C Cardiomyoplasty reviewed: Lessons from the past, prospects for the future

(1)

Cardiomyoplasty reviewed: Lessons from the past, prospects for the future

Stanley Salmons, Jonathan C. Jarvis

Muscle Research Group, Department of Human Anatomy and Cell Biology, University of Liverpool, U.K.

Abstract

Cardiac assistance from autologous skeletal muscle offers a practical surgical solution to end-stage heart failure, one that avoids the problems of donor shortage and immune rejection associated with transplanted organs. Dynamic cardiomyoplasty was an implementation of this approach in which skeletal muscle was wrapped around the host heart. This technique received enthusiastic clinical application for a limited period, yet is now widely regarded as a failed approach. This is an inaccurate assessment, for many patients found the procedure to be beneficial. It is clear in retrospect that clinical outcomes could have been better had the technique been implemented with a more complete knowledge of the underlying basic principles. Here we seek to explain in scientific terms the benefits and shortcomings of cardiomyoplasty, as it was practised in the past, and as it could be practised in the future. It will not be easy to change current perceptions, but the place of cardiomyoplasty in the surgical armamentarium deserves to be reconsidered.

Key Words: cardiac assist, cardiomyoplasty, heart failure, vascular delay, prestimulation, conditioning

Basic Applied Myology 19 (1): 5-16, 2009

C

ardiac transplantation currently offers the best long- term solution to the problem of end-stage heart failure, but the decline in the availability of donor hearts has created a need for alternatives that has yet to be met.

Mechanical pumps of various types have evolved from their initial role as short-term, bridge-to-transplant devices to longer-term application. However, these are expensive, they carry risks of haemolysis and thrombogenesis, and other than for brief periods they encumber the patient with an external power source.

Skeletal muscle is an endogenous source of power. It efficiently converts chemical energy, derived ultimately from the normal intake of food and oxygen, into mechanical work which can, in principle, be harnessed in various ways to assist a failing circulation. The concept embodies amplification: a large amount of mechanical energy is released by the small amount needed to stimulate the motor nerve [62]. The only power source required is a conventional pacemaker-type battery for driving a suitably synchronized neuromuscular stimulator, and such a device could last between 5 and 10 years.

Cardiac assistance from skeletal muscle has a number of other advantages. Unlike transplantation it does not carry the risks, costs and deleterious side-effects of lifelong immunosuppressive therapy. A donor patient (or animal) is not needed and the patient's own heart is

retained under conditions that offer some potential for myocardial recovery. This is an affordable technology that could be considered in parts of the world where other approaches, including mechanical pumps, would be too costly.

Historical background

There is an early history, dating back almost to the beginning of the twentieth century, of the use of skeletal muscle as a passive surgical biomaterial. Around 1960, the first attempts were made to utilize the active properties of muscle in cardiac assistance [35, 36, 52].

These and subsequent pioneering attempts were defeated by the problem of muscle fatigue. The solution to this problem lay in the discovery that skeletal muscles that are exposed to extended periods of increased use undergo functional adaptation [63, 65].

The changes, which develop most fully in response to chronic electrical stimulation, include a marked increase in fatigue resistance. Although an attempt was made to point out the clinical relevance of the discovery [57], it was several years before it was picked up by Dr J.

Macoviak, working in Dr L.W. Stephenson’s laboratory.

Stephenson, then at Philadelphia, and Salmons, then at Birmingham, UK, began a productive and long-lasting collaboration.

(2)

Basic Applied Myology 19 (1): 5-16, 2009

Like the earlier studies, the initial focus was on the use of the diaphragm, specifically as a means of enlarging a hypotrophic right ventricle. The novel feature of the work was the use of chronic electrical stimulation to induce fatigue-resistant characteristics, a procedure for which Stephenson coined the term

“conditioning” (not to be confused with ischaemic preconditioning). Macoviak and Stephenson mentioned the concept during a presentation at the International Society for Artificial Organs, which met in Paris in July 1981, and it attracted considerable interest. The work in Philadelphia sparked interest in other laboratories, leading to a series of mini-symposia and seminars in Philadelphia (hosted by L.W. Stephenson, 1982), Montreal (hosted by R.-C. J. Chiu, 1984) and Birmingham (hosted by S. Salmons, 1985).

By this time interest had switched from the diaphragm to the latissimus dorsi (LD) muscle. This large, flat sheet of muscle had a number of advantages over diaphragm: it was thicker, it was non-essential, and its blood supply appeared to be confined to a single vascular pedicle (although this assumption was later found to be unjustified, as discussed below). The disadvantage, that the muscle was located on the outside of the chest wall, could be overcome by “posting” the mobilized muscle between the ribs. Although other muscles have been investigated since—including rectus abdominis, pectoralis major, serratus anterior, and psoas major—the LD muscle remains the favoured choice.

John Mannion joined Stephenson’s group, and the idea evolved of winding the LD muscle around a mandrel to create an auxiliary heart pump, or skeletal muscle ventricle (SMV). The experimental basis for the technique was sufficiently well established by 1985 that the U.S. National Institutes of Health awarded a grant of nearly $1 million to Stephenson and Salmons to develop it further.

An SMV places a new surface in contact with the circulating blood, and a good deal of the early research focused on eliminating the risk of thrombus formation.

Meanwhile a more conservative solution emerged from Paris when Carpentier and Chachques announced the first successful case of dynamic cardiomyoplasty [9].

Medtronic Inc. met the need for an implantable electrical stimulator triggered by the R-wave of the patient’s ECG; this could deliver a train of impulses to the motor nerve to elicit a contraction from the muscle at the right moment of the cardiac cycle and for the appropriate duration. With the support of Medtronic, the technique began to be implemented clinically on a trial basis. Initially the LD muscle was used to substitute for myocardium that had been resected, but soon the muscle was being applied to the failing intact heart as a reinforcing wrap.

Cardiomyoplasty has the considerable advantage that the procedure does not place new, non-endothelial, surfaces in contact with the circulating blood. However,

it suffers from one serious limitation: the geometry of the wrap is dictated by the size and shape of the host heart. In such a situation, the muscle is constrained to operate far from the peak of its power curve [60, 61].

For this reason work continued in parallel on the SMV, which could be designed to harness the pumping power of the muscle more efficiently. It was found that thrombogenesis could be avoided if strict design criteria were observed to ensure adequate mixing and exchange [67]. An SMV pumped in circulation in a dog for 4 years and was thrombus-free and still providing substantial assist at elective termination [78]. SMVs proved to be capable of a large fraction of the pumping power of the normal native left ventricle [54] and could provide cardiac assistance that was at least equal to that of an intra-aortic balloon pump [55]. Another SMV configuration has been investigated by Guldner [26].

