Beat-to-Beat Analysis of Left Ventricular Pressure-Volume Relation and Stroke Volume by Conductance Catheter and Aortic Modelflow in Cardiomyoplasty Patients
Background Since the clinical introduction of dynamic cardiomyoplasty, a discrepancy has been observed between unchanged measurements of cardiac function and improved clinical outcome.
Methods and Results We performed a beat-to-beat analysis of cardiac performance at rest in nine cardiomyoplasty patients 6 to 24 months after operation. Conductance and micromanometer catheters were placed in left ventricle and aorta and used for measurements over a 15-second period, during which the wrapped latissimus dorsi (LD) muscle was stimulated for 10 seconds in a 1:2 synchronization mode followed by a 5-second period without LD stimulation. The synchronization delay between start of the QRS complex and the LD contraction was changed from 4 up to 125 ms at the patient’s clinical stimulation strength and at an increased supramaximal amplitude. Comparing the LD assisted period to the unassisted period, at the clinical settings no significant changes in stroke volume (SV) as measured by the conductance technique and the aortic Modelflow technique were observed. A significant (P<.05) rise in left ventricular end-diastolic pressure (LVEDP) was observed directly after the assisted 10-second period. The peak ejection rate (PER) of left ventricular volume increased (P<.05), with a mean of 28±23% during the LD stimulated beats. At the patient’s individual best setting, SV of the stimulated beats increased (P<.01) by a mean of 20±15%. Systolic aortic pressure increased (P<.01) by a mean of 7 mm Hg, peak negative dP/dt increased (P<.01), and PER increased, with a mean of 68±24% (P<.01). LVEDP was similar in stimulated and unstimulated beats and increased (P<.05) in the nonpaced 5-second period. The delay for the best setting ranged from 25 to 125 ms; the stimulus strength was 1.5 to 3 V higher than the clinical setting. At the patient’s individual worst setting, SV remained unchanged and PER was higher, with a mean of 30±25% (P<.05). The worst setting was observed at the 1.5- to 3-V-higher stimulus strength; in six patients, it was at a short delay (4 to 25 ms) and in three patients, at the longest delay (100 to 125 ms).
Conclusions By the left ventricular conductance catheter and aortic Modelflow methods, improvement in cardiac function by dynamic cardiomyoplasty was demonstrated in this patient group. The synchronization interval, stimulus strength, and stimulus duration appeared to be critical for obtaining optimal improvement.
Since the clinical introduction of dynamic cardiomyoplasty by Carpentier and Chachques in 1985,1 several studies have reported a discrepancy between clinical outcome and hemodynamic measurements. Whereas clinical improvements according to the New York Heart Association (NYHA) classification2 3 4 were observed in a majority of patients, cardiac output improvement could be demonstrated in only a minority.3 4 In these studies, preoperative patient data were compared with postoperative data.
To elucidate whether dynamic cardiomyoplasty affects cardiac function, analysis of individual cardiac contractions with and without skeletal muscle stimulation should be performed. Some studies, which measured ventricular ejection fraction by echocardiography or by radionuclide angiography, demonstrated an increase in the ejection fraction due to skeletal muscle cardiac assistance3 4 ; another study reported no change.2
The absence of changes in cardiac output or ejection fraction in patients who clinically improved may have several causes.
First, the resolution of the techniques used may have been too small in relation to the magnitude of the hemodynamic changes induced by the wrapped latissimus dorsi (LD) muscle contraction.5
Second, inappropriate synchronization and stimulation settings of the cardiomyostimulator in relation to cardiac contraction may negatively affect cardiac function. For instance, a decrease in stroke volume (SV) was demonstrated in an animal study when LD muscle was stimulated during ventricular depolarization.6 In another animal study,7 increases in SV were demonstrated at an optimal individual synchronization delay range. A follow-up study of 78 cardiomyoplasty patients reported that fixed synchronization delays equally divided in a range from 4 to 125 ms were applied.8
Third, cardiovascular compensatory mechanisms may counteract possible positive effects of LD muscle stimulation when an appropriate output state in the patient has been attained. When the heart as a pump is not the limiting factor in providing adequate organ perfusion, a condition that occurs in most of these patients at rest, an increase in contractility will not result in an increase in cardiac output.
