Reverse Remodeling From Cardiomyoplasty in Human Heart Failure
External Constraint Versus Active Assist
Background Cardiomyoplasty (CM) is a novel surgical therapy for dilated cardiomyopathy. In this procedure, the latissimus dorsi muscle is wrapped around the heart and chronically paced synchronously with ventricular systole. While studies have found symptomatic improvement from this therapy, the mechanisms by which CM confers benefit remain uncertain. This study sought to better define these mechanisms by means of serial pressure-volume relation analysis.
Methods and Results Serial pressure-volume studies were performed by the conductance catheter method in three patients (total to date) with dilated cardiomyopathy (New York Heart Association class III) who underwent CM. Data were measured at baseline (before surgery) and at 6 and 12 months after CM. Chronic left ventricular (LV) systolic and diastolic changes induced by CM were evaluated with the stimulator in its stable pacing mode (every other beat) and after temporarily suspending pacing. CM-stimulated beats were compared with pacing-off beats to evaluate active systolic assist effects of CM. In each patient, CM resulted in a chronic lowering of cardiac end-diastolic volume and an increased ejection fraction. Most notably, the end-systolic pressure-volume relation shifted leftward, consistent with reversal of chronic chamber remodeling. In contrast, the diastolic pressure-volume relation was minimally altered, and the loops shifted down along the same baseline relation. These marked chronic changes in LV function measurable with CM stimulation off contrasted to only minor acute effects observed when the muscle wrap was activated. This suggests that the benefit of CM derived less from active systolic assist than from remodeling, perhaps because of an external elastic constraint.
Conclusions These data, while limited to a small number of patients, suggest that CM can reverse remodeling of the dilated failing heart. While systolic squeezing assist effects of CM may play a role in some patients, our study found that this was not required to achieve substantial benefits from the procedure. We speculate that CM may act more passively, like an elastic girdle around the heart, to help reverse chamber remodeling.
Cardiomyoplasty (CM) is a novel surgical treatment for patients with dilated heart failure.1 2 3 4 5 6 In this procedure, the sheetlike latissimus dorsi muscle is mobilized (while maintaining its vascular supply) and inserted into the mediastinum through a lateral thoracotomy. After pericardiectomy, the muscle is wrapped around both the left and right ventricles. Several weeks are provided for surgical recovery and electrical muscle conditioning, and thereafter the muscle is paced by burst stimulation synchronized to every other cardiac systole. Chronic repetitive stimulation induces biochemical and physiological transformations in the muscle, altering its characteristics toward those of cardiac muscle.7 8 9 These changes include fatigue resistance, prolonged contraction duration, diminution in size, and reduced maximal power.7 8 9 10
The principal mechanism by which CM has been assumed to assist the failing heart is augmentation of systolic ejection by active squeezing of the ventricles. However, while human studies have reported improvement in the clinical symptoms of patients undergoing this therapy,2 3 5 6 evidence for active systolic assist has been inconsistent.10 Some of the variability may stem from the traditional cardiac imaging methods used in most studies, since heart motion and shape changes from CM can limit the reliability of such data. Nonetheless, the results have led some to speculate that more passive external constraining effects of the muscle wrap may be a source of benefit.10 11
To better evaluate the mechanical consequences of CM, we performed serial left ventricular (LV) pressure-volume analysis12 13 14 in all patients undergoing CM at our institution. Data were measured before surgery and at 6 and 12 months after CM. To evaluate chronic changes in cardiac function, data were obtained with the myostimulator temporarily turned off. Acute systolic assist effects of CM were measured by applying stimulation to alternate beats. This brief report describes results from all three patients who have undergone this protocol thus far.
Three patients with idiopathic dilated cardiomyopathy (IDCM) were studied, comprising our total experience to date with clinical cardiomyoplasty. Patients had minimal to no coronary artery or valvular heart disease. A brief clinical summary of each patient follows.
Patient A was a previously healthy and active 32-year-old man, diagnosed with IDCM with pulmonary congestion and hypotension in December 1991. Echocardiography revealed an LV diastolic dimension of 6.5 cm. He was treated medically for 6 months with only moderate symptomatic improvement, remaining in New York Heart Association class III. Due to a history of temporal lobe seizures, he was not considered a candidate for heart transplantation and underwent CM in May 1992.
Patient B, a 62-year-old man, was first diagnosed with IDCM in November 1991, requiring hospitalization for intravenous inotropic and diuretic therapy. LV diastolic dimension was 7.0 cm by echocardiography. Despite medical therapy, his functional status remained class III, and he underwent CM in December 1992.
