of phase
meta
sources
Theoreti
angles
in the
withvariability
or intracellular
the
(15).
focused
as basal
the little
angles
of
(1)
among
could
oftheir
membrane
at a
capaci
I From theDivision
University
by
of Human
School
HD12252
Biology, Department
Dayton,
theNational
of Pediatrics,
WrightState ofMedicine,OH.
2 Supportedgrant fromInstitutesof
Health,
Bethesda,
MD.
3 Address
Human
Xenia
ReceivedApril23,
Accepted
reprint requests
Wright
Yellow
to Richard
University
OH
N Baumgartner,
School
Division of
Biology,
Avenue,
State of Medicine,1005
Springs,
1987.
45387.
for publicationSeptember 1, 1987.
by guest on July 10, 2011
www.ajcn.org
Downloaded from
Page 2
U
U
R?
R0
impedance, or
PHASEANGLEAND BODYCOMPOSITION
17
Resistance
FIG1. Impedance
(R), reactance
conductor
plot
(Xe),
is maximum
illustrating
and
the
angle
relationships
(4). The phase
frequency
between
resis
for atance
biological
phaseangle
at a critical
(fe).
composition
bution
affect
interstitial
To
used
tion
water
this
cylindrical
or associated
body
amount
spaces
bioelectric
measurements
with stature
or offatfree
use is Ohm’s
conductor
withdifferences
individuals,
ofthe
in the
which
through
distri
may offluids
ofshunting
(15).
among
the currentthe
date, studies
ofbioelectric
squared
mass
Law,
to its length
ofbodycomposition have
resistance
as an
The
relates
squared
in combina
of total
(52)
index
theoretical
the
body
(FFM).
which
basis for
volume
divided
ofa
by its
VpL2/R(5)
(2)
Reactance
small
pedance
tance
(7).
composition
study
tance
has beenignored, perhaps
to resistance
conductors
of impedance
of the
studied.
use of measurements
for predicting
because
in the
and
than
angle
The
it is very
total(< 2%)
for most
is a better
Also,
in proportion
biologic
predictor
association
hasnot
to explore
phase angle
im
because resis
reactance
with
the
phase body
of this
of reac
been
the
purpose
was
andbody composition.
Materialsand
Methods
St udy population
Thestudy population
69 white
24 ofthe
were
composition.
lean (< 5%
there
levelsof percent
significantly
population
consistedof 53 white
y. Twentyfour
18 y ofage.
anycriteria
in the population
[%BF])to obese
differences
fat, the adults
children (age
in Table 1.
males,
ages
9
62 y, and
males
individuals
body
very
though
mean
females,
females
selected
Individuals
body
significant
ages 958
were
by
of the
normal
to their
from
and
<
These
notrelating
ranged
fat
(> 50%BF).
the
(age
>
18 y). The
Al
forwerebetween sexes
18 y) were
body
the fatter
is described
than
<
study
Bioelectric impedance
Bioelectric
right
using
MI).
eliminates
impedance
body
101impedance
device
electrode
measurements
for the arm,
analyzer
utilizesa fourelectrode
polarization
were
leg, and
(RJL
takenon the
side
a BIA
This
of thetrunk, whole
body
Systems,
arrangement
the
Detroit,
that
andmeasures
resistance
and
tric
taken
12 h and
using
were taken
sides
bration
400 ?l precision
The
electrodes
wholebody
anterior
face of the
pair
of electrodes
foot and
theanterior
electrode
electrode
pedance
proximal
the
leg
and the other
anterior
suredby attaching
tionson the
surfaceof the shoulder
midpoint ofa
and the source
Measurements
duringquiet
dardize
trunk
rationwas
rationand
participants.
sistance
of ?4.8%
midrespiration.
tanceby
tance.
reactance
current
at ?0930
ofa
800
conductor
?zA
h, after
bladder
cream
the participant
touching
impedance
resistor.
locations
used
were described
measurements
surface ofthe
hand
to injection
50 IdIa.
participant
empty.
tape.
ofan
measurements
had fasted
electrodes
All impedance
supine,the
andthighs
was checked
alternatingdee
wereat and
each
The
for at least
were attached
measurements
relaxed
separated.
daily
with theThe
