Diagnosis of Obesity with Bioimpedance Method

Monika Cichocka1 and Józef Tutaj2

1
Faculty of Electrical Engineering, Automatics, Computer Science and Biomedical
Engineering, University of Science and Technology AGH, Cracow, Poland
[email protected]
2
Faculty of Mechanical Engineering, University of Technology, Cracow, Poland
[email protected]

Abstract. The article describes the bioimpedance method of obesity

diagnosis. Such experiments are very important nowadays, because many
people are obese and have problems connected with it. It is essential to
check parameters such as a total body water (TBW), fat-free body mass
(FFM), fat mass and body fat percentage in order to start a proper
treatment and compose a right diet for the patient. Due to that fact,
a special meter was constructed. The main aim of the project was to
create a cheap, intuitive to operate, easily accessible equipment with high
measuring credibility. The projected meter uses laptops and a LabVIEW
application dedicated to measurements of bioelectrical impedance. Some
results of the meter tests are shown in the article.

Keywords: bioimpedance, dietetics, obesity diagnosis, virtual instru-

ment, LabVIEW.

1 Introduction
Because of the trends which are popular nowadays: a fast lifestyle and care-
lessness in selecting food, more and more people have problems with being
overweight. Therefore, people go to dietetics so that they help them compose
a proper diet that will allow them to maintain a slim figure. For this purpose it
is necessary to determine the composition of the patient’s body by performing
non-invasive, yet effective electrical bioimpedance analysis (BIA)[1].

1.1 Problem of Obesity

The obesity is a serious problem with excessive accumulation of body fat. The
obesity is diagnosed when the body fat exceeds the physiological needs of organ-
isms. A woman is considered to be obedient when her body fat is higher than
25% of a total body weight and in case of men it is 20%. The thing that matters
in not only its amount but also location.
The obesity is a heterogeneous group of conditions with multiple causes. It is
determined by an interaction between a variety of factors:
– environmental the westernization of diet and lifestyles,

J. Korbicz and M. Kowal (eds.), Intelligent Systems in Technical 291

and Medical Diagnostics, Advances in Intelligent Systems and Computing 230,
DOI: 10.1007/978-3-642-39881-0_24,  c Springer-Verlag Berlin Heidelberg 2014
292 M. Cichocka and J. Tutaj

– genetic - genetic diseases associated with mutations in genes,

– pharmacological some drugs can cause weight gain, eg. antidepressants and
anxiolytics,
– biological caused by autonomic nervous system disorders, eg. damage of
the hypothalamus (inflammation or cancer) and ventrolateral nucleus of the
medial hypothalamus,
– psychological people who suffers depression often eat more which causes
overweight [2].

The main reason is eating too much energetic food and a lack of the physical
activity. Thus, it is strongly connected to the problem of overweight.
The obesity causes a variety of health problems. In particular, it is associ-
ated with the development of obstructive sleep apnoea, type 2 diabetes mellitus,
heart disease, etc. A large obesity leads to disability. Obesity is a social issue
in developed countries and in the future there may be an epidemic. It is con-
sidered to be one of the hazards of developed societies civilization. Therefore, it
is necessary to diagnose this problem early and take action to cure the patient.
In order to determine body fat (body composition) and diagnose the obesity,
usually BIA measurement is performed or body mass index (BMI) is calculated.
The latter is easier and does not require any special device, but the obtained
result is approximate and has low reliability of measurement [3].

1.2 Bioimpedance

The bioelectrical bioimpedance is resultant total electrical resistance of tissues.

It is the ratio of the applied voltage U to the AC current I flowing through the
tissues. It consists of resistance R and reactance X, which is capacitance C in
humans.
U (ω) 1
Z(ω) = = R + jX(ω) = R − j (1)
I(ω) ω∗C
The human bioimpedance is a very important diagnostic parameter. Its value
depends on the humans body composition. The electrical properties of the is-
sues changes with the properties of every structural part, eg. intracellular and
extracellular substance, cell membrane and all organelles. Resistance depends
on electrolyte concentration while capacitance depends on potential differences
on cell membranes and varies with the current frequency which was used in
examination [4][5]. If DC or low frequency is used, current cannot go through
a membrane cell. The membrane behaves like an isolator and the extracellular
route is pure resistance. On the other hand, if high frequency is used, a membrane
cell behaves like a capacitor, so capacitance becomes associated with it [6].
In connection with the human tissues structure, the human body can be mod-
eled as an electrical circuit which consists of two parallel branches: one branch
with resistor and the other with resistor and capacitor connected in series [7].
It is why the body composition (especially water weight and fat mass) can be
examined with bioimpedance method.
 Diagnosis of Obesity with Bioimpedance Method 293

Fig. 1. The low frequency current (represented with a continuous arrow) flow extracel-
lular and the high frequency current (represented with a dotted arrow) flow intracellular
which is the reason that the human body can be modeled as resistance and capacitance
connected in parallel[6][8].