Last, a decrease or prevention of acute or chronic ventricular dilatation may be a major beneficial mechanism associated with dynamic cardiomyoplasty.9
To evaluate these possible mechanisms, we studied the acute, direct hemodynamic effects of skeletal muscle stimulation, on a beat-to-beat basis, in nine patients at rest several months after operation. The conductance catheter technique introduced by Baan et al10 was used to measure left ventricular volume instantaneously. The aortic Modelflow method according to Wesseling et al11 was used to measure left ventricular SV.
The hemodynamic effects of cardiomyoplasty were studied during periods of 10-second LD muscle stimulation performed at varied synchronization delays after ventricular activation followed by periods of 5 seconds without LD muscle stimulation. The measurements were carried out at two different LD stimulation strengths: at the previously chosen clinical amplitude and at an increased supramaximal amplitude to obtain complete muscle fiber recruitment.
Nine male patients, 53 to 69 years old, who had undergone dynamic cardiomyoplasty surgery were studied. All patients had dilated cardiomyopathy, related either to ischemia or to an idiopathic origin (Table 1⇓). At the time of the cardiomyoplasty procedure, all patients were in NYHA class III, with a mean radionuclide ejection fraction of 20±8%.
Of the nine patients studied, six had been operated on in Lyon (France) and three in Brescia (Italy). The surgical technique has been described by Chachques et al.12 Briefly, the left LD muscle was mobilized, with the neurovascular pedicle kept intact. After two intramuscular stimulation electrodes (Medtronic SP 5528) were attached, the muscle was introduced into the thoracic cavity and wrapped clockwise around both ventricles. The stimulation electrodes and a sensing electrode attached to the right ventricular wall were connected to a cardiomyostimulator (SP1005, Medtronic). Additional surgery involved aneurysmectomy in two patients and coronary artery bypass graft surgery in three (Table 1⇑).
This study, which had been approved by the medical ethics committees of the respective hospitals, was performed 6 to 24 months after the cardiomyoplasty procedure. Informed patient consent was obtained for insertion of the catheters and for testing at different settings of the cardiomyostimulator.
Patients were sedated and heparinized before catheterization. A Swan-Ganz thermodilution catheter was placed in the pulmonary artery. A double micromanometer transducer catheter (7.5F, Sentron) was inserted via the femoral artery into the left ventricle for the measurement of aortic and left ventricular pressures.
A 12-electrode (dual-field) conductance catheter (7.5F, Webster Labs or 7F, Sentron) was inserted via the other femoral artery into the left ventricle. The correct positioning of the conductance catheter was verified by radiography and by inspection of the segmental conductance signals. In two patients (patients 1 and 2), a 2F micromanometer catheter (Millar Instruments, Inc) was inserted inside the ventricular lumen of the conductance catheter, while aortic pressure was measured by a separate micromanometer catheter (5F, Sentron).
The conductance catheter method uses a Leycom Sigma-5DF signal conditioner–processor (CardioDynamics) to estimate left ventricular volume.10 13 The method is based on measuring the time-varying electrical conductances of five segments of blood in the left ventricle. The five segmental conductances are measured from six adjacent electrodes, and total left ventricular volume is calculated from the sum of the five segmental conductances. The dual-field modification, which has been shown to improve the accuracy of the method, was used in all patients.14 The conductance catheter not only continuously measures the amount of blood in the left ventricle but also the conductance of the myocardium and other surrounding tissues. This parallel conductance offset term was estimated by injection of 5 mL of hypertonic saline (8%) into the pulmonary artery10 without LD stimulation. In three patients, parallel conductance was also estimated during LD stimulation at the highest voltage. Effects of LD stimulation on regional contraction patterns could be studied from the five segmental conductance catheter segments.15
We did not attempt to estimate ventricular function by end-systolic elastance or preload recruitable stroke work in these patients, because this requires a change in loading conditions.16 This was not applied in this group of patients because preoperative conductance catheter and Modelflow measurements were not performed. However, the end-systolic pressure-volume (P-V) points as markers of ventricular performance can be derived from the P-V relations. Moreover, peak ejection rate (PER, negative dV/dtmax) was calculated from the ventricular volume signal as a measure of ventricular pump function analogous to the velocity of circumferential shortening.17
The aortic Modelflow method computes relative SV from aortic pressure using a nonlinear, time-varying three-element Windkessel model.11 The model elements include aortic characteristic impedance, arterial compliance, and systemic vascular resistance. The aortic characteristic impedance and arterial compliance depend on the aortic cross-sectional area, which can be estimated from mean arterial pressure, age, and sex.18 To adapt this estimate of cross-sectional area to that of the individual patient, Modelflow was calibrated by cardiac output measured by thermodilution.