electrodeand
with
arms
at the
callbutnot thebody,
analyzer
The
againstofthea
forplacement
previously
the
foot and ankle
wrist. For
was placed
the other
ofthe
in the same plane
5 cmproximal
wasmeasured
thigh at thesame
butwith the source
pairof electrodes
midlineof the
one pair ofelectrodes
hand and
ofthe
in detail
source and receiving
(16).Briefly,
placed
on the posterior
of the
locations
of electrodes
thigh with
as the gluteal crease
to the measuringelectrode.
by placingone pair
locations
used
andreceiving
on the sternal
neck. Impedanceof the
at the standard
wrist andthe other
withthe measuring
between theacromial
electrode 5 cm medial to this
forthetrunk were
breathing in allparticipants.
impedancemeasurements
demonstratedby datarecorded
maximumexpiration
Maximuminspirationproduced
overthe resistance
Maximum expiration
1.2%.Thephase of respiration
for
the
sur
electrodes wereon
and
and measurement
at the standard
pair
proximal
leg, one
on the
ankle andwas placed
the measuring
andthe source
Trunk
ofelectrodes
for measurement
electrodes
notch
arm
on
midline
im
on the
of
reversed
and
was
the
mea
loca
pair
on the anterior
electrode
process and
midpoint.
takenatmidrespiration
Theneed
withrespect
at maximum
a subsample
anincrease
ofthetrunk
decreased
did not
over
the axilla
the
line
to
to respi
inspi
of
in re
stan
for nine
measured
trunk
affect
at
resis
reac
?‘
Anthropometry
All anthropometry
procedures
Airlie
by two
used
1.0 cmusing
eat 0.1kg on a beambalance
measured
to the
subscapular,
data
similar
Report
and
werecollected
to corresponding
(17).
means
was measured
Weight was
scale. Skinfold
mm with
sites.
following standard
ized
the
made
were
closely techniques in
Consensus
observers
in the
a stadiometer.
All
ofpaired
measurementswere
the
measurements
to the nearest
to the near
thicknesses
atthe
analyses. Stature
measured
were
nearest
lateral
0.2caliperstriceps,
and
calf
Densitometry
Estimates
estimated
volume.
by use ofa
into
corded
microprocessor
weighed
lastthree
volume
nitrogenwashout
of%BFandFFM werederived
corrected
from
body
lung
was
by four
weights
display
participant
and the mean
body
density.
nearest
Model
density
residual
measured
load
were
with
from
Each
steel
underwater
participant’s
chair
of water
to thenearest
weight
underwater
for
weight
suspended
at
0.002
from
Underwater
kg from
to the load cells. Each
10 consecutive
used
to compute
onland
method with
a framecells
a tank 35#{176}C.re
a digital
a
attached was
underwater
weights
was
timesof the
were
Residual
L by
comput
measuredto the
a Gould
0.01
2180
a
by guest on July 10, 2011
www.ajcn.org
Downloaded from
Page 3
18
BAUMGARTNER ETAL
TABLE
Distributions
I
ofbodycompositionvariablesby sex and age*
Age
MaleFemale
>l8y
(n
= 29)
<l8y
(n
= 24)
>l8y
(n
= 44)
<l8y
(n
= 25)
Weight(kg)
Stature(cm)
MSK(mm)
BD (g/cm3)
%BF
FFM(kg)
83.32
181.44
10.14
1.047
22.74
63.72
±
12.33
6.79
1.38
0.016
7.16
6.46
±
±
±
±
±
48.53
159.73
7.87
1.05 1 ±
17.80
39.63
± 13.10
13.91
±
±
1.48
0.02 1
9.91
11.20
±
±
64.67
164.74
13.62
1.030
30.75
43.71
± 15.52
±
5.99
1.43
0.02
9.74
5.93
±
±
1
±
±
48.32
155.88
10.46
1.037
23.86
36.53
±
13.32
12.91
1.47
0.015
7.00
9.62
±
±
±
±
±
erized
ual volumes
either
with
not alterred
density
in the
25 y ofage
weight
spirometer. Previousstudies (18,
not
residual
and
was
with
with
calculated
19) showed
consistently
volumes
calculations
computed
corrections
age and
as the
that
different
measured
of %BF
from
for variation
resid
measuredon landarein
magnitude
the
or direction
in the
1.5%.
equation
offatfree
(2 1). FFM
100
%BF.
from
participantwaterare
by
Siri’s
>
The%BF body
with(20)
densitytissues
was
sex for individuals
product
<
of body
and

