- Email: [email protected]
Dynamic cardiomyoplasty acutely impairs left ventricular diastolic function
Dynamic cardiomyoplasty acutely impairs left ventricular diastolic function In patients with congestive heart failure, medical treatment has a high rate of mortality and morbidity, and transplantation is limited by the availability of donor hearts. Dynamic cardiomyoplasty is being investigated as surgical therapy to improve left ventricular function in these patients. To evaluate the early postoperative effects of this procedure on left ventricular diastolic function, we studied seven dogs through the use of sonomicrometry and micromanometry in a canine model of dynamic cardiomyoplasty. Left ventricular diastolic parameters were determined before wrapping the latissimus dorsi muscle (baseline), after latissimus dorsi muscle wrap but without stimulation, and with synchronous left ventricular contraction-latissimus dorsi muscle stimulation. End-diastolic pressure was increased in both conditions after latissimus dorsi muscle wrap (without stimulation,S ± 1; with stimulation, 6 ± 2 mm Hg; p < 0.05) compared with baseline (3 ± 2 mm Hg). The peak rate of diastolic pressure decay was greater at baseline (1560 ± 370 mm Hgjsec) than after latissimus dorsi muscle wrap, both without (1260 ± 330 mm Hgjsec, p < 0.01) and with (1120 ± 420 mm Hgjsec, p < 0.01) stimulation. The constant of pressure decay was prolonged both without (53 ± 10 seconds, p < 0.05) and with (62 ± 11 seconds, p < 0.01) latissimus dorsi muscle stimulation compared with the baseline (38 ± 5 seconds). Compared with baseline (0.2 ± 0.2 cm- 2), the constant of passive chamber stiffness increased after the latissimus dorsi muscle was wrapped around the heart (1.6 ± 0.7 cm", p < 0.05) and with stimulation (2.1 ± 1.0 cm- 2, p < 0.01). The maximal diastolic filling rate (baseline, 18.1 ± 6.7; without stimulation, 16.6 ± 8.9; with stimulation, 16.6 ± 4.1 cm 2 jsec, not significant) and end-diastolic short-axis area (baseline, 7.3 ± 2.3; without stimulation, 7.4 ± 2.1; with stimulation, 7.5 ± 2.3 cm 2, not significant) were similar among the three conditions. The latissimus dorsi muscle wrap prolonged relaxation and increased left ventricular passive stiffness. Synchronous latissimus dorsi muscle stimulation with left ventricular contraction did not improve diastolic function in this model. The results suggest that in the early postoperative period, dynamic cardiomyoplasty impairs diastolic function. (J THORAC CARDIOVASC SURG 1992;104:1662-71)
William J. Corin, MD, David T. George, PhD, James D. Sink, MD, and William P. Santamore, PhD, Philadelphia, Pa.
COngestive heart failure affects 2.4 million persons in the United States, with an estimated 400,000 new cases annually.' The prognosis for these patients is poor. The reported 5-year mortality is greater than 50%2,3 and, despite intensive medical therapy, remains high.v 5 SurFrom the Philadelphia Heart Institute, Presbyterian Medical Center, Philadelphia, Pa. Received for publication Sept. 23, 1991. Accepted for publication March II, 1992. Address for reprints: William J. Carin, MD, Philadelphia Heart Institute, Presbyterian Medical Center, 39th and Market St., Philadelphia, PA 19104.
12/1/38379 1662
gical therapy by cardiac transplantation improves the quality and duration of life but is available to only a limited number of patients. An alternative approach for the treatment of congestive heart failure, which is currently under investigation, is dynamic cardiomyoplasty.vl? In this surgical procedure the patient's own latissimus dorsi muscle (LDM) is wrapped around the heart to provide direct mechanical assistance. The advantages of the procedure include the availability of skeletal muscle, the absence of immunologic rejection, and the fact that no external power source is required. Previous clinical and experimental studies of dynamic cardiomyoplasty have primarily evaluated left ventricular (LV) systolic function.v'? However, in the
Volume 104 Number 6 December 1992
Cardiomyoplasty impairs diastolic function
166 3
Septal
Cr
tal
Lateral
Crystal
_-..;...:.:.;,g
Fig. 1. Acute dynamic cardiomyoplasty in canine model with use of left LDM. Latissimus muscle is wrapped posteriorly around the two ventricles. Two leads for skeletal muscle stimulation were inserted inside latissimus muscle root, avoiding injury to thoracodorsal vessels.A sensing lead from the surface electrocardiogram (not shown), enabled microcomputer to synchronize latissimus stimulation with LV contraction.
current clinical application of cardiomyoplasty, the LDM remains unstimulated for 2 weeks after the operation to allow the skeletal muscle to revascularize. During the next 2 weeks, the LDM is minimally stimulated.P: 13 Thus, for the initial 4 postoperative weeks, minimal systolic assistance is provided by the LDM wrap. In patients with congestive heart failure, LV diastolic dysfunction also contributes to clinical morbidity: Early diastolic filling is reduced, relaxation is prolonged, and end-diastolic pressure is increased. 14 If the clinical application of dynamic cardiomyoplasty is to be optimized, the immediate effect of this procedure on diastole must be better understood. To address this question, we investigated the immediate postoperative effects of dynamic cardiomyoplasty on LV diastolic function.
Methods Overview. This study was designed to evaluate the acute hemodynamic effect of dynamic cardiomyoplasty in a canine model. The surgical preparation was created in mongrel dogs as previously reported.P We measured LV diastolic function before and after encircling the heart with a latissimus dorsi muscle wrap. Micromanometry and sonomicrometry were used to determine LV pressure and short-axis dimensions and to calculate the early diastolic peak filling rate, the maximum rate of pressure decay, and the constant of pressure relaxation (T). The animals used in this study were cared for in a humane fashion and in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for Care and Use of Laboratory Ani-
.
Anterior
Crystal
Fig. 2. A schematic drawing illustrating sonomicrometry crystal placement on the endocardial surface to measure anteroposterior and septal-free wall dimensions. Crystals were inserted into the left ventricle through the apex, and positioned on the endocardial surface at the base of the heart.
mals'' prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH publication 86-23, revised 1985). This animal investigation was approved by the Institutional Animal Care and Use Committee. Experimental preparation. Acute dynamic cardiomyoplasty was surgically created in seven mongrel dogs (21 to 30 kg) with normal LV function. General anesthesia was induced with intravenous thiopental (Pentothal) sodium (25 rng/kg), and the dogs were intubated. Ventilation was maintained with a volume ventilator (Harvard respiratory pump, Harvard Apparatus Co., Inc., Millis, Mass.) to provide nitrous oxide (60%) and oxygen (40%). Anesthesia was supplemented by intravenous injections of 100 mg thiopental sodium. A limb-lead ECG was monitored throughout the experiment. After adequate anesthesia was confirmed, the left LDM was exposed through a transverse left axillary skin incision, and skin flaps were developed. Dissection of the LDM was carried out carefully to preserve the thoracodorsal pedicle. Pacing leads (MYO/WIRE, 2-0, A&E, Farmingdale, N.J.) were woven into the root ofthe muscle, with the negative lead located near the pedicle. The LDM pacing threshold was measured by stimulation of the unwrapped muscle with one 0.2 msec pulse. The lowest stimulating voltage that produced a visually perceptible muscle contraction was defined as the threshold voltage (V T ) . The thorax was opened by a median sternotomy, and the heart was suspended in a pericardial cradle. Piezoelectric crystals, micromanometry and a fluid-filled catheter were inserted into the LV via the apex (see Instrumentation). After instrumentation was completed and satisfactory signals were confirmed, diastolic function was evaluated by measurement of LV pressure and area. After the initial diastolic evaluation, the LDM wrap was performed. A 5 em arc of the lateral portion of the second rib was resected, and an incision was made to open the fifth intercostal space. The latissimus muscle was then dropped into the chest through
The Journal of Thoracic and Cardiovascular Surgery
1 6 6 4 Carin et al.
