Simultaneous measurement of cardiac output by

British Journal of Anaesthesia 1994; 72: 133-138
EQUIPMENT
Simultaneous measurement of cardiac output by thermodilution,
thoracic electrical bioimpedance and Doppler ultrasound
G. CASTOR, R. K. KLOCKE, M. STOLL, J. HELMS AND I. NIEDERMARK
SUMMARY
To evaluate the accuracy of two non-invasive
techniques for cardiac output (CO) measurement,
we have measured CO simultaneously by thoracic
electrical bioimpedance (TEB), pulsed Doppler
ultrasound (DU) and standard
thermodilution
methods (TD) under different clinical conditions.
Measurements were made in 10patients: (I) during
steady state anaesthesia with controlled IPPV
ventilation (n = 131), spread over the entire
ventilator/ cycle; (II) during apnoea (n = 56); (III)
during spontaneous breathing (n = 152) in the
intensive care unit. Mean (SD) cardiac output values
were: (I) COTD 3.5 (1.0) litre min~1, COTEB 3.4
(0.7) litre min-1, CODU 2.8 (0.7) litre min'1; (II)
COTD 3.6 (0.6) litre min-1, COTEB 3.5 (0.4) litre
min-1, CODU 2.9 (0.7) litre min'1; (III) COTD 7.7
(1.5) litre min-1, COTEB 7.6 (1.9) litre min-1, CODU
5.2(1.4) litre min'1. The mean percentage deviation
of TEB from TD ranged from -2.2% to 1.4% and
that of DU from TD was from -16% to -32%.
There were no statistically significant differences
between TD and TEB, but TD and DU differed
significantly during IPPV, apnoea and spontaneous
ventilation (P < 0.0001). (Br. J. Anaesth. 1994;
72: 133-138)
KEY WORDS
Measurement techniques: cardiac output.
Thermodilution is the most common clinical method
for measurement of cardiac output. However, there
are some theoretical and practical limitations to
dilution techniques and small but significant
associated morbidity and mortality [1, 2]. Consequently, many efforts have been made recently to
develop a non-invasive method with similar accuracy
and reproducibility as for thermodilution. Two of
these alternative methods are thoracic electrical
bioimpedance (TEB) and Doppler ultrasound (DU).
Thoracic electrical bioimpedance ( TEB)
The concept of TEB waveform analysis for
estimating stroke volume was introduced in 1966 by
Kubicek and colleagues [3]. They assumed that the
human thorax is a cylinder with a basic circumference equal to that of the thorax at the level of the
xiphoid [4]. This cylinder has an electrical length
(L), which is the distance between the voltage-
sensing electrodes placed on the neck and the
xiphoid; it is perfused homogeneously with blood of
specific resistivity, p, (units (Q • cm) which varies with
PCV and has a basic impedance, Z o , (Q).
The principle of impedance cardiography is that a
constant small current (2.5 mA at 70 kHz), applied
between the outer current electrodes (1 and 4) on the
human thorax, produces changes in the voltage
sensed with the inner electrodes (2 and 3) when the
thoracic conductivity changes during systole and
diastole. The first derivative of the impedance
waveform, dZ/dt, is related linearly to aortic blood
flow. There exist some modifications of the original
assumption and similar basic analysis, such as the
formula of Sramek [5], who took a truncated cone
instead of a cylinder for the electrically participating
tissue of the thorax and who regarded PCV as a
constant. The detailed derivation of the equation of
Sramek was explained in an article by Bernstein [4].
Doppler ultrasound (DU)
The Doppler effect describes the changes in sound
wave frequencies when the sound and receiver are
moving relative to one another. The waves are
directed towards the moving blood cells and then
reflected back. This results in a frequency differing
from that transmitted and directly proportional to
blood flow velocity. Stroke volume can be calculated
when the area under the velocity curve and the crosssectional area (CSA) of the aorta are known.
A newer pulsed Doppler obviates the need for
measuring CSA separately by insonating the
ascending aorta with a two-element, pulsed, 2-MHz
ultrasound beam from the suprasternal notch. The
ultrasound beam comprises a wide beam, which
insonates the total lumen of the ascending aorta and
GERALD CASTOR*, M.D.; REINHARD K. KxoocEt; M.D.; MARTIN
STOLI4, M.D.; JORGEN HELMS$, M.D.; INGRID NIEDERMARK§,
M.D.; Department of Anaesthesiology, University of Saarland,
66424 Homburg, Germany. Accepted for Publication: July 15,
1993.