Statisticalmethods
Phase
angles for thewhole bodyand for each
body
segment
werecalculated in radians by the formula
(XJR)
‘l? = atan
and
computed
(MSK),
the
ceps,
by
total
18 y of age
Mean
compared
resistance,
MSK,
ling for age
FFM,
rately
and phase
ofphase
position
thropometric
age groups
(0, 1) for sex
sex, age, and phase
converted to degrees
an index
of subcutaneous
of thenatural
lateral
calf
(m) squared
adiposity.
andadults
phaseangles
using
t tests.
reactance,
and W/S2. Ageadjusted
werealso
MSK, andW/S2.
for each
sex and
anglesto determine
angles
with
after adjustment
indices.
andsexes
and including
by
of
multiplying
FFM.
by 57.297.
skinfold
was calculated
of the subscapular,
Weight
calculated
defined
as participants
for the age groups
werecomputed
anglesand
or partialcorrelations
betweenphase
and FFMwere
group
onage,
thestrength
fatfreecomponents
foralternative
Regressions also were
combined usingdummycoded
variables for
angles (22).
S2/R was
thicknessasMean
adiposity, an index as
average
and
stature
body
logarithms
thicknesses.
(W/S2)
Children
were
for the sexes
Correlations
and phase
tn
skinfold
(kg)
an
participants
divided
index was
were
defined
and
as
as
of
<
>
18 y.
were
between
FFM,
control
and
age, %BF,
calculatedangles
regressed
S2/R,
of the
%BF,
sepa
MSK,
%BF
age
W/S2,
associations
ofbodyfat and com
bioelectric
computed
and
with
variables
between
an
the
interactions
Results
Statistics
age
total
There
describing
(<
and
no
thedistributions
18 y) for
withineach sex
by group18 y vs
ofeach
statistically
>
phaseangles of the
body
segmentare presented
significant
in Table 2.
(p
weredifferences
>
although
gles
the
be larger
nificant
sex (males,
for the
arm
By definition,
reactance
Therefore,
0.05)among
the
thesex
leg,
and
arm,
age groups for phaseangles;
trunk,
to be consistently
Also,
in the children
(p
r =
0.44;
trunk were
or leg in each
and
larger
wholebody
in the
the
in the
correlations
females,
r =
and more
sex and age group.
phaseangleis positively
negativelyassociated
variationamong
phasean
tended
females.
malesthan in
tophaseangles
than
fortrunk tended
had
age in each
Phase
than
adults andsig
< 0.05)
negativewith
0.42).
variable
angles
for the larger
associated
with
measured
with
andresistance.
phaseangles at a
(5i
fixed
the
and
segments
in Table
ances
lations
both
were
frequency
associations
between
could
between
body
resistances
be interpreted
resistance
segments.
or reactances
and
and the
withresistances
agegroups,
highlywith
to some
and
extent from
‘
I
reactance
between
are shown
4 for adults.
strong,
for the
resistances
for the
within
body
by sex
Resist
corre
body
the
Resistances
Correlations
for
3 for children
for the
(r
>
0.70)
sexes and
correlated
in Table
leg hadarmpositive
whole in
and
those
forarm
leg.
TABLE
Bioelectric
2
phase angles(degrees)for body
segmentsby age and sex*
Bodysegment
Children
Adults
4
CV%
CV%
Trunk
Male
Female
Leg
Male
Female
Arm
Male
Female
Whole
Male
Female
body
11.17±2.16
10.13±1.63
7.98±1.02
7.57±0.82
6.27±0.85
5.66
±
1.32
6.25
6.21
± 0.68
±0.65
19.34
16.09
12.78
10.83
13.56
23.32
10.88
10.47
10.46±2.75
8.66±2.26
8.16±1.52
7.25± 1.11
7.38±
6.38
1.12
1.65
±
7.01
6.31
± 0.88
±0.69
26.29
26.10
18.63
15.31
15.18
25.86
12.55
10.94
* 1±
SD
by guest on July 10, 2011
www.ajcn.org
Downloaded from
Page 4
PHASE ANGLEANDBODY COMPOSITION
19
* Statisticallysignificant,p > 0.05.
TABLE
Matrix
measurements
3
ofcorrelationsbetween body
segments
or reactance
and the whole
by sex in children*
body
for
ofbioelectric resistance
Body
segment
Resistance or reactancet
Trunk LegArm
Whole
body
Trunk0.66
(0.76)
0.56
(0.70)
0.84
(0.86)
0.39
(0.49)
0.89
(0.84)
0.82
(0.87)
Leg0.40
(0.5 1)
0.44
(0.57)
0.22?
(0.49)
not in parentheses
Arm 0.85
(0.59)
0.92
(0.90)
Whole body
0.89
(0.72)