the defect in the second rib. The most proximal portion of the muscle that reached the heart was applied just below the atrioventricular groove under the left atrial appendage. The costal aspect of the LDM was placed on the epicardium, with the spinal aspect closest to the atrioventricular groove. In this way, the LDM was wrapped posteriorly around the heart, encompassing the left and right ventricles (Fig. I). Instrumentation. Piezoelectric crystals were introduced into the LV cavity via a small puncture in the apex of the heart. To ensure correct positioning of the crystals, a needle was inserted sequentially across the anteroposterior and septal-free wall diameters a t the base ofthe heart. The piezoelectric crystal wires were then retracted from the ventricular cavity through these small holes until significant resistance to further withdrawal was detected. In this manner the piezoelectric crystals were positioned on the endocardial surface of the anterior, posterior, septal, and lateral walls (Fig. 2). At the completion of each experiment, proper sonomicrometry crystal placement was confirmed by visual inspection. Each crystal was located appropriately on the LV endocardial surface. We also inserted a 5F micromanometer (Millar Instruments, Houston, Tex.) and a 7F fluid-filled catheter through the LV apex. The fluid-filled catheter was connected to a pressure transducer (Transpac II, Abbott Laboratories, Salt Lake City, Utah) and was balanced to zero at the level of the left ventricle and to 100 mm Hg with the physiologic monitor (PPG Biomedical, VR-12, Pleasantville, N.Y.). With the same monitor, the micromanometry catheter was calibrated to zero and to the mean pressure of the fluid-filled catheter. Pressure data were digitally converted at 200 Hz (MacAdios II, GW Laboratories, Somerville, Mass.) and stored on computer disks (MAC IIcx, Apple Computers, Cupertino, Calif.) for analysis after completion of the study. The LV short-axis area was calculated from the dimensional data obtained with the sonomicrometer (Triton Industries, San Diego, Calif.). Stimulation protocol. The LDM stimulation leads were connected to a Grass stimulator (model S48, Grass Medical Instruments, Quincy, Mass.) and, with the surface ECG electrodes, to a microcomputer (MAC Ilcx). With software developed in this laboratory, the computer synchronized LDM stimulation with ventricular contraction after an R wave trigger from the surface ECG. The surface ECG and stimulating pulses were continuously monitored to document synchronous LDM stimulation. LV diastolic function was evaluated by measurement of pressure and short-axis area. This was performed before and after wrapping of the LDM, with the following stimulation protocol: normal sinus rhythm without LDM stimulation and a single beat of LDM stimulation during LV systole. The LDM stimulation pulse train included nine pulses at 6 VT (pulse width, 0.2 msec; frequency, 50 Hz). This produced a stimulation duration of 180 msec, nearly identical to that used in a previous studyl'' and to the duration recommended for clinical application." 13 Using this stimulation protocol, we took care to ensure that latissimus muscle stimulation was completed by the end of LV ejection. Data analysis Using software developed in our laboratory, we extracted hemodynamic variables from the digitally stored data. Extrasystolic and postextrasystolic beats were excluded from the analysis. End-diastole was defined as 40 msec after the onset of the QRS waves on the ECG. 16 LV area (A, crrr') was calculated by the following formula:
A
= 7rXY 4
where x and y were diameters (ern) in the LV anteroposterior and septal-free wall dimensions. To assess LV pressure relaxation, we calculated the first time derivative of ventricular pressure (dP/dt, mm Hg/sec) by numeric differentiation of the micromanometer pressure signal using the following formula 17: dP/dt(t) = 200· (-3 . pet - 3) - 2· pet - 2)pet - I) + pet + I) + 2 . pet + 2) + 3 . pet + 3) )/28, where pet) = LV pressure at a given time, t. This formula was also used to calculate the first time derivative of LV area (dA/dt, cm2/sec), with A(t) as the short-axis area at a given time, t. Evaluation of diastolic function. Diastolic function was evaluated by calculation of the maximal -dPI dt, the constant of pressure relaxation, and the peak diastolic filling rate. The constant of isovolumic pressure relaxation (T, sec) was calculated using LV pressure (P, mm Hg) and dP/dt, from maximal -dPI dt to 5 mm Hg above L Vend-diastolic pressure, allowing a shift of the exponential baseline (PB, mm Hg)18: dP/dt = (- liT) . (P - PB)
The peak diastolic filling rate was the maximal rate of ventricular diastolic area increase (dA/dt max ) , calculated by numeric differentiation of the area data. Passive diastolic function was calculated with the use of LV cross-sectional area and pressure, from minimum diastolic pressure to end-diastole. An elastic model with shifting asymptote was used to evaluate LV chamber properties represented as follows!": dP/dA
=
b (A - c)
where P is LV pressure (mm Hg), b is the constant of LV chamber stiffness (I I cm-), A is the LV cross-sectional area (cm-), c is the asymptote of the exponential pressure-area relation (mm Hg), and dP IdA is the instantaneous stiffness (mm Hg/cm-). Values for band c were selected to provide the closest linear curve fit. 19 Statistical analysis. Hemodynamic, dimensional, and diastolic function data before LDM wrap (baseline), after LDM wrap but without stimulation, and with LDM stimulation were compared by repeated-measures one-way analysis of variance. When this analysis indicated a significant difference among the three conditions, Scheffe's test was used to compare the conditions. Differences with a statistical probability less than 0.05 were considered significant. All data are presented as the mean ± I standard deviation.
Results Typical experimental results are presented in Fig. 3. LV pressure and area, the EeG, and the stimulation tracing show beats before and after wrapping of the LDM. The recordings demonstrate that the LDM wrap did not cause any gross changes in diastolic function. The end-diastolic pressure was in the physiologic range, and the end-diastolic area was not changed. Inspection of the LV pressure wave reveals no evidence of constriction, restriction, or tamponade. LDM stimulation occurred during the fourth beat in the postoperative recording and
Volume 104 Number 6 December 1992
'iii
Cardiomyoplasty impairs diastolic function
I665
125
:l:
E 100
SCD
~
III III
e Q. "III
75
50
~
~
'E CD >
25
.; ...I 0
7
~ 6.5 ~
6
III
< 5.5 CD
::i
5
'0 >
S-
e o w
o
1000
2000
Fig. 3 A. Typical simultaneous LV pressure and area recordings, electrocardiogram, and pulse train before LDM wrap. Hemodynamic parameters were evaluated from a beat before LDM wrap, from the beat with LDM stimulation and from the preceding sinus beat without stimulation.
is associated with an increase in peak LV pressure. A typical baseline beat, a beat after LDM wrap but with no stimulation, and a stimulated beat are R-wave matched in Fig. 4. A decrease in the rate of pressure decay and its prolonged duration are evident in the LDM-wrapped beats compared with the baseline beats. Augmentation of systolic pressure in the stimulated LDM beat compared with the nonstimulated beat is also apparent. Hemodynamic data before LDM wrap, after LDM wrap but without stimulation, and with LDM stimulation are summarized in Table I. The heart rate was similar before and after wrapping of the LDM. LV peak systolic pressure was similar at baseline and after the LDM wrap
without stimulation. With LDM stimulation, LV peak systolic pressure increased an average of 17% (p < 0.05) from the nonstimulated beats. LV end-diastolic pressure increased after LDM wrap, both with and without stimulation. LV diastolic function data are presented in Table II. End-diastolic area was unchanged after LDM wrap. The peak rate of diastolic filling (dAjdt) was unchanged after the LDM wrap and with muscle stimulation. LV isovolumic pressure decay slowed after the LDM wrap. Compared with baseline, the constant of chamber stiffness increased after the LDM wrap (p < 0.05) and with LDM stimulation (p < 0.01). The constant of passive chamber
The Journal of Thoracic and Cardiovascular Surgery
Carin et al.
16 6 6
100
Ii
::z:
E
.s
75
e
. Ii: ::J
II II
. .
50
..! ::J
~ C >
25
~ .... 0
7
N' 6.5
E ~ 01 Cl>
6
.:c
5.5
> ....
5
4.5 .
-;:-4
'0
~ 3 c .2 2
01
:i
E
iii
0
6J==~~J= 1000
~
-V
2000
v
Time (msec)
Fig. 3 B. Postoperative record. No gross abnormalities of diastolic function are observed. LDM stimulation occurred on the third beat and was associated with an increase in peak systolic pressure.
Volume 104
Number 6
Cardiomyoplasty impairs diastolic function
December 1992
.. :=
J
tOO 100
-
80
~
'"
~
~
]
---0--
60
~
'-'
Baseline Beat Wrap, -Stim Wrap, «Stirn
~
QI
1:1
QI fIl ~
S 0
40
'"
~
20
Q)
j
~
100
200
400
300
Fig. 4. Typical baseline, wrapped but unstimulated, and wrapped with LDM stimulation beats are superimposed, with the electrocardiogram used for timing. Both of the beats after LDM wrap demonstrate a prolonged period of pressure decay compared with baseline beat, with reduced maximal rate of pressure decay. Increase in systolic pressure with LDM stimulation compared to nonstimulated beat is also evident.
Table I. Hemodynamic data LDMwrap
HR
PSP EDP
..== ~
·25
·50
U
Time (msec)
Baseline
No stimulation
143 ± 13 97 ± 15 3± 2
137 ± 18 91 ± 20 5 ± It
Stimulation
·75
Tau
dP/dt
Diastolic Parameter Fig. 5. Changes from baseline of diastolic parameters are presented for conditions of LDM wrap without stimulation (Wrap, -Stirn), and with stimulation (Wrap, +Stim). Maximal rate of left ventricular pressure decay (dP/dt) decreased after wrap, while constant of pressure decay (tau) increased. End-diastolic area (EDA) maximal rate of area increase (dAidt], and peak normalized filling rate (PNFR) did not differ statistically from baseline.
Table II. Diastolic function data LDMwrap
106 ± 23* 6 ± 2t
< 0.01 versus Baseline.