There has also been a steady development of hybrid devices, which use skeletal muscle as the power source for a mechanical pump [56, 80]. These have the advantage of allowing the muscle to contract in a normal, linear fashion, but major challenges remain: the overall efficiency of energy conversion, cost, haemocompatibility, connective tissue investment, and the risk of infection associated with multiple implanted devices.

There is little doubt that the clinical trialling of cardiomyoplasty helped to sustain interest in all forms of skeletal muscle assistance. Cardiomyoplasty was performed in over ~2000 cases worldwide and 80-85%

patients reported an improved quality of life. On the other hand 15–20% patients showed no benefit from the procedure and fewer than 60% of cardiomyoplasty patients survived for more than 2 years. Interest waned, Medtronic withdrew support for further operations “for commercial reasons”, and this was interpreted as a vote of no confidence in the technique. There is now a widespread perception that the approach failed to live up to its promise. Is this perception justified?

Preserving the Viability of the Muscle Graft Ischaemic damage

Evidence from both animals [2, 16, 23, 29, 43, 48] and human patients [19, 34, 50] suggests that the overall clinical outcome of cardiomyoplasty was adversely affected by deterioration of the LD muscle wrap.

Histological examination of the wrap after cardiomyoplasty revealed evidence of fatty infiltration, fibrosis, and muscle fibre atrophy [29, 30, 38, 42]. The proximal segment was the least affected, and the damage increased progressively towards the distal segment [3, 30, 34, 38].

Trauma to the thoracodorsal nerve due to compression, torsion, or stretching of the neurovascular pedicle, could result in denervation atrophy [42, 49].

However the proximodistal gradient of damage pointed rather to damage of ischaemic origin. In the sheep,

(3)

stimulation of the undisturbed LD muscle was not, on its own, damaging. Mobilizing the LD muscle and replacing it at reduced length, to mimic the loss of physiological resting tension in a cardiomyoplasty wrap, also resulted in little damage. However, when the same mobilization procedure was followed after 2 weeks with a conditioning pattern of stimulation, about half of the muscle fibres were replaced by fat [23]. These results were confirmed by others [3, 29]. Damage, then, was the consequence of subjecting the mobilized, and still partially ischaemic, muscle to the additional metabolic demands of stimulation.

To understand this properly it is necessary to examine the blood supply of the LD muscle. There are two major sources. The first is the thoracodorsal branch of the subscapular artery, which bifurcates or trifurcates at the neurovascular hilum to supply the muscle. This vessel and its branches are preserved when the muscle is mobilized as a pedicled graft. The second is a series of perforating arteries derived from the 9th,10th, and 11th posterior intercostal and 1st, 2nd, and 3rd lumbar arteries, which enter the muscle on its deep surface.

These vessels must be divided in order to mobilize the muscle from its costal, spinal, and pelvic attachments.

The traditional view—now very much out of date—is that the loss of these perforating arteries is not a matter for serious concern. In fact their contribution to blood flow is far from trivial: in dogs, for example, ligation of these perforators resulted in a 70% reduction in blood flow from baseline levels [12, 18]. Furthermore their vascular territory covers as much as two-thirds of the muscle area, including the middle and distal thirds which constitute the functional part of the LD wrap in cardiomyoplasty. This explains why mobilization of the muscle produces ischaemic damage, and thus compromises the potential of the wrap for producing cardiac assistance.

We tried to use a resin-casting technique to demonstrate the blood supply in the rat and the rabbit, but we noticed that injection of resin via either the thoracodorsal or the perforating arteries, with the other route clamped, always filled the entire vascular network [17]. These observations were consistent with earlier, cadaveric studies that pointed to the existence of anastomotic connections between the two vascular territories in this muscle [46, 53, 74, 77, 79]. With the use of fluorescent microspheres we showed that the blood flow due to the thoracodorsal artery decreased in a proximal-to-distal direction, but not to zero; that of the perforating arteries decreased in a distal-to-proximal direction, also not to zero. This implied that the two vascular territories were indeed linked by anastomotic bridges and we went on to obtain unequivocal evidence that they were present in the living organism under physiological conditions [64].

The calibre of the vessels involved is presumably insufficient to allow the middle and distal parts of the

grafted muscle to be well perfused by the intact thoracodorsal artery; this would be exacerbated by spasm and collapse of the vessels resulting from mobilization, cooling, and loss of normal resting tension. If, however, the muscle is stimulated in situ before raising it as a graft (“prestimulation”) the increase in blood flow, accompanied by dilation of existing vessels and formation of new ones, extends to the distal parts of the muscle [72, 73]. Thus in the sheep latissimus dorsi muscle, prestimulation abolished the proximodistal gradient of blood flow that is normally seen in the muscle after mobilization and reattachment [72], and within a few days the blood flow throughout the muscle had recovered to baseline levels [73] (Figure 1). Rabbit latissimus dorsi muscles that were prestimulated at 2.5 Hz for 10 weeks did not show the usual reduction in distal perfusion after they had been mobilized, and were better able to maintain tissue oxygenation during a fatigue test [5]. In rats the loss of viability resulting from mobilization was greatly reduced by prestimulation [83].

Fig. 1 Effect of simulated grafting (mobilization, cooling, and loss of normal resting tension) on sheep latissimus dorsi muscle. Blood flow, expressed as a percentage of values before the procedure (“baseline”), was measured by the fluorescent microspheres technique in proximal, middle, and distal segments of the muscle. Measurement was made immediately after mobilization (“acute”) and 5 days later.

In the unstimulated muscle, the blood flow shows the usual proximodistal gradient of flow.

Five days later, the gradient has largely disappeared but the mean flow is still less than half the baseline value. Prestimulation eliminates the proximodistal gradient even at the acute stage; 5 days later there is no gradient, and the blood flow throughout the muscle is not significantly different to baseline.

(4)

“True” vascular delay

A vascular delay procedure has been used for many years by plastic and reconstructive surgeons to improve survival of skeletal muscle flaps [8]. Vessels other than those in the vascular pedicle are divided and the muscle is allowed to remain in situ for a further period, typically 10-14 days. This procedure results in vasodilatation and rearrangement of the capillaries within the flap and enhances the mean and regional blood flow prior to redeployment. It has been applied experimentally in dogs to improve perfusion of the distal segment of the LD muscle during cardiomyoplasty [1, 11, 12, 31]. This “true” vascular delay should be distinguished from the so-called

“vascular delay” of the usual cardiomyoplasty protocol, in which 2 weeks is allowed to elapse between configuring the muscle wrap and initiating stimulation.