Present addresses:
* Department of Anaesthesiology, Caritas Krankenhaus,
Werkstr. 1, 66763 Dillingen/Saar, Germany.
f Department of Cardiology, Marienhospital, 49377 Vechta.
$ Department of Neurology, University of Saarland, 66424
Homburg.
J Department of Anaesthesiology, Stadtisches Krankenhaus,
67655 Kaiserslautern.
5 Department of Anaesthesiology, Kreiskrankenhaus, 66333
Volklingen.
Correspondence to G.C.*
BRITISH JOURNAL OF ANAESTHESIA
134
a narrow beam which covers a denned area which lies
within this vessel. This enables the meaurement of
both CSA and blood flow velocity from the suprasternal notch. The flow lumen area of the wide beam,
which totally encompasses the ascending aorta,
depends on the angle of insonation and the cross
sectional area of the aorta. The average blood flow
velocity is measured using the Doppler shift frequency of the wide beam; as this beam covers the
whole lumen of the aorta, it is an average value of the
whole flow lumen area. So, calculation of CO is
independent of the angle of insonation. When the
angle increases, the velocity increases and the lumen
area decreases proportionally. This independence,
which is new compared with other Doppler measurements of CO, is important for the practical use of this
method.
When compared with standard invasive methods,
the results of TEB and Doppler methods range from
poor correlation [6-10] to good agreement [11-16].
The purpose of this study was to evaluate the
accuracy of CO measurements obtained simultaneously by TEB and DU in comparison with the
standard TD method in different clinical conditions.
MATERIAL AND METHODS
The study was approved by the local Ethics Committee and the patients gave written, informed
consent. We studied 10 patients (four male), ASA
III and IV, aged 35-65 yr, who had a Swan-Ganz
catheter (93A-131-7F, Baxter Edwards Laboratories, Irvine, CA) inserted during neurosurgical
removal of intracranial tumour or aneurysm. The
catheter was inserted via the right internal jugular
vein and placed in the pulmonary artery, as confirmed radiologically. Two pairs of electrodes were
applied (5 cm apart) to the neck and lower thoracic
aperture at the level of the xiphoid and connected to
an impedance cardiograph (NCCOM3-R7, BoMed
Med., Irvine, CA). Measurements started after
incision of the dura mater at steady-state conditions
of arterial pressure and heart rate. TEB and TD
measurements (Cardiac Output Computer 9520A,
Baxter Edwards Laboratories, Irvine, CA with a
closed injectate system) were made simultaneously.
Ice-cold 5 % glucose 10 ml was injected via a closed
system for TD measurements. Measurement of CO
was performed by a time-temperature integration
curve of the indicator-blood mixture. As it is the
standard method used by Baxter, integration was
terminated automatically when the curve returned to
30% of its peak, because this point is related to
minimal variability. Deadspace correction was made
according to instructions given by Edwards Laboratories. Measurement procedure was as follows:
during injection of the ice-glucose until the presentation of the result by the TD computer, COTEB
values were derived by beat-to-beat mode, normally
four to six values, depending on the heart rate, and
the mean taken for comparison. Thereafter, the
Doppler transducer (Quantascope-Vital Science,
Denver, CO) was placed at the suprasternal notch.
Normally, a measuring time of 15—30 s was necessary
to obtain one averaged CODU value, depending on
the signal quality, which was then taken for comparison with the two other methods. One single TD
measurement was compared with the averaged TEB
and DU values.
Simultaneous measurement of TEB and DU leads
to prolonged disturbance of the impedance signal.
The metallic Doppler transducer absorbs a large
portion of the current between the two inner sensing
electrodes of the impedance cardiograph; this
reduces the dZ/dtmax signal and thus calculation
of the TEB cardiac output results in decreased CO
values. The investigator using the Doppler signal
was unaware of any of the TD and TEB
measurements. CO measurements took place during
the entire ventilatory cycle (I) and during endexpiratory apnoea (II). Mechanical ventilation was
set at a ventilatory frequency of 12 b.p.m. and a
minute volume of 100 ml kg"1. Measurement of TD,
TEB and DU was repeated on the next day in the
intensive care unit during spontaneous breathing
(III) after tracheal extubation.