Baseline
EDA dAjdtma x b
dPjdt max T
stiffness was significantly increased after LDM wrap but was not changed further with stimulation. The maximal rate of LV diastolic pressure decay was reduced after LDM wrap, both with and without LDM stimulation (Fig. 5). After wrapping of the LDM, the constant of pressure decay increased by 39%compared with baseline; with LDM stimulation, the constant was prolonged by 63%. To evaluate the rate of LV pressure decay from peak LV pressure to 5 mm Hg above end-diastolic pressure, Fig. 6 presents dP /dt data as a function of LV pressure in typical baseline, nonstimulated, and LDM-stimulated beats.i" This figure demonstrates that during this period the rate of LV pressure decline was reduced after the LDM wrap. Fig. 7 demonstrates the significant inverse relationship (p < 0.01) between peak systolic pressure and the max-
PNFR
dNdt
EDA
NA
LDM. Latissimus dorsi muscle; HR. heart rate (beats/min); PSP. peak LV systolic pressure (mm Hg); EDP. end-diastolic pressure (rnrn Hg): NA. not applicable. *p < 0.01 LDM wrap versus LDM wrap with stimulation. tp
....... ~
d
4::
I 667
7.3 ± 2.3
18.1 0.2 -1560 38
± ± ± ±
6.7 0.2 370 5
No stimulation 7.5 ± 16.3 ± L6 ± -1260 ± 53 ±
2.3 8.9 0.7* 330t 10*
Stimulation 7.4±2.1 16.6 ± 4.1 2.1 ± LOt -1120 ± 420t 62 ± lIt
EDA, End-diastolic area (cm-): dA/dt",m. maximal rate of diastolic area increase (em' /sec); b. constant of chamber stiffness (I/em'); dP/dt"",,, maximal rate of LV diastolic pressure decay (mm Hg/sec): T. constant of diastolic pressure decay (second).
*p < 0.05 tp
versus baseline.
< 0.01 versus baseline.
imal rate of pressure decay for the baseline and stimulation conditions. The 95% confidence band for the baseline is also presented. Each of the data points during LDM stimulation fall outside this 95% confidence band, demonstrating that, across the observed range of systolic pressure, the peak rate of pressure decay was depressed • with LDM stimulation. Discussion Dynamic cardiomyoplasty has been proposed as a surgical method to improve LV function in patients with
The Journal of Thoracic and Cardiovascular Surgery
I 6 6 8 Carin et al.
2400 Baseline
2200
Wrap. -Stirn Wrap. -Stim
Jj ...
,-..
2000
tJ
QI
rn
1800
tIl
1600
~
e
-lOCO
~
?d
'"~
1400
....e
1200
~
:s
.»
~
"tlI Left Ventricular Pressure (mrn Hg)
6 Baseline, R =0.86 • Wrap, +Stim, R= 0.76
70
Fig. 6. Relationship of LV pressureand dPj dt are presented for a baseline beat and for beats after LDM wrap both with (+Stim) and without (-Stirn) musclestimulation. The segment of isovolumic relaxationoccurson the lowerleft-handportionof the curves,beginning at minimumdPjdt and continuing to the left and up as both pressure and dPjdt decline. Compared to baseline, the rate of pressuredecay was depressed throughout diastole after the LDM wrap.
congestive heart failure. Some clinical studies have demonstrated improved systolic function, with regard to ejection fraction and stroke work, after cardiomyoplasty,6,8,11-13,21,22 but this is not a universal finding. 13, 23-25 Similarly, some experimental studies have shown improved LV systolic performance.F' although others have shown only a minimal benefit. 27,28 For the first 2 weeks after the heart is surgically wrapped, the LDM remains unstimulated to permit skeletal muscle revascularization.f-L' During the next 2 weeks, LDM training begins with only minimal stimulation (a single muscle twitch on every other heartbeat). As a result, there is minimal systolic assistance during the first 4 postoperative weeks of dynamic cardiomyoplasty.9,29 The significance of systolic impairment is widely recognized in patients with congestive heart failure. However, diastolic dysfunction also contributes to the clinical signs and symptoms of dyspnea and reduced exercise tolerance. I 4 Thus the effects of dynamic cardiomyoplasty on diastolic function are important to the clinical outcome of this procedure. Few previous studies have evaluated the effect of chronic cardiomyoplasty on diastole, and none has investigated the acute effects. The present study examined diastolic function immediately after cardiomyoplasty and is the only investigation to examine diastolic function acutely after surgical procedures. We compared diastolic
-o-_--+_----l
400+80
90
100
110
-+-_ _ '120
130
140
Peak Systolic Pressure (mm Hg)
150
Fig. 7. The significant inverse relationship between peak systolic pressureand maximal rate of pressuredecayis confirmed for baseline and after LDM wrap with stimulation. The 95% confidence band for baseline data across the range of LV systolic pressure is presented. Each of the data points for LDM stimulation fell below the confidence band, indicating a depressed rate of pressuredecay.
function at three stages: without an LDM wrap, with the wrap alone, and with LDM stimulation. The results of this study suggest that dynamic cardiomyoplasty causes no gross abnormalities in LV diastolic function acutely. LV end-diastolic pressure and area were in the physiologic range, and there was no evidence of ventricular constriction. However, sensitive indexes of diastolic function demonstrated a decline in the rate of diastolic pressure decay, a prolongation of the pressure relaxation constant, and an increase in passive chamber stiffness. Furthermore, stimulation of the LDM wrap did not improve LV diastolic function compared with the diastolic function with nonstimulated beats. The peak rate of isovolumic pressure decay is dependent on several factors, including systolic pressure, heart rate, and LV contractility. 3D The peak systolic pressure was similar before and after LDM wrap. As demonstrated in Fig. 7, there was a close correlation between peak systolic pressure and peak pressure decay, both at baseline and after LDM with stimulation. The data demonstrate the significant prolongation of pressure decay immediately after cardiomyoplasty. Although atrial pacing was not performed, the heart rate at baseline was similar to the rate after surgical intervention and therefore would not be expected to have caused the observed changes. The constant of isovolumic pressure decay is not
Volume 104 Number 6 December 1992
affected by changes in ventricular load or heart rate;'! but it changes with inotropic stimulation, myocardial ischemia, and hypertrophy.Vv" Inotropic medication was not administered, and there was no evidence of myocardial ischemia during the investigation. Therefore, the present results are likely to be a reflection of prolonged relaxation associated with cardiomyoplasty, and not secondary to changes in other hemodynamic parameters. Passive chamber stiffness significantly increased after LDM wrap but was not elevated further with muscle stimulation. Increased chamber stiffness has been demonstrated clinically to be associated with LV hypertrophy'? and fibrosis.P In the present study, after LDM wrap there was a significant increase in muscle mass surrounding the LV cavity. The mechanical situation after cardiomyoplasty may therefore be analogous to ventricular hypertrophy, in which LV wall thickness has increased. In this study, done immediately after the surgical procedure, there was insufficient time for either myocardial or skeletal muscle fibrosis to occur. There are several mechanisms that may contribute to these observations, including LDM mechanical restraint of the left ventricle and an increase in muscle mass around the LV chamber. In this study, the increase in diastolic function was unlikely to have been caused by a volume infusion, which was limited to less than 100ml during the period of evaluation of diastolic function. In this study the LDM was wrapped as previously reported from this laboratory I 5 and in clinical and experimental reports.f Although fiber orientation can be replicated easily, the tightness of fit of the wrap around the heart is more variable. For the LDM to provide optimal mechanical assistance to the heart, the skeletal muscle must be preloaded; that is, the muscle needs to be stretched during diastole. Previous investigators have described methods to prevent the wrap from impeding LV filling,12 but they did not evaluate skeletal muscle lengthening or LV filling immediately after surgery. In the current study care was taken not to overstretch the LDM,12 but our results are consistent with a diastolic restraining effect of the LDM wrap on the left ventricle. Immediately after surgical intervention, the LDM cannot fully adapt to the heart size, and the tight wrap has a deleterious effect on diastolic function. In addition, the skeletal muscle wrap is attached to the heart with sutures; these points of attachment might restrain diastolic filling. Thus the physical contact between the muscles and constraint of the left ventricle by the wrap may partially explain the results of the present investigation. Clinically, in concentric hypertrophy caused by systemic hypertension or aortic stenosis, LV diastolic pressure is substantially increased relative to the diastolic volume.l" In addition, passive stiffness increases, the peak
Cardiomyoplasty impairs diastolic function
16 6 9
rate of pressure decay is reduced, and the constant of pressure relaxation is prolonged, even in patients with normal systolic function.P 36 The diastolic dysfunction observed immediately after cardiomyoplasty might also be due to the increase in muscle mass around the left ventricle. With the application of skeletal muscle around the heart, the amount of muscle surrounding the LV chamber increased. We observed conditions similar to those associated with LV hypertrophy: an increase in passive stiffness, a decline in the peak rate of pressure decay, and prolonged relaxation. Comparison with literature. Two previous studies have examined diastolic function in a chronic model of cardiomyoplasty. To our knowledge, the present study is the first to evaluate the effect of cardiomyoplasty on passive diastolic function. In a study from this laboratory, radionuclide ventriculography was used to obtain images of the left ventricles of dogs with doxorubicin (Adriamycinj-induced congestive failure and an LDM wrap.P After a IO-weektraining period ofthe skeletal muscle, LV function during continuous LDM stimulation was compared with function without stimulation. The results showed an increase in the peak diastolic filling rate during synchronous latissimus stimulation compared with the filling rate without stimulation. In a preliminary report, Lee and colleagues'? evaluated LV diastolic function in dogs with normal systolic function after 6 weeks of LDM training. Using two-dimensional echocardiography, they found that the LV peak filling rate was increased during LDM stimulation compared with the filling rate without stimulation. The diastolic filling rate of the wrapped left ventricle was not compared with the unwrapped chamber in either of these studies. It has been stated that LV depression is a prerequisite for dynamic cardiomyoplasty to augment systolic function." Augmented diastolic filling was demonstrated in both normal and depressed ventricles in chronic models during LDM stimulation. In the present study of the acute aspects of cardiomyoplasty, LV systolic pressure increased with latissimus stimulation. These findings are consistent with previous studies showing an improvement in systolic function. I 2 In contrast, diastolic function in the normal LV was impaired immediately after cardiomyopIasty. Limitations. Differences between the present study and clinical practice may limit extrapolation of the shortterm results of clinical cardiomyoplasty. The present study investigated the acute effects on diastolic function with use of a canine model in which baseline LV function was normal. Dynamic cardiomyoplasty is performed on patients with congestive heart failure that is characterized by depressed LV systolic and diastolic function. The effect of the LDM wrap on a chronically dilated, dysfunctional
I 6 70
Carin et at.
