Steendijk crit care med 2000 28 1599

Ovid: Evaluation of a new transcardiac conductance method for continuous on-line m...

Page 1 of 9

Critical Care Medicine Issue: Volume 28(5), May 2000, pp 1599-1606 Copyright: © 2000 Lippincott Williams & Wilkins, Inc. Publication Type: [Apparatus and Techniques] ISSN: 0090-3493 Accession: 00003246-200005000-00057 Keywords: cardiac function, cardiac output, conductance catheter, contractile state, end-systolic volume, end-diastolic volume, hemodynamics, monitoring, on-line measurements, sheep, thoracic impedance [Apparatus and Techniques]

Evaluation of a new transcardiac conductance method for continuous on-line measurement of left ventricular volume Steendijk, Paul PhD; Lardenoye, Jan-Willem MD; van der Velde, Enno T. PhD; Schalij, Martin J. MD, PhD; Baan, Jan PhD

Author Information From Leiden University Medical Centre, Department of Cardiology, Cardiac Physiology Laboratory, The Netherlands.

Supported, in part, by the Leiden University Medical Centre.

Address requests for reprints to: Paul Steendijk, PhD, Leiden University Medical Centre, Department of Cardiology, PO Box 9600, 2300 RC Leiden, The Netherlands.

Abstract

Objective: To evaluate a new, less invasive, conductance method to measure continuous on-line left ventricular volume. End-systolic and end-diastolic volumes obtained with this transcardiac conductance method were compared with simultaneous measurements using the conventional intracardiac conductance catheter.

Design: Controlled animal study.

Setting: Research laboratory in a university hospital.

Subjects: Six sheep.

Interventions: Anesthetized sheep were instrumented and inotropic condition was varied by beta-receptor stimulation (5 µg/kg/min of dobutamine) and beta-receptor blockade (1 mg/kg of propranolol). In each condition (control, dobutamine, repeat control, propranolol), ventricular volume was varied over a wide range by gradual preload reduction using a vena caval balloon catheter.

Measurements and Main Results: We compared the two methods by performing linear regression analysis on simultaneous end-systolic and end-diastolic volumes obtained during gradual caval occlusions. We statistically analyzed the intercepts, slopes, and correlation coefficients of the regression equations relating the transcardiac and conductance catheter measurements to determine the effects of interanimal variability, inotropic condition, and cardiac phase on the relationship between the two methods.

The results show an excellent linear correlation between the two methods (mean intercept, -1.82 ± 1.24 mL; mean slope, 0.787 ± 0.024 and r2 = .94). Both slope and intercept of the relationship between the two methods show a significant interanimal and cardiac phase related variability but no significant dependence on inotropic condition.

Conclusions: The significant interanimal variability indicates that the new method requires individual calibration in each subject. However, the small variability of the regression coefficients with changes in condition indicates that after initial calibration, end-systolic and end-diastolic volume can be followed accurately even in the presence of large changes in volume and inotropic state. This new method may facilitate quantitative continuous assessment of cardiac function in clinical practice, for example, in the intensive care unit.

Estimates of ventricular volume and derived variables such as stroke volume, ejection fraction, filling rate, and ejecting rate are used widely to quantify the cardiac status of patients. There is consensus that absolute ventricular volume is a key measurement because it enables one to assess myocardial contractile state relatively independent of loading condition (1). This is important because, as a rule, in pathologic conditions neurohumoral-induced changes in vascular tone and intrinsic myocardial contractility seek to maintain central arterial blood pressure and perfusion to critical organs. Likewise, many drugs affect both the myocardial contractility and pre- and afterload by changes in vessel tone. In addition, left ventricular end-systolic volume has been shown to be the major determinant of survival after recovery from myocardial infarction (2, 3). Recently, the Global Utilization of Streptokinase and t-PA for Occluded Coronary Arteries (GUSTO)-I study showed that end-systolic volume index measured as early as 90 mins into reperfusion therapy is a strong predictor of early and late mortality (4). However, current methods to measure volume (echocardiography, radionuclide ventriculography, magnetic resonance imaging, angiography) are not suited for continuous monitoring of patients, for example, in the intensive care unit. In practice, monitoring of volumetric variables is restricted to cardiac output by thermodilution. Extensive work has been done to derive volumetric variables using thoracic impedance measurements. The main advantages of this technique are that it is noninvasive, provides on-line signals, and causes minimal discomfort to the patient. At best, however, the method gives an estimate of changes in cardiac output. Moreover, recent reviews indicate that its accuracy remains controversial (5, 6). The main reason for the difficulties in interpreting the thoracic impedance signal is that the signal originates from systemic arterial and venous blood volume variations as well as from changes in the orientation of erythrocytes in these vessels (7). The direct contribution of changes in ventricular volume per se was shown to be very small (8, 9). The conductance catheter method previously developed in our laboratory uses the conductive properties of blood and tissues to measure volumetric variables. In addition to measuring cardiac output, this technique, when properly calibrated, gives absolute on-line ventricular volume (10, 11). Whereas with thoracic impedance measurements, the myocardial wall tends to "shield" the influence of intracardiac volume changes from the recorded signal, the situation is reversed with the conductance catheter that is placed inside the ventricular cavity. Various comparative studies have demonstrated the accuracy of this technique for instantaneous assessment of absolute left ventricular volume and stroke volume (10-16). The disadvantage of the conductance catheter is that it is highly invasive, which restricts its application to patients in the catheterization lab or the operating room.