We measured the blood flow after mobilization in LD muscles that had been subjected to either prestimulation or true vascular delay. The techniques are somewhat different in principle: prestimulation is designed to enhance the vasculature associated with all sources, whereas true vascular delay is designed to enhance the vasculature associated with the single remaining blood supply. The reason why prestimulation is effective is that it allows the thoracodorsal artery to maintain continuity with the vascular tree of the perforating arteries, even after these have been divided. In fact the direct comparison between prestimulation and true vascular delay showed that both procedures ameliorated the fall in distal blood flow after mobilization, but that prestimulation had the greater effect of the two [82].

Prestimulation has a number of advantages. It is not wholly dependent on vascular neogenesis but enhances distal perfusion through existing vessels. It is potentially less invasive than true vascular delay, which calls for extensive dissection, and is therefore less likely to create adhesions that interfere with subsequent surgery.

Finally, a protocol incorporating prestimulation would also initiate conditioning at a pre-operative stage, preparing the muscle for cardiac patterns of work.

As it was practised, clinical cardiomyoplasty used a combination of a “vascular delay” and a gradually escalating conditioning protocol, which introduced an 8- week postoperative lag before the benefits of the procedure were fully available to the patients. The resultant high mortality could be overcome only by rigorous patient selection. Prestimulation enables such problems to be avoided, because a degree of assistance can be provided from the time of surgery.

Practical prestimulation

The least invasive method of implementing prestimulation would be through the use of electrodes placed on the skin surface. In practice this has drawbacks. It is unlikely that the patient would be able to tolerate the level of stimulation needed to recruit the

whole muscle. Reducing the amplitude to a more comfortable level would leave part of the muscle unrecruited. The corresponding muscle fibres would not then be conditioned and would be susceptible to fatigue and damage when the muscle graft was stimulated supramaximally in its final configuration.

A better course of action would be to implant a stimulator and electrode by a minimally invasive method. For ease of access, and to minimize the amount of dissection, a single epimysial electrode, with an insulated backing to prevent stimulus spread, could be placed just deep to the anterior border of the LD muscle near to where the thoracodorsal nerve branches, rather than close to the free portion of the nerve. The stimulator would be used in monopolar mode, the metal case being the indifferent electrode. There is, of course, no need for ECG-sensing electrodes at this stage, as stimulation is asynchronous. The LD tendon would still be attached to the humerus, so a burst pattern of stimulation would elicit a strong contraction which would move the upper arm. However, continuous low- frequency stimulation (e.g. 0.5 Hz) would have the same conditioning effect but deliver less forceful contractions, and stimulation could be turned off at night.

Avoiding 1:1 contraction

There is one final issue that needs to be addressed in the context of the viability of the wrap. When a muscle contracts strongly, its blood supply is arrested by the high intramuscular pressure. If the contractions are repeated too frequently there is insufficient time for relaxation and reperfusion. Van Doorn and her colleagues demonstrated a rise in venous lactate and reactive hyperaemia following cessation of stimulation when the LD muscle was made to contract more frequently than 1:2 with the heart rate. This is unequivocal evidence that a 1:1 régime of stimulation produces partially anaerobic, and therefore unsustainable, conditions [81]. Unfortunately some clinical centres implemented 1:1 régimes, and the damage they reported was taken as proof that graft survival would always be a problem.

Clearly, therefore, the muscle should never be allowed to contract more frequently than 1:2 with the heart and, to allow a safe margin, less frequently than that.

Conditioning and activation Goal of conditioning

“Conditioning” of a skeletal muscle by electrical stimulation induces the adaptive changes that enable it to perform cardiac work. The clinical protocol introduced for cardiomyoplasty employed short bursts of stimulation at 30 Hz. The number of impulses in each train was increased progressively, in a largely intuitive attempt to escalate the challenge to the muscle over a number of weeks [13, 25]. This protocol was widely

(5)

adopted, not because it had a firm scientific basis, but because of commercial pressures and the need to achieve FDA approval. It was, in fact, far from optimal.

Conditioning took 8 postoperative weeks, delaying the benefits of the surgical procedure to a sick patient. More importantly still, the endpoint was a slow, fatigue- resistant (type 1) muscle.

The goal of conditioning and activation for cardiac assistance should be a fatigue-resistant muscle that has fast contractile characteristics (type 2A) [32, 59].

Transformation to a slow phenotype reduces dramatically the power available from the grafted muscle. It also poses problems of synchronization: an overconditioned wrap is slower-contracting than the myocardium so it cannot reinforce the cardiac wall during systole; during diastole it may relax too slowly, interfering with ventricular filling.

The adaptive changes in vascularity and oxidative capacity responsible for the increased resistance to fatigue depend on the aggregate number of impulses reaching the muscle, independent of the pattern with which the impulses are delivered [40, 41, 71]. These changes are more readily induced than the transition to the slow myosin isoform and can be achieved with a less metabolically demanding regime [5, 33, 47, 70]. In the rabbit, a period of 10 Hz stimulation as short as 30 minutes a day is sufficient to achieve a fatigue-resistant muscle in which more than 90% of the fibres are of type 2A [40].

Species differences

Much of what we now know about the adaptive response of mammalian skeletal muscle to a change in use comes from studies in rats or rabbits. Caution needs to be exercised in attempting to extrapolate from these results because adaptive changes in larger species are brought about with stimulation patterns containing fewer impulses.

In the rabbit, for example, continuous stimulation at 2 Hz produces a stable population of type 2A fibres; in the dog or sheep, an identical pattern results in complete transformation to Type 1.

We have recently studied the response to stimulation in the pig, an animal whose body size is comparable to that of humans. The stimulator was programmed to deliver pulses of 210 µs duration at a frequency of 30 Hz, on for 0.19 s and off for 6 s, 24 h/day, which in terms of the number of pulses is equivalent to continuous stimulation at a frequency close to 1 Hz.

This pattern conditioned the muscle to type 2A in about 6 weeks, but after that the proportion of slow (type 1) fibres continued to creep upwards. As already stated, this aspect of the response is dictated by the aggregate amount of stimulation rather than its pattern. A régime designed to achieve transformation to fast, fatigue- resistant characteristics should therefore deliver no more than the daily equivalent of continuous 0.5 Hz, or even 0.25 Hz.

Studies in rabbits show that patterns delivering a smaller aggregate number of impulses induce a more gradual transformation, as well as a different endpoint [47]. Thus there is a case for accelerating the initial stages of the conditioning process with a more intensive pattern, and reverting to the final pattern after a few weeks in order to achieve the desired long-term characteristics.

