Innovation

Single prediction equation for bioelectrical impedance analysisin adults aged 2259 years

B. R. PATIL*{, D. P. PATKAR{, S. A. MANDLIKx, M. M. KUSWARKARx and G. D. JINDALx

{Department of Instrumentation Engineering, Bharati Vidyapeeth College of Engineering,Navi Mumbai, 400 614, India

{MRI Centre, Dr Balabhai Nanavati Hospital, Vile Parle (West), Mumbai, 400 056, IndiaxBio-Medical Section, Electronics Division, Modular Laboratories, B.A.R.C., Mumbai, 400 085, India

(Received 17 October 2010; revised 14 December 2010; accepted 24 November 2010)

The purpose of this study was to validate a single bioelectrical impedance analysis (BIA)

equation in healthy Indian subjects aged 2259 years with a body mass index (BMI)

between 16.8 and 47.3 kg m72. Healthy subjects (34 men and 30 women) were measured

by two methods: bone mineral content (BMC) was measured by a commercial body

composition analyser and bioelectrical impedance at various frequencies was measured

by a newly developed bioelectrical impedance measurement system. As correlations were

high and prediction error was low, a single equation was developed using all subjects

as follows: BMC73.5268 (0.02796 h) (0.01456w) (1846 (h2/Zbody50)) (1.086 (w6 h2/Zbody6.25)) (0.00326 (age)) (0.1036 (sex); men 1, women 0).BMC measured from commercial instrument InBody720 was 2.552+ 0.457 kg. BMCpredicted by equation was 2.554+ 0.447 kg (R 0.976, adjusted R2 0.948, standarderror of estimate 0.104 kg, total error 0.09987 kg). The results of this study show thatthe newly developed multi-frequency bioelectrical impedance measurement system with

the single prediction BIA equation can be used in screening the subjects suspected with

osteoporosis and for follow-up study of the patient under the therapy for osteoporosis.

For validation of commercial instrument InBody720, BMC of 22 healthy subjects was

measured by InBody720 and dual-energy X-ray absorptiometry. High correlation

(R 0.9531) and low error (total error 0.0913 kg) was found between these twomethods.

Keywords: Body parameters; Bioelectrical impedance analysis; Bone mineral content;

Osteoporosis

1. Introduction

Osteoporosis is a disease in which the bones deteriorate and

become weak and fragile. Around 500 million populations

worldwide are estimated to have osteoporosis. Osteoporo-

sis in women is increasingly being recognized as an

important health issue. In the early stage of this disease

there may be no symptoms. It is frequently painless until a

bone breaks. In fact, a fractured bone may be the rst

symptom of osteoporosis. Osteoporosis is also character-

ized by an abnormal loss of bone mineral content, which

leads to a tendency to non-traumatic bone fractures or to

*Corresponding author. Email: babapatil2004@yahoo.co.in

Journal of Medical Engineering & Technology, Vol. 35, No. 2, February 2011, 109114

Journal of Medical Engineering & TechnologyISSN 0309-1902 print/ISSN 1464-522X online 2011 Informa UK, Ltd.

http://www.informahealthcare.com/journalsDOI: 10.3109/03091902.2010.543751

J Med

Eng

Tec

hnol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Bhar

ati V

idya

peet

h D

eem

ed U

nive

rsity

on

01/1

4/11

For p

erso

nal u

se o

nly.

structural deformations of bone [1]. Osteoporosis is a

widespread medical condition aecting middle-aged and

older people [2]. The accurate estimation of the BMC has

been an important diagnostic indicator for determining the

status of osteoporosis and for the follow-up study of

patients under therapy for osteoporosis. Various BMC

measurement techniques/methods are available, such as

dual-energy X-ray absorptiometry (DEXA), ultrasound

densitometry and quantitative computed tomography

(QCT). DEXA is the most well-known method. However,

the use of DEXA is limited because of its practicability

and cost.