In the present study, we introduce a transcardiac conductance method that uses a catheter with a tip-electrode placed in the superior vena cava in combination with standard epithoracic electrocardiogram (ECG) electrodes to measure ventricular volume. By using the catheter electrode and an ECG electrode near the xyphisterial joint, we set up an electric field that is mainly confined to the heart and resembles the conductance catheter field. As with the conductance catheter technique, six additional electrodes were used to pick up five voltage differences. These electrodes were placed in line on the left side of the thorax, roughly covering the projection of the heart.

Clearly, this new method is much less invasive than the intracardiac conductance catheter because it does not require placement of a catheter in the ventricular cavity; this should greatly expand the applicability in patients. We tested the transcardiac conductance method by comparing its signals with those measured simultaneously with a conventional conductance catheter placed in the left ventricle. Experiments were performed with six anesthetized sheep in which pharmacologic and hemodynamic interventions were performed to vary ventricular volumes over a wide range.

http://ovidsp.tx.ovid.com/sp-3.3.1a/ovidweb.cgi

2011-03-10

Ovid: Evaluation of a new transcardiac conductance method for continuous on-line m...

Page 2 of 9

MATERIALS AND METHODS

Animals. The study was approved by the animal research committee of the University of Leiden. Six sheep (19-32 kg, age 3-6 months) were sedated, intubated, and ventilated with oxygen and room air. The animals were anesthetized with ketamine (4-10 mg/kg/hr iv), xylazine (1 mg/kg im), and atropine (0.5 mg/kg sc). Atracurium (0.25 mg/kg iv) was used as a muscle relaxant. PaO2, PaCO2, and pH were checked every 30 mins and kept within normal ranges by adjusting respiration.

Conductance Catheter. The conductance catheter technique was described in detail previously (11). Briefly, a catheter with ten electrodes is positioned along the longitudinal axis of the left ventricle. The electrode distance is chosen such that with electrode 1 within the apex, electrode 9 is situated just above the aortic valve. Through the two most proximal and the two most distal electrodes, two 20 kHz currents (current ratio 1:0.25) opposite in polarity are applied, creating a dual electric field in the ventricular cavity (14, 17). The six interposed electrodes are used to measure the conductances of five intraventricular segments. Total left ventricular conductance, G(t), is calculated as the sum of these five segmental conductances. Time-varying total left ventricular volume is calculated as V(t) = (1/[alpha]) · (rho · L2) · (G(t) - Gp), where [alpha] is the slope factor, rho is the specific resistivity of the blood measured from a blood sample using a special cuvette, L is the electrode spacing, and Gp is the parallel conductance.

Transcardiac Conductance. Changes of ventricular volume can be registered by conductance measurements on the chest provided that the applied electric field mainly encompasses the left ventricle. This hypothesis is based on the fact that the equipotential planes (at right angles to the field lines) extend through the ventricular wall and its surrounding tissues (18). Theoretically, it should make no difference whether a voltage is measured within the cardiac cavity or on the outside of the heart or chest wall as long as the point of measurement lies on the same equipotential plane. To test this hypothesis, we set up an electric field as close as possible to the left ventricle without manipulating in the ventricle itself using one electrode on a catheter placed in the vena cava superior situated just above the right atrium and an epithoracic electrode placed ~1 cm lateral of the xyphisternal joint. We placed six pick-up electrodes in line on the left side of the thorax to register five segmental conductances from which, analogous to the conductance catheter technique, total left ventricular volume is calculated (Fig. 1).

Figure 1. Overview of the instrumentation. The distal electrode on a pacing catheter in the superior vena cava is used in combination with a subcutaneous needle close to the xyphisternal joint to set up the electric field for the transcardiac conductance measurements (in most studies the needle was replaced by an ECG electrode). Six epithoracic electrodes on the left side of the thorax are used to pick up the resulting voltages. A conductance catheter and a pressure catheter are placed in the left ventricle. To inject hypertonic saline for determination of parallel conductances, a catheter is placed in the pulmonary artery. A balloon occlusion catheter is positioned in the inferior vena cava for temporary preload reduction.

Parallel Conductance. An important factor for both methods is that the electric field is not restricted to the ventricular blood volume, but rather current also passes through the ventricular wall, other cardiac chambers, and in fact to some extent through all electrically conductive structures surrounding the heart. As a consequence, the total conductance signal is the sum of two factors: the conductance of the blood in the left ventricle, and the "parallel" conductance of the surrounding structures. For the conductance catheter, the parallel conductance constitutes ~50% of the total signal; however, we anticipate that for transcardiac conductance it forms a substantially larger fraction. We assumed that parallel conductance (Gp) is a constant factor. Baan et al. (11, 19) devised the so-called saline dilution method to determine this factor. With this method, a small bolus of hypertonic saline (typically 2-3 mL of 10% saline) is injected into the pulmonary artery, which transiently changes the conductivity of the blood without affecting parallel conductance. By analyzing the conductance signal acquired during passage of this bolus through the left ventricle, parallel conductance can be determined. Note that parallel conductance values for the conductance catheter and for transcardiac conductance are not equal, because the electric fields have a different geometry. However, both factors can be determined from a single injection by analyzing the appropriate signals.

