Biomedical Engineering Reference
In-Depth Information
impedance electrical model (often with the reactance
component
X
neglected), or the parallel equivalent have
been used. Several indexes have been introduced in order
to increase the accuracy. Gender, age and anthropometric
results such as total body weight and height are param-
eters used. An often used index is
H
2
/R
segm
, where
H
is
the body height and
R
segm
the resistance of a given seg-
ment. Because of the 1/
R
segm
term this is therefore ac-
tually a
conductance
index. Calibration can be done by
determining the
k
-constants in the following equations:
H
2
R
segm
TBW ¼ k
1
þ k
2
(4.1.10)
Figure 4.1-11 Human body divided into five impedance
segments, octopolar electrodes.
H
2
R
segm
TBW ¼ k
1
þ k
2
W þ k
3
(4.1.11)
then varied in succession. The system will be highly
sensitive for the detection of asymmetrical limb bioim-
pedance. Standardization of the type of electrodes used
and their placement is a major concern (Kyle et al.,
2004). Cornish et al. (1999) provided a set of standard
electrode sites for bioimpedance measurements.
Calibration is also a major concern in BIA. Calibration
can be done with more accurate but cumbersome
methods such as using deuterium, underwater weighing
or dual energy X-ray absorption. However, dilution
methods have their own errors (
>
2 l for TBW) and yield
different results (e.g. 4% difference between the deu-
terium-TBW method and the
18
O-TBW method). Al-
though body impedance reflects tissue hydration, soft
tissue mass (lean and fat) can also be empirically derived
by correlation in healthy subjects because the
compart-
ments of soft tissue are correlated with each other through
physiological constants
. However, physiological constants
become flawed in patients with fluid disorders, which
accounts for most conflicting results in the literature
(Kyle et al,. 2004).
Body
position
is important because it influences the
distribution both of blood and the fluids in the stomach/
intestine tissue. Direct body segment to body segment
skin contact must be avoided in order to obtain stable
readings. The feet should therefore be kept at a distance
from each other, and the arms should be held out from
the chest. Scharfetter et al. (1998) also analyzed the
artefacts produced by stray capacitance during whole
body or segmental Bioimpedance Spectroscopy (BIS),
and proposed a model for simulating the influence of
stray capacitance on the measured data.
BIA as a tool for assessment of the hydration of soft
tissue may be divided along
three methods of body fluid
volume assessment
.
The first
and the most validated method is prediction
of TBW from whole body impedance measurements at
a single frequency, often 50 kHz. Either the series
Such equations are not directly derived from bio-
physical laws, but have been empirically selected because
they give the best correlation. The correlation according
to eq.
(4.1.10)
can be better than 0.95, and it can be
slightly improved by also taking into account the body
weight
W
, eq.
(4.1.11)
. Hundreds of validation studies
with isotope dilution have established a solid relation
between whole-body impedance at 50 kHz and body
fluid volume (Kyle et al., 2004). Complex impedance
data can be given also as modulus and phase. Phase has
been used as an index of
nutrition.
This is true only in
comparison between vectors with the same modulus. For
instance, short vectors with a small phase angle are as-
sociated with edema whereas long vectors with an in-
creased phase angle indicate dehydration.
Fat free mass
is
predicted either from TBW (TBW/0.73) or through
specific regression equations including the same variables
as TBW, with different partial regression coefficients
(Sun et al., 2003; Kyle et al., 2004). Fat mass is calculated
by difference. The prediction error of best equations
while suitable for epidemiological studies is too high for
the clinical use (standard error of the estimate in the
order of 3-4 l for TBWand 3-4 kg for the fat free mass)
(Sun et al., 2003).
The second
method is the use of BIS following the
Cole model approach (many groups call it a Cole-Cole
model but that is a permittivity model), early used by
Cornish et al. (1993).
R
values are extrapolated at ex-
treme limit frequencies (0 and infinity) for prediction of
TBWand extracellular water (ECW) respectively, and by
difference intracellular water (ICW).
Body cell mass
is
then predicted as a function of the ICW (De Lorenzo
et al., 1997). The results of Lozano et al. (1995) showed
that there is a sharp disequilibrium between the in-
tracellular and extracellular compartments in the very
first dialysis period and they stressed the importance of
continuously monitoring segmental
impedance during