Bioelectrical impedance analysis (BIA) is a user-

friendly, safe, simple, inexpensive and non-invasive

technology for clinical and non-clinical purposes. Mea-

surements can be made quickly and can be repeated at

short time intervals. BIA has the potential to provide

more information regarding the functioning of the human

body. To date the full potential of this technology has

not been realized. Commercial instruments available for

BIA measure electrical parameters in dierent body

segments and yield clinical information regarding total

body water, fat free mass, and body fat, etc. Many

investigators have developed empiric BIA equations for

prediction of fat-free mass, total body water and body

fat [314]. Most of these equations have been validated in

relatively young healthy adults against several body-

composition techniques [10]. Studies have shown that

BIA formulae developed for healthy, normal-weight sub-

jects are not suitable for obese subjects [14, 15]. In

longitudinal studies, the use of dierent BIA formulae in

the same subject who becomes overweight introduces a bias

into body-composition studies, making one question

whether the dierences in body composition are due to

changes in the BIA formula or due to changes in body

composition. Thus, it would be advantageous to use a

single formula that is applicable in both normal and

overweight subjects and permits estimation of body

composition. Roubeno et al. [14] and others concluded

that BIA equations are subject to errors that cannot be

determined a priori unless they are validated in the specic

population in which they are to be applied [1618]. Thus,

BIA equations must be validated in a representative

population sample against a reference method before they

can be accepted as accurate. The purpose of the present

investigation was to validate a single BIA equation for

the assessment of BMC, against the commercial device

InBody720 (Biospace, Seoul, Korea) which was validated

with DEXA [19]. DEXA is a reference method for BMC

that has been validated against many independent methods.

A single BIA equation that is valid in subjects with dierent

BMIs for use in longitudinal studies would be a signicant

advantage in screening subjects suspected with osteoporosis

and for follow-up study of patients under therapy for

osteoporosis.

2. Subjects and methods

2.1. Subjects

The study group consisted of 64 individuals (34 men and 30

women) in the age group of 22 to 59 years. All subjects were

born in India and resided in Mumbai. Although subjects

were randomly selected, statistical analysis showed not

much dierence in (h), (w) and BMI between subjects and a

control group of healthy age-matched 264 men and 279

women. Body height was measured to the nearest 0.5 cm.

The purpose of the study was explained to all subjects and

their oral consent was taken. Each subject was measured by

two methods: bone mineral content (BMC) was measured

by a commercial instrument (InBody720, Biospace) and

bioelectrical impedance at various frequencies was mea-

sured by a newly developed bioelectrical impedance

measurement system.

2.2. Segmental multi-frequency bioelectrical impedance

analysis (MF-BIA) instrument (InBody720)

The BMC of each subject was obtained from the segmental

multi-frequency bioelectrical impedance analysis instru-

ment (InBody720). From segmental impedance values,

using proprietary equations, the instrument gives an

estimated value of BMC. The correctness of this estimated

value is validated against DEXA measurement. This

commercial product is approved by FDA (Food and Drug

Administration, United States, May 2003). The instrument

was operated at frequencies of 1, 5, 50, 250, 500 and 1000

kHz which were pre-set by the manufacturer and intro-

duced into the body in the ascending order of frequency.

This device uses contact electrodes, located in the handgrips

and the footpads. Subjects were asked to stand with the ball

and heel of each foot on two metallic electrodes on the oor

scale and hold handrails with metallic grip electrodes in

contact with the palm and thumb [20, 21]; they were

instructed to keep their arms fully extended and abducted

approximately 208 laterally. For 1 kHz programmedfrequency, an alternating current of 100 mA and for otherprogrammed frequencies, an alternating current of 500 mAwas applied between a pair of carrier electrodes, and

voltage was measured across a pair of sensing electrodes.

For cross validation of InBody720, the BMC of healthy 22

subjects was measured by InBody720 and dual-energy

X-ray absorptiometry (Lunar Prodigy, DPX-IQ, General

Electric Healthcare, Belgium, Europe). High correlation

(R 0.9531) and low error (total error 0.0913 kg) werefound between these two methods (gure 1).

2.3. Developed multi-frequency bioelectrical impedance

measurement system

The body impedance of each subject at various frequencies

was measured and recorded by the new system, using the

110 B. R. Patil et al.

J Med

Eng

Tec

hnol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Bhar

ati V

idya

peet

h D

eem

ed U

nive

rsity

on

01/1

4/11

For p

erso

nal u

se o

nly.