2 Related Works

Two basic method of bioelectrical impedance analysis and dedicated BIA devices
can be distinguished. Single Frequency Bioelectrical Impedance Analysis (SF-
BIA) uses only one frequency (usually it is 50 kHz) and current 0,8-1 mA. It
is applied to assessment of body composition in healthy subjects [1]. The very
precise measurements of patients in postoperative period is performed by Multi
Frequancy Bioelectrical Impedance Analysis (MF-BIA) method. A variety of
signal frequencies is used: the range is 0-500 kHz (mostly 0, 1, 5, 50, 100, 200 or
500 kHz, but at frequencies below 5 kHz and above 200 kHz poor reproducibility
has been noted [7].
There are four standard methods of bioimpedance measurements. Two of them
use current source and two use voltage source. The other classification criterion
is a number of sensors placed on the patient body. The bipolar methods use only
two sensors (eg. hand-to-hand, hand-to-foot or foot-to-foot) while tetrapolar
methods use four sensors (hand-to-hand-to-foot-to-foot) [4].
For dietetics diagnosis purpose there is commonly used a single frequency
measurement of the whole body with tetrapolar system [1]. Unfortunately, the
bioimpedance meters are very expensive so the cost of examination is also quite
big and not everybody can afford it.

3 Projected Bioimpedance Meter

The aim of the meter project was to construct a cheap, intuitive and easily ac-
cessible equipment with high measuring credibility. Therefore the authors of the
paper oryginally invented and constructed a system for bioimpedance measure-
ments using widely accessible laptops, simple electronic circuits and intuitive
software (an application in LabVIEW) [10].
Every medical equipment has to be safe for a patient as well as for a physician,
so a galvanic separation is needed. Using a laptop running on battery power do
294 M. Cichocka and J. Tutaj

not require any additional protection. There is a logical barrier and the battery
is loaded only when the equipment is shut down. Two basic functional blocks of
the bioimpedance meter can be distinguished:
– hardware (laptop with external sound card, signal adaptor - voltage divider
and electrodes),
– software (the application dedicated to measurements of bioelectrical
impedance which was developed in LabVIEW).

Fig. 2. Block diagram of the projected bioimpedance meter [10]

The above mentioned components are described with details in next subsections.

3.1 Hardware
The meter was created using the bipolar voltage method. In order to decrease
cost of the construction, it was decided to use standard inputs and outputs avail-
able in every laptop: MIC IN and PHONES OUT. They were examined using
HUNG CHANG 5804 40MHz oscilloscope and HAMEG HM8131-2 15MHz gen-
erator. Because of their nonlinear amplitude-frequency characteristic, it would
be necessary to implement some extra complicated calibration functions in the
software. Therefore, external sound card Behringer UCA202 was examined and
used (it is not expensive and its inputs/outputs). Unfortunately each laptop
has different internal parameters which cause that different signal amplification
values are obtained [10]. To minimize artefacts caused by various computer ar-
chitecture, calibration and parameterization of software have to be performed
every time the device is connected to a new laptop for the first time. However,
using LINE IN input and LINE OUT output does not require any additional
signal preprocessing in software, because these outputs and inputs are linear.
The use of LINE IN and LINE OUT introduced restrictions in generated signal
frequency. The minimal signal frequency value which will not be filtered is 100
Hz and the maximum one is 10 kHz regarding to the Nyquist-Shannon sampling
theorem (sampling frequency of A/C and C/A converters in sound card is 22
 Diagnosis of Obesity with Bioimpedance Method 295

Fig. 3. The projected bioimpedance meter consists of a laptop connected to an external

sound card with LINE OUT output and LINE IN input.[10].

kHz). There are also imposed limits of the input signal magnitude which value
depends on the laptops type.
A voltage signal is generated and the voltage signal has to be acquired because
of the output/input character. Regarding to the figure 2, it was necessary to
create an electronic module which will be a kind of diagnostic signal adaptor
a voltage divider with an additional resistor. Thus, the voltage value of the
LINE IN signal is proportional to the resistance value and the current value
which flows through this resistor (it depends on the measured bioimpedance).
Such solution enables computing the current value as a ratio of the additional
resistors value and the measured voltage value.
The big problem in medical diagnostics and biomedical equipment is caused
by noise. It is essential to provide proper equipments and cables shielding and
to use good quality sensors. In this case, there were used shielded cables from
electrocardiography device and Ag/AgCl single-use electrodes.