left ventricle may differ from its effect on a normal ventricle. The conditioned skeletal muscle has different characteristics than the nonconditioned muscle. In the present study, the LDM was not conditioned before stimulation. This contrasts with the clinical setting, in which the LDM will be conditioned during the first 3 postoperative months. LV diastolic function was evaluated after stimulation during a single heartbeat. Clinically, 6 weeks after the operation, the muscle wrap is stimulated with every heartbeat. Clinical relevance. Current treatment options for chronic heart failure have significant limitations. Medical therapy carries a significantly high rate of mortality and morbidity, and cardiac transplantation is limited by the shortage of available donor hearts. Dynamic cardiomyopia sty has been proposed as alternative therapy for these patients. Improvement in LV systolic function has been demonstrated in some early studies of dynamic cardiomyoplasty, but its optimal application is not known. In patients with congestive failure, LV diastolic dysfunction contributes to the clinical morbidity. Little is known about the effects of dynamic cardiomyoplasty on diastolic function before the institution of significant systolic assistance. The results of this study suggest that shortly after cardiomyoplasty, diastolic function may be impaired. We acknowledge the superb secretarial assistance of Ms. Jan Merolli and technical support of John Michele and Gioia DiLoreto. REFERENCES I. Furberg CD, Yusuf S. Effect of vasodilators on survival in congestive heart failure. Am J Cardiol 1985;55:1109-13. 2. Smith WM. Epidemiology of congestive heart failure. Am J Cardiol 1985;55:3A-8A. 3. CONSENSUS Trial Study Group. Effects of enalapril on mortality in severe congestive heart failure. N Engl J Med 1987;316:1429-35. 4. SOLVD Investigators. Effect of enalapril on survival in patients with reduced left ventricular ejection fractions and congestive heart failure. N Engl J Med 1991;325:293-302. 5. Cohn IN, Johnson G, Ziesche S, et al. A comparison of enalapril with hydralazine-isosorbide dinitrate in the treatment of chronic congestive heart failure. N Engl J Med 1991;325:303-10. 6. Carpentier A, Chachques .IC, Myocardial substitution with a stimulated skeletal muscle: first successful clinical case [letter]. Lancet 8840:1267,1985. 7. Carpentier A, Chachques .IC, The use of stimulated skeletal muscle to replace diseased human heart muscle. In: Chiu RC-J, ed. Biomechanical cardiac assist. Mount Kisco, New York: Futura Publishing Co., Inc., 1986:85-102. 8. Chachques JC, Grandjean P, Schwartz K, et al. Effect of
The Journal of Thoracic and Cardiovascular Surgery
latissimus dorsi dynamic cardiomyoplasty on ventricular function. Circulation 1988;78(Pt 2):III203-16. 9. Chachques JC, Grandjean PA, Carpentier A. Latissimus dorsi dynamic cardiomyoplasty. Ann Thorac Surg 1989;47:600-4. 10. Sink JD, Soberman MS. Latissimus dorsi dynamic cardiomyoplasty: a procedure in evolution. Emory Univ J Med 1988;2:223-32. II. Molteni L, Almada H, Ferreira R. Synchronously stimulated skeletal muscle graft for left ventricular assistance. J THORAC CARDIOVASC SURG 1989;97:439-46. 12. Lee DF, Dignan RJ, Parmer JM, et al. Effects of dynamic cardiomyoplastyon left ventricular performance and myocardial mechanics in dilated cardiomyopathy. J THORAC CAROlOVASC SURG 1991;102:124-31. 13. Hagege AA. Desnos M, Chachques JC, et al. Preliminary report: follow-up after dynamic cardiomyoplasty. Lancet 1990;335:1122-4. 14. Gaasch WHo Diastolic mechanism in heart failure. Heart Failure 1985;1:195-202. 15. Cheng W, Justicz AG, Soberman MS, Alazrake NP, Santamore WP, Sink JD. The effect of dynamic cardiomyoplasty on left ventricular systolic and diastolic functions in a canine model of chronic heart failure. J THORAC CAROlOVASC SURG 1992;103:1207-13. 16. Grossman W, Stefadourous MA, McLaurin LP, Rolett EL, Young DT. Quantitative assessment of left ventricular diastolic stiffness in man. Circulation 1973;47:567-74. 17. Marble AE, McIntyre CM, Hastings-James R, Hor CWo A comparison of digital algorithms used in computing the derivative of left ventricular pressure. IEEE Trans Biomed Eng 1981;28:524-8. 18. Eichhorn P, Grimm J, Koch R, Hess OM, Carroll J, Krayenbuehl HP. Left ventricular relaxation in patients with left ventricular hypertrophy secondary to aortic valve disease. Circulation 1982;77:1395-404. 19. Corin WJ, Murakami T, Monrad ES, Hess OM, Krayenbuehl HP. Left ventricular passive diastolic properties in chronic mitral regurgitation. Circulation 1991;83:797-807. 20. Slinker BK, Goto Y, LeWinter MM. Systolic direct ventricular interaction affects left ventricular contraction and relaxation in the intact dog circulation. Circ Res 1989;65:307-15. 21. Moreira LFP, Stolf NAG, Jatene AD. Benefits of cardiomyoplasty for dilated cardiomyopathy. Semin Thorac Cardiovasc Surg 1991;3:140-4. 22. Lee KF, Dyke CM, Dignan RJ, et al. Effects of dynamic cardiomyoplastyon left ventricular dynamic geometry in a canine model of dilated cardiomyopathy. Surg Forum 1990;41:263-5. 23. Chiu RCJ. Biomechanical cardiac assist: quo vadimus? Semin Thorac Cardiovasc Surg 1991;3:160-3. 24. Hill A, Chiu RC-J. Dynamic cardiomyoplasty for treatment of heart failure. Clin Cardiol 1989;12:681-8. 25. Magovern JA, Furnary AP, Christlieb IY, Kao RL, Park SB, Magovern GJ. Indications and risk analysis for clinical cardiomyoplasty. Semin Thorac Cardiovasc Surg 1991; 3:145-8.
Volume 104 Number 6 December 1992
26. Moreira LFP, Stolf NAG, Bocchi EA, et al. Latissimus dorsi cardiomyoplasty in the treatment of patients with dilated cardiomyopathy. Circulation 1990;82(Pt 2):IV25763. 27. Anderson WA, Anderson JS, Acker MA, et al. Skeletal muscle grafts applied to the heart. Circulation 1988;78:(Pt 2):III 180-90. 28. Kao RL, Christlieb IY, Magovern GJ, ParkSB, Magovern GJ Jr. The importance of skeletal muscle fiber orientation for dynamic cardiomyoplasty. J THORAC CARDIOVASC SURG 1990;99:134-40. 29. Moreira LFD, Stolf NAG, Jatene AD. Hemodynamic benefits of cardiomyoplasty in clinical and experimental myocardial dysfunction. In: Chiu RCJ, Bourgeois 1M, eds. Transformed muscle for cardiac assist and repair. Mt. Kisco, NY: Futura, 1990:179-88. 30. Weisfeldt ML, Scully HE, Frederiksen J, et al. Hemodynamic determinants of maximum negative dP / dt and periods of diastole. Am J Physiol 1974;277:613-21. 31. Weiss JL, Frederiksen JW, Weisfeldt ML. Hemodynamic determinants of the time-course of fall in canine left ventricular pressure. J Clin Invest 1976;58:751-60.