* Correlations are for boys;correlations in paren
theses are for girls.
t Correlations
relations
t Not
above
diagonal
diagonal
are for resistance
are for reactancewith reactance; cor
belowwithresistance.
statisticallysignificant,p > 0.05.
for
with
leg,
women.
strong,
body
correlated
the
actances
lation
whole
Correlations
each
and sex.
positively
for the
were
thetrunk,however,
body
wholebody
Similarly,
positive
in allfour
highly
trunk generally
for the
for reactance
body
were not
in the
correlated
boys
in
significantly
the
or
the
for the
the
arm
Reactances
strongly
leg, and
reactance
significant
andreactance
in Table
5 by
reactance were
< 0.05)
for the
in the boys and
correlated in the
whole
or
resistances or with
men
and
arm,
resistances
reactances
correlations
groups,
with
were
whole
in
notstatistically
between
segment
Resistance
correlated
body
segments
notsignificantly
the
arm
in
leg had
whole
were
the
for the
with
reactances
for
correlated
arm,
trunk
reactance
and
those
for
theleg. for
renot with
body,
the
and
and
the corre
in
men.
within
group
the
was in the
resistance
shown bodyare
age
and
(p
significantly,
whole
thegirls
women
body
but
and
they
thefor
TABLE
Matrix
measurements
4
ofcorrelations between bodysegments
or reactance
andthe whole
by sex in adults*
bodyfor
ofbioelectnicresistance
Body
segment
Resistance or reactancet
Trunk LegArm
Wholebody
Trunk
0.60
(0.72)
0.70
(0.66)
0.49
(0.70)
0.3l?
(0.59)
0.85
(0.95)
0.43
(0.61)
Leg
0.2Sf
(0.l7)?
0.34?
(0.044
0. l0?
(0.08)f
Arm
0.80
(0.67)
0.94
(0.88)
Whole
body 0.91
(0.85)
*Co?lations not in parentheses are for men;correlationsin paren
thesesare for women.
t Correlations
relations below
:1:Not statistically
above
diagonal
diagonal
are
are for reactance
for resistance
significant,p > 0.05.
with reactance;
resistance.
cor
with
TABLES
Correlations
segment
between
for the whole
resistance and
by age group
reactancewithin
and
eachbody
and bodysex
Body segmentChildrenAdults
Trunk
Male
Female
Leg
Male
Female
Arm
Male
Female
Whole
Male
Female
0.62
0.74
#{216}#{149}35*
0.50
0.73
0.63
0.27*
0.71
0.76
0.55
0.22*
0.20*
body
0.76
0.56
0.38
0.71
* Not statisticallysignificant,p > 0.05.
arm
lations
whole
in the extremities.
wereindependent
tancein the
resistance
to a degree
Correlations
fromdensitometry
age group.
positively
thewomen.
tivecorrelation
reactance
with%BF
Ageadjusted
or in the menfor
that
principally
Resistance
to some
ofthe
reactance
in all segments
of
thetrunk,
bioelectnc
leg, or arm.
impedance
These corre
for suggested
body
the
wasassociated
and
extent
especially
were independent
ofthe
resistance
are shown
for the
with%BF
Reactancefor the
with %BF
wassignificantly
in both themen
correlations
with impedance
in the
and
adults.
ofeach
men.
with
each
significantly
boys, the
a significant
and wholebody
negatively
women.
body
reactance
ofresistance
in the
trunk
reac
Also,
other
rest body
and
body
in the
and
in Table
trunk
in the
leg had
in the
and
and the
among
reactance %BF
and
and
and
nega
6 for sex
Resistance
correlated
was
men,
men
correlated
composition
TABLE
Correlations
6
ofresistanceand reactance for bodysegments
with %BF
Bodysegment
%BF
Chil
Male
dren
FemaleMale
Adults
Female
Trunk
Resistance
Reactance
Leg
Resistance
Reactance
Arm
Resistance
Reactance
Whole
Resistance
Reactance
0.67*
0. 15
0.170.56*
0.24
0.44*
0.190. 13
0.250.12 0.280.02
0.29
0.10 0. 19
0.34
0.09
0. 11
0.04
0.
0.2
10 0.05
0.271
body
0.230.050.300.17
0.050.08
by guest on July 10, 2011
www.ajcn.org
Downloaded from
Page 5
20
BAUMGARTNER ETAL
t Statistically
efficient.
significant,
p < 0.001. SD, standarddeviation of co
TABLE
Matrix
7
ofageadjustedcorrelations amongbody composition
variables and bioelectric impedance phaseanglesby sex*
%BF FFM W/52
MSK S2/R
?,
4?
?w
%BF
ITM
W/S2
MSK
S2/Z