Data are presented as mean (SD) and as mean
differences with SD of the differences (SDD). They
were analysed for statistical significance using the
Wilcoxon matched pairs signed rank test for nonparametric data. Statistically significant differences
between TD and TEB and between TD and DU
were assigned at P < 0.05.
A bias analysis was performed for cardiac outputs
obtained using TD and TEB and using TD and DU.
According to the statistical method of Bland and
Altman [17], we plotted the mean of TD and TEB
and the mean of TD and DU against the percentage
deviation of TEB from TD and of DU from TD. As
confidence and bias are the same when expressed as
percentages, we modified the Bland-Altman method
using percentage deviations instead of absolute
values. Two SD was chosen as confidence interval. In
this manner, the limits of agreement of each noninvasive method compared with TD can be taken as
bias±2SD.
RESULTS
During IPPV, 131 simultaneous measurements of
TD, TEB and DU were analysed in 10 patients. The
mean of all TD CO measurements was 3.5 (SD
1.0) litre min"1, that of TEB 3.4 (0.7) litre mnr 1 and
that of DU 2.8 (0.7) litre min-1. During IPPV, the
respective mean percentage deviation of TEB from
TD was 1.4 (SDD 16.2)% and that of DU from TD
- 1 6 (21)% (fig. 1).
During apnoea, the mean of all TD measurements
was 3.6 (0.6) litre min-1, that for TEB 3.5
(0.4) litre min"1 and that for DU 2.9 (0.7) litre min"1.
The mean percentage deviation of TEB from TD
was -2.2 (11.2)% and that of DU from TD - 18.7
(16.3)% (fig. 2).
In the intensive care unit, during spontaneous
ventilation, the mean of TD measurements was 7.7
(1.5) litre min"1, that for TEB 7.6 (1.9) litre mur 1
and that for DU 5.2 (1.4) litre min"1. The mean
percentage deviation of TEB from TD was —2.1
(11.0)% and of DU from TD -32.4 (12.4)%
(fig- 3).
There was no statistically significant difference
NON-INVASIVE VS INVASIVE CARDIAC OUTPUT MEASUREMENT
135
8080
•
£
2SD
Q
Mean
Q
c
o
co
0-
c
o
2SD
to
402SD
oMean
' I •
v -40
> -40 -4
D
Q
•2SD
-80
-80
2
3
4
5
6
7
Mean COTD + COTEB (litre min" 1 )
Mean COTD + CO DU (litre min" 1 )
FIG. 1. Bias plots of CO values obtained by TD and TEB, and by TD and DU during IPPV. Plots represent
percentage deviation of TEB from TD, or DU from TD. Solid lines = mean difference between the two methods and
two standard deviations, as the random error of variability between the two techniques.
80 -i
80-|
40 -
£40-1
Q
(-
Q
•2 SD
• 2 SD
I
CD
0-
• Mean
o
-2SD
CO
Q
Q
0
c
o
CsO
v -40 -J
a
-80-
Mean
• 2
SD
-80
2
3
4
5
Mean COTD - COTEB (litre mirT1)
2
3
Mean COTD
CO DU (litre min" 1 )
FIG. 2. Bias plots of CO values obtained by TD and TEB, and by TD and DU during apnoea. Plots represent
percentage deviation of TEB from TD, or DU from TD. Solid lines = mean difference between the two methods and
two standard deviations, as the random error of variability between the two techniques.
80-i
40Q
£
40"
3
Q
c
o
0-
2 SD
CO
LU
h—
Mean
o-
c
eviat
o
2SD
2SD
Mean
-40"
-402SD
-80
5
7
9
11
Mean COTD + COTEB (litre min"1)
13
-80
11
Mean C0 T D + CODU (litre min^1)
13
FIG. 3. Bias plots of CO values obtained by T D and TEB, and by T D and D U during spontaneous breathing. Plots
represent precentage deviation of TEB from T D , or D U from T D . Solid lines = mean difference between the two
methods and two standard deviations, as the random error of variability between the two techniques.
between TD and TEB during the different
conditions of measuring, but there was between TD
and DU and TEB and DU during IPPV, apnoea and
spontaneous ventilation (P < 0.0001).