Conditioning is completely reversible on cessation of stimulation [6, 7, 22, 58, 69]. The pattern of stimulation that is used to activate the cardiomyoplasty wrap must therefore also maintain the conditioned state of the muscle, and must not lead to overconditioning to the slow phenotype. In practice this can be achieved in three main ways:

(i) Stimulation with optimized bursts. Stimulus bursts containing a high-frequency component are needed not only to generate the best force but also to maintain the mass of the muscle in the long term [21, 24, 32, 71]. It is conventional, but quite unphysiological, to stimulate electrically using constant-frequency bursts. These may be replaced with bursts in which the interpulse intervals are optimized to produce the greatest force per impulse [39]. In this way fewer pulses can be delivered without compromising performance.

(ii) Heart synchronization ratio. The muscle may be activated on a smaller proportion of cardiac cycles. In terms of the girdling action of the wrap (see below) this would not compromise the clinical benefit. In fact it was quite usual in Russia to use heart synchronization ratios between 1:4 and 1:16, apparently with good results [15].

(iii) Demand stimulation. Stimulation may be restricted to part of the day—for example, during waking hours, or at times of increased demand, detected by a rise in heart rate above a specified threshold.

Recent clinical studies provide clear evidence of increased contractile speed (and therefore power) of the grafted muscle when stimulation is switched from a conventional to an intermittent régime [10] (and see Rigatelli et al., this issue).

In practice it is the demand stimulation protocol that has the greatest potential for making a substantial reduction in the aggregate number of impulses. Figure 2 shows the relationship between heart assist ratio and hours of assist per day that would deliver the equivalent of 0.5 Hz continuously for three values of impulse per burst (Figure 2a) and for three values of heart assist ratio (Figure 2b). Graphs of this type are easily generated and provide a valuable guide to determining the desired pattern.

Understanding the mechanism

At first glance it may seem that a wrap of skeletal muscle around the heart would behave much like an additional layer of myocardium, contributing to systolic contraction and the ejection of blood. Although this is a persuasive notion, it is based on a simplistic view of muscle biomechanics.

(6)

Fig. 2 Making a rationale choice of stimulation régime. The graphs illustrate the relationship between heart assist ratio and hours of assist per day that would deliver the equivalent of 0.5 Hz continuously. If, for example, one wished to assist 1 beat in 6, stimulation should be limited to no more than 10 hours per day if fast contractile properties are to be retained. The illustration is based on 6 impulses per burst and a heart rate of 72 beats/min (full lines).

Figure 2(a) also shows the effect of a variation of heart rate from 60 beats/min (lower dashed line) to 90 beats/min (upper dashed line).

Figure 2(b) shows the effect of varying the number of impulses per burst from 5 (lower dashed line) to 7 (upper dashed line). It can be seen that the effect of these variations is not great; the most important parameters are those on the axes.

The muscular wrap is heavily loaded and therefore cannot shorten fast enough to develop appreciable power [61]. This situation is exacerbated by overconditioning, under which conditions the muscle acquires slow contractile characteristics. Slowness means that the wrap cannot follow the more rapid contraction of the cardiac wall, let alone reinforce it.

The wrap cannot even develop the full force of an isometric contraction during the permissible duration of the stimulation burst, which must be limited to synchronize its action with the cardiac cycle. If we add to this the loss of functional muscle mass due to ischaemic damage, particularly in the middle and distal parts of the flap, it would seem that there are few grounds for believing that such a configuration was ever capable of producing significant systolic assistance.

Not surprisingly, early practitioners had difficulty in demonstrating consistently the expected improvements in cardiac output, systemic arterial pressure, pulmonary capillary wedge pressure, and ejection fraction that should have accompanied systolic enhancement. The protagonists of cardiomyoplasty were nevertheless reluctant to abandon the concept of a “systolic squeeze”

as the underlying mechanism and many surgeons and cardiologists, seeing little haemodynamic evidence of it in their own results, assumed that the procedure had been unsuccessful.

Meanwhile about 85% of patients were reporting symptomatic benefits of dynamic cardiomyoplasty, with an improvement of between 1 and 2 classes on the New York Heart Association scale. Clearly the procedure had worked in some way, albeit not in the manner predicted.

There is now a good deal of evidence to indicate that the muscular wrap functioned as a support, reducing the wall stress experienced by the myocardium, restoring a more orderly contraction, and preventing further distension of the ventricles. What, then, was the role of stimulation? Although the grafted LD muscle remains innervated by the thoracodorsal nerve, the wrap is under less than normal physiological tension; it is therefore uncertain how much endogenous activity it receives from the host nervous system. Stimulation would be needed to avoid flaccidity and disuse atrophy and to maintain close apposition of the wrap to the heart wall, particularly at the beginning of systole.

This role of stimulation is consistent with the fact that beneficial effects of cardiomyoplasty were observed in the early stages of the technique, when the only synchronous stimulator available was a cardiac pacemaker that delivered a single pulse. It took a paper from R.-C. Chiu’s group to point out that the twitch contraction developed by skeletal muscle in response to a single stimulus is very brief in relation to cardiac systole and develops only a fraction of the available force [20]. Even with a purpose-built cardiomyostimulator delivering the necessary burst

(7)

stimulation it was found that the therapeutic benefits were not compromised by less frequent activation of the wrap [15]. More direct evidence was obtained by recording pressure-volume loops under clinical conditions, in which stimulation of the wrap produced no systolic effects [66]. But Kass and his colleagues recorded a progressive reduction in end-diastolic volume in cardiomyoplasty patients and this was consistent with reverse remodelling, with the muscular wrap acting as a ‘ratcheting girdle’ [37].

Cardiomyoplasty may also be beneficial in supplementing the blood flow from the coronary circulation through collateral vessels that bridge between the muscular wrap and the myocardium [4, 45].

The vessels appear to be attracted especially to ischaemic sites in the myocardium. This extramyocardial neovascularization was sought many years ago by Russian surgeons, who used a passive muscle flap for the purpose.

A combination of prestimulation, to minimize ischaemic damage, and demand activation, to avoid overconditioning, would result in a skeletal muscle graft that was several times more powerful and could be synchronized more effectively with the cardiac cycle.

Under these conditions there would seem to be a much better prospect of delivering consistent and measurable systolic augmentation. The same conditions would favour the long-term girdling action of the wrap and the potential for extramyocardial revascularization, both of which depend on maintaining close contact between a healthy, well-vascularized wrap and the heart wall.