0.67
0.48
0.66

0.740.59 0.29
0.46
0.38
0.42
0.62
0.34
0.57
0.39

0.77
0.89
0.44


0.51
0.57
0.65
0.41
0.39

0.26

0.55
0.40

0.28

0.9 1
0.37
0.49
0.49
0.50




0.52 0.33
0.41
0.44
0.52
0.79
0.46
0.55
0.61
0.87
0.53
4’I
4’,

0.76
0.61
0.54

0.38

0.82
0.84
0.34

0.81
* Males below diagonal,females abovediagonal.Nonsignificant correlations(p > 0.05)not shown.
and
R was
significant
%BF
males
0.34).
the
nificant
three
negative
gles
in the
although
in the
adjusted
slightly
(R
tion
that
that
and
0.52;
Results
?
are
forced
for the
enter
counted
ance
0.45).
after
the
tions
< 0.001)
ling
0. 10; females,
tion including
the variance
(rmse)
tion explained
rmse of
callysignificant.
bioelectric
correlated
variables
strongly
are shown
with
in Table
and
with
with
the males
7 by sex.
hadsmall
in each
S2/R
was
angles
%BF
angles
hadsignificant
females.
with FFM
correlations,
< 0.05) and
magnitude
(R
=
of %BF
In themales
wasslightly
0.42) butslightly
Ageadjusted
also weresignificant
0.65).
of%BFon age,
8.Age,MSK,
in thatorder
and wholebody
algorithm.Age
(p
portion
sexes(males,
R2
not add significantly
for variationfrom
variable to be entered
and explainedstatistically
of thevariance
MSK, and W/S2in both
R2
In the
age,MSK, W/S2, and
in%BF with a root
?6%. In thefemales
60% of the total
?5%.
The interaction
S2/
but
sex.
fe
FFM
positive
not
the
Ageadjusted
leg,
and negative
segments
correlations
were not
females;
small,
males.
correlation
greater
= 0.67) or MSK
of%BF
of %BF
with MSK
MSK
females,
ofregressions
shown
into
trunk,
using
for a significant
in %BF
W/S2
accounting
only bioelectric
of %BF
portions
forage,
correlations W/S2
wascorrelated
correlation
significantly in the
smalland in
(r
=
correlations
whole
in the
the
with %BF
significantly
the
significant
females
of%BF
thecorrelations
(R
(r
=
0.57)
W/S2
(r
= 0.62).
in each sex
R
=
of phase
with
Phase
body
in the
for
sig
all
trunk,and thebody
males.
whole
were
for
and
?
Phasean
correlated
however,
were
In the
or S2/R
corresponding
(p
the
with
positive
of the
0.74)
with
the
greater
age
was
W/S2
?
than
= 0.66).correla
with
with
?
than
than
=
less
(r
correlationsfor
4’t
(males,
R
=
MSK,
and
and
were
W/52,
W/S2
phase
allowed
and
ofthe
= 0.44;
to
age and
in the
significant
in %BF
sexes
males the
? explained
meansquared
corresponding
variance
termswere
and
were
angles
inTable
the regression
leg,
stepwise
arm, to
a MSK ac
< 0.001)van
in both
did
females,
the
MSK.
R2
R2
=
total
4? was
predic
(p
aftercontrol
(males,
full
R2
=
= 0. 15).equa
55%
error
equa
with
statisti
of
of the
in%BF
not
Thecontribution
in FFM
methods.
in FFM
females
anglesin the
prediction
Again, the
Theslopes
were
(Table8).
%BFwas
males),
55.8% ofthe
rmse of6.2%.
3 1.5%of the
gression
MSK,W/S2,
of
was
phase
evaluated
S2/R
males
3.7kg)
did
aftercontrolling
terms
regression
significantly
Therefore,
regressed
on
age, and
4?
total variance
In contrast,
total variance
equation controlling
and
4? explained
angles to explanation
similar
92%
kg)
inclusion
significantly
for age
significant.
on age,
between
werecombined
0 for females,
equation
in thesample
age and?
with an rmse
hierarchically
61%of the
of
variance
sion
ance
the
also usingregres
van
83%
phase
to the
52/R.
Age
in the
(rmse
regressions
of FFM
interaction
ofthe
not
andexplained
(rmse
but
not
of the
and
=
4.6 in
=
the
add
of
and
werenot
of%BFMSK,
the
and
‘I?t
different
groups
(coded
regression
in %BF
sex,
sexes
and
1 for
the
sex
. Theexplained
withan
explained
of 7.8%.
for
total
only
A re
age,sex,
variance
TABLE
Results
and
8
ofregression
of%BFfrom densitometryon age, MSK,W/S2,
f,
Independent