We chose the method of Bland and Altman to
average the respective values of TD and TEB and of
TD and DU in order to minimize the inaccuracies in
the assumption that TD may not be a "gold
standard" of measuring CO [18]. As shown in figure
1, agreement between COTEB and CO-rj, was best
during IPPV; it was reduced slightly during apnoea
and spontaneous breathing. For CODU also, the best
agreement was during IPPV, with the largest bias
during spontaneous breathing.
During IPPV, apnoea and spontaneous breathing,
TEB overestimated CO compared with TD at
BRITISH JOURNAL OF ANAESTHESIA
136
smaller values of CO, whereas there was an underestimation at greater CO compared with TD during
spontaneous breathing.
DU seemed to underestimate CO continuously
compared with TD during all conditions.
DISCUSSION
According to Bland and Altman, bias analysis (the
mean difference between two methods) is more
appropriate than correlation and regression analysis
for comparing two methods that measure the same
variable. The bias represents the systematic error
between two measurement techniques, whereas SD of
the bias represents die random error of variability
between the two different techniques. Variability in
COTD, CO TEB , or CODU measurements can affect the
differences between the methods. (For witiiintechnique variabilities, see the work of Wong and
colleagues [19].) This variability may be caused by
systematic and random errors based on the different
principles for determination of CO. TEB and DU
measure CO of the left heart, whereas a thermodilution catheter is placed in the right heart. Left and
right ventricular stroke volume are not necessarily
equal [20, 21]. There is a greater constancy on the
left side [22,23], as die influence of ventilation,
above all during artificial ventilation, is less on the
left ventricle. Right ventricular stroke volume
changes significantly during inspiration and
expiration as a result of decreased venous return and
changing intrathoracic pressure [24, 25]. These noncontinuous flow conditions render the TD method
and all dilution methods more inaccurate compared
with continuous flow models [26, 27] because the
Stewart-Hamilton equation is not correct for varying
flow conditions.
Mackenzie and co-workers [28] assessed the
accuracy and reproducibility of thermodilution in an
artificial circulation during continuous or pulsatile
flow at several flow rates, using three different
cardiac output computers: Instrumentation Laboratories IL 701, Edwards 9520A and Gold SP 1435.
Pulsatile instead of continuous flow produced a
pronounced and similar increase in the variability of
measurement of all three devices, with the SE of the
prediction being approximately double that with
continuous flow. The accuracy of the measurements
was reduced, with flow rates being overestimated.
With pulsatile flow, injectate at ice temperature and
a 10-ml injectate, the Edwards 9520A computer gave
least variability.
According to Stetz and colleagues [18], who
reviewed the published literature and evaluated the
reproducibility and accuracy of the TD method, it is
difficult to determine the accuracy of this technique,
especially for clinical CO determination. Both the
Fick and die dye dilution methods are dilution
techniques with the same theoretical errors. With a
single TD measurement, a difference of about 22 %
is needed, particularly in the intensive care setting,
to demonstrate significant trends in CO. As a result,
TD measurements in determining CO of 3.5 and
4.5 litre min"1 during therapeutic intervention may
represent only reproducibility error, rather dian
dierapeutic success. Even with triplicate measurements, changes of 15 % in CO may not be significant.
Stetz's group used the standard error of the mean
(SEM%) (SEM/average CO) for examination of the
reproducibility of the TD technique. It varied from
2.0% to 5.0% for three measurements per
determination and from 3.5% to 8.7% for single
measurements. The SEM% values predict that the
TD method may need to differ by 6.0-15.0%
(3XSEM%) for triplicate measurements and by
9.5-26.1% for single measurements, in order to
establish a significant change.
Other problems with TD are injectate volume and
temperature, injection velocity (that may itself
produce accelerations of flow in the pulmonary
artery system [29]), heat loss in the catheter system
and in the vessel wall, and placement of the
tiiermistor in the pulmonary artery. As a consequence, thermodilution cannot be regarded as the
gold standard.
Compared widi TD, TEB slightly overestimates
cardiac output in the normal range during spontaneous ventilation and IPPV. In increased cardiac
output states there is a tendency to overestimation
during spontaneous ventilation, whereas during
IPPV, TEB seems to underestimate cardiac output.
There are two major problems of die TEB
method: the correct signal processing of die critical
parameters Zo, LVET and dZ/drmax [30] and die
empirically derived equation. The assumption by
Sramek [5] that the volume of the diorax, as
electrically participating tissue, is calculated as a
truncated cone instead of as Kubicek's cylinder, may
improve die accuracy [4] or not [31], but does not
correct overestimation of stroke volume produced by
die TEB mediod in spontaneous breadiing patients.