Devices

Nearly all the cardiomyoplasty operations performed were dependent on cardiomyostimulators supplied by Medtronic Inc. The commercial decision by Medtronic to withdraw from the field was damaging, as it led to the widespread conclusion that the technique was inherently flawed. Ironically, it was the drive to meet regulatory requirements that discouraged departures from a fixed protocol, isolating clinical practice from progress in the underlying basic science. The suspension of new cardiomyoplasty procedures now provides the opportunity for incorporating the advances in basic knowledge and improved protocols described here, with the prospect of a more positive and consistent outcome.

This is, however, a Catch-22 situation, for these benefits cannot be demonstrated without clinically approved stimulators, and manufacturers may be reluctant to enter into such development in view of the past history.

Fortunately new devices are now emerging, such as the LD-PACE II from CCC Uruguay. This may enable cardiomyoplasty and related research to be revived.

A major cause of mortality in patients who had undergone cardiomyoplasty was cardiac arrhythmia.

Preliminary results in which defibrillators were implanted as well as cardiomyostimulators indicated that a number of patients could be saved in this way.

When considering new initiatives in the device area, thought should be given to incorporating a defibrillation function.

Cellular cardiomyoplasty

Cardiomyoplasty is less invasive, and a good deal less costly, than heart transplant surgery. Nonetheless there is an understandable desire, on the part of both patients and surgeons, to seek options that are even less invasive. One such approach is to introduce stem cells into infarcted or peri-infarcted regions of myocardium, a technique that has been referred to as ‘cellular cardiomyoplasty’. Progress has been reviewed [14] and the potential problems discussed [75, 76, 84].

A full consideration of this field is beyond the scope of this paper, but it is worth mentioning issues which, in the current enthusiasm for cell-based solutions, are in danger of being overlooked.

The term “cellular cardiomyoplasty” implies functional augmentation of a failing heart resulting from the introduction, survival, differentiation, and integration of contractile tissue. We will not, therefore, be concerned here with solutions based purely on attempts to revascularize the infarcted area, or on the influence of paracrine secretions from the implanted cells.

Candidate cells

It is quite wrong to assume—although such a belief persists in some quarters—that any cells implanted into a cardiac environment will automatically turn into cardiomyocytes; indeed, in some circumstances they may actually contribute to scar tissue formation. The sources of potentially useful cells are therefore limited.

Autologous skeletal muscle myoblasts may be obtained and expanded in vitro prior to their introduction [27]. But myoblasts become muscle cells, not cardiomyocytes, and they do not become integrated into existing myocardium. Their presence may even interfere with normal cardiac rhythmicity [68].

Furthermore, in the absence of innervation the regenerated tissue may be expected to suffer the atrophy and necrosis that is characteristic of denervated muscles, even if it survives in the early stages.

Suitable precursor cells derived or purified from bone marrow are few in number, may not generate cardiomyocytes [51] and may not be retained to any significant extent in the heart when injected into the coronary circulation [28].

Embryonic stem cells are a potential source of cardiomyocytic precursors. However, in addition to developing immunogenicity, there is a danger that such cells will retain their potential for unlimited division or for generating other, unwanted tissues [44]. Purification or genetic modification may help to overcome these limitations.

The existence of a small population of autologous cardiac progenitor cells suggests a promising approach,

(8)

but doubts remain about their capacity for expansion in vitro and their subsequent ability to integrate electrically and mechanically with uninjured myocardium.

Candidate routes

Given an appropriate volume of suitable cells, how is it to be delivered?

Surgical introduction into the infarcted area risks placing the cells in an environment that is either unsupportive, because of local ischaemia, or hostile, promoting inflammation and apoptosis. Injection into the coronary circulation depends on appropriate extravasation and migration of the cells, and may cause embolism.

Function

Although cellular cardiomyoplasty, using autologous skeletal myoblasts, has been claimed to improve the function of experimentally infarcted hearts, it is difficult to see how an unstimulated island of tissue, isolated electrically and mechanically from myocardium, can contribute actively to cardiac performance. Any improvement is more likely to have been the result of local vasculogenesis and increased compliance of the infarcted area due to displacement of the more rigid scar tissue.

Cellular or tissue cardiomyoplasty?

It would be foolhardy to pretend that these and other problems will not be overcome in the long term.

Nevertheless, much further research will be needed before such techniques can be shown to be safe and beneficial to patients. As Taylor has stated: “The promise for cell transplantation is too great to be spoiled by ill-designed attempts that forget to account for the biology of both the cells and the myocardium.” [76]

In contrast to the few milligrams of contractile tissue that may be achieved by current cell-based techniques, tissue cardiomyoplasty involves the surgical redeployment of hundreds of grams of existing, mature contractile tissue. We should not be diverting all our energies into cell-based therapies when this technique offers a more immediate solution.

Conclusion

Cardiomyoplasty is not a failed approach. The therapeutic benefits have been underestimated because of an imperfect understanding of the underlying mechanism and shortcomings in the protocol for mobilizing and activating the graft. Correcting these mistakes will result in a muscle graft that is more viable and more powerful. There will then be a good prospect that such a procedure could provide systolic assist in the short term and reverse remodelling and extramyocardial revascularization in the long term. The need for alternatives to cardiac transplantation is urgent and currently there appears to be no other technique of the same maturity and promise.

The reappearance of cardiomyoplasty in the surgical armamentarium should also stimulate interest in other forms of cardiac assist, including hybrid assist devices and SMVs, which have the potential for harnessing the pumping power of skeletal muscle even more effectively.

Declaration

This review contains reference to animal experiments.

Where these were conducted within the authors’

laboratory, procedures conformed strictly to the Animals (Scientific Procedures) Act 1986, which governs animal experimentation in the UK, and all animals were obtained from a registered source.

Acknowledgements

The authors acknowledge the research contributions of their colleagues in the Liverpool Muscle Research Group and in Dr. L.W. Stephenson’s laboratory, Detroit, USA. Much of the authors’ contribution to the field was supported by the British Heart Foundation.

Address Correspondence to:

Emeritus Professor Stanley Salmons, Muscle Research Group, Department of Human Anatomy and Cell Biology, University of Liverpool, The Sherrington Buildings, Ashton Street, Liverpool L69 3GE, UK.

Email s.salmons@liverpool.ac.uk References

[1] Ali AT, Santamore WP, Chiang BY, Dowling RD, Tobin GR, Slater AD: Vascular delay of the latissimus dorsi provides an early hemodynamic benefit in dynamic cardiomyoplasty. Ann Thorac Surg 1999; 67: 1304-11.