Slope Factor [alpha]. After correction for parallel conductance, the volume signal derived by the conductance catheter is directly proportional to actual ventricular volume but generally underestimates true volume by a fixed percentage. To correct the underestimation, the slope factor [alpha] is introduced. In practice, [alpha] is determined by comparing the conductance-derived volume (or stroke volume) with an independent measurement such as angiography or thermodilution. In animals such as dogs or sheep, [alpha] is typically 0.8 (11, 14). However, it may vary substantially between individual subjects. In this study we used the conductance catheter as the criterion and did not assess [alpha] (i.e., it was assumed to be 1).

Instrumentation. We obtained an electrocardiogram using four subcutaneous needle electrodes. Sheaths were placed in left and right carotid arteries and jugular veins for the introduction of catheters. A standard 5-Fr bipolar pacing catheter was placed in the vena cava superior with the distal electrode just above the right atrium to serve as current electrode for the transcardiac conductance method. A standard pediatric ECG electrode placed ~1 cm lateral of the xyphisternal joint served as the second current electrode (in the first experiment, a short subcutaneous needle was used). Six additional ECG electrodes were positioned in-line on the left side of the thorax between the fifth and the eighth rib to serve as voltage-sensing electrodes. To ensure good electrical contact, the skin was shaven. The positions of the electrodes were optimized by inspection of the segmental volume signals. A 5-Fr Berman catheter (Arrow International, Reading, PA) was placed in the pulmonary artery via the left jugular vein to inject hypertonic saline for determination of parallel conductances. A 5-Fr pressure catheter (Millar Instruments, Houston, TX) was inserted in the left carotid artery and advanced into the left ventricle. A dual-field conductance catheter (Webster Laboratories, Baldwin Park, CA) was inserted in the right carotid artery and its tip advanced to the apex of the left ventricle. We performed conductance and blood resistivity measurements by using two Leycom Sigma-5 DF signal-processors (CardioDynamics, Zoetermeer, The Netherlands). To eliminate interference, the processor used for the conductance catheter measurements used an excitation frequency of 21 kHz, whereas the processor for the transcardiac conductance measurements was modified to operate at 16 kHz. A 7-Fr balloon catheter (Cordis, Miami, FL) was placed in the vena cava inferior just above the diaphragm via the right femoral vein to perform vena caval occlusions. All catheters were placed under fluoroscopic guidance. An overview of the instrumentation is shown in Fig. 1.

http://ovidsp.tx.ovid.com/sp-3.3.1a/ovidweb.cgi

2011-03-10

Ovid: Evaluation of a new transcardiac conductance method for continuous on-line m...

Page 3 of 9

Protocol. Measurements were performed in four conditions: control, enhanced contractile state (5 µg/kg/min of dobutamine), repeat control, and reduced contractile state (1 mg/kg of propranolol). In each condition, blood resistivity (rho) was determined, at least two consecutive saline injections were performed to determine parallel conductances, and simultaneous measurements of ECG, left ventricular pressure, volume by transcardiac conductance, and volume by conductance catheter were acquired during a gradual caval occlusion. To assess reproducibility, we repeated the caval occlusion run. Because respiration may affect parallel conductance and actual end-diastolic volume, all data were acquired during apnea at end-expiration.

Data Collection and Statistical Analysis. Recordings of ECG, left ventricular pressure, the five segmental volume signals by transcardiac conductance, and the five segmental volume signals by conductance catheter were digitized at 12-bit accuracy and a sample frequency of 200 Hz and stored on hard disk for later analysis. Total volume for both methods was calculated off-line by summation of the segmental signals and subtraction of parallel conductance. We analyzed data acquired during caval occlusions as follows: First, end-diastole was defined using the time of the peak of the QRS complexes, and end-systole was defined as the time at which the upper left corner of the pressure-volume loop is reached by calculating the maximal ratio of pressure and volume as described previously (20, 21). Next, we compared the data obtained by transcardiac conductance (VTCC) and by conductance catheter (VCC) by performing linear regressions for the end-systolic and the end-diastolic volumes (VTCC = A + B·VCC). To assess the influence of interanimal variability, of condition (control, dobutamine, repeat control, propranolol), and of cardiac phase (end-systole or end-diastole) on the slope (B), the intercept (A), and the correlation coefficient (r2) of the relationship between transcardiac conductance and conductance catheter volumes, we used a multiple linear regression implementation of a repeated measures analysis of variance with dummy variables (22). The specific regression model was y = bO + [SIGMA] biS·Si + [SIGMA] biC·Ci + bP·P, where y is the dependent variable (intercept A, slope B, or correlation coefficient r2). The n - 1 dummy variables Si account for between-animal differences allowing the n animals to have a different mean value (effects coding). The Ci dummy variables code the inotropic condition: (C1,C2,C3) = (-1,-1,-1) represents control, (1,0,0) is dobutamine, (0,1,0) is repeat control, and (0,0,1) is propranolol. The dummy variable P codes the cardiac phase: P = 1 for end-systole, P = -1 for end-diastole. Consequently, using this coding, the overall mean value of the dependent is bO, the mean value for the end-systolic volumes is bO + bP and for end-diastole bO - bP, and so forth. As an example, the predicted end-systolic value with dobutamine is given by bO + b2C + bP. The interanimal variabilities and the condition variabilities are estimated as the SD of the corresponding set of coefficients. F tests were performed for the model as a whole, for the set of animal variables, and for the set of condition variables. If the F test for the set of conditions revealed significance, we used Dunnett's test for multiple comparisons against control to assess the effects of the various conditions. Parallel conductances for conductance catheter and transcardiac conductance were determined by the hypertonic saline method in each animal, at each condition. To assess the mean value, interanimal variability, and dependence on condition of the parallel conductance correction volumes, we used the same multiple linear regression model as mentioned previously excluding the cardiac phase term P. We performed data acquisition and analysis by using Conduct-PC (CardioDynamics, Zoetermeer, The Netherlands) and custommade software. Statistical significance was defined as p < .05.