Patient C, a 55-year-old man, was diagnosed with IDCM in 1992. His echocardiographic diastolic dimension was 7.2 cm. He had required repeat hospitalizations for recurrent atrial fibrillation and pulmonary congestion and remained in functional class III despite cardioversion to normal sinus rhythm and maximal medical therapy. He underwent CM in June 1993. Several episodes of atrial fibrillation occurred after the CM, and these were also successfully treated by cardioversion. Each episode was treated shortly after the onset of the arrhythmia. All data reported for this patient were obtained in normal sinus rhythm.
All patients were receiving chronic treatment with a diuretic (furosemide or bumetanide), digoxin, and an angiotensin-converting enzyme inhibitor (enalapril or captopril) at the time of CM. Patients B and C also were treated with amiodarone for ventricular and supraventricular arrhythmias. With the exception of an increased amiodarone dose in patient C for atrial fibrillation, doses of the other medications were either unchanged or reduced after CM conditioning. None of the patients had a diagnosis of alcohol or other substance abuse. Other potentially correctable factors (for example, active myocarditis, ischemic heart disease, valvular disease, or metabolic disease) were ruled out at the time of entry into the protocol.
After initial baseline study, each patient underwent a left muscle wrap procedure, using the technique of Carpentier et al.1 Two weeks were provided for surgical recovery, and this was followed by a 3-month stimulation training period, during which time the voltage output of the myostimulator was gradually increased. After this period, the cardiomyostimulator (SP1005, Medtronic, Inc) was set to deliver six 210-millisecond pulses spaced within 185 milliseconds, synchronized with early systole of every other contraction.
Cardiac catheterization studies were performed just before and 6 and 12 months after CM in each patient. Catheterization included a routine right heart study, from which cardiac output, right atrial, and pulmonary capillary wedge pressures were measured. Left heart study included contrast ventriculography and conductance catheter study to assess LV pressure-volume relations.12 13 14 The volume catheter signal was calibrated separately at each observation time point of the study. Catheter signal end-diastolic and end-systolic volumes were set equal to EDVcath=(COtd/HR)/EFlvg and ESVcath=EDVcath−(COtd/HR), where EFlvg and COtd were ejection fraction and cardiac output derived from contrast ventriculogram and thermodilution catheter, respectively.
Pressure-volume data and relations were first determined at resting preload volume with the myostimulator on in its stable pacing mode (CM activation every other beat). The myostimulator was then acutely turned off for several minutes, and the data collection was repeated. In this nonstimulated mode, preload was then transiently reduced by obstruction of inferior vena caval blood return to assess end-systolic and diastolic pressure-volume relations (ESPVR and DPVR, respectively).12 13 14 End-systolic pressure-volume data were fit by perpendicular regression to derive the ESPVR, using an iterative method to determine the volume axis intercept.13 This relation provided an index of systolic chamber function. The DPVR was derived by a monoexponential curve fit of late diastolic data from these variably preloaded cardiac cycles.12 This provided a measure of chamber passive properties. To examine active cardiac assist effects of CM stimulation, alternate beats during CM stimulation were averaged to yield two mean cardiac cycles (an “on” and “off” beat, respectively). These data were further compared with those obtained with myostimulation acutely suspended. In addition to the cardiac catheterization data, exercise capacity was assessed by upright bicycle ergometry with metabolic testing. The maximal oxygen consumption achieved at peak exertion is reported. The adequacy of each exercise test was assured by confirming that each patient exceeded the anaerobic threshold during the test.
Table 1⇓ provides hemodynamic data for the three patients at each time point. All data were measured at normal sinus rhythm. For 6- and 12-month post-CM follow-up, data measured both with the myostimulator on in its chronic configuration (stimulating every other beat) and after acutely suspending stimulation (numbers in parentheses) are provided. Cardiac index and wedge pressure were only determined initially with the stimulator on. Pulmonary capillary wedge pressure, LV end-diastolic pressure and volume, and right atrial pressure all declined after CM in each patient. Acute cessation of myostimulation minimally influenced these cardiac pressures and volumes. Ejection fraction and exercise capacity, the latter indexed by maximal oxygen consumption, both rose after CM. Cardiac index changed minimally in patients A and C, whereas it declined somewhat in patient B.
Fig 1⇓ displays LV pressure-volume loops and end-systolic and diastolic pressure-volume relations for each patient at baseline (upper panels) and 1 year after CM (lower panels). The data were measured with myostimulation acutely suspended to assess chronic changes in chamber function. Pressure-volume relations measured before CM are reproduced in each respective lower panel to facilitate comparisons.
