Clinical Science (1982) 63,107-1 13
107
Cardiac output measured by transthoracic impedance
cardiography at rest, during exercise and at various lung
volumes
A . T . E D M U N D S , S. G O D F R E Y * A N D M A R I O N T O O L E Y
Department of Paediatrics and Neonatal Medicine, Hammersmith Hospital. London
(Received 27 July114 December 1981; accepted 5 March 1982)
Summary
1. Cardiac output measured by transthoracic
impedance cardiography has been compared with
simultaneous measurements made by the indirect
Fick CO, rebreathing method in nine adults and
14 children. All were healthy normal volunteers.
Sixty-six comparisons were made at rest and
during steady exercise at work loads up to 100
W.
2. Impedance measurements of cardiac output
were consistently higher than indirect Fick
measurements of cardiac output, but after application of a correction factor related to packed
cell volume there was close correlation between
the results obtained by the two methods (r =
0-94).
3. The mean coefficient of variation of impedance measurements of cardiac output was
13% at rest and 5% during steady-state exercise.
4. Changes of lung volume due to breath
holding or resulting from addition of an expiratory resistance did not affect the measurement of cardiac output by impedance.
5. Transthoracic impedance cardiography is a
rapid, non-invasive technique for measurement of
cardiac output. It requires very little active
co-operation from the subject. The method would
probably give reliable results for patients with
respiratory illnesses such as acute asthma or
bronchiolitis, during which changes of lung
volume may be expected to occur.
* Present address: Department of Paediatrics,
Hadassah University Hospital, Mount Scopus,
Jerusalem, Israel.
Correspondence: Dr A. T. Edmunds, Royal Hospital for Sick Children, Sciennes Road, Edinburgh
EH9 lLF, Scotland, U.K.
0143-5221/82/080107-07$01.50
Key words: cardiac output, exercise, impedance
cardiography, indirect Fick, lung volumes.
Introduction
Impedance cardiography is a non-invasive technique which has numerous potential applications in
clinical practice. Cardiac output is measured by
detection of the changes in transthoracic electrical impedance associated with the cardiac cycle
[l]. The method and its potential applications
have recently been reviewed by Mohapatra [21. It
relies on the over-simplified assumption that the
thorax is a homogeneous column of blood, the
volume of which increases by the left ventricular
stroke volume during each systole [31. Because of
this the results are largely empirical and measurements made under various physiological conditions must be verified by comparison with
established standard methods. Comparison with
several other methods has been made in adults
and children with various results 13-11]. As far
as we are aware cardiac output measured by the
impedance technique has not previously been
compared with simultaneous measurements made
by the indirect Fick CO, rebreathing method. We
report such a comparison made in normal
subjects at rest and during exercise. We have also
studied the effects of alteration of lung volume
caused by breath holding and by breathing
through an expiratory resistance on impedance
measurements of cardiac output in normal
subjects. This has been done in an attempt to
determine the potential value of the impedance
method for measuring cardiac output in patients
with lung disease.
0 1982 The Biochemical Society and the Medical Research Society
108
A . T. Edmunds, S. Godfrey and M . Tooley
TABLE1. Anthropometric data of 23 subjects
Mean values SE are shown.
Subject
Males
Females
Age
(years)
Height
(cm)
Weight
(kg)
~
Adults
Children
4
5
9
5
29 ? 1.1
11,5?0.8
Methods
Subjects
Comparison of cardiac output measured by
impedance with simultaneous indirect Fick CO,
measurements was made in studies of 23 normal
healthy volunteers. Nine were adults (five women
and four men), mean age 29 f SD 1 years.
Fourteen were children (nine boys and five
girls), mean age 11.8 rt 0.8 years. Their
anthropometric data are shown in Table 1. All
consented to participate in the study after full
explanation, and were free to withdraw at any
stage.
