An autologous biologic pump motor

J THoRAc CARDIOVASC SURG 92:733-746,1986

An autologous biologic pump motor Latissimus dorsi skeletal muscle ventricles were constructed in six beagles. They first underwent a period of vascular delay and of electrical preconditioning over several weeks. The skeletal muscle ventricles were then connected to a totaUy implantable mock circulation that allowed for the chronic measurement of pressures and flows produced by the muscle. The skeletal muscle ventricles were actuated by stimulation of the motor nerve with an implanted generator that delivered brief pulse trains. The skeletal muscle ventriclespumped continuously against an afterload of 80 mm Hg with a preload of 40 to 50 mm Hg at a rate of 54 times per minute. At initiation of pumping, systolic pressure was 135 ± 24 mm Hg and flow was 464 ± 116 m1/min. After 2 weeks of continuous pumping, the systolic pressure was 104 ± 1 mm Hg and continuous flow was 206 ± 16 m1/min. Two of the skeletal muscle ventricles pumped continuously for 5 and 9 weeks, respectively. At the end of that time one was stiU capable of generating pressure up to 205 mm Hg and the other, 160 mm Hg. These results suggest that a chronic auxiliary skeletal muscle ventricle is a feasible approach to the treatment of end-stage cardiac failure.

Michael A. Acker, M.D. (by invitation), Robert L. Hammond, B.S. (by invitation), John D. Mannion, M.D. (by invitation), Philadelphia, Pa., Stanley Salmons, Ph.D. (by invitation), Birmingham. England, and Larry W. Stephenson, M.D., Philadelphia, Pa.

One possible therapeutic approach to augment the failing heart would be to construct a ventricle from living, contracting, autologous tissue. Such a biologic motor could be used as an auxiliary ventricular assist to augment or replace a failing ventricle. Although this idea is theoretically appealing, skeletal muscle-powered cardiac assist devices have thus far been hindered by muscle fatigue. loS In previous acute experiments, we6 have demonstrated that skeletal muscle ventricles (SMVs) can be made more fatigue resistant by a combination of vascular delay, chronic electrical conditioning, and multilayered construction. The purpose of this study was to determine whether such an SMV is capable of continuous and chronic work in an awake animal.

From the Harrison Department of Surgieal Research, Department of Surgery, Division of Cardiothoracie Surgery, University of Pennsylvania, School of Medieine, Philadelphia, Pa., and the Department of Anatomy, University of Birmingham, Birmingham, England. Supported by National Institutes of Health Grant HLBI34778. Read at the Sixty-sixth Annual Meeting of The Ameriean Association for Thoracic Surgery, New York, N. Y., April 28-30,1986. Address for reprints: Dr. Miehael Acker, Harrison Department of Surgical Research, 313 Medical Education Bldg., University of Pennsylvania,' Philadelphia, Pa. 19104.

Methods and materials Latissimus dorsi SMVs were constructed in six adult male beagles whose weights ranged from 9 to 13 kg. After a period of vascular delay and electrical preconditioning, the SMVs were connected to a totally implantable mock circulation system that allowed independent control of the preload (ventricular filling pressure) and afterload of each dog's SMV. Except when measurements were made, no tubes or wires crossed the animal's skin. The SMVs were stimulated to contract, in a pulsatile fashion, continuously 24 hours a day. Pressure and flow measurements were recorded either daily or every other day for each SMV. SMV construction. The animals were operated on under sterile conditions, in accordance with the "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences.' Anesthesia was induced (pentobarbital 30 mg/kg) and an endotracheal tube placed. Preoperative and postoperative antibiotic prophylaxis (cephalothin 500 mg intramuscularly or orally, every 8 hours) was administered to each animal until all drains were removed. Postoperatively, pain was controlled with morphine 0.1 mg/kg intramuscularly for the first 24 hours and then with aspirin by mouth as needed. A left flank incision was made and the collateral blood vessels to the latissimus dorsi muscle were ligated 733

7 3 4 Acker et al.

Fig. 1. Construction of the SMV.

and divided. The muscle was mobilized from its attachments to the lower ribs, thoracodorsal fascia, trapezius, teres major, and platysma muscles. The insertion on the humerus was also detached such that only the neurovascular bundle (thoracodorsal artery, vein, and nerve) connected the muscle to the body of the dog. By a technique similar to the one previously developed in our laboratory" (Fig. 1), the muscle was wrapped around a Teflon mandrel. The mandrel was cone shaped with a length of 8 em, a volume of 17 mI, and a diameter of 1.9 ern at its widest end. The muscle was wrapped around this mandrel in spiral fashion so that the proximal blood vessels were on the outside. A total of 2 to 2.5 wraps of muscle were obtained. A Teflon collar was attached to the top of the mandrel to which the muscle wraps were sutured. A specially modified Medtronic electrode was placed around the proximal thoracodorsal nerve" and connected to a Medtronic Itrel unipolar pulse generator (Model 7421, Medtronic, Inc., Minneapolis, Minn.), which was placed below the rectus muscle. The wound was closed in layers. After construction of the SMV, the dog was left to