To obtain reliable estimates of cardiac output by thermodilution, a computer-controlled injection system was used.19 20 Cardiac output was determined by use of a cardiac output computer (COM-2, Baxter) and injections of 10 mL ice-cold glucose 5%.
Data Acquisition and Analysis
ECG (extremity leads), aortic pressure, left ventricular pressure, and left ventricular volume signals were digitized at a sampling rate of 200 Hz on a 80386 personal computer. The dedicated data-acquisition and -analysis software package conduct-pc (CardioDynamics) was applied for conductance catheter–related data analysis. The continuous cardiac output software package Modelflow (TNO Biomedical Instrumentation, Amsterdam, the Netherlands) was applied for calculation of left ventriuclar stroke volume.
After insertion of the catheters, baseline measurements were performed with the LD stimulator off. Blood resistance and parallel conductance were measured before baseline measurements. Episodes of 15 seconds’ duration were recorded while the patients performed expiratory breath-holding to prevent respiration-related changes in venous return. The LD stimulator was synchronized in a 1:2 ratio during the first 10 seconds leading to one assisted beat for every two ventricular contractions, immediately followed by a nonpaced episode of 5 seconds after blocking of the LD stimulator.
The LD stimulation protocol started with the patient’s clinical setting, at a burst rate of 30 Hz, a burst duration of 185 ms, an individual stimulus strength between 3 and 8 V, and an individual delay synchronized to the QRS complex.
Stimulation was performed at delays of 4, 25, 50, 75, 100, and 125 ms. This series was then repeated at a 1.5- to 3-V-higher supramaximal stimulus strength. Whenever an obvious best setting was observed during the measurements, this stimulator setting was repeated at a stimulus duration of 240 ms. Comparisons were made between LD assisted beats and nonassisted beats of the 10-second period and beats of the nonpaced 5-second period.
Statistical analysis was done with Student’s test for paired variates; comparison between three data groups was also performed with randomized block ANOVA. Values are presented as mean±SD; significance was assumed at P<.05.
The mean duration of the catheterization procedure including the experimental study was 105±10 minutes. Occasionally, ectopic beats were observed during the introduction of the catheters and during the study protocol at higher stimulation amplitudes and the longest delays.
The study group originally consisted of 13 patients. In 2 patients, arrhythmic cardiac contractions prevented a proper analysis of the P-V loops; another patient had Valsalva maneuvers during the 15-second sample periods; and in 1 patient, the conductance catheter signals were not suitable for proper analysis. In all these cases, however, SV estimations by Modelflow revealed changes due to LD stimulation similar to those in the other 9 patients.
The mean duration of the breath-holding episodes was 15±2 seconds. A baseline cardiac index ranging from 1.4 up to 3.0 L/min was observed without cardiomyostimulation (Table 1⇑).
Fig 1⇓ shows representative recordings of the conductance catheter of patient 7. The stimulated beats revealed a decrease in end-systolic volume and an increase in SV compared with unstimulated beats, and concomitant higher systolic aortic pressures. Fig 2⇓ shows left ventricular segmental volume changes of patient 1, and the contribution of specific segments to the total increase in SV can be studied.
In three patients, the parallel conductance was similar during LD stimulation (234±28 mL) and without LD stimulation (230±27 mL).
Table 2⇓ presents the results of nine patients at their clinically used LD stimulator setting. Mean SV determined by the Modelflow method (SVmf) and the conductance catheter method (SVcon) of the stimulated beats both increased (P<.05), with a mean of 3% compared with the unstimulated beats in between. Mean SV of the nonpaced 5-second period was not significantly changed compared with the mean SV of the earlier 10-second 1:2 stimulated period. Systolic aortic pressure, dP/dtmax, and end-systolic volume remained unchanged during the LD stimulation period. Left ventricular end-diastolic pressure (LVEDP) was similar in the stimulated and unstimulated beats and increased (P<.05) in the subsequent nonpaced 5-second period. End-systolic volume, dP/dtmax, remained unchanged during LD stimulation. Peak negative dP/dt was higher (P<.05) in the stimulated beats compared with the alternating unstimulated beats (during the 10-second LD stimulated period) and unchanged compared with the nonpaced period. PER was higher, with a mean of 28±23% (P<.05) during the stimulated beats compared with the unstimulated beats of both periods. The clinically used delays were at 25 and 50 ms, and stimulus strengths ranged from 4 to 9 V.