Cardiomyoplasty impairs diastolic function
16 7 1
32. Carroll JD, Lang RM, Neumann AL, BorowKM, Rajfer SI. The differential effects of positive inotropic and vasodilator therapy on diastolic properties in patients with congestive cardiomyopathy. Circulation 1986;74:815-25. 33. Eichhorn P, Grimm J, Koch R, Hess 0, Carroll J, Krayenbuehl HP. Left ventricular relaxation in patients with left ventricular hypertrophy secondary to aortic valve disease. Circulation 1982;65:1395-404. 34. Hirota Y. A clinical study of left ventricular relaxation. Circulation 1980;62:756-63. 35. Hess OM, Ritter M, Schneider J, Grimm J, Turina M, Krayenbuehl HP. Diastolic stiffness and myocardial structure in aortic valve disease before and after valve replacement. Circulation 1984;69:855-65. 36. Lorell BH, Grossman W. Cardiac hypertrophy: the consequences for diastole. J Am Coli CardioI1987;9:1189-93. 37. Lee KF, Parmar 1M, Nixon lV, et al. Effect of dynamic cardiomyoplastyon left ventricular diastolic chamber function. Circulation 1990;82(Pt 2):III712.
William J. Corin, MD, David T. George, PhD, James D. Sink, MD, and William P. Santamore, PhD, Philadelphia, Pa.
COngestive heart failure affects 2.4 million persons in the United States, with an estimated 400,000 new cases annually.' The prognosis for these patients is poor. The reported 5-year mortality is greater than 50%2,3 and, despite intensive medical therapy, remains high.v 5 SurFrom the Philadelphia Heart Institute, Presbyterian Medical Center, Philadelphia, Pa. Received for publication Sept. 23, 1991. Accepted for publication March II, 1992. Address for reprints: William J. Carin, MD, Philadelphia Heart Institute, Presbyterian Medical Center, 39th and Market St., Philadelphia, PA 19104.
12/1/38379 1662
gical therapy by cardiac transplantation improves the quality and duration of life but is available to only a limited number of patients. An alternative approach for the treatment of congestive heart failure, which is currently under investigation, is dynamic cardiomyoplasty.vl? In this surgical procedure the patient's own latissimus dorsi muscle (LDM) is wrapped around the heart to provide direct mechanical assistance. The advantages of the procedure include the availability of skeletal muscle, the absence of immunologic rejection, and the fact that no external power source is required. Previous clinical and experimental studies of dynamic cardiomyoplasty have primarily evaluated left ventricular (LV) systolic function.v'? However, in the
Volume 104 Number 6 December 1992
Cardiomyoplasty impairs diastolic function
166 3
Septal
Cr
tal
Lateral
Crystal
_-..;...:.:.;,g
Fig. 1. Acute dynamic cardiomyoplasty in canine model with use of left LDM. Latissimus muscle is wrapped posteriorly around the two ventricles. Two leads for skeletal muscle stimulation were inserted inside latissimus muscle root, avoiding injury to thoracodorsal vessels.A sensing lead from the surface electrocardiogram (not shown), enabled microcomputer to synchronize latissimus stimulation with LV contraction.
current clinical application of cardiomyoplasty, the LDM remains unstimulated for 2 weeks after the operation to allow the skeletal muscle to revascularize. During the next 2 weeks, the LDM is minimally stimulated.P: 13 Thus, for the initial 4 postoperative weeks, minimal systolic assistance is provided by the LDM wrap. In patients with congestive heart failure, LV diastolic dysfunction also contributes to clinical morbidity: Early diastolic filling is reduced, relaxation is prolonged, and end-diastolic pressure is increased. 14 If the clinical application of dynamic cardiomyoplasty is to be optimized, the immediate effect of this procedure on diastole must be better understood. To address this question, we investigated the immediate postoperative effects of dynamic cardiomyoplasty on LV diastolic function.
Methods Overview. This study was designed to evaluate the acute hemodynamic effect of dynamic cardiomyoplasty in a canine model. The surgical preparation was created in mongrel dogs as previously reported.P We measured LV diastolic function before and after encircling the heart with a latissimus dorsi muscle wrap. Micromanometry and sonomicrometry were used to determine LV pressure and short-axis dimensions and to calculate the early diastolic peak filling rate, the maximum rate of pressure decay, and the constant of pressure relaxation (T). The animals used in this study were cared for in a humane fashion and in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for Care and Use of Laboratory Ani-
.
Anterior
Crystal
Fig. 2. A schematic drawing illustrating sonomicrometry crystal placement on the endocardial surface to measure anteroposterior and septal-free wall dimensions. Crystals were inserted into the left ventricle through the apex, and positioned on the endocardial surface at the base of the heart.
mals'' prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH publication 86-23, revised 1985). This animal investigation was approved by the Institutional Animal Care and Use Committee. Experimental preparation. Acute dynamic cardiomyoplasty was surgically created in seven mongrel dogs (21 to 30 kg) with normal LV function. General anesthesia was induced with intravenous thiopental (Pentothal) sodium (25 rng/kg), and the dogs were intubated. Ventilation was maintained with a volume ventilator (Harvard respiratory pump, Harvard Apparatus Co., Inc., Millis, Mass.) to provide nitrous oxide (60%) and oxygen (40%). Anesthesia was supplemented by intravenous injections of 100 mg thiopental sodium. A limb-lead ECG was monitored throughout the experiment. After adequate anesthesia was confirmed, the left LDM was exposed through a transverse left axillary skin incision, and skin flaps were developed. Dissection of the LDM was carried out carefully to preserve the thoracodorsal pedicle. Pacing leads (MYO/WIRE, 2-0, A&E, Farmingdale, N.J.) were woven into the root ofthe muscle, with the negative lead located near the pedicle. The LDM pacing threshold was measured by stimulation of the unwrapped muscle with one 0.2 msec pulse. The lowest stimulating voltage that produced a visually perceptible muscle contraction was defined as the threshold voltage (V T ) . The thorax was opened by a median sternotomy, and the heart was suspended in a pericardial cradle. Piezoelectric crystals, micromanometry and a fluid-filled catheter were inserted into the LV via the apex (see Instrumentation). After instrumentation was completed and satisfactory signals were confirmed, diastolic function was evaluated by measurement of LV pressure and area. After the initial diastolic evaluation, the LDM wrap was performed. A 5 em arc of the lateral portion of the second rib was resected, and an incision was made to open the fifth intercostal space. The latissimus muscle was then dropped into the chest through
The Journal of Thoracic and Cardiovascular Surgery
1 6 6 4 Carin et al.