The Journal of Thoracic and Cardiovascular Surgery

recover with no further procedures during the next 3 weeks. Electrical preconditioning. After the 3·week vascular delay period, the stimulator was activated with a Medtronic 7431 programmer. Dogs 1 and 6 were preconditioned with a burst frequency of 25 Hz and a duty cycle of 312 rnsec on, 812 rnsec off, which resulted in 54 SMV contractions per minute. The individual electrical pulses were of 210 j./Sec duration and had a supramaximal amplitude ranging from 1 to 3 volts. Dogs 2, 3, 4, and 5 were preconditioned with a 5 Hz burst (duty cycle: 312 msec on, 812 msec ofJ). Preconditioning stimulation was continuous for 6 to 9 weeks. Implantable mock circulation system. The mock circulation device consisted of two components (Fig. 2). The first component was a segmented polyurethane (Biomer, Ethicon, Inc., Somerville, N. J.) bladder manufactured by Promeon Division, Medtronic, Inc., Minneapolis, Minnesota, which was fixed to the end of a polymethyl methacrylate conduit. It was 8 em long with a volume of 20 mi. The second component was a similar polyurethane bladder fixed within a hermetically sealed acrylic cannister. The two components were connected with a silicone rubber collar. The cannister had a known volume (VI) of between 155 and 175 ml and was specially molded to fit snugly to the chest wall. Each component of the device had a pressure port similar to that used for long-term intravenous therapy. These Vascular Access Ports (Norfolk Medical Products, Skokie, Ill.) were secured under the skin and were accessed percutaneously. The pouch pressure port allowed direct adjustment of filling volumes and measurement of the pressure produced by the SMV. The cannister pressure port allowed direct adjustment and measurement of the pressure within the cannister. Upon implantation of the device during the operation, the bladders and conduit were completely filled with saline. Air was then injected into the cannister and the fluid-filled conduit was vented so that the cannister bladder collapsed completely. A valve was effectively produced by the collapsed bladder within the cannister, thus isolating the pressure in the cannister (PI) from the pressure in the SMV when the SMV was at rest. This permitted independent adjustment of filling pressures (preload) within the SMV and independent adjustment of the cannister air pressure (afterload). When the SMV contracts, it begins to empty only after it has generated sufficient pressure to overcome the resting pressure (PI) within the cannister. When ventricular pressure exceeds the cannister pressure (PI)' the cannister bladder begins to accept fluid and distends. As

Volume 92 Number 4 October 1986

Autologous biologic pump motor 7 3 5

pouch pressure port

Fig. 2. Mock circulation device.

the effective air volume within the cannister decreases, the air pressure within the cannister must increase according to Boyle's law. At the completion of the SMV's contraction and ejection, the air within the cannister has a new smaller volume (V 2) and a new higher pressure (P 2) . As the SMV relaxes, the higher cannister air pressure (P2) collapses the cannister bladder and returns the fluid to the SMV. The cannister had then returned to its initial volume (VI) and its initial pressure (PI)' Because VI was known at the time of construction of the cannister and both PI and P 2 were measured directly, V2 could be determined according to the Boyle's law (PIV I = P 2V2 at constant temperature). The ambient atmospheric pressure had to be added to both P; and P 2 so that

v, = (P, + Patm) X V,/(P, + Patm) The stroke volume of the SMV was simply V I - V 2• Before each mock circulation system was implanted, the cannister was calibrated empirically. Known volumes in 1 ml increments were added to the cannister bladder up to the capacity of the bladder (20 ml) and the changes in the cannister pressure were recorded. All cannisters exhibited a pressure rise of 4.5 to 5.5 mm Hg per milliliter of stroke volume. There was a good correlation between the volume injected and that calculated theoretically from pressure changes in the cannister. After the final measurements were taken when the dog was sacrificed, some cannisters were recalibrated. No differences between the pre-implant calibration and the post-termination calibration were observed. In one acute pilot experiment (dog 7), a mock circulation system was connected to an SMV immediately after it had been acting as a diastolic counterpulsator, continuously for 4 hours, within the animal's own

Fig. 3. The mock circulation system connected in situ to the SMV.

systemic circulation. The pressures and flows produced by the same SMV against the mock circulation were compared to those produced against the systemic circulation. Implantation of the mock circulation device. After several weeks of electrical preconditioning, a second sterile operation was performed. A second left flank incision was made dorsal to the previous one. An extensive subcutaneous pocket was made for the mock circulation device. The portion of the mandrel extending out from the SMV was dissected free. The SMV itself was left completely undisturbed. The mandrel was cut free from the Teflon collar and extracted. The polyurethane pouch bladder was inserted into the SMV cavity. The sewing ring of the bladder was then sutured to the Teflon anulus of the SMV. The cannister was placed into the subcutaneous pocket and connected to the acrylic conduit. The pressure ports were sutured to the midline dorsal fascia and the wound was closed in layers (Fig. 3). Air was added to the cannister so that the resting cannister pressure (afterload) was 80 mm Hg. The amount of saline was adjusted so that the filling pressure in the SMV (preload) was 40 mm Hg. Measurements. After implantation of the mock circulation system and while the dog was still under anesthesia, thresholds and supramaximal voltages were determined. Resting compliance of the SMV was determined by emptying the ventricle and then recording the resulting pressure changes as saline was added in 1 ml increments against a cannister pressure of 80 mm Hg. In every case the previously implanted electrical burst stimulator was then activated at the following settings: 25 Hz; duty cycle 312 msec on, 812 msec o..o;·supramaximal amplitude ranging from 1 to 3 volts. These settings resulted in 54 SMV contractions per minute. Continuous recordings of the cannister and SMV pressure traces were obtained at an afterload of 80 mm Hg and a

The Journal of

736

Acker et al.