Table 3⇓ presents the results of the nine patients at the individual best setting. The best and the worst settings were defined as largest and smallest increases in SV due to LD stimulation as measured by both the Modelflow and conductance methods at the higher applied stimulus strength.
At the patient’s best setting, SVmf and SVcon increased significantly (P<.01) during LD stimulation compared with the alternating unstimulated beats. The SV poststimulated period was significantly lower (P<.01), with a mean of 8% compared with the preceding 1:2 stimulated 10-second period. Systolic aortic pressure was higher (P<.01) during LD stimulation, with a mean of 7 mm Hg, whereas diastolic pressure was lower (P<.01), with a mean of 2 mm Hg. Left ventricular end-systolic volume decreased in eight patients during LD stimulation, while end-diastolic volume was unchanged in four patients, slightly increased in four patients (patients 3, 6, 7, and 8), and decreased in one patient (patient 1). LVEDP was similar in stimulated and unstimulated beats and increased significantly (P<.05) in the nonpaced 5-second period. Maximal dP/dt was unchanged between stimulated and unstimulated beats. Peak negative dP/dt was higher (P<.01 and P<.05) during the stimulated beats compared with the unstimulated beats during the stimulated and nonpaced periods, respectively. The PER of the stimulated beats was significantly higher (P<.01), with a mean of 68±24% compared with the unstimulated beats of the stimulated and nonpaced periods.
In all patients, the optimal stimulus strength was at the 1.5- to 3-V-higher voltage. In one patient (patient 2) the best setting occurred at the longer applied stimulus duration of 240 ms. Fig 3⇓ presents P-V loops of all nine patients at their best settings. In eight patients, no signs of abnormal left ventricular relaxation during LD stimulation could be observed. Also, middiastolic pressure was not significantly changed during the LD stimulated beats. In two patients (patients 2 and 8), an abnormal isovolemic contraction phase, suggesting mitral valve regurgitation, can be observed. The P-V loops during LD stimulation clearly demonstrated a leftward shift of the end-systolic P-V points in each patient, increased SVs and systolic LV pressure, and therefore stroke work at similar end-diastolic volumes.
At the patient’s worst setting (Table 4⇓), SVmf and SVcon were not significantly changed during LD stimulation. In four patients, SVmf and SVcon decreased during LD stimulation. End-systolic volume, dP/dtmax, and peak negative dP/dt remained unchanged during LD stimulation. Systolic aortic pressure was slightly higher (P<.05) during LD stimulation. LVEDP during the nonpaced period was higher (P<.05) compared with the stimulated and unstimulated beats of the stimulated period. The PER of the stimulated beats was higher (P<.05), with a mean of 30±25% compared with the unstimulated beats of both periods. In six patients, the worst setting was at a short delay and in three patients, at a relatively long delay.
The regression equation of the percentage change in the SVcon and SVmf methods was SVcon=1.34±0.96 SVmf, r=.96, n=27.
This study demonstrates the sensitivity of both the left ventricular conductance catheter technique and the aortic Modelflow method in quantifying the contribution of the wrapped LD muscle in cardiomyoplasty patients.
The feasibility of the conductance catheter technique has been demonstrated in patients during cardiac surgery.16 An excellent correlation between SVcon and SVmf changes was observed in the present study. The resolution of these two objective and independent methods surpasses the changes in SV observed at the best settings. According to Applegate et al,21 the conductance catheter accurately measures volume with high time resolution over single cardiac cycles, provided that the preload changes are within a physiological range. LD stimulation might have effects on the parallel conductance of left ventricular volume. However, in three patients, similar parallel conductance values with and without LD stimulation were observed. Moreover, the combined results of SVcon and SVmf reveal that effects of a change in parallel conductance on SV, if present, are small.