Two important changes were observed in each patient. The first was that chamber volumes declined with a leftward shift of the pressure-volume loops. Diastolic pressures also fell, yet the pre- and post-CM diastolic pressure-volume relations (lower pressure-volume loop boundary) remained very similar to baseline. The decline in diastolic pressure and leftward shift of the data were largely established at 6 months after CM (see Table 1⇑), with only modest further changes after 1 year.
The second major finding was a leftward shift in the ESPVR after CM. This shift was more consistent and prominent than a slope change in the relation. Individual ESPVR slopes and volume offsets for each patient are provided in Table 2⇓. The volume axis position was quantified by determining the end-systolic volume along the ESPVR at a common end-systolic pressure of 120 mm Hg (Ves120). As shown in Fig 1⇑, the most consistent change among the three patients was a progressive decline in Ves120, which fell between 22% to 42% of baseline among the patients after 1 year of CM.
To examine the direct effect of acute CM activation on systolic and diastolic function, pressure-volume data were also obtained with myostimulation applied to every other beat. This is the normal pattern of chronic CM stimulation maintained for each patient. Alternate contractions were averaged to generate a mean response for the on versus off conditions. Both sets of data were compared with a third loop measured with the stimulator fully off. In contrast to the chronic changes demonstrated without myostimulation in Fig 1⇑, acute “squeezing” effects from CM activation appeared to be minimal. Skeletal muscle stimulation produced slight lateral displacements of the heart (being tugged toward the left shoulder) in two of the patients (B and C). In patient A, we did not observe any extra motion of the heart during myostimulation or even subtle fluctuations in right heart pressures that would indicate active CM contraction. Most significantly, myostimulation had minimal effects on the pressure-volume data. Fig 2⇓ displays the three different steady-state pressure-volume loops for each patient, and in each instance, the loops were very similar.
Cardiomyoplasty was not initially conceived as a means to reverse chamber dilation in heart failure but rather as a method to mechanically assist the heart during ejection. Over time, this could improve cardiac function and perhaps lead to a decline in chamber volume. While admittedly limited, our experience to date suggests a somewhat different view. While there was evidence for substantial reversal of chronic chamber remodeling associated with systolic functional improvement, this was achieved in each patient with minimal indication that the CM wrap itself provided active systolic assist.
Chronic dilation and remodeling are valuable adaptations that allow the weakened heart to achieve near-normal systolic pressures and flows at increased but still tolerable diastolic pressures. These changes are manifest by rightward shifts in diastolic and systolic pressure-volume relations. While initially adaptive, continued chamber dilation ultimately limits cardiac reserve and is a major risk factor for mortality.15 16 Recent clinical trials in patients with dilated cardiomyopathy have reported that angiotensin-converting enzyme inhibitors can attenuate this progressive cardiac enlargement and improve survival.16
The present data suggest that CM can similarly reverse chronic chamber remodeling. The leftward shift of the ESPVR observed in each patient reflected this change. This shift was critical to enabling the post-CM ventricle to generate similar cardiac outputs and systolic pressures to baseline but at substantially reduced diastolic filling. While the leftward ESPVR shift no doubt indicated an improvement in LV systolic pump performance associated with chamber remodeling, interpretations of the slope changes are less clear. Altered chamber geometry and wall stress resulting from the muscle wrap could themselves alter the ESPVR slope without requiring underlying changes in myocardial contractility.
In contrast to the marked chronic changes in LV chamber function from CM, we found minimal evidence of acute systolic assist from myostimulation. While our sample size was clearly too small to draw firm conclusions, the results did show that squeezing the heart to enhance ejection was not mandatory for achieving a substantial benefit. We speculate that the muscle wrap provides an elastic constraint to the epicardial surface, limiting cardiac dilation much like a girdle. Unlike the pericardium, which is highly compliant when understretched but becomes very stiff when expanded beyond rest volumes, skeletal muscle stiffens more gradually as it is lengthened.17 Thus, the CM may better accommodate enhanced cardiac filling when this is called for, without markedly increasing diastolic pressures. Chronic repetitive stimulation reduces muscle fiber caliber,18 and this may help to gradually shrink the CM around the heart. Such a constraining action would not necessarily increase diastolic pressures so long as the muscle wrap remained compliant over its operating dimensions. This hypothesis is consistent with our observations that the DPVRs were altered little after CM.
The notion that an external muscle wrap might provide an elastic constraint without inducing constrictive physiology is supported by a recent experimental study.19 Capouya et al19 applied a nonstimulated muscle wrap to the canine heart and then rapidly paced the heart to induce dilated cardiomyopathy. The previously wrapped hearts developed less chamber enlargement and less decline in ejection fraction. Whether skeletal muscle itself is required or whether an artificial elastic “sock” placed around the heart could achieve similar effects remains an intriguing question.