variable
Coefficient*
Change
in R2
F
Intercept
Age(y)
MSK
W/S2
23.69
0.14±0.10
0.73
0.49
1.54±0.53
± 8.21
0.07
0.37
0.01
0.10
3.31
34.27t
0.53
6.96t
(mm)
(kg/m2)
± 0.34
± 0.36
4?t(o)
TotalR2
SEE
Intercept
Age(y)
MSK
W/S2(kg/m2)
= 0.55
5.99%
=
41.14±8.44
0.06
0.46
0.01
1.96±0.46
± 0.06
± 0.24
± 0.38
0.05
0.40
<0.00
0.15
2.85
37.39t
<0.00
l8.Olt
(mm)
4?t(o)
TotalR2
SEE
= 0.60
= 5.08%
*1±
SD
by guest on July 10, 2011
www.ajcn.org
Downloaded from
Page 6
60
50
0
0
0a
30
y
?
3.026x
+
54.904
A2 0.56
0
20
10
0
8
0? 
098
0
0
0
0
0
4
68 10 12 1416
Trunk
Phase Angle
(degr..s)
18
20
E
0
U
U
0
.?
a
Trunk
Resistance(ohms)
PHASE ANGLE AND BODYCOMPOSITION
21
.?,
FIG 2. Plot
phase
of percent
angle
122).
body fat fromunderwaterweighing
and
by bio
ageselectric
958y(n=
forthetrunk, including malesfemales
in
planed
variance
variables
The
in Figure
gression
ship
retical
A polynomial
a significantly
< 0.01),
R2
Figure
thetrunk
sion lines
(> median)
(±SD)
± 0.77)
± 0.58)
wassignificantly
group
group
lower
(p
76.5±
Themean
was significantly
(8.4±
displacement
sources
ing body
%BF with
a statistically
in %BF
were
unadjusted
2 for
line
between
minima
an rmse of5.5%. In
portion
< 0.001)
this equation
of the
after
? ex
significant
= 0.05,
residual
(R2p
theseother
controlled for.
relationship
total
been
‘??t and
for %BF
regression
better
the improvement
of%BF
Although
to the
maybe nonlinear
(essential
model using
fit thanthe
to ?is illustrated
a straight
true relation
due
for reactance.
? provided
model
(F
small (increment
thesample.re
hasfitted
%BF
data,the
to theo
fat)and
? and
linear
was
=
8.7,
p
butin
0.03).
3 shows
at the
describing
vs low (< median)
theregression
was significantly
for the high
a plot
fixed
ofreactance
frequency
the average
versusresistance
with
within
for
of 50
association
groups
thelow
greater
group
(p
<
0.005)
± 4.4 (2) than
andresistance
in the low
in the high
(I)t in thelow
higherthan
2.0#{176}). Considerable
was evident
ofvaniation in cl?t among
compositionare
kHzregres
high
slope
(0.26
(0.10
for %BF.
%BF
the
0.05).
in
in the
was
group
group
group
in the
as well
plot, suggesting
participants
The
offor group
slope
Reactance
thelow
high
significantly
(mean
(90.7
(1 1.3 ± 2. 10)
high %BF
as horizontal
than
(p
%BF
<
higher%BF
%BF(mean
(13.3
± SD,
±
0.001)
than
(±SD)
15.4
?1)3.7
<
%BF
%BF
%BF
that
vertical
in the
± SD,
±
21).
13.2)
group
that the
with differ
complex.
Discussion
This study shows
ofthe
that
bioelectric
equations
ofadiposity
phase
%BFcan
phase
using
can
In
be predicted
angle
conventional
be improved
previous
from
trunk measurements
and
thropometric
addition
ofthe
that prediction an
indicesby the
of trunkangle. studies,the
estimation
ited
%BF
weight.
R2s and
to equations
proved
trast,
tion
It is clear
the length
tudinal
conductor
index
correlations
the arm
pothesis
S2/R
to a direct
Explanation
and %BF
bemultiple
biological
the frequency
correlations
actance
trunk
factthat
impedance.
tance
thebioelectric
ofwomen
for the
with %BF
trunk
with %BF.
thehigh
mean
relations
of%BF
accuracy
calculated
Direct
large
frombioelectnic
of the
as 100
of%BF
of estimation;
anthropometric
of%BF
do not
age and
the theory
conductor
is proportional
so it is not
of totalbody
ofwholebody
and theleg but
of Settle et al (23)
and TBWand
relationship.
of the
in this study
causes
conductor
of the
among
indicate that
aredifferent
is obscured
In addition,
andreactance
properties
andchildren
trunkwassignificantly
except in the
tendedto be uncorrelated
Meanresistance
%BFgroup
reactancewas
between?
impedance wasurn
by
was
the estimation
(weight
ofFFM,where
FFM)/