3.2 LabVIEW Application

Control and measurement systems developed in LabVIEW are commonly used
in engineering. This design platform and development system for a visual pro-
gramming from National Instruments enables developing applications virtual
instruments. When you use them, they are similar to real equipment [11][12].
That is the reason why LabVIEW was chosen to develop an application for
bioimpedance measurements.
The author of this paper created the application dedicated to measurements
of bioelectrical impedance. The application generates, acquires, processes and
analyses the diagnostic signal in bioimpedance measurements.
The main programme is based on the Multiple Loop Architecture using Func-
tional Global Variable.. It enables correct data flow control between programme
elements which work in parallel. Three loops have to be performed at the same
time:
– Event Handler Loop,
– Generation Loop,
– Acquisition and Analysis Loop.
296 M. Cichocka and J. Tutaj

The Event Handler loop was created using event structure. It defines an event
(a code fragment [12]) which should be performed if one of the buttons on the
user interface (Start, Pause or Stop) is pushed.
The information which event should be performed and the diagnostic signal
data are stored and sent between the loops by specially created global variable.
It consists of a data cluster including data which is essential to handling Gen-
eration Loop and Acquisition and Analysis Loop. Moreover, there are contained
parameters of the generated and aquired signals. Values of the generated signal
are to be found there too.

Fig. 4. State diagram of Case Structure in Acquisition and Analysis Loop presents the
way how the states can change. The AcquisitionInit state initalizes the whole process,
which is performed during Acquire state.

The Generation Loop and The Acquisition and Analysis Loop were created
using an architecture which is similar to a state machine architecture as it uses
a Case Structure. The difference is in a way of sending information by enum type
control between each circuits [11][12]. Instead of a Shift Register, the aforemen-
tioned functional global variable was used. Both loops have an idle, initialization,
the basic operation and stop states. The basic operation in Generation Loop is
to generate a sinusoidal signal with set frequency and magnitude and then to
put it in on the computer sound card outputs: PHONES OUT or LINE OUT.
The Acquisition and Analysis Loop enables a signal acquisition from the com-
puter sound card inputs: MIC IN or LINE IN, and then its analysis using an
implemented algorithm. The acquisition and the analysis parts are combined to-
gether. It is caused by laptops memory and calculation capabilities [10]. If these
two functions were to be separated, noticeable changes in the program structure
would be needed and hardware requirements would increase.
 Diagnosis of Obesity with Bioimpedance Method 297

The crucial element of the signal processing block in the application is a cal-
cutation of an instantaneous bioimpedance value. It enables creating a plot of
how bioimpedance varies in time (an impedance curve). Moreover, the obtained
data is analysed in terms of fat content in the body. The algorithms were im-
plemented to calculate some diagnostic parameters: total body water (TBW),
fat-free body mass (FFM), fat mass and body fat percentage [1].

Fig. 5. User interface of the application dedicated to the bioimpedance measurements

[10]

The specialized software was prepared in LabVIEW 2011 Student Edition and
then compiled in LabVIEW 2010 Professional Development System. The instal-
lation file with LabVIEW Run-Time Engine was prepared so that the installa-
tion is simple and quick, application costs are low and less computer memory is
needed [11].
298 M. Cichocka and J. Tutaj

4 Results

Some tests of the configured device were made. Due to the fact that every laptop
has different internal parameters, it was necessary to calibrate the meter every
time when a laptop was used for a very first time. Firstly, the meter was examined
using reference decade resistance. The measurement error was lower than 2%
and it is an acceptable result. No matter what signal frequency was used, the
results were the same. It is because the pure resistance does not depend on the
frequency.