RESULTS

A typical example of steady-state transcardiac conductance and conductance catheter signals is shown in Fig. 2. Figure 3 shows the signals after injection of hypertonic saline into the pulmonary artery for the assessment of parallel conductances. Note that parallel conductance for transcardiac conductance is substantially larger than for the conductance catheter: In this study the mean values were 234.7 ± 60.5 mL and 58.5 ± 17.6 mL, respectively. Figure 4 displays an example of transcardiac conductance, conductance catheter, and left ventricular pressure signals during a caval occlusion. For this intervention, the end-systolic and end-diastolic volumes were determined and the relations between the two methods were quantified by linear fits (Fig. 5). This analysis was performed in each condition. The slopes, intercepts, and correlation coefficients of the linear fits were subjected to a multiple linear regression analysis to determine the interanimal variability, the effects of cardiac phase, and the effects of the various conditions on these variables. The results are given in Table 1. The mean intercept (A) of the relation (VTCC = A + B·VCC) between the two conductance methods was -1.82 ± 1.24 mL. The F test for the animal coefficients indicates that the intercept varied significantly between animals: The interanimal variability, estimated as the SD of the animal coefficients, was 9.23 mL. Cardiac phase also significantly affected the intercept: The value of bP = -5.04 indicates that the mean intercept for the end-systolic points was -6.86 mL (= bO + bP), whereas it was 3.22 mL (= bO - bP) for the end-diastolic points. The various conditions did not significantly affect the intercept. The mean slope (B) of the relation between the two methods was 0.787 ± 0.024. The differences between animals were significant: The interanimal variability of the slope was 0.317.

The mean slope was also significantly different for end-systolic and end-diastolic points: 0.956 and 0.618, respectively. Like the intercept, the slope was not significantly affected by any of the conditions. The mean correlation coefficient was .94, whereas the F test for the model indicated that none of the variables significantly affected the correlation coefficient. Figure 6 graphically depicts the intercept, slope, and correlation coefficient for the end-systolic and end-diastolic relation in the various conditions.

Figure 2. Simultaneous volume signals in a sheep by transcardiac conductance (VTCC) and conductance catheter (VCC) during three cardiac cycles (both volume signals have been corrected for parallel conductance). Note the similarity in shape and phase of the waveforms and the fact that, consistent with our general findings, absolute VTCC underestimates VCC, especially for end-diastole. As a reference, the bottom panel shows left ventricular pressure.

http://ovidsp.tx.ovid.com/sp-3.3.1a/ovidweb.cgi

2011-03-10

Ovid: Evaluation of a new transcardiac conductance method for continuous on-line m...

Page 4 of 9

Figure 3. Transcardiac conductance (VTCC) and conductance catheter (VCC) signals during injection of 3 mL of hypertonic saline into the pulmonary artery to determine parallel conductances. The apparent increase in volume is caused by the passage of blood with increased conductivity through the left ventricle. Parallel conductance is determined by plotting end-diastolic volume vs. end-systolic volume for the beats in the "saline wash-in" phase and by calculating the intersection of the regression line through these points with the line-of-identity. In this case parallel conductance was 264 mL for VTCC and 44 mL for VCC. PLV, left ventricular pressure.

Figure 4. To obtain a wide range in volumes, preload was gradually reduced by balloon occlusion of the inferior vena cava. The figure shows volume by transcardiac conductance (VTCC), volume by conductance catheter (VCC), and left ventricular pressure (PLV) during occlusion and release of the caval balloon. Note the extra systolic beat at the start of the occlusion.

Figure 5. To quantify the relationship between volumes derived simultaneously by transcardiac conductance (VTCC) and by conductance catheter (VCC), the end-systolic (ESV) and end-diastolic volumes (EDV) were determined and fitted separately with a linear regression line. Data in this figure were obtained from the caval occlusion shown in Fig. 4.

http://ovidsp.tx.ovid.com/sp-3.3.1a/ovidweb.cgi

2011-03-10

Ovid: Evaluation of a new transcardiac conductance method for continuous on-line m...