The three-element Windkessel model provides a reasonable representation of afterload to estimate SV when real impedance spectra are incorporated in the model.22 In a human study, this was also demonstrated by Modelflow,11 a nonlinear three-element Windkessel model with incorporated static elastic properties of the human aorta.18 A direct effect of LD muscle stimulation on arterial impedance is not likely because the LD muscle is wrapped around the heart. An “indirect” change of aortic geometry could result from a heart lifting effect of LD stimulation, which might possibly lead to a change in aortic input impedance. Such an effect should be present especially during stimulation at the early delays and should give different results for conductance and Modelflow SVs. However, such deviations were not present during the worst setting (Table 4⇑).
SV was measured during relatively short periods of about 15 seconds when assisted and unassisted cardiac contractions were compared. These short periods were selected to prevent effects of hemodynamic compensatory reflexes, which may appear 10 to 15 seconds after a significant hemodynamic change. At the clinically used setting, no significant change in SV was observed in comparing the assisted period to the unassisted period.
A mean increase in SV of 20% of the assisted beats compared with the nonassisted beats in between was observed at the patient’s individual best stimulator setting. Comparing SV of all the beats of the assisted period with the nonpaced period revealed a mean increase <10% (8%). Thus, in some patients, the increase in SV was partly at the cost of the unstimulated beats in between. Extrapolation of the increase in SV at this best setting to a longer period may result in an effective improvement of cardiac output of 8%, provided that compensatory reflexes do not occur.
The delay between onset of the QRS complex and contraction of the wrapped LD muscle appeared to be critical, with SV even decreasing at the individual worst delay settings (Table 4⇑). The optimal delay ranged from 25 to 125 ms in the nine patients studied, while the worst delay setting was early (4 and 25 ms) in six patients and relatively late (100 and 125 ms) in the other three. The optimal delay may be related to several factors. First, LD muscle contraction should be synchronized to the ejection phase to induce a larger SV. Second, because intraventricular conduction is disturbed in many patients with chronic heart failure, the proper onset of skeletal muscle contraction will also be dependent on the individual duration of intraventricular conduction. In the present study, three of the nine patients had bundle branch block and one had a DDD pacemaker. Third, the QRS synchronization delay should be triggered by the steep phase of the QR interval. In one patient (patient 6), the synchronization delay was triggered by the RS part of the ECG, revealing that a given delay is not always referring to a fixed reference point. In this case, the 75-ms delay was comparable to a delay of 135 ms in the other patients.
Last, the presence of aneurysms or dyskinetic areas can evoke paradoxical movements of the ventricular wall that may require different timing of LD muscle contraction. Segmental volume analysis of patient 1, with an unresected apical aneurysm, showed a contribution of the apical segments to the volume decrease during LD stimulation (Fig 2⇑), whereas the apex showed a paradoxical volume change during the ejection phase without stimulation. In this case, the paradoxical apical movement is presumably forced into phase with the other segments during LD stimulation.
Left Ventricular Pressure
In general, it can be expected that the diastolic phase of the left ventricle will be hampered by the wrapped LD muscle, even without LD muscle contraction.23 In the present study, no significant difference was observed between the LVEDP of the stimulated and alternating unstimulated beat at any setting used. At all settings, however, the LVEDP was significantly higher during the nonpaced 5-second period. A possible intrathoracic pressure-increasing effect due to a Valsalva-like maneuver at the end of the 15-second breath-hold period must be considered, but this should decrease venous return immediately.
At the best setting, peak negative dP/dt was even significantly higher, and diastolic pressure did not change during the LD stimulated beats, indicating the absence of a deleterious effect on diastolic function.
The increase in systolic aortic pressure is in agreement with the increased SV and higher PER during LD stimulation. In contrast, left ventricular dP/dtmax was not significantly changed in the best, worst, and clinical settings. An increase in dP/dtmax is not indicative of an increased SV during LD stimulation because, in most circumstances, the maximum pressure rise develops before the opening of the aortic valve.24 On the other hand, peak ejection rate during the systolic ejection phase increased significantly during LD stimulation in the clinical, worst, and best settings. As mentioned by Lee et al,25 the peak velocity of circumferential shortening is decreased in dilated cardiomyopathy as a consequence of intrinsic myocardial disease and increased afterload.