whole body impedance measurement approach. The system

operated at frequencies of 6.25, 12.5, 25, 50 and 100 kHz

which were pre-set and introduced into the body in the

ascending order of frequency. Before testing, standardiza-

tion of a subjects posture is not required. Subjects were

asked to sit on the chair with connected electrodes. For

each of the programmed frequencies, an alternating current

of 500 mA was applied between silver band currentelectrodes, which are located in the palm and feet on same

side. The two sensing electrodes are located in the wrist and

ankle on the same side. The amplitude of the sensed voltage

drop is directly proportional to the impedance of the body

(Zbody) between the sensing electrodes. Figure 2 shows a

schematic block diagram of the multi-frequency bioelec-

trical impedance measurement system. The Biomedical

Instrumentation Processor Board (BMIPB) is an interface

board used to interface the system with compatible PC. The

100 kHz square wave signal is derived from BMIPB, which

is further divided by the synchronous up/down counter and

produces the four square wave singles of 50, 25, 12.5 and

6.25 kHz frequencies. For smoothening, each square wave

signal is ltered with the help of second order low pass lter

followed by a band pass lter. This produces pure

sinusoidal alternating signals. These signals are then

multiplexed with the help of an analogue multiplexer.

Any one of the signals can be selected at the output of the

multiplexer by the operator through the user interface

panel. The selected signal is applied to the VI (voltage to

current) converter. The sinusoidal alternating current of

constant magnitude produced by VI converter is passed

through the body with the help of isolation transformer.

The voltage signal developed along the current path is

sensed with the help of sensing electrodes and amplied

using a precision instrumentation amplier. The wide band

pass lter removes the superimposed noise and produces a

pure sinusoidal alternating voltage, which is proportional

to the impedance of the body under investigation. Further

the output of the band pass lter is rectied, ltered and

buered to obtain a pure DC voltage which is also

proportional to the body impedance. The injected sinusoi-

dal alternating current and the voltage signal developed

along the current path were multiplied by analogue

multiplier. The two componentsdouble frequency sinu-

soidal signal and DC signal (mean value)were separated

by the second order low pass lter.

2.4. Statistics

Descriptive statistics were calculated for (h), (w), (age),

(sex), BMI and BIA parameters, including body impedance

(Zbody), (Zbody/h), (Zbody/w), (h2/Zbody) and (w6 h

2/Zbody)

at 6.25, 12.5, 25, 50 and 100 kHz frequencies. The values

were expressed as mean+ standard deviation (SD). Simpleregressions were calculated to test correlations between

BMC obtained from InBody720 and BIA parameters

measured from developed bioelectrical impedance measure-

ment system. Student t-test was used to test dierences

between methods. BMC measured by commercial instru-

ment InBody720 was used as the criterion measurement.

Stepwise multiple regression analysis was used to derive a

prediction equation by BIA. Predictor variables, entered

into the BIA model in the order of highest correlation

coecient and smallest standard error of estimation (SEE)

were (h2/Zbody50), h, w, (w6 h2/ ZBody6.25), (sex) and (age)

and a single equation was developed using the entire

sample. In addition to correlation and regression techni-

ques, error analysis was performed. SEEs were calculated

and used as errors of prediction for InBody720 derived

BMC and predicted BMC by BIA equation. The total error

(TE) was calculated as:

TE P

ni BMC1i BMC2i 2

n 2

s; 1

where (BMC1) is the observed value of BMC by In-

Body720 and (BMC2) is the predicted value of BMC by

equation. The total error of measurement estimates the

magnitude of the error for a given measurement and is

dened as the dierence between measurements for the

individual (i), i 1, . . . , n, where n is the number ofindividuals [22]. The total error was compared with the

SEE. A small dierence between total error and SEE

indicates high accuracy of the prediction. To assess the

agreement between the two clinical measurements the

dierence between the values was plotted against their

means because the mean was the best available estimate of

the true value. This analysis allows for the calculation of

bias (estimated by the mean dierences), the 95%

condence interval for the bias, and the limits of agreement

(two SDs of the dierence) [23]. Statistical signicance was

set at p 5 0.05 for all tests.

Figure 1. Relationship between bone mineral content

(BMC) measured with DEXA and the commercial instru-

ment InBody 720.

Bioelectrical impedance analysis in adults 111

J Med

Eng

Tec

hnol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Bhar

ati V

idya

peet

h D

eem

ed U

nive

rsity

on

01/1

4/11

For p

erso

nal u

se o

nly.

3. Results

A total of 64 healthy adults aged 2259 years were recruited

as subjects. Table 1 shows the anthropometric character-

istics and table 2 shows the BIA characteristics of subjects

grouped by dierent frequencies. The prediction equation

developed from all subjects is shown in table 3. The order

of entry of predictor variables was (h2/Zbody50), h, w, (w6h2/Zbody6.25), (sex) and (age) for the BIA model (table 4).