Table 1. Results of the device test using decade resistors and laptop Lenovo 3000 N20

Real set resistance [kΩ], The resistance value measured Relative error of
the device class: 0,05 by the virtual instrument [kΩ] the meter [in %]
10 9,90 1,00
50 50,81 1,62
100 101,9 1,90
150 152,27 1,51
200 204,57 1,29
250 254,15 1,66
300 299,85 0,05

Moreover, the device was tested on 23-year-old woman using electrodes at-
tached to the top of her hands. For example, for signal with amplitude peak-
to-peak 1,5V and frequency 500 Hz the bioimpedance was equal to 332 kΩ
(TBW=36,30 kg, FFM=49,71 kg, Fat Mass=2,27 kg, Body Fat=4,37%) and
for signal with frequency 10 kHz the bioimpedance was equal to 315 kΩ
(TBW=37,95 kg, FFM=51,98 kg, Fat Mass=0,02 kg, Body Fat=0,03%). The
difference between these two results is seen if view of the fact that when the
signal frequency is higher, the capacitance component appears.

5 Conclusion

In the paper the bioimpedance meter designed to the obesity diagnosis was de-
scribed. Regarding to the obtained results, it turned out that it is technically and
economically possible to construct bioimpedance device using special software
in LabVIEW and hardware including laptops and sound cards. It is planned to
configure the device with a tablet or a smartphone instead of a laptop. Some
changes in the application will also be made. The possibility to connect to the
patients database and save the measurements results will be introduce. It will
facilitate the physicians work.
There is an issue is with the frequency limits imposed by the sound card input
and output. Usually, as it was show in section 2, bioimpedance measurements
in dietetics are performed using high frequency signal, but in this solution it
 Diagnosis of Obesity with Bioimpedance Method 299

was impossible. The device can generate only some of the frequencies used in
MF-BIA: 0, 1, 5 and 10 kHz. Due to this fact, the created device can be used in
diagnosis of obesity only in a limited way.

Acknowledgments. The work described in this paper was supported by Min-

istry of Science and Higher Education in Republic of Poland (project no. M-
4/84/2013/DS).

References
1. Lewitt, A., Ma̧dro, E., Krupienicz, A.: Podstawy teoretyczne i zastosowania analizy
impedancji bioelektrycznej (BIA). Endokrynologia Otylość Zaburzenia Przemiany
Materii 3, 79–84 (2007)
2. Kopelman, P.G.: Obesity as a medical problem. Nature 404, 635–643 (2000)
3. Marra, M., Pasanisi, F., Scalfi, L., Colicchio, P., Chelucci, M., Contaldo, F.: The
prediction of basal metabolic rate in young adult, severely obese patients using
single-frequency bioimpedance analysis. Acta Diabetologica 40, 139–141 (2003)
4. Grimmes, S., Martinsen, O.G.: Bioimpedance & bioelectricity Basics. Academic
Press, Elsevier (2008)
5. Wtorek, J., Nowakowski, A., Palko, T.: Pomiary bioelektoimpedancyjne. In:
Nalȩcz, N.: Biocybernetyka i Inżynieria Biomedyczna 2000, Biopomiary, tom. Aka-
demicka Oficyna Wydawnicza EXIT, Warsaw (2001)
6. New imaging methods for the detection of cancer biomarkers,
http://nano.mdx.ac.uk/bioimpedance/
7. Kyle, U., Bosaeus, I., De Lorenzo, A., Deurenberg, P., Elia, M., Gomez, J.M., et al.:
Bioelectrical impedance analysis part I: review of principles and methods. Clinical
Nutrition 23, 1226–1243 (2004)
8. International Society for Electrical Bio-impedance, http://www.isebi.org
9. Major-Goluch, A., Miazgowski, T., Krzyanowska-winiarska, B., Safranow, K.,
Hajduk, A.: Porwnanie pomiarw tuszczu u modych zdrowych kobiet z prawidowa̧
masa̧ ciaa za pomoca̧ impedancji bioelektrycznej i densytometrii. Endokr Oty Zab
Przem Mat. 6(4), 189–195 (2010)
10. Cichocka, M.: Projekt miernika bioimpedancji z wykorzystaniem środowiska Lab-
VIEW. Thesis on University of Technology in Cracow (supervisor: Tutaj, J.) (2013)
11. National Instruments, http://www.ni.com
12. Tlaczala, W.: Środowisko LabVIEW w eksperymencie wspomaganym komput-
erowo, Wydawnictwo Naukowo-Techniczne, Warsaw (2002)