the defect in the second rib. The most proximal portion of the muscle that reached the heart was applied just below the atrioventricular groove under the left atrial appendage. The costal aspect of the LDM was placed on the epicardium, with the spinal aspect closest to the atrioventricular groove. In this way, the LDM was wrapped posteriorly around the heart, encompassing the left and right ventricles (Fig. I). Instrumentation. Piezoelectric crystals were introduced into the LV cavity via a small puncture in the apex of the heart. To ensure correct positioning of the crystals, a needle was inserted sequentially across the anteroposterior and septal-free wall diameters a t the base ofthe heart. The piezoelectric crystal wires were then retracted from the ventricular cavity through these small holes until significant resistance to further withdrawal was detected. In this manner the piezoelectric crystals were positioned on the endocardial surface of the anterior, posterior, septal, and lateral walls (Fig. 2). At the completion of each experiment, proper sonomicrometry crystal placement was confirmed by visual inspection. Each crystal was located appropriately on the LV endocardial surface. We also inserted a 5F micromanometer (Millar Instruments, Houston, Tex.) and a 7F fluid-filled catheter through the LV apex. The fluid-filled catheter was connected to a pressure transducer (Transpac II, Abbott Laboratories, Salt Lake City, Utah) and was balanced to zero at the level of the left ventricle and to 100 mm Hg with the physiologic monitor (PPG Biomedical, VR-12, Pleasantville, N.Y.). With the same monitor, the micromanometry catheter was calibrated to zero and to the mean pressure of the fluid-filled catheter. Pressure data were digitally converted at 200 Hz (MacAdios II, GW Laboratories, Somerville, Mass.) and stored on computer disks (MAC IIcx, Apple Computers, Cupertino, Calif.) for analysis after completion of the study. The LV short-axis area was calculated from the dimensional data obtained with the sonomicrometer (Triton Industries, San Diego, Calif.). Stimulation protocol. The LDM stimulation leads were connected to a Grass stimulator (model S48, Grass Medical Instruments, Quincy, Mass.) and, with the surface ECG electrodes, to a microcomputer (MAC Ilcx). With software developed in this laboratory, the computer synchronized LDM stimulation with ventricular contraction after an R wave trigger from the surface ECG. The surface ECG and stimulating pulses were continuously monitored to document synchronous LDM stimulation. LV diastolic function was evaluated by measurement of pressure and short-axis area. This was performed before and after wrapping of the LDM, with the following stimulation protocol: normal sinus rhythm without LDM stimulation and a single beat of LDM stimulation during LV systole. The LDM stimulation pulse train included nine pulses at 6 VT (pulse width, 0.2 msec; frequency, 50 Hz). This produced a stimulation duration of 180 msec, nearly identical to that used in a previous studyl'' and to the duration recommended for clinical application." 13 Using this stimulation protocol, we took care to ensure that latissimus muscle stimulation was completed by the end of LV ejection. Data analysis Using software developed in our laboratory, we extracted hemodynamic variables from the digitally stored data. Extrasystolic and postextrasystolic beats were excluded from the analysis. End-diastole was defined as 40 msec after the onset of the QRS waves on the ECG. 16 LV area (A, crrr') was calculated by the following formula:
A
= 7rXY 4
where x and y were diameters (ern) in the LV anteroposterior and septal-free wall dimensions. To assess LV pressure relaxation, we calculated the first time derivative of ventricular pressure (dP/dt, mm Hg/sec) by numeric differentiation of the micromanometer pressure signal using the following formula 17: dP/dt(t) = 200· (-3 . pet - 3) - 2· pet - 2)pet - I) + pet + I) + 2 . pet + 2) + 3 . pet + 3) )/28, where pet) = LV pressure at a given time, t. This formula was also used to calculate the first time derivative of LV area (dA/dt, cm2/sec), with A(t) as the short-axis area at a given time, t. Evaluation of diastolic function. Diastolic function was evaluated by calculation of the maximal -dPI dt, the constant of pressure relaxation, and the peak diastolic filling rate. The constant of isovolumic pressure relaxation (T, sec) was calculated using LV pressure (P, mm Hg) and dP/dt, from maximal -dPI dt to 5 mm Hg above L Vend-diastolic pressure, allowing a shift of the exponential baseline (PB, mm Hg)18: dP/dt = (- liT) . (P - PB)
The peak diastolic filling rate was the maximal rate of ventricular diastolic area increase (dA/dt max ) , calculated by numeric differentiation of the area data. Passive diastolic function was calculated with the use of LV cross-sectional area and pressure, from minimum diastolic pressure to end-diastole. An elastic model with shifting asymptote was used to evaluate LV chamber properties represented as follows!": dP/dA
=
b (A - c)
where P is LV pressure (mm Hg), b is the constant of LV chamber stiffness (I I cm-), A is the LV cross-sectional area (cm-), c is the asymptote of the exponential pressure-area relation (mm Hg), and dP IdA is the instantaneous stiffness (mm Hg/cm-). Values for band c were selected to provide the closest linear curve fit. 19 Statistical analysis. Hemodynamic, dimensional, and diastolic function data before LDM wrap (baseline), after LDM wrap but without stimulation, and with LDM stimulation were compared by repeated-measures one-way analysis of variance. When this analysis indicated a significant difference among the three conditions, Scheffe's test was used to compare the conditions. Differences with a statistical probability less than 0.05 were considered significant. All data are presented as the mean ± I standard deviation.
Results Typical experimental results are presented in Fig. 3. LV pressure and area, the EeG, and the stimulation tracing show beats before and after wrapping of the LDM. The recordings demonstrate that the LDM wrap did not cause any gross changes in diastolic function. The end-diastolic pressure was in the physiologic range, and the end-diastolic area was not changed. Inspection of the LV pressure wave reveals no evidence of constriction, restriction, or tamponade. LDM stimulation occurred during the fourth beat in the postoperative recording and
Volume 104 Number 6 December 1992
'iii
Cardiomyoplasty impairs diastolic function
I665
125
:l:
E 100
SCD
~
III III
e Q. "III
75
50
~
~
'E CD >
25
.; ...I 0
7
~ 6.5 ~
6
III
< 5.5 CD
::i
5
'0 >
S-
e o w
o
1000
2000
Fig. 3 A. Typical simultaneous LV pressure and area recordings, electrocardiogram, and pulse train before LDM wrap. Hemodynamic parameters were evaluated from a beat before LDM wrap, from the beat with LDM stimulation and from the preceding sinus beat without stimulation.
is associated with an increase in peak LV pressure. A typical baseline beat, a beat after LDM wrap but with no stimulation, and a stimulated beat are R-wave matched in Fig. 4. A decrease in the rate of pressure decay and its prolonged duration are evident in the LDM-wrapped beats compared with the baseline beats. Augmentation of systolic pressure in the stimulated LDM beat compared with the nonstimulated beat is also apparent. Hemodynamic data before LDM wrap, after LDM wrap but without stimulation, and with LDM stimulation are summarized in Table I. The heart rate was similar before and after wrapping of the LDM. LV peak systolic pressure was similar at baseline and after the LDM wrap
without stimulation. With LDM stimulation, LV peak systolic pressure increased an average of 17% (p < 0.05) from the nonstimulated beats. LV end-diastolic pressure increased after LDM wrap, both with and without stimulation. LV diastolic function data are presented in Table II. End-diastolic area was unchanged after LDM wrap. The peak rate of diastolic filling (dAjdt) was unchanged after the LDM wrap and with muscle stimulation. LV isovolumic pressure decay slowed after the LDM wrap. Compared with baseline, the constant of chamber stiffness increased after the LDM wrap (p < 0.05) and with LDM stimulation (p < 0.01). The constant of passive chamber
The Journal of Thoracic and Cardiovascular Surgery
Carin et al.
16 6 6
100
Ii
::z:
E
.s
75
e
. Ii: ::J
II II
. .
50
..! ::J
~ C >
25
~ .... 0
7
N' 6.5
E ~ 01 Cl>
6
.:c
5.5
> ....
5
4.5 .
-;:-4
'0
~ 3 c .2 2
01
:i
E
iii
0
6J==~~J= 1000
~
-V
2000
v
Time (msec)
Fig. 3 B. Postoperative record. No gross abnormalities of diastolic function are observed. LDM stimulation occurred on the third beat and was associated with an increase in peak systolic pressure.
Volume 104
Number 6
Cardiomyoplasty impairs diastolic function
December 1992
.. :=
J
tOO 100
-
80
~
'"
~
~
]
---0--
60
~
'-'
Baseline Beat Wrap, -Stim Wrap, «Stirn
~
QI
1:1
QI fIl ~
S 0
40
'"
~
20
Q)
j
~
100
200
400
300
Fig. 4. Typical baseline, wrapped but unstimulated, and wrapped with LDM stimulation beats are superimposed, with the electrocardiogram used for timing. Both of the beats after LDM wrap demonstrate a prolonged period of pressure decay compared with baseline beat, with reduced maximal rate of pressure decay. Increase in systolic pressure with LDM stimulation compared to nonstimulated beat is also evident.
Table I. Hemodynamic data LDMwrap
HR
PSP EDP
..== ~
·25
·50
U
Time (msec)
Baseline
No stimulation
143 ± 13 97 ± 15 3± 2
137 ± 18 91 ± 20 5 ± It
Stimulation
·75
Tau
dP/dt
Diastolic Parameter Fig. 5. Changes from baseline of diastolic parameters are presented for conditions of LDM wrap without stimulation (Wrap, -Stirn), and with stimulation (Wrap, +Stim). Maximal rate of left ventricular pressure decay (dP/dt) decreased after wrap, while constant of pressure decay (tau) increased. End-diastolic area (EDA) maximal rate of area increase (dAidt], and peak normalized filling rate (PNFR) did not differ statistically from baseline.
Table II. Diastolic function data LDMwrap
106 ± 23* 6 ± 2t
< 0.01 versus Baseline.
Baseline
EDA dAjdtma x b
dPjdt max T
stiffness was significantly increased after LDM wrap but was not changed further with stimulation. The maximal rate of LV diastolic pressure decay was reduced after LDM wrap, both with and without LDM stimulation (Fig. 5). After wrapping of the LDM, the constant of pressure decay increased by 39%compared with baseline; with LDM stimulation, the constant was prolonged by 63%. To evaluate the rate of LV pressure decay from peak LV pressure to 5 mm Hg above end-diastolic pressure, Fig. 6 presents dP /dt data as a function of LV pressure in typical baseline, nonstimulated, and LDM-stimulated beats.i" This figure demonstrates that during this period the rate of LV pressure decline was reduced after the LDM wrap. Fig. 7 demonstrates the significant inverse relationship (p < 0.01) between peak systolic pressure and the max-
PNFR
dNdt
EDA
NA
LDM. Latissimus dorsi muscle; HR. heart rate (beats/min); PSP. peak LV systolic pressure (mm Hg); EDP. end-diastolic pressure (rnrn Hg): NA. not applicable. *p < 0.01 LDM wrap versus LDM wrap with stimulation. tp
....... ~
d
4::
I 667
7.3 ± 2.3
18.1 0.2 -1560 38
± ± ± ±
6.7 0.2 370 5
No stimulation 7.5 ± 16.3 ± L6 ± -1260 ± 53 ±
2.3 8.9 0.7* 330t 10*
Stimulation 7.4±2.1 16.6 ± 4.1 2.1 ± LOt -1120 ± 420t 62 ± lIt
EDA, End-diastolic area (cm-): dA/dt",m. maximal rate of diastolic area increase (em' /sec); b. constant of chamber stiffness (I/em'); dP/dt"",,, maximal rate of LV diastolic pressure decay (mm Hg/sec): T. constant of diastolic pressure decay (second).