(h2/Zbody50) accounted for 84.12% of the variability

(SEE 0.182 kg) of the equation whereas (h) aloneaccounted for only 74.51% of the variability (SEE0.231kg) and (w) alone accounted for 53.74% of the

variability (SEE 0.311 kg). Inclusion of height (h), weight(w), (age) and (sex) without BIA parameters accounted for

92.62% of the variability with a SEE of 0.124 kg. Thus

inclusion of BIA parameters clearly improved the prediction

power and decreased the SEE compared with anthropo-

metric parameters only. Figure 3 shows the correlation and

mean dierence, according to Bland and Altman [23], using

the prediction equation in all subjects [23]. Subjects (data

not shown) with BMIs above 27 kg m72 were analysed

separately to determine whether greater error occurred with

the BIA equation in larger subjects. BMC measured for 22

subjects with BMIs above 27 kg m72 from commercial

instrument InBody720 was 2.701+ 0.5473 kg. BMCpredicted by equation was 2.706+ 0.5173 kg (R 0.978,total error 0.114 kg). Thus, it is possible to estimate BMCwith the same equation for overweight and obese subjects.

Figure 2. Schematic block diagram of the multi-frequency bioelectrical impedance measurement system.

Table 1. Anthropometric Characteristics of healthy subjects.

Men Women Combined

n 34 30 64

(age) (years) 36.15+ 8.72 43.9+ 7.83 39.78+ 9.12h (cm) 167.1+ 6.47 155.2+ 5.08 161.50+ 8.33w (kg) 73.23+ 14.44 62.08+ 8.79 68.01+ 13.26BMI (kg m72) 26.2+ 4.88 25.9+ 4.18 26.1+ 4.53

nnumber of subjects; hheight; wweight; BMI body mass index

112 B. R. Patil et al.

J Med

Eng

Tec

hnol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Bhar

ati V

idya

peet

h D

eem

ed U

nive

rsity

on

01/1

4/11

For p

erso

nal u

se o

nly.

Table 2. Bioelectrical impedance analysis characteristics of healthy subjects (n 64).Frequency (kHz)

6.25 12.5 25 50 100

(Zbody) (O) 658.31+ 101.10 563.98+ 91.41 499.44+ 78.74 460.19+ 73.31 425.22+ 62.87(Zbody /h) (O cm

71) 4.09+ 0.73 3.51+ 0.66 3.11+ 0.57 2.87+ 0.53 2.65+ 0.46(Zbody/w) (O kg

71) 10.17+ 3.06 8.72+ 2.66 7.72+ 2.33 7.11+ 2.15 6.56 + 1.89(h2/Zbody) (m

2 O71) 0.0041+ 0.0009 0.0048+ 0.001 0.0054+ 0.001 0.0058+ 0.001 0.0063+ 0.001(w6 h2/Zbody)(kg m2 O71)

0.289 + .137 0.338+ 0.162 0.382 + 0.179 0.415+ 0.196 0.447+ 0.211

nnumber of subjects; Zbody body impedance; hheight; wweight.

Table 3. Prediction equation for BMC using all subjects(n 64).

BMC73.5268 (0.02796 h) (0.01456w) (1846 (h2/Zbody50) (1.086 (w6 h2/Zbody6.25)) (0.00326 (age)) (0.1036 (sex);men 1, women 0)

Observed BMC by commercial instrument (InBody720) 2.552+0.457 kg

Predicted BMC 2.554+ 0.447 kg (R 0.976, adjusted R2 0.948,SEE 0.104 kg, TE 0.09987 kg)

nnumber of subjects; BMC bone mineral content; Zbody50bodyimpedance at 50 kHz; h height; wweight; Zbody6.25body im-pedance at 6.25 kHz; R validity coecient; SEE standard error ofthe estimate; TE total error.

Table 4. Contribution and order of entry of variables to thebioelectrical impedance analysis model and anthropometric

model for bone mineral content (n 64 subjects).