Page 5 of 9

Table 1. Statistical analysis of relationship between volumes derived by transcardiac conductance and conductance catheter: VTCC = A + B·VCC

Figure 6. A multiple linear regression model was used to assess the influence of interanimal variability, cardiac phase, and inotropic condition on the intercept (A), slope (B), and correlation coefficient (r2) of the relation between transcardiac conductance and conductance catheter (VTCC = A + B·VCC). The figure graphically displays the results.

CON, control; DOB, dobutamine; REP, repeat control; PROP, propranolol. For intercept and slope the differences between end-systole (ESV, open bars) and end-diastole (EDV, crosshatched bars) were significant, but the differences between the conditions were not. For the correlation coefficient, no significant differences were found.

Parallel conductance for both methods was determined in each animal, at each condition. The statistical analysis is given in Table 2. Consistent with previous studies, the parallel conductance for the conductance catheter varied significantly between animals. The mean value of 58.5 mL showed an interanimal variability of 17.3 mL; the variability between conditions was also significant but substantially smaller in magnitude, 3.1 mL. For transcardiac conductance, the mean value was 234.7 mL, with a significant interanimal variability of 59.3 mL and a significant condition variability of 11.7 mL.

Table 2. Statistical analysis of parallel conductance correction volume for transcardiac conductance (VC-TCC) and conductance catheter (VC-CC)

DISCUSSION

Transcardiac Conductance Versus Conductance Catheter Volumes. The results in the present study indicate that there is an excellent linear correlation between transcardiac conductance and intracardiac conductance. However, the quantitative relationship between the two methods varied substantially between animals. Comparison of conductance catheter derived volume with various independent techniques also has shown substantial interindividual variabilities (11, 14, 23). This variation was thought to be mainly caused by a mismatch between catheter and ventricular long axis (19). In addition, nonuniformity of the intracavitary electric field may cause underestimation of true volume and is likely to be more pronounced in larger hearts (17). Similar effects will be present for the transcardiac conductance measurements: There were obviously anatomical variations, and the positions of both the current and the pick-up electrodes varied somewhat between animals. The second main source of variability was cardiac phase. The relationships for systolic volumes were significantly different from the diastolic relationships. The offset of the relationship between the two methods depends mainly on the estimation of parallel conductance (for both methods), which was obtained with the saline dilution method. This technique was shown to be reliable for intracardiac conductance (11, 12, 23), but it has not been tested previously for transcardiac conductance; because parallel conductance typically represented ~90% of the

http://ovidsp.tx.ovid.com/sp-3.3.1a/ovidweb.cgi

2011-03-10

Ovid: Evaluation of a new transcardiac conductance method for continuous on-line m...

Page 6 of 9

total transcardiac signal, even a relatively small error may produce large errors in the absolute volume. In addition, the conventional saline dilution method gives a single value for parallel conductance independent of cardiac phase. Parallel conductance for the conductance catheter has been shown to be fairly constant during the cardiac cycle (24-26), but this is not necessarily the case with transcardiac conductance. Both the movement of the heart during the cardiac cycle and, for example, the filling of the right ventricle (which constitutes part of the parallel conductance) may cause such variations. The influence of the former effect may be difficult to predict, but the latter would make actual parallel conductance larger for end-diastole than for end-systole. Because we used the same parallel conductance value for end-diastole and end-systole, this may partly explain why the intercept for end-diastolic relations was significantly larger than for end-systolic relations. The statistical analysis indicates that the parallel conductance for the transcardiac conductance method varies significantly between animals. The interanimal variability amounted to 25% of the overall mean value, thus clearly necessitating assessment of this factor in individual subjects. The variability between conditions was substantially less (5% of the mean) but was also statistically significant. Therefore, repeated assessments of parallel conductance within a subject may be required. Consistent with previous results (12, 27), very similar findings were obtained for the conductance catheter.

The slope of the relationship between transcardiac conductance and conductance catheter was also dependent on the cardiac phase: The slope for the end-systolic volumes was almost 1 (0.96), whereas for the end-diastolic volumes it was only 0.62. Apparently, changes in end-systolic volume are picked up with a higher sensitivity than end-diastolic volume. This is likely related to a more favorable position of the heart during systole with respect to the applied electric field and the pick-up electrodes. It easily can be shown mathematically that, as a direct result of the slope for end-diastolic volumes being less than that for end-systolic volume, the slope of the relationship comparing stroke volumes would always be less than that for end-diastolic volume. At a 50% ejection fraction, the slope for stroke volumes would be only about half the slope for the end-diastolic volumes. This explains the common finding that changes in stroke volume during vena cava occlusions were clearly less pronounced with transcardiac conductance than with the conductance catheter (Fig. 4).

Finally, we found that inotropic stimulation (both positive and negative) did not affect the relationship between the two methods. This is an important finding, because it indicates that within a specific subject, end-systolic and end-diastolic volumes can be measured and followed reliably even if ventricular volume varies substantially because of changes in loading or contractile state.