[2] Anderson WA, Andersen JS, Acker MA, Hammond RL, Chin AJ, Douglas PS, Khalafalla AS, Salmons S, Stephenson LW: Skeletal muscle grafts applied to the heart: a word of caution.

Circulation 1988; 78 (suppl III): 180–190.

[3] Anderson WA, Ianuzzo CD, Klabunde RE:

Cardiomyoplasty: studies on goat latissimus dorsi blood flow and muscle damage following surgical dissection and chronic electrical stimulation. J Card Surg 1996; 11: 237-246.

[4] Bailey WF, Jr., Magno MG, Buckman PD, DiMeo F, Langan T, Armenti VT, Mannion JD:

Chronic stimulation enhances extramyocardial collateral blood flow after a cardiomyoplasty.

Ann Thorac Surg 1993; 56: 1045-52.

[5] Barron DJ, Etherington PJ, Winlove CP, Jarvis JC, Salmons S, Pepper JR: Muscle transformation in cardiomyoplasty: the effect of conditioning and mobilisation on perfusion, oxygenation and fatigue resistance in the latissimus dorsi muscle. Eur J Cardiothorac Surg 1998; 13: 588-98.

[6] Brown JMC, Henriksson J, Salmons S:

(9)

Restoration of fast muscle characteristics following cessation of chronic stimulation:

physiological, histochemical and metabolic changes during slow-to-fast transformation. Proc Biol Sci 1989; 235: 321-346.

[7] Brown WE, Salmons S, Whalen RG:

Mechanisms underlying the asynchronous replacement of myosin light chain isoforms during stimulation-induced fibre-type transformation of skeletal muscle. FEBS Lett 1985; 192: 235-238.

[8] Callegari PR, Taylor GI, Caddy CM, Minabe T:

An anatomic review of the delay phenomenon: I.

Experimental studies. Plast Reconstr Surg 1992;

89: 397-407; discussion 417-8.

[9] Carpentier A, Chachques J-C: Myocardial substitution with a stimulated skeletal muscle:

first successful clinical case. Lancet 1985; i:

1267.

[10] Carraro U, Docali G, Barbiero M, Brunazzi C, Gealow K, Casarotto D, Muneretto C: Demand dynamic cardiomyoplasty: improved clinical benefits by non-invasive monitoring of LD flap and long-term tuning of its dynamic contractile characteristics by activity-rest regime. Basic Appl Myol 1998; 8: 11-15.

[11] Carroll SM, Carroll CMA, Stremel RW, Heilman SJ, Tobin GR, Barker JH: Vascular delay of the latissimus dorsi muscle: an essential component of cardiomyoplasty. Ann Thorac Surg 1997; 63:

1034-1040.

[12] Carroll SM, Heilman SJ, Stremel RW, Tobin GR, Barker JH: Vascular delay improves latissimus dorsi muscle perfusion and muscle function for use in cardiomyoplasty. Plast Reconstr Surg 1997; 99: 1329-1337.

[13] Chachques JC, Carpentier A: Postoperative management., in Carpentier A, Chachques JC, Grandjean P (eds): Cardiomyoplasty. Mount Kisco, NY, Futura Publishing Inc, 1991, pp 131- 138.

[14] Chachques JC, Shafy A, Duarte F, Cattadori B, Goussef N, Shen L, Carpentier A: From dynamic to cellular cardiomyoplasty. J Card Surg 2002;

17: 194-200.

[15] Chekanov V: After cardiomyoplasty there is no need to stimulate the muscle with every heart contraction. Ann Thorac Surg 1994; 57: 1370- 1371.

[16] Cheng W, Michele JJ, Spinale FG, Sink JD, Santamore WP: Effects of cardiomyoplasty on biventricular function in canine chronic heart failure. Ann Thorac Surg 1993; 55: 893–901.

[17] Craven AJ, Jarvis JC, Salmons S:

Vascularisation of the latissimus dorsi muscle for cardiac assist. J Anat 1994; 185: 706–707.

[18] Cruz MP, Michele JJ, Mannion JD, Magno M, George DT, Santamore WP: Cardiomyoplasty.

Latissimus dorsi muscle function and blood flow during isolation. ASAIO J 1997; 43: 338-44.

[19] Davidse JHL, van der Veen FH, Lucas CMHB, Penn OCKM, Daemen MJAP, Wellens HJJ:

Structural alterations in the latissimus dorsi muscles in three patients more than 2 years after a cardiomyoplasty procedure. Eur Heart J 1998;

19: 310-318.

[20] Dewar ML, Drinkwater DC, Wittnich C, Chiu RC-J: Synchronously stimulated skeletal muscle graft for myocardial repair. An experimental study. J Thorac Cardiovasc Surg 1984; 87: 325- 331.

[21] Eerbeek O, Kernell D, Verhey BA: Effects of fast and slow patterns of tonic long-term stimulation on contractile properties of fast muscle in the cat. J Physiol 1984; 352: 73-90.

[22] Eisenberg BR, Brown JMC, Salmons S:

Restoration of fast muscle characteristics following cessation of chronic stimulation. The ultrastructure of slow-to-fast transformation. Cell Tissue Res 1984; 238: 221-230.

[23] El Oakley RM, Jarvis JC, Barman D, Greenhalgh DL, Currie J, Downham DY, Salmons S, Hooper TL: Factors affecting the integrity of latissimus dorsi muscle grafts: implications for cardiac assistance from skeletal muscle. J Heart Lung Transplant 1995; 14: 359-365.

[24] Ferguson AS, Stone HE, Roessmann U, Burke M, Tisdale E, Mortimer JT: Muscle plasticity:

comparison of a 30-Hz burst with 10-Hz continuous stimulation. J Appl Physiol 1989; 66:

1143-1151.

[25] Grandjean PA, Lori Austin RN, Chan S, Terpestra B, Bourgeois IM: Dynamic cardiomyoplasty: clinical follow-up results. J Card Surg 1991; 6: 80–88.

[26] Guldner NW, Klapproth P, Grossherr M, Brugge A, Sheikhzadeh A, Tolg R, Rumpel E, Noel R, Sievers HH: Biomechanical hearts: muscular blood pumps, performed in a 1-step operation, and trained under support of clenbuterol.

Circulation 2001; 104: 717-22.

[27] Hagege AA, Carrion C, Menasche P, Vilquin JT, Duboc D, Marolleau JP, Desnos M, Bruneval P:

Viability and differentiation of autologous skeletal myoblast grafts in ischaemic cardiomyopathy. Lancet 2003; 361: 491-2.