It should be noted that while the shift in ESPVR position most likely reflected reverse remodeling effects of CM, the exact placement of the loops along the volume axis and DPVR depended on other factors as well. Variations in hydration status at the time of catheterization, for example, would also contribute. This might have been pertinent to patient B, who displayed low filling pressures at 1 year after CM, along with somewhat reduced cardiac output. This patient has continued to display low filling pressures (right atrial pressure, 2 mm Hg; pulmonary capillary wedge pressure, 7.6 mm Hg) 2 years after CM, however, suggesting that this was not just concurrent overdiuresis. Most importantly, the left shift of the ESPVR would not be explainable on the basis of temporary variations in volume loading.
Careful patient selection probably contributed to the CM results achieved in the present study. Prior investigations have demonstrated that New York Heart Association class IV patients and those with very dilated ventricles generally do poorly after the myoplasty procedure.4 5 10 11 First, these patients suffer increased morbidity and mortality from the surgery itself. Second, extremely dilated hearts are often difficult to adequately wrap with the skeletal muscle, and excessive strain on the muscle may render it more ischemic. Last, the end-stage failing heart with greater myocyte loss and replacement fibrosis may be less capable of reverse chamber remodeling compared with a heart in the earlier stages of disease. Previous clinical CM studies often included more severe class IV patients,2 5 6 and this may help to explain the overall lack of volume changes reported.
Firm conclusions cannot be drawn from data obtained in only three patients. However, the present results for both ESPVR and DPVR display a remarkable consistency among the patients, suggesting a real effect of CM. Based on this admittedly limited sample, we speculate that the muscle wrap can provide an elastic constraining effect to the heart, which in turn can partially reverse chronic chamber remodeling of heart failure. Active systolic assist also may contribute to improved cardiac function and ejection in some patients, but this does not appear to be a requirement for achieving benefit from CM. The present data also suggest a novel direction for heart failure therapy, based on the premise that chronic reversal of chamber dilation may be achieved by application of a dynamic external constraint.
This study was supported by NIH-NCRR, Osler 5 GCRC grant RR00035, Public Health Service grant HL-47511, and an AHA Established Investigator Award (D.A.K.).
- Received December 7, 1994.
- Revision received February 2, 1995.
- Accepted February 8, 1995.
- Copyright © 1995 by American Heart Association
Moriera LFP, Stolf NAG, Bocchi EA, Pereira-Barretto 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.
Moreira LF, Seferian PJ, Bocchi EA, Pêgo-Fernandes PM, Stolf NAG, Pereira-Barretto AC, Jatene AD. Survival improvement with dynamic cardiomyoplasty in patients with dilated cardiomyopathy. Circulation. 1991;84(suppl III):III-296-III-302.
Mannion JD, Bitto T, Hammond RL, Rubinstein NA, Stephenson LW. Histochemical and fatigue characteristics of conditioned canine latissimus dorsi muscle. Circ Res. 1986;58:298-304.
Salmons S, Sreter FA. Significance of impulse activity in the transformation of skeletal muscle type. Nature. 1967;263:30-34.
Sreter FA, Gergely J, Salmons S, Romanul F. Synthesis by fast muscle of myosin light chains characteristic of slow muscle in response to long-term stimulation. Nature. 1973;241:17-19.
El Oakley RM, Jarfis JC. Cardiomyoplasty: a critical review of experimental and clinical results. Circulation. 1994;90:2085-2090.
Chachques JC, Grandjean P, Schwartz K, Mihaileanu S, Fardeau M, Swynghedaum B, Fontaliran F, Romero N, Wisnewsky C, Perier P, Chauvaud S, Bourgeois I, Carpentier A. Effect of latissimus dorsi dynamic cardiomyoplasty on ventricular function. Circulation. 1988;78(suppl III):III-203-III-216.
Kass DA, Midei M, Brinker J, Maughan WL. Influence of coronary occlusion during PTCA on end-systolic and end-diastolic pressure-volume relations in man. Circulation. 1990;81:447-460.
Konstam MA, Rousseau MF, Kronenberg MV, Udelson JE, Melin J, Stewart D, Dolan N, Edens TR, Ahn S, Howe DM, Kilcoyne L, Metherall J, Benedict C, Yusuf S, Pouleur H, SOLVD Investigators. Effects of the angiotensin-converting enzyme inhibitor enalapril on the long-term progression of left ventricular dysfunction in patients with heart failure. Circulation. 1992;86:431-438.
Strumpf RK, Humphrey JD, Yin FCP. Biaxial mechanical properties of passive and tetanized canine diaphragm. Am J Physiol. 1993;265:H469-H475.