The increase in peak ejection rate determined as −dV/dtmax during LD muscle stimulation is also clearly demonstrated in Fig 2⇑, in which the net systolic ejection time is shorter and SV is larger during the stimulated heartbeats.
During LD stimulation at the best setting, the leftward shift of the end-systolic pressure-volume points, the increase in PER with a mean of 68%, the increase of peak negative dP/dt, the decreased LVEDP compared with the clinical setting, and the increased SV all indicate a higher contractile state of the combined left ventricular and wrapped LD muscle. Moreover, all these changes occurred at a similar preload in stimulated and unstimulated beats, whereas the afterload of the stimulated beats as derived from the diastolic aortic pressure was higher.
Analysis of the end-systolic pressure-volume relation using a preload reduction maneuver26 27 has been reported in animal studies. However, this technique will probably reveal increases in end-systolic elastance and shifts to a lower end-systolic volume only when the applied stimulator settings are tuned to the patient’s individual best settings.
Measurements of LD muscle shortening from x-ray films demonstrated that the contraction-relaxation cycle of a stimulated LD muscle may last up to 600 ms in a cardiomyoplasty patient.28 29 Ideally, the contraction phase of the LD muscle should coincide with the ejection phase, and the relaxation phase should not interfere with the subsequent diastolic filling phase.
The increase in PER with a mean of 30% at the worst setting without an obvious increase in SV suggests that effective LD muscle shortening lasted too short a time. In six patients, the worst setting was observed at a short delay (4 and 25 ms), suggestive of an ineffective LD muscle contraction starting before and finishing too early during the systolic ejection phase. This may impair systolic ejection because of a tetanic or relaxing LD muscle causing an increase in (pseudo) afterload at the end of the ejection phase. In the other three patients, the worst settings were observed at the late delays, suggesting ineffective LD muscle contraction occurring too late during the ejection phase.
In Fig 4⇓, changes in delay combined with the contraction cycle of the LD muscle and left ventricular pressure and volume are presented schematically, assuming an interval of 50 ms between start of the stimulus and LD contraction.
This study shows that in patients 6 to 24 months after cardiomyoplasty surgery, significant hemodynamic effects were present during LD muscle stimulation when the patients were at rest. This was demonstrated on a beat-to-beat basis by two operator-independent techniques with a high time resolution: the conductance catheter technique and the aortic Modelflow method. Beat-to-beat analyses allow the hemodynamic effects to be observed before cardiovascular compensatory reflexes become operative. The cardiomyostimulator settings appeared to be critical in obtaining an increase or decrease in SV during stimulation. At the optimal stimulator setting, the increased contractile state during LD stimulation of the left ventricle–LD muscle combination is demonstrated by the increased peak ejection rate, leftward shift of the end-systolic P-V points, increased negative dP/dtmax, increased SV, and lower end-diastolic left ventricular pressure at similar preload and afterload conditions compared with the unstimulated beats. On the basis of the 68% increase in PER at the best setting, one may expect a substantially larger SV, provided that an ideally timed (delay, stimulus duration) skeletal muscle contraction at a given amplitude can be obtained.
At the clinically used settings, we could not demonstrate an improvement in SV, but LVEDP increased immediately after the 10-second stimulation period. Clinical improvement (NYHA class) therefore may have been caused by active prevention of cardiac dilatation.
In the present study, patients were evaluated at rest, whereas hemodynamic assist during exercise is of clinical importance. Therefore, the relevance of the findings described should be confirmed while patients perform exercise.
This study was financially supported by the Bakken Research Center, Maastricht, the Netherlands. We wish to thank T. van der Nagel, C. Struble, and L. Bonandi for excellent assistance. We also express thanks to M. Lahaije and L. Frissen for assistance in preparing the manuscript.
- Received July 7, 1994.
- Revision received October 10, 1994.
- Accepted October 31, 1994.
- Copyright © 1995 by American Heart Association
Delahaye F, Jegaden O, Montagna P, Desseigne P, Blanc P, Vedrinne C, Tonboul P, Saint-Pierre A, Perinetti M, Rossi R, Itti R, Mikaeloff P. Latissimus dorsi l in severe congestive heart failure: the Lyon experience. J Card Surg. 1991;65:106-112.