[28] Hofmann M, Wollert KC, Meyer GP, Menke A, Arseniev L, Hertenstein B, Ganser A, Knapp WH, Drexler H: Monitoring of bone marrow cell homing into the infarcted human myocardium.

Circulation 2005; 111: 2198-202.

[29] Ianuzzo CD, Ianuzzo SE, Carson N, Feild M, Locke M, Gu J, Anderson WA, Klabunde RE:

Cardiomyoplasty: degeneration of the assisting skeletal muscle. J Appl Physiol 1996; 80: 1205- 1213.

(10)

[30] Ianuzzo CD, Ianuzzo SE, Feild M, Locke M:

Cardiomyoplasty—preservation of the latissimus dorsi muscle. J Card Surg 1995; 10: 104-110.

[31] Isoda S, Yano Y, Jin Y, Walters HL, Kondo J, Matsumoto A: Influence of a delay on latissimus dorsi muscle flap blood flow. Ann Thorac Surg 1995; 59: 632-638.

[32] Jarvis JC, Mayne CN, Salmons S: Basic studies on skeletal muscle for cardiac assistance. J Card Surg 1991; 6 (Suppl.): 204-209.

[33] Jarvis JC, Sutherland H, Mayne CN, Gilroy SJ, Salmons S: Induction of a fast-oxidative phenotype by chronic muscle stimulation:

mechanical and biochemical studies. Am J Physiol Cell Physiol 1996; 270: C306-312.

[34] Kalil-Filho R, Bocchi E, Weiss RG, Rosemberg L, Bacal F, Moreira LFP, Stolf NAG, Magalh~aes AAC, Belotti G, Jatene A, Pileggi F:

Magnetic resonance imaging evaluation of chronic changes in latissimus dorsi cardiomyoplasty. Circulation 1994; 90: II102-6.

[35] Kantrowitz A: Functioning autogenous muscle used experimentally as an auxiliary ventricle.

Trans ASAIO 1960; 6: 305-310.

[36] Kantrowitz A, McKinnon W: The experimental use of the diaphragm as an auxiliary myocardium. Surg Forum 1959; 9: 266-268.

[37] Kass DA, Baughman KL, Pak PH, Cho PW, Levin HR, Gardner TJ, Halperin HR, Tsitlik JE, Acker MA: Reverse remodeling from cardiomyoplasty in human heart failure. External constraint versus active assist. Circulation 1995;

91: 2314-2318.

[38] Kratz JM, Johnson WS, Mukherjee R, Hu J, Crawford FA, Spinale FG: The relation between latissimus dorsi skeletal muscle structure and contractile function after cardiomyoplasty. J Thorac Cardiovasc Surg 1994; 107: 868-78.

[39] Kwende MMN, Jarvis JC, Salmons S: The input- output relationships of skeletal muscle. Proc Biol Sci 1995; 261: 193–201.

[40] Lopez-Guajardo A, Sutherland H, Jarvis JC, Salmons S: Dynamics of stimulation-induced muscle adaptation: insights from varying the duty cycle. J Muscle Res Cell Motil 2000; 21:

725-735.

[41] Lopez-Guajardo A, Sutherland H, Jarvis JC, Salmons S: Induction of a fatigue-resistant phenotype in rabbit fast muscle by small daily amounts of stimulation. J Appl Physiol 2001; 90:

1909-18.

[42] Lucas C, Van Der Veen F, Kloosterman E, Habets J, Wellens H: Long term dynamic cardiomyoplasty in the goat is accompanied by major degenerative changes in the wrapped latissimus dorsi muscle. J Am Coll Cardiol 1993;

21: 469A.

[43] Lucas CMHB, van der Veen FH, Cheriex EC,

Lorusso R, Havenith M, Penn OCKM, Wellens HJJ: Long-term follow–up (12 to 35 weeks) after dynamic cardiomyoplasty. J Am Coll Cardiol 1993; 22: 758-767.

[44] Lyon A, Harding S: The potential of cardiac stem cell therapy for heart failure. Curr Opin Pharmacol 2007; 7: 164-70.

[45] Mannion JD, Magno MG, Buckman PD, DiMeo F, Greene R, Bowers M, McHugh M, Menduke H: Acute electrical stimulation increases extramyocardial collateral blood flow after a cardiomyoplasty. Ann Thorac Surg 1993; 56:

1351-1358.

[46] Mathes SJ, Nahai F: Classification of the vascular anatomy of muscles: experimental and clinical correlation. Plast Reconstr Surg 1981;

67: 177–187.

[47] Mayne CN, Sutherland H, Jarvis JC, Gilroy SJ, Craven AJ, Salmons S: Induction of a fast- oxidative phenotype by chronic muscle stimulation: histochemical and metabolic studies.

Am J Physiol Cell Physiol 1996; 270: C313-320.

[48] Misawa Y, Mott BD, Lough JO, Chiu RC-J:

Pathologic findings of latissimus dorsi muscle graft in dynamic cardiomyoplasty: clinical implications. J Heart Lung Transplant 1997; 16:

585-595.

[49] Monnet E, Orton EC, Child G, Getzy D, Jacobs G, Metelman L: Neuromuscular function of the latissimus dorsi muscle in goats after dynamic cardiomyoplasty. Pacing Clin Electrophysiol 1999; 22: 1625-1633.

[50] Moreira LFP, Bocchi EA, Stolf NAG, Pileggi F, Jatene AD: Current expectations in dynamic cardiomyoplasty. Ann Thorac Surg 1993; 55:

299–303.

[51] Murry CE, Soonpaa MH, Reinecke H, Nakajima H, Nakajima HO, Rubart M, Pasumarthi KB, Virag JI, Bartelmez SH, Poppa V, Bradford G, Dowell JD, Williams DA, Field LJ:

Haematopoietic stem cells do not transdifferentiate into cardiac myocytes in myocardial infarcts. Nature 2004; 428: 664-8.

[52] Nakamura K, Glenn WL: Grafts of diaphragm as a functioning substitute for the myocardium. J Surg Res 1964; 4: 435-439.

[53] Radermecker MA, Triffaux M, Fissette J, Limet R: Anatomical rationale for use of the latissimus dorsi flap during the cardiomyoplasty operation.

Surg Radiol Anat 1992; 14: 5–10.

[54] Ramnarine I, Capoccia M, Ashley Z, Sutherland H, Russold M, Summerfield N, Salmons S, Jarvis J: Comparison between the power output of a skeletal muscle ventricle and the left ventricle in the same circulation. Artif Organs 2004; 28: 766-767.