estimated
S2/R
the addition
measurements
marginally
appreciably
are controlled
ofbioelectric
squared divided
to the
surprisingthat
waterandof FFM
resistance
notin the
that the
FFM maybe fortuitous
regressions
errors
using
on producedlow
of S2/R
im
predictions
phase
ofFFM
only
add
S2/R
(10).
to the
In con
predicangles
after
from
ofthe
impedance
for.
0
impedancethat
by its longi
volume
is an accurate
7). The
resistances
support
association
and
of the
S2/R
(5, high
within
trunkthe hy
between
not due
association
is more
of variation
that
current
body
the
from
by the
betweenphaseangle
maydifficult
in phase
difficult
cannot
segments
bioelectric
thoseof the
measurement
thecorrelations
within body
of men
in all body
because there
withinanglea
areto separate
be varied
for resistance
properties
arm
when
The
and
of
leg,
(15).
re
the
and
of wholebody
between
suggest
differ from
segments.
positively
whereasreactance
or negatively
wassignificantly
in the low %BF
significantlylower.
and%BFweresignificantly
thea
resis
that
those
segments
may
Resistance
correlated and
girlsfor the
correlated
higher
whereas
in
thangroup
As a result,cor
negative
FIG
females
groups
medians.
27%.)
3. Plot ofreactance
ages
with percent
(Median
by resistance
122).
fat greater
in males
for the trunk,
lines
andless than
21%;median
including
been
the sexspecific
%BFin females
males
and958 y(n
=
Regression
than
havefitted for
body
%BF
=
=
by guest on July 10, 2011
www.ajcn.org
Downloaded from
Page 7
22
BAUMGARTNER ET AL
and
the
indicated
variation
thickness
mass.
of intraabdominal
angles
tissue
distribution.
Theoretically,
duce increased
observed
fora
intraabdominal
could reduce
producing
(1 5).Disproportionate
tance and,
anglesmay
factors.
sues decreases
adipose
Therefore,
between
thedistribution
compartments
Thesechanges
andmay
subcutaneous
would be necessary
in the trunk
currentand
ume from
Several
phase angles
tion.Using
presentstudy,
andreactance
tients with
Resistance,
alysisbut
and for
relative
greater
changes
correlated
creased
nyan
et al (27)
changes
body because
andother
failure.
changes
showed
frequency
frequency
cellularportion
Thepresent
mean
high
? was higher in the
multiple
for
is independent
low %BF
regression
groupthan
analyses
portion
age,
and
the volume
ratios
to index
in
%BF
that
in %BF
of
It is possible
group.
?
The
accounts
that
a significantof
of sex,
tissue,
mean
body subcutaneousadipose
4? is affected
tissue.
might
that
adipose
the
by
Also, of phase
adiposefor bodysegments be used
increased
shunting
decreases
frequency.
fat
cell
increases
extracellular
of the
both
Increased
wellas
and tissue
inboth
changes
changes
result
oftotal
thevolume
(25)
observed
and %BF
of body
associated
primarily
changes
adipose tissues.
to measure
over a range
to obtain intraabdominal
computedtomography.
previous studies
are sensitive
equipment
Spence
at a fixed
renal failure
reactance,and
the relative
phaseangle than
change in phase
thanthat for
in thecomponents
with changes
5%after dialysis.
alsodemonstrated
in phaseangles
of therapy
medications)
These changes
in total body
that theratio
(5 kHz)and
(100 kHz) can
oftotal
study shows
fluid
which
and
could
would
reactance
pro
current,
resistance
be
as in
fixedsubcutaneous and
asinterstitial
interface
reactance
in
in
ofthe
body
fat
permeability,
and resistance
resistance
the observed
combination
waterin fatfree
ofinterstitial
increases
in the present
couldreflect
amongfat
increased
tooccur
in intraabdominal
Totest this
resistance
offrequencies
adipose
in muscle
wall
and reac
phaseconsequently,
be the
percentage
(24)
tissues
the
phase
net of these
Thetis
and water in
increases
associations
angles
within adiposity.
study
changes
and
adiposity.
inthe
as well
hypothesis
andreactance
of alternating
tissue
in
water fatfree
with
appear trunk
involve as
it
vol
support the hypothesis
in water
employed
the
the
after
increased
greater
resistance.
for thechest
wholebody
ofimpedance
in total body
Ina similar
statistically
at a fixed frequency
(sodium restriction,
in patients with
were inversely
water.In addition,
of impedance
impedance
be usedas an index
body water.
that4? can
that
to differences
similar
et al (26)
frequency
before
phase
change
for
angle
the
distribu
in the
resistance
chests
hemodialysis.
after
forreactance
Moreover,
(+32%)
(+20%).
wereinversely
water, which
study, Subrama
significant
for the whole
diuretics,
congestive
correlated
Jenin
measured
measured
ofthe
to that
measured
forof pa
and
angle
was
di
the
was
These
de
heart
with
et al (6)
at a low
at a high
extra
be used to predict
%BF.
regression
considerably
over,
diction
age,
intraabdominal
byconventional
phase
of resistance
as a simple
body
cations
niques
trunk
weight,
of percent
search.
bioelectric
may be more
thatpotential
looked
andreactance
ies could
tween
Theerror
estimation
of theprediction
equations
than
of ?significantly
accounting
MSK,suggesting
adiposetissue
anthropometry.
trunk,
reactance
comparatively
be useful
conventional
impractical.
be usedalso
skinfold thicknesses