One of die reasons for this overestimation is die
assumption diat the ejection velocity during systole
is constant during die entire left ventricular ejection
time in die equations of Kubicek and colleagues [3]
and Sramek [5], whereas in reality die ejection
velocity decreases to zero after reaching its peak,
represented as the ohmic counterpart dZ/drmax in
die impedance-stroke volume calculations. This is
supported by the findings of Colocousis, Huntsman
and Curreri [32], in animal experiments measuring
CO by continuous-wave Doppler technique, tiiat at
small stroke volumes die ejection period is abbreviated and die velocity record is symmetrical, while
at large values of SV diere is a prolongation of
ejection and loss of symmetry. Odier investigators
[33,34] observed changes in ejection shape and
duration in response to altered haemodynamic conditions or heart rate. During mechanical ventilation,
tiiis velocity profile is altered, as TEB overestimates
CO compared widi TD in low flow conditions and
underestimates CO during sepsis [35]. Alteration of
Zo as basic impedance which represents die basic
fluid volume of die diorax may be excluded, as our
own investigations [13] and diose of Edmunds,
Godfrey and Tooley [36] demonstrated no loss of
accuracy during different lung volumes. The limitations for TEB are aortic regurgitation, tachyarrhythmiaSj open heart surgery and extreme obesity.
Compared with TD, DU underestimates CO con-
NON-INVASIVE VS INVASIVE CARDIAC OUTPUT MEASUREMENT
sistently, above all in increased cardiac output states
during spontaneous ventilation, as described also for
transtracheal Doppler [9]. Technical problems arose
from the fact that our Doppler device averaged CO
over a 12-s interval, which implies more than one
ventilatory cycle. Moreover, DU measurements
took place after termination of the TD and TEB
measurements.
Methodological problems in the Doppler method
include the assumption that the aorta is circular and
that there is no diameter change during systole [37].
As the cross-sectional area is a quadratic function of
the radius, small errors in measurement of diameter
may produce large errors in CO values [38].
Pulsations of the non-rigid aorta during systole
produce 5-17 % changes in CSA from its diastolic to
systolic pressure extremes [39]. In the presence of
turbulence which may occur with anaemia, tachycardia, etc., the mean velocity measured by the
Doppler beam is unrepresentative of the forward
flow that represents CO [14]. Another phenomenon,
known as aliasing, occurs with the pulsed Doppler in
high flow states and makes accurate measurement of
blood velocity impossible [6]. Doppler frequency
shift can be measured accurately only if it is less than
50 % of the sampling frequency.
Practical considerations
The DU method is user-dependent and requires a
long period of training [40]. It often takes a long time
to obtain a satisfactory signal. The recorded Doppler
signals selected for calculation of cardiac output
should be of optimal quality. With bad positioning of
the probe, the large beam does not cover the whole
lumen of the aorta and it may be impossible to obtain
an adequate signal in some patients [12]. Other
sources of error are a bad ultrasound contact through
the skin or the electric activity of surgical cautery.
Continuous monitoring is not possible, so it is
impossible to detect sudden changes in CO.
Our results on the accuracy of DU compared with
TD relate only to the dual beam Doppler.
In contrast, the TEB method is easy to perform.
The advantage of using normal ECG electrodes
instead of aluminium strips makes it comfortable for
the patient and it can be performed by nursing staff
after a short period of instruction.
Sources of errors in the TEB method include
incorrect electrode placement. Small alterations in
the position of the sensing inner electrodes 2 and 3
produce changes in CO of 5—10%; decreased
distance leads to overestimation, whereas increasing
distance produces underestimation of CO. Incorrect
input of height and weight of the patient in the
computerized system has similar effects. Normally,
2-3 cm errors in the adult range of the length L of
the human thorax (25—32 cm), which is implicated in
the Sramek equation [5] in the third power, would
produce errors of 20-30% in stroke volume calculation. As L is standardized as a percentage of the
patient's height H (0.17 H)3 in the Sramek equation,
the actual distance between the inner electrodes is
less important; the actual error in calculation is
reduced to about 10%. Both lead to a constant error
in COTEB.