Calibration of Transcardiac Conductance in the Clinical Setting. Basically, we showed that for both end-systolic volume and end-diastolic volume, there is an excellent linear relationship between transcardiac conductance and the conductance catheter, a method that has been validated extensively (10-19). Because these relationships were not dependent on changes in contractile state, the signals (even without any further calibration) can be used to follow changes in end-systolic and end-diastolic volume and as such could be clinically useful. One step further is to calibrate the signals and interpret them in absolute or relative terms. Calibration requires some independent technique to measure volumetric variables, which in the intensive care unit is generally limited to thermodilution or echocardiography. Because the relationships for endsystolic and end-diastolic volume are different, they should be calibrated separately. In addition, this finding also implies that the ratio between transcardiac conductancederived stroke volume and, for example, thermodilution-derived stroke volume cannot be used to estimate the slope of the relationship between transcardiac conductancederived absolute volume and true volume. To calibrate, one needs at least two data points (conditions). Theoretically, one option would be to determine volume by echocardiography at baseline and, for example, after a volume load and compare these estimates with the simultaneous transcardiac conductance-derived volumes. However, this is obviously not a very practical approach. Fortunately, the hypertonic saline method can be used because parallel conductance basically represents the conductance at (hypothetical) zero true volume. Furthermore, because this is such a specific "data-point" (zero true volume), after correction for parallel conductance the transcardiac conductance signal is directly proportional to true volume (VTRUE = F·VTCCVc-corrected). In most clinical situations it will be sufficient to monitor only the relative (percentage) changes in end-systolic and end-diastolic volume, in which case there is no need to assess the factor F. In those cases where absolute volume measurements are required, the factor F can be determined by comparing the transcardiac conductance volume with echocardiographic volume (F = VECHO/VTCCVc-corrected). Because the slope factor has been shown to be unaffected by changes in contractile state, a single steady-state measurement would be sufficient.

Factors that are likely to determine the actual values of the calibration factors include position of the electrodes and the patient's body size and fat content. Those factors are unlikely to cause major within-patient variability during the limited observation period (several days at most) and mainly add to the between-patient variability.

Other factors, like edema and pleural effusions, may potentially change the calibration factors within a specific patient during the monitoring period. However, in principle, multiple recalibrations can be done in the intensive care unit to deal with those problems. Clearly, future work is needed to elucidate these points.

Differences With Thoracic Impedance. Noninvasive conductance methods, typically using band-electrodes around the neck and the lower thorax, have been used extensively to measure a time-varying quantity of blood in the body; these methods have been used mainly as an empirical technique to measure stroke volume. To derive an analytic relationship between thoracic impedance and cardiac volume is virtually impossible because of the inhomogeneity and time-varying blood contents of the thorax. It has been shown that the impedance cardiogram originates mainly from systemic arterial blood volume variations and from the changes in orientation of the erythrocytes in these vessels and in the caval and pulmonary veins (7, 9, 28). This may explain why, despite extensive work, the accuracy of the technique remains controversial (5, 6). Some attempts have been made to change the electrode configuration such that the recordings directly represent ventricular volume changes. Geddes (29) and Patterson (30) reported on the use of an esophageal pick-up electrode coupled with external band electrodes to measure ventricular stroke volume; however, those authors only showed typical impedance waveforms and presented no quantitative comparative data. The approach in the present article differs from conventional thoracic impedance measurements in several aspects. First, we apply the electric field such that it is mainly confined to the heart, thereby greatly reducing the influence of volume changes and blood flow in the large vessels. In addition, in this setup we found the phasic component of the signal to be typically 5% to 10% of the total signal, whereas in thoracic impedance measurement the change in the signal during the cardiac cycle is usually only ~1%. Second, the voltages are picked up using a linear array of electrodes covering the projection of the heart on the left side of the thorax. The electrodes produce five segmental conductance signals, and the total conductance signal is calculated as the sum of the segmental conductances. Because we used closely spaced electrodes, the intracavitary electric field in the corresponding segment can be considered reasonably uniform, which enables the use of a simple physical model to relate conductance to volume (i.e., the same as for the conductance catheter). Moreover, because we use the summation of the segmental signals, the total conductance is much less dependent on the exact placement of the electrodes: For example, if the heart moves as a whole without changing its volume, the individual segmental signals may change, but these changes will, at least theoretically, cancel out in the summation if the array "covers" the whole heart. Finally, whereas thoracic impedance techniques generally aim to estimate stroke volume, the transcardiac conductance method provides a means to measure absolute end-systolic and end-diastolic volume. Clearly, the approach in this study is more invasive than thoracic impedance measurements, but a central venous catheter will be acceptable in many clinical situations, and complications or technical difficulties during insertion of the catheter are unlikely. Although in this study we used angiography for positioning, in clinical practice right atrial electrocardiography may help in accurate placement (31).

Differences With Thermodilution. Thermodilution is the most widely used technique for clinical bedside hemodynamic monitoring, and because we believe transcardiac conductance potentially could be used in the same field, the following is intended to compare some aspects of both techniques. With regard to accuracy, transcardiac conductance obviously requires more extensive validation especially in a clinical situation; however, despite widespread use there is still controversy concerning accuracy in determining cardiac output by thermodilution (32). Patient safety and discomfort are comparable for both techniques. An important advantage of thermodilution is that no calibration in individual subjects is required, whereas transcardiac conductance will give only relative changes in volume unless an initial calibration is performed. The transcardiac conductance technique, on the other hand, is an online technique and gives beat-to-beat information, whereas thermodilution requires repeated injections in steady-state conditions to obtain a time-averaged cardiac output. The recent development of continuous thermodilution cardiac output (33) will reduce operator interaction, but the delay of ~10 mins between a change in cardiac output and the displayed value remains. However, the most important potential advantage of transcardiac conductance is its ability to monitor absolute volume (changes) in contrast to mere stroke volume as obtained by thermodilution. Absolute left ventricular volume, especially end-systolic volume, has been shown to provide important additional diagnostic and prognostic information, and continuous monitoring of such variables may help guide therapeutic interventions (1-4, 34).

http://ovidsp.tx.ovid.com/sp-3.3.1a/ovidweb.cgi

2011-03-10

Ovid: Evaluation of a new transcardiac conductance method for continuous on-line m...