Model and variables

Cumulative variables

used in model Variables

Ad.R2 SEE P Ad.R2 SEE P

BIA

(h2/Zbody50) 84.12 0.182 0.0001 84.12 0.182 0.0001

h 91.87 0.130 0.0001 74.51 0.231 0.0001w 93.70 0.115 0.0001 53.74 0.311 0.0001(w6 h2/Zbody6.25) 94.39 0.108 0.0001 62.89 0.278 0.0001(sex) 94.57 0.106 0.0001 49.49 0.325 0.0001(age) 94.80 0.104 0.0001 6.24 0.442 0.0001

BMI

h 74.51 0.231 0.0001

hw 91.82 0.131 0.0001hw (age)

92.73 0.123 0.0001

hw (age) (sex)

92.62 0.124 0.0001

n number of subjects; BIA bioelectrical impedance analysis;BMIbody mass index; hheight; wweight; BMCbone mineralcontent; Zbody50body impedance at 50 kHz; Zbody6.25 bodyimpedance at 6.25 kHz; SEE standard error of the estimate;P signicance of contribution of each additional individual para-meter to the stepwise multiple regression model; Ad.R2% adjustedsquared value of the validity coecient.

Figure 3. Correlations (a) and dierences (b) of bone

mineral content (BMC) in all subjects estimated by

commercial instrument InBody720 and developed BIA

equation. The dierence (BMC obtained by InBody720

BMC calculated by BIA equation per BlandAltman) is

plotted against the mean of the measurements of BMC by

commercial instrument InBody720 and developed BIA

equation. SEE, standard error of the estimate; R, validity

coecient; TE, total error. Circles indicate men; triangles

indicate women.

Bioelectrical impedance analysis in adults 113

J Med

Eng

Tec

hnol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Bhar

ati V

idya

peet

h D

eem

ed U

nive

rsity

on

01/1

4/11

For p

erso

nal u

se o

nly.

4. Discussion

BIA has been developed for eld use and has shown great

potential for use in estimating body composition and it is

easy, non-invasive, and inexpensive. In osteoporosis, bones

become porous and bone mineral content reduces as

minerals such as calcium leach out. Bones aected by

osteoporosis are less dense than the normal bones, resulting

in reduction in their density and an increase in their

resistivity. Due to the change in resistivity of bones, the

bioelectrical parameters of the body changes and provide

quantitative information about the BMC.

All subjects in this study were reported to be healthy. The

BIA equation developed in this study is valid for healthy

adults 2259 years old. Slightly lower accuracy should be

expected in subjects older than 60 years. Validity of the BIA

equation in subjects older than 60 years is unknown and

requires further validation. It is also necessary in subjects

with BMIs below 17 kg m72.

5. Conclusion

The results of this study show that the newly developed

system with the single prediction BIA equation validated

against commercial instrument (InBody720, Biospace, Seoul,

Korea) can be used to predict BMC in subjects aged 2259

years and with BMIs ranging from 16.8 to 47.3 kg m72. This

low cost, portable and simple system will help in screening

subjects with suspected osteoporosis and in follow-up studies

of patients under therapy for osteoporosis.

Acknowledgements

This studywas supportedby theElectronicsDivision,Bhabha

Atomic Research Centre (BARC),Mumbai and Dr Balabhai

Nanavati Hospital, Mumbai. We are thankful to Aarti

Karkera for helping in data collection and S. K. Athawale

for proof-reading. We are also grateful to G. P. Srivastava,

Director, E & I Group, R. K. Patil, Associate Director, E & I

Group (C) and C. K. Pithawa, Head, Electronics Division of

BARC for their support and encouragement.

Declaration of interest: The authors declare that they have

no conict of interest.

References

[1] Wasserman, S.H. and Barzel, U.S., 1987, Osteoporosis: The state of

the art in 1987 a review. Seminars in Nuclear Medicine, 17, 283292.

[2] National Institute of Health, 1984, Consensus conference: osteoporo-

sis. Journal of the American Medical Association, 252, 799802.

[3] Segal, K.R., Van Loan, M., Fitzgerald, P.I., Hodgdon, J.A. and Van

Itallie, T.B., 1988, Lean body mass estimation by bioelectrical

impedance analysis: a four-site cross over validation. American Journal

of Clinical Nutrition, 47, 7.

[4] Kushner, R.F. and Schoeller, D.A., 1986, Estimation of total body

water by bioelectrical impedance analysis. American Journal of Clinical

Nutrition, 44, 417.

[5] Lukaski, H.C., 1985, Assessment of fat-free mass using bioelectrical

impedance measurements of the human body. American Journal of

Clinical Nutrition, 41, 810.

[6] Lukaski, H.C., 1986, Validation of tetrapolar bioelectrical impedance

measurements to assess human body composition. Journal of Applied

Physiology, 60, 1327.