137
REFERENCES
1. Bdhrer H, Wehlage DR, Stuth EEA. Komplikationen der
Swan-Ganz-Katheterisierung. Intensivmedizin 1990; 27:
155-162.
2. Elliot CG, Zimmerman GA, Clommer TP. Complications of
pulmonary artery catheterization in the care of critically ill
patients. Chest 1979; 76: 647-652.
3. Kubicek WG, Karnegis JN, Patterson RP, Witsoe DA,
Mattson RH. Development and evaluation of impedance
cardiac output system. Aerospace Medicine 1966; 37:
1208-1212.
4. Bernstein DP. A new stroke volume equation for thoracic
electrical bioimpedancc: Theory and rationale. Critical Care
Medicine 1986; 14: 904-909.
5. Sramek BB. Non-invasive technique of measuring cardiac
output by means of electrical impedance. In: Proceedings of
the 5th International Conference on Electrical Bioimpedance,
Tokyo, 1981; 39.
6. Donovan KD, Dobb GJ, Newman MA, Hockings BES,
Ireland M. Comparison of pulsed Doppler and thermodilution methods for measuring cardiac output in critically ill
patients. Critical Care Medicine 1987; IS: 853-857.
7. Keim HJ, Wallace JM, Thurston H, Case DB, Dreyer J.
Impedance cardiography for determination of stroke index.
Journal of Applied Physiology 1976; 41: 797-799.
8. Preiser JC, Daper A, Parquier JN, Contempre, Vincent JL.
Transthoracic electrical bioimpedance versus thermodilution
technique for cardiac output measurement during mechanical
ventilation. Intensive Care Medicine 1989; 15: 221-223.
9. Siegel LC, Fitzgerald DC, Engstrom RH. Simultaneous
intraoperative measurement of cardiac output by thermodilution and transtracheal doppler. Anesthesiology 1991; 74:
664-669.
10. Weber J, Heidelmeyer CF, Kubatz E, Bruckner JB. Die
Bestimmung des Herzzeitvolumens unter PEEP-Beatmung
mit dem nitchtinvasiven Bioimpedanzgerat "NCCOM3" im
Vergleich zur Thermodilutionsm-methode. Eine Untersuchung an narkotisicrtcn Hunden. Anaesthesist 1986; 35:
744-747.
11. Castor G, Helms J, Niedermark I, Molter G, Motsch J,
Altmayer P. Influence of the ventilatory cycle on cardiac
output determination during mechanical ventilation. (Comparison of noninvasive electrical bioimpedance to the standard
thermodilution method). American Journal of Noninvasive
Cardiology 1991; 5: 57-61.
12. Looyenga DS, Liebson PR, Bone RC, Balk RA, v. Messer J.
Determination of cardiac output in critically ill patients by
dual beam doppler echocardiography. Journal of the American
College of Cardiology 1989; 13: 340-347.
13. Cerny JC, Ketslakh M, Poulos CL, Dechert RE, Bartiett RH.
Evaluation of the Velcom-100 pulse Doppler cardiac output
computer. Chest 1991; 100: 143-146.
14. Mehta N, Iyvawe VI, Cummin ARC, Baylay S, Saunders
ICB, Bennett ED. Validation of a Doppler technique for
beat-to-beat measurement of cardiac output. Clinical Science
1985; 69: 377-382.
15. Rose JS, Nanna M, Shahbudin H, Rahimtoola SH, Elkayam
U, McKay C, Chandraratna PAN. Accuracy of determination
of changes in cardiac output by transcutaneous continuouswave Doppler computer. American Journal of Cardiology
1984; 54: 1099-1101.
16. Spinale FG, Smith AC, Crawford FA. Relationship of
bioimpedance to thermodilution and echocardiographic
measurements of cardiac function. Critical Care Medicine
1990; 18: 414-^118.
17. Bland JM, Altman DG. Statistical methods for assessing
agreement between two methods of clinical measurement.
Lancet 1987; 1:307-310.
18. Stetz CW, Miller RG, Kelly GE, Raffin TA. Reliability of the
thermodilution method in the determination of cardiac output
in clinical practise. American Review of Respiratory Diseases
1982; 126: 1001-1004.
19. Wong DH, Tremper KK, Stemmer EA, O'Connor D, Wilkur
S, Zaccari J, Reeves C, Weidoff P, Trujillo RJ. Noninvasive
cardiac output: Simultaneous comparison of two different
methods with thermodilution. Anesthesiology 1990; 72:
784-792.