Impedance cardiac output
Impedance cardiac output was measured with
the Minnesota Impedance Cardiograph model
304A. This utilizes four disposable aluminized
strip electrodes. Two were placed around the
neck, one with its lower border at the level of the
supra-sternal notch anteriorly and the spinous
process of the seventh cervical vertebra posteriorly and one parallel to the latter, but at least 3 cm
above it. The lower pair of electrodes was applied
horizontally, one with its upper border at the tip
of the xiphisternum and the second parallel and at
least 3 cm below it. A constant sinusoidal current
of 4 mA r.m.s. with a frequency of 100 KHz was
applied between the outer two electrodes. The
potential difference, Z,, between the two inner
electrodes was recorded via a high input impedance linear amplifier. The potential varies
during the cardiac cycle. For the calculation of
stroke volume the thorax is assumed to be a
homogeneous cylinder of blood, the diameter of
which changes uniformly during the cardiac
cycle. The measured change of potential reflects
the increase in volume of the cylinder during
systole. This increase in volume, AV, can be
calculated by using the formula of Kubicek et al.
[31.
173 ? 4
141f4
Surface area
(m')
~~
64 ? 3
31?3
1.75 f 0.06
1.11f0.06
where AV represents stroke volume (ml), P is the
resistivity of blood (ohmlcm), L is the mean
distance between the recording electrodes, 2, is
the basal thoracic impedance (ohm), (dZldt)max.
is the maximum rate of fall of impedance in
ohmsls and T is the ventricular ejection time (s).
In this study the outputs taken from the impedance cardiograph were Z,, dZldt, an ECG
trace and a phonocardiogram trace. A typical
record is illustrated in Fig. 1.
P was calculated according to packed cell
volume from the data of Mohapatra et al. [121.
Packed cell volume was measured on finger-prick
blood specimens spun for 3 min on a Hawksley
haematocrit centrifuge. L was calculated as the
mean of the distances between the lower edge of
the upper and upper edge of the lower recording
electrodes in the mid-line anteriorly and posteriorly, 2, was taken as the mean of the maximum
and minimum values on the digital display during
the period of the record. The subjects breathed
freely during measurements. Therefore the
baseline for the measurement of (d2ldt)max. was
taken as the start of the steep upstroke irrespective of its position relative to the calibration zero
[2, 101. T was calculated as the time between the
start of the steep upstroke and the start of the
second heart sound, which coincides with the
lowest point on the dZldt trace. Beats were
analysed only if there was no obvious movement
artifact. Traces from subjects at rest, during
treadmill walking and during moderate exercise
on the bicycle ergometer were rarely distorted,
and most could be analysed without difficulty. At
the highest workloads movement of the baseline
in association with respiration became more
prominent and a proportion of the traces were
not suitable for analysis. When this happened the
measurement was immediately repeated, while
the steady-state exercise was continued unaltered.
Cardiac output was calculated from the mean
stroke volume measured over 8-12 consecutive
beats during one or more complete respiratory
cycles:
Cardiac output = Mean AV x HR
where HR is heart rate (beatslmin).
Impedance cardiography
Indirect Fick CO, rebreathing cardiac output
Measurements of cardiac output by the indirect Fick CO, rebreathing method were made
as described by Godfrey [131. As the subjects
were all normal, end-tidal CO, concentration was
measured and assumed to have a partial pressure
equal to that of arterial blood (Paco,). Subjects
breathed through a low resistance, low dead
space (50 ml) valve. Inspired ventilation was
measured with a dry gas meter and the expired
gas flushed through a 4 litres mixing chamber.
Gas was continuously sampled either from the
mouthpiece or from the outlet of the mixing
chamber. It was passed through an infrared
analyser to measure the concentration of CO,
and a paramagnetic oxygen analyser to measure
the concentration of 0,. The analysers were
calibrated frequently during study periods with
gases of known concentration, which had been
analysed chemically by means of the Haldane
technique. Data were continuously recorded on an
ink jet recorder. CO, rebreathing was performed
with a sliding valve assembly. By means of
this the subject could be switched instantaneously
from breathing air through the valve, to rebreathe
from a previously prepared bag of CO, in 0,.
The bag CO, concentration was made slightly
higher than the predicted mixed venous CO,
concentration (Hco,) and the volume was
adjusted according to the subject’s tidal volume
to ensure good equilibration with alveolar gas in
less than one circulation time. fico, was
109
measured as the plateau concentration of CO,
achieved. If no satisfactory plateau was recorded
it was estimated by using the extrapolation
method of Denison [13, 141. In the calculation of
cardiac output, measured values of HCO,
were
used for all subjects with no correction for the
down-stream effect described by Jones et al. 1151.