bodyfat using
resultsof this
impedance
complex
applications
in previousresearch.
for body
clarifyas well
bioclectricimpedance
is similar
using
equation
tothat for
MSK
using
improved
variation
anassociation
cannot
Thus,
from
trunk,
accurate
particularly
anthropornetric
Bioelectric
with measurements
to improve
these variables
studyindicate
for estimating
than is generally
existthat
Measurement
segmentsfor a range
as improve
and body
or W/S2 and
?
the
from
betteran
. More
theinclusion
of %BF
W/S2,
pre
sex,
with
after for
and
thatbe measured
the bioelectric
measurements
may
indexof percent
for clinical
angle ofthe
and
and
It may
in which
are
may
and
calculated
forthebeused
fat.appli
tech
forphaseangle the
of stature,
estimations
survey
the
composition
recognized
have been
of resistance
of frequenc
associations
composition.
in re
of The thatuse
body
and
over
thebe
#{163}3
References
1. Brazier
technic
MAB.
and
The
study
impedance
of
angle testfor thyrotoxicosis.
J Sung
Part1:
normals. WestObstet Gynecol
1935;43:4294l.
Barnett
ods for measuring
relation
Gynecol
Horton
man body.
Barnett
clinical
HofferEC,
whole
body
AcadSci 1970;
6. Jenin
compartments
Space
7.Lukaski
fatfree
human
8.Kushner
electric
9.PascoJA,
nc and
1985;39C:3659.
10. Guo5, Roche
tion from
1 1. Lukaski
bioelectric
J Appl
12.SegalKR,
position
Appl Physiol l985;58:l56571.
2. A. The basic factors
involved
function.
conformation.
in proposed
II Resistance
electrical
and
J Sung
meth
thyroid
physical
Qfactorin
to sex
1937;45:3807.
JW, Van
J Franklin
A, Bagno
measure
Meador
impedance
and West Obstet
3.Ravenswaay AC.
Electrical
impedanceofthe hu
Inst l935;20:55772.
S. Thephysiological
ofphase angle.
CK,Simpson
and
170:45261.
J, Roullet
byelectrical
Med l975;46:1525.
Johnson PE,
using bioelectrical
AmJ Clin Nutr
SchoellerDA.
impedanceanalysis.
RutishauserlHE.
electricalimpedance
4.mechanisms involved in the
Am J Physiol
DC.
total
body
1936; 114:36682.
relationship
watervolume.
5.A between
AnnNY
P, LenoirC, et al. Determination
impedance
of bodyfluid
Aviatmeasurements.
Environ
HC,
mass
body.
Bolonchuk WW,et al. Assessment
measurements
of
impedance
1985;4l:8l07.
Estimation
Am J Clin
Bodyfat estimated
measurements.
of the
RF, oftotal
Nutr
bodywaterby bio
1987;44:4l724.
from
Hum
anthropomet
NutrClinNutr
AF,ChumleaWC,
Biol
et al. Body
1987;59(2):22l35.
HallCB,
to assess
compositionpredic
impedance.
HC,
impedance
Physiol
Gutin
by electrical
Hum
BolonchukWW,et al.Validation
body
of the
method
132732.
E, et al. Estimation
impedance
humancomposition.
1986;60:
B, Presta ofhuman
a comparative
bodycom
methods:study.J
by guest on July 10, 2011
www.ajcn.org
Downloaded from
Page 8
PHASE ANGLEANDBODYCOMPOSITION
23
13.Shore
sis in body
Sports
Keller
estimation
1986;
Ackmann
ments
Chumlea
used
bioelectric
Lohman
ization
1988.
Sawka
mersion
in man.
Girandola
Effects
composition
Sin WE.
ofmethods.
P, TaylorWG. Thevalidity
analysis
ofbioelectric
in young
impedance
males.
analy
composition
1986;
adult Med Sci
Exerc
B, Katch
18(2):Sl6(abstr).
F!. Validity
of body
7(abstr).
JJ, Seitz MA.
in biologic tissue.
WC,Baumgartner
to estimatefatfree
impedance.
TG, Martorell
reference manual.
14.ofbioelectnic
lean
resistive
Med
impedance
Sports
for
fat inmales.SciExerc
18(2):Sl
15.Methods
Crit
ofcomplex
Biomed
Roche
from
J Clin Nutr
R, Roche
Champaign
impedance
1984;
AF.Specific
body
l988;48:7lS.
Anthropometric
IL: Human
measure
Rev
RN,
Eng1 1:28131 1.
16.resistivity
measuresmasssegmental of
Am
17. AF. standard
Books,kinetics
18.WE,
on residual
Ergonomics
RN,
ofwater
Weber H, Knowlton
lung
l978;2l:8994.
Wiswell
immersion
analysis.
composition
In: Brozek
RG.
and
The
body
effect
density
oftotalbodysub
volumesmeasurements
19.RA, Mohles
on lung
Physiol
from
JG, RomeroGT,
Barnes
WS.
volumes:
1977;43:2769.
fluidspaces
A; eds. Techniques
implications for body
J Appl
20. Bodyand density,analysis
J, Henschelfor measur
ing body
1961:22344.
Lohman
constants
and
SpOrt
32557.
Sokal
1981.
Settle
ment:
Cancer
Wedgwood
Sci l963;llO:l4152.
MartinAD,
fat by skinfold
Obes1984;8:1725.
SpenceJA,
in total body
bioelectric
l979;25:S
Subramanyan
in congestive
Physicians
composition.Washington,DC: National AcademyPress,
21.TG.
for
science
Applicability
children
reviews.
of body
youths.
Vol
composition
In:Randolf
14. New
techniques
KB,
Macmillan,
and
and ed. Exercise
York:1986:
22. RR, Rohlf FJ.Biometry.SanFrancisco: WHFreeman,
23. RG,
whole
l980;2:7280.
Foster
body
KR,
impedance
Epstein BR,
and
Mullen
body
JL.Nutritional
compartments.
assess
Nutr fluid
24.RJ. Inconstancy ofthelean bodymass.Ann NY Acad
25. Ross WD,Drinkwater
assumptions
DT,
and
et al. Prediction
cadaver
of body
mt
odiper?evidence.
j
26. BaligaR, Nyboer
cardiac
J, et al. Changes
output
analysis. Trans
during
fluid
hemodialysis
as detectedwaterand chestby
impedance
1S.
AmSoc Artiflntern Organs
27. R, Manchanda
heart
failure:
India1980;28:25762.
SC, NyboerJ,
a pre and
et al. Total
treatment
bodywater
post study.J Assoc
by guest on July 10, 2011
www.ajcn.org
Downloaded from