Moreira LEF, Stolf NAG, Bocchi EA, Pereira-Baretto AC, Meneghetti JC, Giorgi MCP, Moraes AV, Leite JJ, da Luz PL, Jatene AD. Latissimus dorsi cardiomyoplasty in the treatment of patients with dilated cardiomyopathy. Circulation. 1990;82(suppl IV): IV-257-IV-263.
Lee KF, Dignan RJ, Parmar YM, Dyke CM, Benton G, Yeh T, Abd.-Elfattah AS, Wecksler AS. Effects of dynamic cardiomyoplasty on left ventricular performance and myocardial mechanics in dilated cardiomyopathy. J Thorac Cardiovasc Surg. 1991; 162:124-131.
Molteni L, Almada H, Ferreira R, Ortega D. Assessment of the optimal time interval between QRS and single pulse stimulation in dynamic cardiomyoplasty. In: Chiu RCI, Bourgeois IM, eds. Transformed Muscle for Cardiac Assist and Repair. Mount Kisco, NY: Futura Publishing Co Inc; 1990:189-196.
Grandjean PA, Autin L, Chan S, Terpstra B, Bourgeois IM. Dynamic cardiomyoplasty: clinical follow-up results. J Card Surg. 1991;6S:80-88.
Chachques JC, Acar C, Tapia M, Guibourt P, Fiemeyer A, Bensasson D, Berrebi A, Grare P, Bechara M, Baron JF, Carpentier M. Résultats à moyen terme de la cardiomyoplasty. Arch Mal Coeur. 1994;87:49-56.
Baan J, van der Velde ET, De Bruin HG, Smeenk GJ, Koops J, Van Dijk AD, Temmerman D, Senden PJ, Buis B. Continuous measurement of left ventricular volume in animals and humans by conductance catheter. Circulation. 1984;70:812-823.
Wesseling KH, Janssen JRC, Settels JJ, Schreuder JJ. Computation of aortic flow from pressure in humans using a nonlinear, three-element model. J Appl Physiol. 1993;74:2566-2573.
Burkhoff D, van der Velde ET, Kass D, Baan J, Maughan WL, Sagawa K. Accuracy of volume measurement by conductance catheter in isolated, ejecting canine hearts. Circulation. 1985;72: 440-447.
Steendijk P, van der Velde ET, Baan J. Left ventricular stroke volume by single and dual excitation of the conductance catheter in dogs. Am J Physiol. 1993;264(Heart Circ Physiol 33):H2198-H2207.
Van der Velde ET, van Dijk AD, Steendijk P, Diethelm L, Chagas A, Lipton MJ, Glantz SA, Baan J. Left ventricular segmental volume by conductance catheter and cine-CT. Eur Heart J. 1992;13(suppl E):15-21.
Van der Linden LP, van der Velde ET, Bruschke AVG, Baan J. Comparison between force-velocity and end-systolic pressure-volume characterization of intrinsic LV function. Am J Physiol. 1990;259(Heart Circ Physiol 28):H1419-H1426.
Langewouters GJ, Wesseling KH, Goedhard WJA. The static elastic properties of 45 human thoracic and abdominal aortas in vitro and the parameters of a new model. J Biomech. 1984;17: 425-435.
Jansen JRC, Wesseling KH, Settels JJ, Schreuder JJ. Continuous cardiac output monitoring by pulse contour during cardiac surgery. Eur Heart J. 1990;11SI:26-32.
Applegate RJ, Cheng CP, Little WC. Simultaneous conductance catheter and dimension assessment of left ventricle volume in the intact animal. Circulation. 1990;81:638-648.
Burkhoff D, Alexander J, Schipke J. Assessment of Windkessel as a model of aortic input impedance. Am J Physiol. 1988;255(Heart Circ Physiol 24):H742-H753.
Quinones MA, Gaasch WH, Alexander JK. Influence of acute changes in preload, afterload, contractile state and heart rate on ejection and isovolumic indices of myocardial contractility in man. Circulation. 1976;53:293-302.
Lee KF, Dyke CM, Wechsler AS. Theoretical considerations in the use of dynamic cardiomyoplasty to treat dilated cardiomyopathy. J Card Surg. 1991;6S:119-123.
Van der Veen FH, Lucas CMHB, Lorusso R, van der Nagel T, Penn OCKM, Wellens HJJ. A new method to select stimulus strength after cardiomyoplasty. J Card Surg. 1991;6S:259-264.