Page 7 of 9

CONCLUSION

The results of this study indicate that with the new transcardiac conductance method, a clear cardiac volume signal can be registered similar in shape and phase with the signal registered by the conductance catheter. A quantitative analysis indicates an excellent linear correlation, but the relationship between the two methods varies significantly between animals and also depends on cardiac phase. Between-subject comparisons of end-systolic or end-diastolic volumes, therefore, require initial calibration with an independent technique. However, within a subject, end-systolic and end-diastolic volumes can be followed accurately even over a large volume range such as induced by inotropic or loading interventions. The potential to measure on-line absolute volumes is likely to be valuable in hemodynamic monitoring and to add important diagnostic and prognostic information not provided by cardiac output monitoring.

REFERENCES

1. Braunwald E: Heart Disease: A Textbook of Cardiovascular Medicine. Fifth Edition. Philadelphia, WB Saunders, 1997 [Context Link]

2. Kostuk WJ, Kazamias TM, Gander MP, et al: Left ventricular size after acute myocardial infarction. Serial changes and their prognostic significance. Circulation 1973; 47: 1174-1179 ExternalResolverBasic Full Text via SFX Ovid Full Text Bibliographic Links [Context Link]

3. White HD, Norris RM, Brown MA, et al: Left ventricular end-systolic volume as the major determinant of survival after recovery from myocardial infarction. Circulation 1987; 76: 44-51 ExternalResolverBasic Full Text via SFX Ovid Full Text Bibliographic Links [Context Link]

4. Migrino RQ, Young JB, Ellis SG, et al: End-systolic volume index at 90 to 180 minutes into reperfusion therapy for acute myocardial infarction is a strong predictor of early and late mortality. The Global Utilization of Streptokinase and t-PA for Occluded Coronary Arteries (GUSTO)-I Angiographic Investigators. Circulation 1997; 96:116-121 [Context Link]

5. Jensen L, Yakimets J, Teo KK: A review of impedance cardiography. Heart Lung 1995; 24:183-193 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

6. Fuller HD: The validity of cardiac output measurement by thoracic impedance: A meta-analysis. Clin Invest Med 1992; 15: 103-112 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

7. Visser KR, Lamberts R, Zijlstra WG: Investigation of the origin of the impedance cardiogram by means of exchange transfusion with stroma free haemoglobin solution in the dog. Cardiovasc Res 1990; 24:24-32 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

8. Bonjer F, Van den Berg J, Dirken M: The origin of the variations of body impedance occurring during the cardiac cycle. Circulation 1952; 6:415-420 ExternalResolverBasic Full Text via SFX Ovid Full Text Bibliographic Links [Context Link]

9. Patterson RP, Kubicek WG, Witsoe DA, et al: Studies on the effect of controlled volume change on the thoracic electrical impedance. Med Biol Eng Comput 1978; 16:531536 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

10. Baan J, Aouw Jong TT, Kerkhof PLM, et al: Continuous stroke volume and cardiac output from intra-ventricular dimensions obtained with impedance catheter. Cardiovasc Res 1981; 15:328-334 [Context Link]

11. Baan J, Van der Velde ET, De Bruin H, et al: Continuous measurement of left ventricular volume in animals and humans by conductance catheter. Circulation 1984; 70: 812-823 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

12. Burkhoff D, Van Der Velde E, Kass D, et al: Accuracy of volume measurement by conductance catheter in isolated, ejecting canine hearts. Circulation 1985; 72:440-447 [Context Link]

13. Pattynama PMT, De Roos A, Van Der Velde ET, et al: Magnetic resonance imaging analysis of left ventricular pressure-volume relations: Validation with the conductance method at rest and during dobutamine stress. Magn Reson Med 1995; 34:728-737 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

14. Steendijk P, Van Der Velde ET, Baan J: Left ventricular stroke volume by single and dual excitation of conductance catheter in dogs. Am J Physiol 1993; 264(Heart Circ Physiol 33):H2198-H2207 [Context Link]

15. Applegate RJ, Cheng CP, Little WC: Simultaneous conductance catheter and dimension assessment of left ventricle volume in the intact animal. Circulation 1990; 81:638648 ExternalResolverBasic Full Text via SFX Ovid Full Text Bibliographic Links [Context Link]

16. Cassidy SC, Teitel DF: The conductance volume catheter technique for measurement of left ventricular volume in young piglets. Pediatr Res 1992; 31:85-90 ExternalResolverBasic Full Text via SFX Ovid Full Text Request Permissions Bibliographic Links [Context Link]

17. Steendijk P, Van Der Velde ET, Baan J: Single and dual excitation of the conductance-volume catheter analysed in a spheroidal mathematical model of the canine left ventricle. Eur Heart J 1992; 13(Suppl E):28-34 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

http://ovidsp.tx.ovid.com/sp-3.3.1a/ovidweb.cgi

2011-03-10

Ovid: Evaluation of a new transcardiac conductance method for continuous on-line m...