[55] Ramnarine IR, Capoccia M, Ashley Z, Sutherland H, Russold M, Summerfield N,

(11)

Salmons S, Jarvis JC: Counterpulsation from the skeletal muscle ventricle and the intraaortic balloon pump in the normal and failing circulations. Circulation 2006; 114: I10-I15.

[56] Reichenbach SH, Gustafson KJ, Egrie GD, Weidman JR, Farrar DJ, Hill JD: Evaluation of a skeletal muscle energy convertor in a chronic animal model. ASAIO J 2000; 46: 482-5.

[57] Salmons S: On the feasibility of using diaphragm muscle as a myocardial substitute. Med Biol Eng 1975; 13: 608-609.

[58] Salmons S: On the reversibility of stimulation- induced muscle transformation, in Pette D (ed):

The Dynamic State of Muscle Fibres. Berlin, Walter de Gruyter, 1990, pp 401-414.

[59] Salmons S: Optimizing the benefits of cardiomyoplasty. Br J Hosp Med 1992; 49: 137.

[60] Salmons S, Jarvis JC: Cardiomyoplasty: the basic issues. Cardiac Chronicle 1990; 4: 1-7.

[61] Salmons S, Jarvis JC: Cardiomyoplasty: a look at the fundamentals, in Carpentier A, Chachques JC, Grandjean PA (eds): Cardiomyoplasty.

Mount Kisco, NY, Futura Publishing Company, Inc., 1991, pp 3-17.

[62] Salmons S, Jarvis JC: Cardiac assistance from skeletal muscle: a critical appraisal of the various approaches. Br Heart J 1992; 68: 333-338.

[63] Salmons S, Sréter FA: Significance of impulse activity in the transformation of skeletal muscle type. Nature 1976; 263: 30–34.

[64] Salmons S, Tang ATM, Jarvis JC, Degens H, Hastings M, Hooper TL: Morphological and functional evidence, and clinical importance, of vascular anastomoses in the latissimus dorsi muscle of the sheep. J Anat 1998; 193: 93-104.

[65] Salmons S, Vrbová G: The influence of activity on some contractile characteristics of mammalian fast and slow muscles. J Physiol 1969; 201: 535-549.

[66] Schreuder JJ, van der Veen FH, van der Velde ET, Delahaye F, Alfieri O, Jegaden O, Lorusso R, Jansen JRC, van Ommen V, Finet G, Wellens HJJ: Beat-to-beat analysis of left ventricular pressure-volume relation and stroke volume by conductance catheter and aortic modelflow in cardiomyoplasty patients. Circulation 1995; 91:

2010-2017.

[67] Shortland AP, Jarvis JC, Salmons S:

Haemodynamic considerations in the design of a skeletal muscle ventricle. Med Biol Eng Comput 2003; 41: 529-535.

[68] Smits PC, van Geuns RJ, Poldermans D, Bountioukos M, Onderwater EE, Lee CH, Maat AP, Serruys PW: Catheter-based intramyocardial injection of autologous skeletal myoblasts as a primary treatment of ischemic heart failure:

clinical experience with six-month follow-up. J Am Coll Cardiol 2003; 42: 2063-9.

[69] Sréter FA, Elzinga M, Mabuchi K, Salmons S, Luff AR: The N -methylhistidine content of myosin in stimulated and cross-reinnervated skeletal muscles of the rabbit. FEBS Lett 1975;

57: 107-111.

[70] Sutherland H, Jarvis JC, Kwende MMN, Gilroy SJ, Salmons S: The dose-related response of rabbit fast muscle to long-term low-frequency stimulation. Muscle Nerve 1998; 21: 1632-46.

[71] Sutherland H, Jarvis JC, Salmons S: Pattern dependence in the stimulation-induced type transformation of rabbit fast skeletal muscle.

Neuromodulation 2003; 6: 176-189.

[72] Tang ATM, Jarvis JC, Hooper TL, Salmons S:

Observation and basis of improved blood flow to the distal latissimus dorsi muscle: a case for electrical stimulation prior to grafting.

Cardiovasc Res 1998; 40: 131-7.

[73] Tang ATM, Jarvis JC, Hooper TL, Salmons S:

Cardiomyoplasty: the benefits of electrical prestimulation of the latissimus dorsi muscle in situ. Ann Thorac Surg 1999; 68: 46-51.

[74] Tatjanchenko VK, Sherstennikov EN: Arterial architectonics of musculus latissimus dorsi in man and dog. Archiv Anatomii Gistologii i Embriologii 1978; 74: 28–33.

[75] Taylor DA: Cellular cardiomyoplasty with autologous skeletal myoblasts for ischemic heart disease and heart failure. Curr Control Trials Cardiovasc Med 2001; 2: 208-210.

[76] Taylor DA: Cell-based myocardial repair: how should we proceed? Int J Cardiol 2004; 95 Suppl 1: S8-12.

[77] Taylor GI, Palmer JH: The vascular territories (angiosomes) of the body: experimental study and clinical applications. Br J Plast Surg 1987;

40: 113–141.

[78] Thomas GA, Hammond RL, Greer K, Lu H, Jarvis JC, Shortland AP, Pullan DM, Salmons S, Stephenson LW: Functional assessment of skeletal muscle ventricles after pumping for up to four years in circulation. Ann Thorac Surg 2000; 70: 1281-9; discussion 1290.

[79] Tobin GR, Schusterman M, Peterson GH, Nichols G, Bland KI: The intramuscular neurovascular anatomy of the latissimus dorsi muscle: the basis for splitting the flap. Plast Reconstr Surg 1981; 67: 637–641.

[80] Trumble DR, Magovern JA: Capturing in situ skeletal muscle power for circulatory support: a new approach to device design. ASAIO J 2003;

49: 480-5.

[81] van Doorn CAM, Bhabra MS, Hopkinson DN, Barman D, Cranley JJ, Hooper TL: Latissimus dorsi muscle blood flow during synchronized contraction—implications for cardiomyoplasty.

Ann Thorac Surg 1996; 61: 603-609.

[82] Woo EB-C, Jarvis JC, Hooper TL, Salmons S:

(12)

Avoiding ischemia in latissimus dorsi muscle grafts: electrical prestimulation versus vascular delay. Ann Thorac Surg 2002; 73: 1927-1932.

[83] Woo EB-C, Tang ATM, Jarvis JC, Hasleton PS, Salmons S, Hooper TL: Improved viability of latissimus dorsi muscle grafts after electrical

prestimulation. Muscle Nerve 2002; 25: 679-684.

[84] Yacoub M, Suzuki K, Rosenthal N: The future of regenerative therapy in patients with chronic heart failure. Nat Clin Prac Cardiovasc Med 2006; 3 Suppl 1: S133-5.