Measurements of steady-state oxygen consumption (Yo,) and cardiac output were made
with the subject seated at rest on a bicycle
ergometer and performing exercise at one or
more workloads ranging from 16.5 to 100 W.
Impedance cardiograph recordings were made
during the measurement of steady-state Po,
immediately before the performance of the CO,
rebreathing manoeuvre, so that comparison of
paired measurements could be made.
Effects of change of lung volume and addition of
expiratory resistance
The effect of change of lung volume on the
measurement of cardiac output by impedance
was determined in seven resting subjects. Measurements were made during breath holding with
glottis open at functional residual capacity
(FRC), total lung capacity (TLC) and at residual
volume (RV). For each subject vital capacity was
first measured by spirometry. During the experiment the subject breathed through a mouthpiece
and three-way tap with outlets to room air and to
a water-sealed spirometer. Cardiac output was
measured first at FRC, then at TLC and finally
110
A . T. Edmunds, S. Godfrey and M . Tooley
at RV after a maximum expiratory manoeuvre
into the spirometer.
The effect of breathing against an expiratory
resistance on the measurement of cardiac output
by impedance was determined in six normal
adults. With the subject seated in a constantpressure whole-body plethysmograph and breathing in time with a metronome, a variable
expiratory resistance was adjusted until FRC
increased by 50% of the resting inspiratory
capacity. Subsequently outside the plethysmograph, again with the subject breathing in
time with the metronome, a steady-state resting
cardiac output was measured by impedance and
indirect Fick methods. The measurements were
then repeated with the previously determined
resistance in the expiratory line between the
mouthpiece and the mixing chamber.
The reproducibility of the measurement of
cardiac output by impedance was determined at
rest in 11 subjects and during steady-state
treadmill walking in three subjects. Six to ten
measurements were made over a period of half an
hour and the coefficient of variation was
calculated.
Results
The linear relationship between the data obtained
by simultaneous measurements of cardiac output
by indirect Fick CO, rebreathing and by impedance is shown in Fig. 2. The data for nine
adults and 14 children are plotted separately.
Impedance measurements of cardiac output
were consistently higher than indirect Fick
measurements of cardiac output. This difference
was more marked for adults than for children.
The ratio of impedance cardiac output values to
indirect Fick values was significantly related to
packed cell volume (Fig. 3). This relationship
could be used as a correction factor to reduce the
calculated values of impedance cardiac output.
From the regression line in Fig. 3 the impedance
cardiac output/Fick cardiac output ratio was
determined according to the measured packed
cell volume. Impedance cardiac output was then
divided by this ratio, which was used as the
correction factor. The relation between indirect Fick cardiac output and impedance cardiac
output for all subjects after correction for packed
cell volume in this way is shown in Fig..4. There
was a linear relationship between Vo, and
cardiac output measured by both methods, as
shown in Fig. 5.
The reproducibility of the measurement of
cardiac output by impedance cardiography in 11
subjects sitting at rest and in three subjects during
. .
.'./
22 23
21
20
-
-
Indirect Fick Q (Vmin)
FIG. 2. Linear regression of simultaneous paired
measurements of cardiac output (Q) by impedance
cardiography vs indirect Fick CO, rebreathing.
Measurements were made in resting subjects and
Line of
during exercise. a, Adults; , children. -,
best fit for adults: y = 1 . 3 4 ~+ 1.79 (k2.1 SE), r =
0.92, n = 34. - - -, Line of best fit for children: y =
1.22X + 0.61 (f1.2 SE), r = 0.88, n
32. --,
Line of identity.
34 35 36 37 3 8 39 40 41 42 43 44 45 46 47
PCV (%)
FIG. 3. Linear regression for the ratio of cardiac
output (Q) measured by impedance and measured by
indirect Fick CO, rebreathing vs packed cell volume
(PCV). y = 0 . 0 3 ~+ 0.28 (k0.26 SE), r = 0.39,
n = 66.
steady-state treadmill walking is shown in Table
2. The mean coefficient of variation at rest was
13% and during exercise was 5%. Analysis of
variance demonstrated that after removal of the
difference between subjects the standard
Impedance cardiography
//
*=
i
.i
..
g 12;1 1 g 10x 9.s 8 E
i
-2
1
2 3 4
4
1
v
7-
'E
0
111
5 6 7 8 9 10111213141516
6 54 3 -
2 -
0.2
0
0.4
0.6
Indirect Fick Q (I/min)
FIG. 4. Linear regression of simultaneous paired
measurements of cardiac output (Q) by impedance
cardiography corrected for packed cell volume (PCV)
and by indirect Fick CO, rebreathing. Measurements
were made in resting subjects and during exercise.