Page 8 of 9

18. Mur G, Baan J: Computation of the input impedances of a catheter for cardiac volumetry. IEEE Trans Biomed Eng 1984; 31:448-453 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

19. Baan J, Van der Velde ET, Steendijk P, et al: Calibration and application of the conductance catheter for ventricular volume measurement. Automedica 1989; 11:357-365 ExternalResolverBasic [Context Link]

20. Kono A, Maughan WL, Sunagawa K, et al: The use of left ventricular end-ejection pressure and peak pressure in the estimation of the end-systolic pressure-volume relationship. Circulation 1984; 70:1057-1065 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

21. Van der Velde ET, Burkhoff D, Steendijk P, et al: Nonlinearity and load sensitivity of the end-systolic pressure-volume relation of canine left ventricle in vivo. Circulation 1991; 83:315-327 ExternalResolverBasic Full Text via SFX Ovid Full Text Bibliographic Links [Context Link]

22. Glantz SA, Slinker BK: Primer of Applied Regression and Analysis of Variance. New York, McGraw Hill, 1990 [Context Link]

23. Kass DA, Midei M, Graves W, et al: Use of a conductance (volume) catheter and transient inferior vena caval occlusion for rapid determination of pressure-volume relationships in man. Cathet Cardiovasc Diagn 1988; 15:192-202 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

24. Lankford ED, Kass DA, Maughan WL, et al: Does volume catheter parallel conductance vary during a cardiac cycle? Am J Physiol 1990; 258(Heart Circ Physiol 27): H1933H1942 [Context Link]

25. White PA, Chaturvedi RR, Shore D, et al: Left ventricular parallel conductance during cardiac cycle in children with congenital heart disease. Am J Physiol 1997; 273(1 Pt 2):H295-H302 [Context Link]

26. Szwarc RS, Mickleborough LL, Mizuno S, et al: Conductance catheter measurements of left ventricular volume in the intact dog: Parallel conductance is independent of left ventricular size. Cardiovasc Res 1994; 28:252-258 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

27. Klautz RJM, Teitel DF, Steendijk P, et al: Interaction between afterload and contractility in the newborn heart: Evidence of homeometric autoregulation in the intact circulation. J Am Coll Cardiol 1995; 25:1428-1435 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

28. Geddes LE, Baker LE: Thoracic impedance changes following saline injection into right and left ventricles. J Appl Physiol 1972; 33:278-281 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

29. Geddes LA: Intraoesophageal monitoring of the electrocardiogram and respiration in anaesthetised animals. Med Biol Eng Comput 1984; 22:90-91 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

30. Patterson RP: Possible technique to measure ventricular volume using electrical impedance measurements with an oesophageal electrode. Med Biol Eng Comput 1987; 25:677-679 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

31. McGee WT, Ackerman BL, Rouben LR, et al: Accurate placement of central venous catheters: A prospective, randomized, multicenter trial. Crit Care Med 1993; 21:11181123 ExternalResolverBasic Full Text via SFX Ovid Full Text Request Permissions Bibliographic Links [Context Link]

32. Sherman MS, Kosinski R, Paz HL, et al: Measuring cardiac output in critically Ill patients: Disagreement between thermodilution-, calculated-, expired gas-, and oxygen consumption-based methods. Cardiology 1997; 88:19-25 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

33. Edwards JD: Continuous thermodilution cardiac output: A significant step forward in hemodynamic monitoring. Crit Care Med 1997; 25:381-382 ExternalResolverBasic Full Text via SFX Ovid Full Text Request Permissions Bibliographic Links [Context Link]

34. Kass DA, Grayson R, Marino P: Pressure-volume analysis as a method for quantifying simultaneous drug (amrinone) effects on arterial load and contractile state in vivo. J Am Coll Cardiol 1990; 16:726-732 ExternalResolverBasic Full Text via SFX Bibliographic Links [Context Link]

Key Words: cardiac function; cardiac output; conductance catheter; contractile state; end-systolic volume; end-diastolic volume; hemodynamics; monitoring; on-line measurements; sheep; thoracic impedance

IMAGE GALLERY

Select All

http://ovidsp.tx.ovid.com/sp-3.3.1a/ovidweb.cgi

Export Selected to PowerPoint

2011-03-10

Ovid: Evaluation of a new transcardiac conductance method for continuous on-line m...

c Figure 1 d e f g

Page 9 of 9

c Figure 3 d e f g

c Figure 2 d e f g

c Table 1 d e f g

c Figure 4 d e f g

c Figure 5 d e f g Â

c Table 2 d e f g

c Figure 6 d e f g Back to Top Copyright (c) 2000-2010 Ovid Technologies, Inc.

Terms of Use

Support & Training

About Us

Contact Us

Version: OvidSP_UI03.03.01.103, SourceID 53860

http://ovidsp.tx.ovid.com/sp-3.3.1a/ovidweb.cgi

2011-03-10