*p < 0.05 tp
versus baseline.
< 0.01 versus baseline.
imal rate of pressure decay for the baseline and stimulation conditions. The 95% confidence band for the baseline is also presented. Each of the data points during LDM stimulation fall outside this 95% confidence band, demonstrating that, across the observed range of systolic pressure, the peak rate of pressure decay was depressed • with LDM stimulation. Discussion Dynamic cardiomyoplasty has been proposed as a surgical method to improve LV function in patients with
The Journal of Thoracic and Cardiovascular Surgery
I 6 6 8 Carin et al.
2400 Baseline
2200
Wrap. -Stirn Wrap. -Stim
Jj ...
,-..
2000
tJ
QI
rn
1800
tIl
1600
~
e
-lOCO
~
?d
'"~
1400
....e
1200
~
:s
.»
~
"tlI Left Ventricular Pressure (mrn Hg)
6 Baseline, R =0.86 • Wrap, +Stim, R= 0.76
70
Fig. 6. Relationship of LV pressureand dPj dt are presented for a baseline beat and for beats after LDM wrap both with (+Stim) and without (-Stirn) musclestimulation. The segment of isovolumic relaxationoccurson the lowerleft-handportionof the curves,beginning at minimumdPjdt and continuing to the left and up as both pressure and dPjdt decline. Compared to baseline, the rate of pressuredecay was depressed throughout diastole after the LDM wrap.
congestive heart failure. Some clinical studies have demonstrated improved systolic function, with regard to ejection fraction and stroke work, after cardiomyoplasty,6,8,11-13,21,22 but this is not a universal finding. 13, 23-25 Similarly, some experimental studies have shown improved LV systolic performance.F' although others have shown only a minimal benefit. 27,28 For the first 2 weeks after the heart is surgically wrapped, the LDM remains unstimulated to permit skeletal muscle revascularization.f-L' During the next 2 weeks, LDM training begins with only minimal stimulation (a single muscle twitch on every other heartbeat). As a result, there is minimal systolic assistance during the first 4 postoperative weeks of dynamic cardiomyoplasty.9,29 The significance of systolic impairment is widely recognized in patients with congestive heart failure. However, diastolic dysfunction also contributes to the clinical signs and symptoms of dyspnea and reduced exercise tolerance. I 4 Thus the effects of dynamic cardiomyoplasty on diastolic function are important to the clinical outcome of this procedure. Few previous studies have evaluated the effect of chronic cardiomyoplasty on diastole, and none has investigated the acute effects. The present study examined diastolic function immediately after cardiomyoplasty and is the only investigation to examine diastolic function acutely after surgical procedures. We compared diastolic
-o-_--+_----l
400+80
90
100
110
-+-_ _ '120
130
140
Peak Systolic Pressure (mm Hg)
150
Fig. 7. The significant inverse relationship between peak systolic pressureand maximal rate of pressuredecayis confirmed for baseline and after LDM wrap with stimulation. The 95% confidence band for baseline data across the range of LV systolic pressure is presented. Each of the data points for LDM stimulation fell below the confidence band, indicating a depressed rate of pressuredecay.
function at three stages: without an LDM wrap, with the wrap alone, and with LDM stimulation. The results of this study suggest that dynamic cardiomyoplasty causes no gross abnormalities in LV diastolic function acutely. LV end-diastolic pressure and area were in the physiologic range, and there was no evidence of ventricular constriction. However, sensitive indexes of diastolic function demonstrated a decline in the rate of diastolic pressure decay, a prolongation of the pressure relaxation constant, and an increase in passive chamber stiffness. Furthermore, stimulation of the LDM wrap did not improve LV diastolic function compared with the diastolic function with nonstimulated beats. The peak rate of isovolumic pressure decay is dependent on several factors, including systolic pressure, heart rate, and LV contractility. 3D The peak systolic pressure was similar before and after LDM wrap. As demonstrated in Fig. 7, there was a close correlation between peak systolic pressure and peak pressure decay, both at baseline and after LDM with stimulation. The data demonstrate the significant prolongation of pressure decay immediately after cardiomyoplasty. Although atrial pacing was not performed, the heart rate at baseline was similar to the rate after surgical intervention and therefore would not be expected to have caused the observed changes. The constant of isovolumic pressure decay is not
Volume 104 Number 6 December 1992
affected by changes in ventricular load or heart rate;'! but it changes with inotropic stimulation, myocardial ischemia, and hypertrophy.Vv" Inotropic medication was not administered, and there was no evidence of myocardial ischemia during the investigation. Therefore, the present results are likely to be a reflection of prolonged relaxation associated with cardiomyoplasty, and not secondary to changes in other hemodynamic parameters. Passive chamber stiffness significantly increased after LDM wrap but was not elevated further with muscle stimulation. Increased chamber stiffness has been demonstrated clinically to be associated with LV hypertrophy'? and fibrosis.P In the present study, after LDM wrap there was a significant increase in muscle mass surrounding the LV cavity. The mechanical situation after cardiomyoplasty may therefore be analogous to ventricular hypertrophy, in which LV wall thickness has increased. In this study, done immediately after the surgical procedure, there was insufficient time for either myocardial or skeletal muscle fibrosis to occur. There are several mechanisms that may contribute to these observations, including LDM mechanical restraint of the left ventricle and an increase in muscle mass around the LV chamber. In this study, the increase in diastolic function was unlikely to have been caused by a volume infusion, which was limited to less than 100ml during the period of evaluation of diastolic function. In this study the LDM was wrapped as previously reported from this laboratory I 5 and in clinical and experimental reports.f Although fiber orientation can be replicated easily, the tightness of fit of the wrap around the heart is more variable. For the LDM to provide optimal mechanical assistance to the heart, the skeletal muscle must be preloaded; that is, the muscle needs to be stretched during diastole. Previous investigators have described methods to prevent the wrap from impeding LV filling,12 but they did not evaluate skeletal muscle lengthening or LV filling immediately after surgery. In the current study care was taken not to overstretch the LDM,12 but our results are consistent with a diastolic restraining effect of the LDM wrap on the left ventricle. Immediately after surgical intervention, the LDM cannot fully adapt to the heart size, and the tight wrap has a deleterious effect on diastolic function. In addition, the skeletal muscle wrap is attached to the heart with sutures; these points of attachment might restrain diastolic filling. Thus the physical contact between the muscles and constraint of the left ventricle by the wrap may partially explain the results of the present investigation. Clinically, in concentric hypertrophy caused by systemic hypertension or aortic stenosis, LV diastolic pressure is substantially increased relative to the diastolic volume.l" In addition, passive stiffness increases, the peak
Cardiomyoplasty impairs diastolic function
16 6 9
rate of pressure decay is reduced, and the constant of pressure relaxation is prolonged, even in patients with normal systolic function.P 36 The diastolic dysfunction observed immediately after cardiomyoplasty might also be due to the increase in muscle mass around the left ventricle. With the application of skeletal muscle around the heart, the amount of muscle surrounding the LV chamber increased. We observed conditions similar to those associated with LV hypertrophy: an increase in passive stiffness, a decline in the peak rate of pressure decay, and prolonged relaxation. Comparison with literature. Two previous studies have examined diastolic function in a chronic model of cardiomyoplasty. To our knowledge, the present study is the first to evaluate the effect of cardiomyoplasty on passive diastolic function. In a study from this laboratory, radionuclide ventriculography was used to obtain images of the left ventricles of dogs with doxorubicin (Adriamycinj-induced congestive failure and an LDM wrap.P After a IO-weektraining period ofthe skeletal muscle, LV function during continuous LDM stimulation was compared with function without stimulation. The results showed an increase in the peak diastolic filling rate during synchronous latissimus stimulation compared with the filling rate without stimulation. In a preliminary report, Lee and colleagues'? evaluated LV diastolic function in dogs with normal systolic function after 6 weeks of LDM training. Using two-dimensional echocardiography, they found that the LV peak filling rate was increased during LDM stimulation compared with the filling rate without stimulation. The diastolic filling rate of the wrapped left ventricle was not compared with the unwrapped chamber in either of these studies. It has been stated that LV depression is a prerequisite for dynamic cardiomyoplasty to augment systolic function." Augmented diastolic filling was demonstrated in both normal and depressed ventricles in chronic models during LDM stimulation. In the present study of the acute aspects of cardiomyoplasty, LV systolic pressure increased with latissimus stimulation. These findings are consistent with previous studies showing an improvement in systolic function. I 2 In contrast, diastolic function in the normal LV was impaired immediately after cardiomyopIasty. Limitations. Differences between the present study and clinical practice may limit extrapolation of the shortterm results of clinical cardiomyoplasty. The present study investigated the acute effects on diastolic function with use of a canine model in which baseline LV function was normal. Dynamic cardiomyoplasty is performed on patients with congestive heart failure that is characterized by depressed LV systolic and diastolic function. The effect of the LDM wrap on a chronically dilated, dysfunctional
I 6 70
Carin et at.