M, Adults; 0 , children. -, Line of best fit:
y = 0.93~+ 0.31 (&1.17 SE), r = 0.94, n = 66.
__- ,Line of identity.
15
14
-
;.
.z2
Discussion
In early studies of impedance cardiography a
constant value of 150 ohm/cm was assumed for
P 13-6, 91. More recently various values of P,
related to the subjects' packed cell volume, have
been used [7, 8, 10, 111. A number of different
relationships between P and packed cell volume
have been reported 1111. We have used the
relationship reported by Mohapatra ef al. 1121
because it was obtained with fresh specimens of
capillary blood from normal newborn babies,
children and adults. Costeloe et al. 1101 found
1.2
1.4
1.6
13
1.2
1.4
1.6
-
1211 10-
U
*
98 7-
e
.-
6 -
.0
deviation of the residual differences for the resting
measurements was 10%. For the measurements
during steady-state exercise it was 4.9%. Thus
the 95% confidence limits for measurement of
cardiac output by impedance cardiography were
f10% during moderate steady-state exercise.
Comparison of 19 paired estimates of cardiac
output made in seven subjects breath holding at
TLC and RV is shown in Fig. 6. A paired t-test
showed no significant differences between the
measurements made at the two lung volumes (P
> 0.05).Comparison of paired measurements in
six subjects made at rest and while breathing
against an expiratory resistance is shown in Fig.
7. Again a paired f-test showed no significant
difference between measurements made at rest
and those made while breathing through the
resistance (P> 0.05).
0.8 1.0
Po, (I/min)
-
0
E
54 3 -
21 -
0
0.2
0.4
0.6
0.8
1.0
Po, (I/min)
FIG. 5 . Linear regression for cardiac output vs fro,
at rest and during exercise. (a) Impedance cardiac
output (Q) corrected for packed cell volume. M, Adults
[ y = 6 . 2 ~+ 3.7 ( + l a 3
SE), r = 0.93, n = 341;
0 , children [ y = 5 . 4 ~+ 2.7 (k1.1 SE), r =.0.79,
n = 321. (b) Indirect Fick cardiac output (Q). H,
Adults [ y = 7x + 3.1 (t1.0 SE), r = 0.97,n = 341;
0 , children [ y = 6x + 2.7 (L-0.8 SE), r = 0.9,
n = 321.
this gave an overestimation of cardiac output in
infants, and derived a correction factor related to
packed cell volume to bring their impedance
measurements into line with simultaneous
measurements of pulmonary capillary flow obtained by the nitrous oxide rebreathing method.
Many factors including chest shape and the
resistivity of tissue must affect measurement of
cardiac output by impedance. Hence it is unlikely
that the same correction factor would apply to
newborn babies and to older children or adults.
We therefore derived a new correction factor for
our subjects. It is again related to packed cell
A . T. Edmunds, S . Godfrey and M . Tooley
112
TABLE2. Reproducibility of measurements of cardiac output by
impedance cardiography in 11 subjects at rest and three subjects
during steady-state exercise
Mean cardiac outputs f SD are shown.
Subject
no.