20. Franklin DL, Van Citters RL, Rushmer RF. Balance between
BRITISH JOURNAL OF ANAESTHESIA
138
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
right and left ventricular output. Circulation Research 1962;
10: 17-26.
Guz A, Hoffman LIE, Charlier A, Wierich WR, Zanger L,
Benke J. Simultaneous observations of right and left
ventricular stroke volumes in conscious dogs. Federation
Proceedings 1961; 20: 133.
Hoffman JIE, Guz A, Charlier AA, Wilcken DEL. Stroke
volume in conscious dogs: Effect of respiration, posture and
vascular occlusion. Journal of Applied Physiology 1965; 20:
865-877.
Santamorc WP, Heckman JL, Bove AA. Cardiovascular
changes from expiration to inspiration during IPPV.
American Journal of Physiology 1983; 245: H307-312.
Armcngol J, Man GCW, Balsys AJ, Wells AL. Effects of the
respiratory cycle on cardiac output measurements. Reproducibility of data enhanced by timing the thermodilution
injections in dogs. Critical Care Medicine 1981; 9: 852-854.
Morgan BC, Dillard DH, Guntheroth WG. Effect of cardiac
and respiratory cycle on pulmonary vein flow, pressure and
diameter. Journal of Applied Physiology 1966; 21: 1276-1280.
Cropp GJA, Burton AC. Theoretical considerations and
model experiments on the validity of indicator dilution
methods for measurements of variable flow. Circulation
Research 1965; 18: 26-^8.
Scheuer-Leeser M, Morguet A, Rene H, Irmich W. Some
aspects to the pulsation error in blood-flow calculations by
indicator-dilution techniques. Medical and Biological
Engineering and Computing 1977; IS: 118-123.
Mackenzie JD, Haites NE, Rawles JM. Method of assessing
the reproducibility of blood flow measurement: factors
influencing the performance of thermodilution cardiac output. British Heart Journal 1986; 55: 14-24.
Fronefc A, Ganz V. Measurement of flow in single blood
vessels in changing cardiac output by local thermodilution.
Circulation Research 1960; 8: 175-179.
Nagel JH, Shyn SP, Reddy SP, Hurwitz BE, McCabe PM,
Schneiderman N. New signal processing techniques for
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
improved precision of noninvasive impedance cardiography.
Annals of Biomedical Engineering 1989; 17: 517-534.
De Mey C, Enterling D. Noninvasive assessment of cardiac
performance by impedance cardiography: Disagreement
between two equations to estimate stroke volume. Aviation,
Space and Environmental Medicine 1988; 59: 57-62.
Colocousis JS, Huntsman LL, Curreri PW. Estimation of
stroke volume changes by Ultrasonic Doppler. Circulation
1977; 56: 914-917.
Noble MIM, Trenchard D, Guz A. Changing heart rate on
cardio-vascular function in the conscious dog. Circulation
Research 1966; 19: 206-213.
Rushmer RF. Recent advances in cardiovascular physiology.
Anesthesia and Analgesia 1966; 45: 383-389.
Castor G, Motsch J, Altmayer P, Helms J, Molter G,
Niedcrmark I. Clinical evaluation of cardiac output measurement by thoracic electrical bioimpedance. European Journal
of Pharmacology 1989; 36: 38.
Edmunds AT, Godfrey S, Tooley M. Cardiac output
measured by transthoracic impedance cardiography at rest,
during exercise and at various lung volumes. Clinical Science
1983; 63: 107-113.
Bernstein DP. Noninvasive cardiac output, Doppler
flowmctry, and gold-plated assumptions. Critical Care Medicine 1987; 15: 886-887.
Chandrarama A, Nanna M, McKay C, Nimalasuriya A,
Swinney R, Elkayam U, Rahimtoola SH. Determination of
cardiac output by transcutaneous continuous-wave Ultrasonic
Doppler computer. American Journal of Cardiology 1984; 53:
234-237.
Greenfield JC, Patel DJ. Relation between pressure and
diameter in the ascending aorta of man Circulation Research
1962; 10: 778-781.
Loeppky JA, Hoekenga DE, Greene ER, Luft UC. Comparison of noninvasive pulsed Doppler and Fick
measurements of stroke volume in cardiac patients. American
Heart Journal 1984; 107: 339-346.