left ventricle may differ from its effect on a normal ventricle. The conditioned skeletal muscle has different characteristics than the nonconditioned muscle. In the present study, the LDM was not conditioned before stimulation. This contrasts with the clinical setting, in which the LDM will be conditioned during the first 3 postoperative months. LV diastolic function was evaluated after stimulation during a single heartbeat. Clinically, 6 weeks after the operation, the muscle wrap is stimulated with every heartbeat. Clinical relevance. Current treatment options for chronic heart failure have significant limitations. Medical therapy carries a significantly high rate of mortality and morbidity, and cardiac transplantation is limited by the shortage of available donor hearts. Dynamic cardiomyopia sty has been proposed as alternative therapy for these patients. Improvement in LV systolic function has been demonstrated in some early studies of dynamic cardiomyoplasty, but its optimal application is not known. In patients with congestive failure, LV diastolic dysfunction contributes to the clinical morbidity. Little is known about the effects of dynamic cardiomyoplasty on diastolic function before the institution of significant systolic assistance. The results of this study suggest that shortly after cardiomyoplasty, diastolic function may be impaired. We acknowledge the superb secretarial assistance of Ms. Jan Merolli and technical support of John Michele and Gioia DiLoreto. REFERENCES I. Furberg CD, Yusuf S. Effect of vasodilators on survival in congestive heart failure. Am J Cardiol 1985;55:1109-13. 2. Smith WM. Epidemiology of congestive heart failure. Am J Cardiol 1985;55:3A-8A. 3. CONSENSUS Trial Study Group. Effects of enalapril on mortality in severe congestive heart failure. N Engl J Med 1987;316:1429-35. 4. SOLVD Investigators. Effect of enalapril on survival in patients with reduced left ventricular ejection fractions and congestive heart failure. N Engl J Med 1991;325:293-302. 5. Cohn IN, Johnson G, Ziesche S, et al. A comparison of enalapril with hydralazine-isosorbide dinitrate in the treatment of chronic congestive heart failure. N Engl J Med 1991;325:303-10. 6. Carpentier A, Chachques .IC, Myocardial substitution with a stimulated skeletal muscle: first successful clinical case [letter]. Lancet 8840:1267,1985. 7. Carpentier A, Chachques .IC, The use of stimulated skeletal muscle to replace diseased human heart muscle. In: Chiu RC-J, ed. Biomechanical cardiac assist. Mount Kisco, New York: Futura Publishing Co., Inc., 1986:85-102. 8. Chachques JC, Grandjean P, Schwartz K, et al. Effect of
The Journal of Thoracic and Cardiovascular Surgery
latissimus dorsi dynamic cardiomyoplasty on ventricular function. Circulation 1988;78(Pt 2):III203-16. 9. Chachques JC, Grandjean PA, Carpentier A. Latissimus dorsi dynamic cardiomyoplasty. Ann Thorac Surg 1989;47:600-4. 10. Sink JD, Soberman MS. Latissimus dorsi dynamic cardiomyoplasty: a procedure in evolution. Emory Univ J Med 1988;2:223-32. II. Molteni L, Almada H, Ferreira R. Synchronously stimulated skeletal muscle graft for left ventricular assistance. J THORAC CARDIOVASC SURG 1989;97:439-46. 12. Lee DF, Dignan RJ, Parmer JM, et al. Effects of dynamic cardiomyoplastyon left ventricular performance and myocardial mechanics in dilated cardiomyopathy. J THORAC CAROlOVASC SURG 1991;102:124-31. 13. Hagege AA. Desnos M, Chachques JC, et al. Preliminary report: follow-up after dynamic cardiomyoplasty. Lancet 1990;335:1122-4. 14. Gaasch WHo Diastolic mechanism in heart failure. Heart Failure 1985;1:195-202. 15. Cheng W, Justicz AG, Soberman MS, Alazrake NP, Santamore WP, Sink JD. The effect of dynamic cardiomyoplasty on left ventricular systolic and diastolic functions in a canine model of chronic heart failure. J THORAC CAROlOVASC SURG 1992;103:1207-13. 16. Grossman W, Stefadourous MA, McLaurin LP, Rolett EL, Young DT. Quantitative assessment of left ventricular diastolic stiffness in man. Circulation 1973;47:567-74. 17. Marble AE, McIntyre CM, Hastings-James R, Hor CWo A comparison of digital algorithms used in computing the derivative of left ventricular pressure. IEEE Trans Biomed Eng 1981;28:524-8. 18. Eichhorn P, Grimm J, Koch R, Hess OM, Carroll J, Krayenbuehl HP. Left ventricular relaxation in patients with left ventricular hypertrophy secondary to aortic valve disease. Circulation 1982;77:1395-404. 19. Corin WJ, Murakami T, Monrad ES, Hess OM, Krayenbuehl HP. Left ventricular passive diastolic properties in chronic mitral regurgitation. Circulation 1991;83:797-807. 20. Slinker BK, Goto Y, LeWinter MM. Systolic direct ventricular interaction affects left ventricular contraction and relaxation in the intact dog circulation. Circ Res 1989;65:307-15. 21. Moreira LFP, Stolf NAG, Jatene AD. Benefits of cardiomyoplasty for dilated cardiomyopathy. Semin Thorac Cardiovasc Surg 1991;3:140-4. 22. Lee KF, Dyke CM, Dignan RJ, et al. Effects of dynamic cardiomyoplastyon left ventricular dynamic geometry in a canine model of dilated cardiomyopathy. Surg Forum 1990;41:263-5. 23. Chiu RCJ. Biomechanical cardiac assist: quo vadimus? Semin Thorac Cardiovasc Surg 1991;3:160-3. 24. Hill A, Chiu RC-J. Dynamic cardiomyoplasty for treatment of heart failure. Clin Cardiol 1989;12:681-8. 25. Magovern JA, Furnary AP, Christlieb IY, Kao RL, Park SB, Magovern GJ. Indications and risk analysis for clinical cardiomyoplasty. Semin Thorac Cardiovasc Surg 1991; 3:145-8.
Volume 104 Number 6 December 1992
26. Moreira LFP, Stolf NAG, Bocchi EA, et al. Latissimus dorsi cardiomyoplasty in the treatment of patients with dilated cardiomyopathy. Circulation 1990;82(Pt 2):IV25763. 27. Anderson WA, Anderson JS, Acker MA, et al. Skeletal muscle grafts applied to the heart. Circulation 1988;78:(Pt 2):III 180-90. 28. Kao RL, Christlieb IY, Magovern GJ, ParkSB, Magovern GJ Jr. The importance of skeletal muscle fiber orientation for dynamic cardiomyoplasty. J THORAC CARDIOVASC SURG 1990;99:134-40. 29. Moreira LFD, Stolf NAG, Jatene AD. Hemodynamic benefits of cardiomyoplasty in clinical and experimental myocardial dysfunction. In: Chiu RCJ, Bourgeois 1M, eds. Transformed muscle for cardiac assist and repair. Mt. Kisco, NY: Futura, 1990:179-88. 30. Weisfeldt ML, Scully HE, Frederiksen J, et al. Hemodynamic determinants of maximum negative dP / dt and periods of diastole. Am J Physiol 1974;277:613-21. 31. Weiss JL, Frederiksen JW, Weisfeldt ML. Hemodynamic determinants of the time-course of fall in canine left ventricular pressure. J Clin Invest 1976;58:751-60.
Cardiomyoplasty impairs diastolic function
16 7 1
32. Carroll JD, Lang RM, Neumann AL, BorowKM, Rajfer SI. The differential effects of positive inotropic and vasodilator therapy on diastolic properties in patients with congestive cardiomyopathy. Circulation 1986;74:815-25. 33. Eichhorn P, Grimm J, Koch R, Hess 0, Carroll J, Krayenbuehl HP. Left ventricular relaxation in patients with left ventricular hypertrophy secondary to aortic valve disease. Circulation 1982;65:1395-404. 34. Hirota Y. A clinical study of left ventricular relaxation. Circulation 1980;62:756-63. 35. Hess OM, Ritter M, Schneider J, Grimm J, Turina M, Krayenbuehl HP. Diastolic stiffness and myocardial structure in aortic valve disease before and after valve replacement. Circulation 1984;69:855-65. 36. Lorell BH, Grossman W. Cardiac hypertrophy: the consequences for diastole. J Am Coli CardioI1987;9:1189-93. 37. Lee KF, Parmar 1M, Nixon lV, et al. Effect of dynamic cardiomyoplastyon left ventricular diastolic chamber function. Circulation 1990;82(Pt 2):III712.