No. of
observations
Mean cardiac
output
(I/min)
~
Coefficient of
variation
Mean coefficient
of variation
(%)
(%)
~~~
At rest
1
2
3
4
5
6
3.9 f 0.7
4.2 f 0.4
3.0 f 0.3
6.0 0.4
6.2 f 0.7
8
6
6
9
9
+
2.7 f 0.7
5.9 f 0.6
8
9
7
8
8
5.4 f 0.2
8.0f 0.4
3.9 f 0.4
5.4 f 0.6
10
10
10
7.8 f 0.4
13.5 It 0.6
13.2 f 0.6
8
8
8
9
10
11
13
10
4
5
10
II
Treadmill
walking
12
13
14
:}
5
5
7-
7 -
z.-
6h
E
2
>
e:
z
C
3
2
5-
65-
4 -
Q
g
z
3-
E 2.,a
1 ~
0
1
2
3
4
5
6
7
8
9
Q measured at TLC (I/min)
FIG. 6. Relation between 19 paiied measurements of
impedance cardiac output (Q) made at residual
volume and total lung capacity in seven subjects at
rest. - - -, Line of identity; 20% limits are shown by
unbroken lines.
volume and, probably because the range of
packed cell volume in our subjects was relatively
small, it could be expressed in linear form. After
applying the above correction there was very
close correlation (r = 0.94) between the simultaneous measurements of cardiac output made by
the impedance and the indirect Fick methods
(Fig. 4). This demonstrates that impedance
cardiography can provide measurements of cardiac output in normal subjects at rest and during
moderate exercise which are as accurate as those
obtained with the indirect Fick CO, rebreathing
0
1
2
3
4
5
6
7
Q at rest (I/min)
FIG. 7. Relation in six subjects between paired
measurements of cardiac output (Q) while breathing
against expiratory resistance and at rest. Measurements were made simultaneously by impedance (.)
and by indirect Fick (0)in each subject. - - -, Line of
identity; 20% limits are shown by unbroken lines.
method. This accuracy was further confirmed by
the finding of the expected linear relationship
between cardiac output and Vo, (Fig. 5). The
equations for the regression lines of cardiac
output related to VO,for both adults and children
obtained by impedance cardiography and by
indirect Fick C O , rebreathing were similar to
previously published equations derived from
direct Fick measurements [ 161, indirect Fick
Impedance cardiography
measurements 131 and dye-dilution measurements [16, 171. The scatter of the indirect Fick
results about the regression line was very similar
to that in previous studies [ 13, 181 and that of the
impedance data was only slightly greater.
Analysis of variance demonstrated that between
individual subjects there were significant differences in the slope of the relationship between
impedance cardiac output and indirect Fick CO,
rebreathing measurements of cardiac output.
These differences were accounted for by the
increasing scatter of measurements at the highest
workloads on the bicycle ergometer. This scatter
was not greater than expected when the
coefficient of variation of the impedance measurements was taken into account. Therefore all data
have been used in the calculation of the regression equations in Figs. 4 and 5. The mean
coefficient of variation of impedance measurements at rest was 13% with 95% confidence
limits of +_20%.This is similar to previously
published values [3, 61. During steady-state
treadmill walking the mean coefficient of
variation was reduced to 5% with 95% confidence limits for the measurement of f 10%. This
compares favourably with published values for
reproducibility for 0, direct Fick [191, dyedilution [19, 201 and CO, rebreathing indirect
Fick measurements [2 11. The reproducibility of
measurements of impedance cardiac output was
not determined at the highest workloads on the
bicycle ergometer, but movement of the baseline
was prominent in this situation and a proportion
of the traces could not be analysed because of
movement artifact. Thus the 95% confidence
limits would almost certainly have been greater
than f10%.
Change of lung volume as a result of breath
holding or induced by breathing against an
expiratory resistance did not materially affect the
measurement of cardiac output by impedance
(Figs. 6 and 7). Thus impedance cardiography
would probably be reliable for both clinical and
experimental purposes in patients with respiratory illnesses such as acute asthma or
bronchiolitis, during which changes of lung
volume may be expected to occur. It has the
added advantage that it is rapid, non-invasive and
it can easily be carried out at the bedside.
Acknowledgments
We thank Mr A. 0. Hughes for assistance with
statistical interpretation. A.T.E. was supported
113
throughout the study by a project grant from the
Medical Research Council.
References
I l l BONJER, F.H., VAN DER BERG, J.W. & DIRKEN,M.N.J.
(1952) The origin of the variations of body impedance
occurring during the cardiac cycle. Circulation, 6,415-420.
S.N. (1981) Non-invasive Cardiovascular Moni[21 MOHAPATRA,
toring by Electrical Impedence Technique, pp. 1-185, 82.
Pitman Medical, London.