[7] Van Loan, M. and Mayclin, P., 1987, Bioelectrical impedance

analysis. Is it a reliable estimator of lean body mass and total body

water? Human Biology, 2, 299.

[8] Jackson, A.S., Pollock, M.L., Graves, J.E. and Mahar, M.T., 1988,

Reliability and validity of bioelectrical impedance in determining body

composition. Journal of Applied Physiology, 64, 529.

[9] Lukaski, H.C. and Bolonchuk, W.W., 1988, Estimation of body uid

volumes using tetrapolar bioelectrical measurements. Aviation, Space,

and Environmental Medicine, 59, 1163.

[10] Graves, J.E., Pollack, M.L., Calvin, A.B., Van Loan, M. and Lohman,

T.G., 1989, Comparison of dierent bioelectrical impedance analyzers

in the prediction of body composition. American Journal of Human

Biology, 1, 603.

[11] Heitmann, B.L., 1990, Prediction of body water and fat in adult danes

from measurement of electrical impedance. A validation. International

Journal of Obesity, 14, 417.

[12] Deurenberg, P., van der Kooy, K., Evers, P. and Hulshof, T., 1990,

Assessment of body composition by bioelectrical impedance

in a population aged 60 y. American Journal of Clinical Nutrition,

51, 3.

[13] Deurenberg, P., van der Kooy, K., Leenen, R., Westrate, J.A. and

Seidell, J.C., 1991, Sex and age specic prediction formulas for

estimating body composition from bioelectrical impedance: a cross-

validation study. International Journal of Obesity, 15, 17.

[14] Roubeno, R., Baumgartner, R.N., Harris, T.B., et al., 1997,

Application of bioelectrical impedance analysis to elderly population.

Journal of Gerontology, 52A, M129.

[15] Stolarczyk, L.M., Heyward, V.H., Van Loan, M.D., Hicks, V.L.,

Wilson, W.L. and Reano, L.M., 1997, The fatness-specic bio-

electrical impedance analysis equations of Segal et al: are they

generalizable and practical? American Journal of Clinical Nutrition, 66,

817.

[16] Pichard, C., Kyle, U.G., Janssens, J.P., Burdet, L., Rochat, T.,

Slosman, D.O., Fitting, J-W., Thiebaud, D., Roulet, M., Tschopp,

J.M., Landry, M., Schutz, Y., 1997, Body composition by x-ray

absorptiometry and bioelectrical impedance in chronic respiratory

insuciency patients. Nutrition, 13, 9528.

[17] Kyle, U.G., Pichard, C., Rochat, T., Slosman, D.O., Fitting, J-W.,

Thiebaud, D., 1998, New bioelectrical impedance formula for patients

with respiratory insuciency: comparison to dual-energy X-ray

absorptiometry. European Respiratory Journal, 12, 960.

[18] Pichard, C., Kyle, U., Gremion, G., Gerbase, M. and Slosman, D.,

1997, Body composition by X-ray absorptiometry and bioelectrical

impedance in elite female runners. Medicine & Science in Sports &

Exercise, 29, 1527.

[19] http://www.inbody./resources/userles/File/720manual.pdf; chapter

no.2;page no.2-16; Last accessed date: 30 Nov.2010.

[20] Deurenberg, P., 1994, International consensus conference on impe-

dance in body composition. Age & Nutrition, 5, 142145.

[21] Bedogni, G., Malavolti, M., Severi, S., Poli, M., Mussi, C., Fantuzzi,

A.L., Battistini, N., 2002, Accuracy of an eight-point tactile- electrode

impedance method in the assessment of total body water. European

Journal of Clinical Nutrition, 56, 11431148.

[22] Wellens, R., Chumlea, W.C., Guo, S., Roche, A. F., Reo, N.V.,

Siervgel, R.M., 1994, Body composition in white adults by dual

energy X-ray absorptiometry, densitometry, and total body water.

American Journal of Clinical Nutrition, vol.59, no.3, 547555.

[23] Bland, J.M. and Altman, D.G., 1986, Statistical methods for assessing

agreement between two methods of clinical measurement. Lancet, i,

307.

114 B. R. Patil et al.

J Med

Eng

Tec

hnol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Bhar

ati V

idya

peet

h D

eem

ed U

nive

rsity

on

01/1

4/11

For p

erso

nal u

se o

nly.