131 KUBICEK,
W.G., KARNEGIS,
J.N., PATTERSON,
R.P., WITSOE,
D.A. & MATTSON,R.H. (1966) Development and evaluation
of an impedance cardiac output system. Aerospace Medicine,
37,1208-1212.
J.C. (1968) Determination of
141 HARLEY,A. & GREENFIELD,
cardiac output in man by means of impedance plethysmography.Aerospace Medicine, 39,248-252.
151 KINNEN,E. (1970) Cardiac output from transthoracic impedance variations. Annals of the New York Academy of
Sciences, 170,747-765.
F.M. &
161 BAKER,L.E., JUDY,W.V.,GEDDES,L.E., LANGLEY,
HILL,D.W. (1971) Measurement of cardiac output by means
of electrical impedance. Cardiovascular Research Centre
Bulletin, 9,135-145.
Z., EHMKE,D.A., DURMISS,
R.E., LEAVERTON,
P.E.
[71 LABABIDI,
& LAVER,R.M. (1971) Evaluation of impedance cardiac
output in children. Paediatrics, 47,870-879.
J.C., MAHER,J.T., REEVES,J.T., CRUZ,J.C.,
181 DENNISTON,
CYMERMAN,
A. & GROVER,R.F. (1976) Measurement of
cardiac output by electrical impedance at rest and during
exercise. Journal of Applied Physiology, 40,9 1-95.
J.H., ORO, L. & EKELUND,
L.G.
191 GABRIEL,S., ATTERHOG,
(1976) Measurement of cardiac output in patients with
myocardial infarction. Comparative evaluation of impedance
and dye dilution methods. Scandinauian Journal of Clinical
and Laboratory Investigation, 36,29-34.
1101 COSTELOE,
K., STOCKS,J. & GODFREY,S. (1977) Cardiac
output in the neonatal period using impedance cardiography.
Paediatric Research, 11, 1171-1 177.
[ I 11 ARENSON,
H.M. (1978) The use of impedance techniques for
the assessment of bloodflow during anaesthesia. Ph.D. Thesis,
London University.
1121 MOHAPATRA,S.N., COSTEWE,K. & HILL, D.W. (1977)
Blood resistivity and its implications for the calculation of
cardiac output by the thoracic electrical impedance technique.
Intensive Care Medicine, 3,63-67.
S. (1974)Exercise Testing in Children, pp. 42-50,
1131 GODFREY,
47,95-101. W.B. Saunders Co. Ltd, London.
(141 DENISON,D. (1968) Mixed venous blood gas tensions and
respiratory stress in man. Ph.D. Thesis, London University.
I151 JONES,N.L., CAMPBELL,
E.J.M., MCHARDY,
G.J.R., HIGGS,
B.E. & CLODE,M. (1967) The estimation of carbon dioxide
pressure of mixed venous blood during exercise. Clinical
Science, 32,311-327.
1161 WADE, O.L. & BISHOP,J.M. (1962) Cardiac Output and
Regional Blood Flow, pp. 5 1-72. Blackwell Scientific Publications, Oxford.
1171 ASTRAND,P.O., CUDDY,T.E., SALTIN,B. & STENBERG,
J.
f 1964) Cardiac outnut durine submaximal and maximal work.
Journal of Applied i'hysiolo&, 19,268-274.
I181 JONES, N.L., CAMPBELL,E.J.M., EDWARDS,R.H.T. &
ROBERTSON,
D.G. (1975) Clinical Exercise Testing, pp.
33-35. W.B. Saunders Co. Ltd, London.
1191 COURNAND,
A,, RILEY,R.L., BREED,E.S., BALDWM, E.F. &
RICHARDS,D.W. (1945) Measurement of cardiac output in
man using a technique of catheterisation of the right auricle or
ventricle. Journal of Clinical Investigation, 24,106-1 16.
1201 G F ~ B YG,., NILSSON,N.J. & SANNE, H. (1966) Serial
determination of cardiac output at rest. British Heart Journal,
28,118-121.
1211 ZEIDIPARD,E., SILVERMAN,
M. & GODPREY,S. (1972)
Reproducibility of indirect (CO,) Fick method for calculation
of cardiac output. Journal of Applied Physiology, 33,14 1-143.