Thoracic and Cardiovascular Surgery

Table I. Duration of continuous pumping and reasons for termination Dog No.

Fig. 4. Mock circulation. Day 8 of continuous pumping. Trace B is the pressure generated by the SMV. At rest the preload is 40 mm Hg. At peak systole the pressure is 110 mm Hg. On relaxation the pressure falls initially to zero because the SMV is now empty. As fluid is returned from the cannister bladder, the diastolic pressure increases until it has reached its preset preload (40 mm Hg). Trace A is the pressure within the cannister (afterioad) against which the SMV is pumping. The afterioad at rest is 80 mm Hg. At peak systole the pressure within the cannister matches that produced by the SMV. On relaxation the pressure falls, as fluid is returned to the SMV, until the initial afterioad pressure has been reached (80 mm Hg). The change in the pressure of trace A represents flow as described previously.

preload of 40 mm Hg for the first 30 minutes. After this, recordings under various preloads, afterloads, and burst patterns were obtained. The parameters were then returned to the chronic settings, i.e., an afterload equal to 80 mm Hg, a preload equal to 40 to 50 mm Hg, a burst frequency of 25 Hz, and 54 SMV contractions per minute. The SMV was stimulated to pump continuously at these settings without interruption. Measurements were obtained either daily or every other day. The cannister pressure (afterload) varied slightly with changes in atmospheric pressure and was adjusted daily or every other day as needed. Except for the days of implantation and sacrifice, recordings were obtained in the awake animal. Huber needles (20 gauge) were inserted percutaneously into the vascular access ports after the skin had been infiltrated with 1% lidocaine. A few animals were mildly sedated with acepromazine maleate (0.5 to 1.0) rng/kg) during pressure measurements. The SMVs were allowed to pump continuously until technical problems arose.

Days of continuous work

Reasonfor termination

35

Small leak in bladder, day 15; gradual increase in leak Ingrowth of connective tissue around anulus, obstructing flow Conduit disconnected from cannister Anulus separated from SMV Herniation of SMV bladder at anulus Outflow obstruction from day I; at termination, anulus separated and bladder in SMV twisted

2

17

3

8

4 5

13 15

6

63

Results The beagles tolerated the latissimus dorsi SMV well and were able to move about freely, with no apparent physical impairment or discomfort. Because the muscle was completely detached from the humerus, contraction of the SMV was independent of the left foreleg, allowing ambulation without difficulty. The length of time for which each dog's SMV pumped against the mock circulation system and the reasons for termination are given in Table I. This does not include the 6 to 9 weeks of continuous isometric contraction during muscle preconditioning before insertion of the mock circulation device. A typical pressure trace from an SMV working against an implanted mock circulation system set at the chronic settings stated previously is shown in Fig. 4. The pressure and flow produced against the mock circulation and that produced against the systemic circulation within the same dog at almost the same time are shown in Fig. 5. The pressures and flows generated by the SMV in the two different systems were similar. In addition, the flow calculated by the change in the mock circulation pressure using Boyle's law correlated well with that measured by the electromagnetic flow probe (C). The pressures and flows produced by dogs 1, 2, 3,4, and 5 over 2 weeks of continuous pumping 54 times per minute with a 25 Hz burst frequency against an afterload of 80 mm Hg and a preload of 40 to 50 mm Hg are shown in Fig. 6. The systolic pressure at implantation was 135 ± 24 mm Hg, at 8 days it was 107 ± 8 mm Hg, and at 2 weeks 104 ± 1 mm Hg (Fig. 6, A). The flow (stroke volume X 54 beats/min) at implantation was 464 ±

Volume 92 Number 4 October 1986

Autologous biologic pump motor

737

Fig. 5. Traces Band b are the pressures produced by the SMV against the mock circulation and systemic circulation, respectively, in the same dog and at almost the same time. When the SMV pumped against the systemic arterial circulation it was synchronized with the heart beat I :3 and was stimulated to contract during diastole. Note diastolic augmentation occurring (Trace a). Traces A and a are the pressure changes produced in the mock circulation cannister and carotid arterial pressure, respectively. Traces C and c are the electromagnetic flow recordings. The flow generated by the SMV when connected to the mock circulation was similar to that generated when the SMV pumped against the animal's own system circulation. Traces D and d represent the 25 Hz burst stimulation for 312 msec.

116 ml/rnin, at 8 days it was 256 ± 28 ml/rnin, and at 2 weeks 206± 61 nil/min (Fig. 6, B). Maximum systolic pressures and maximum flows for dogs 1, 2, 3, 4, and 5 measured over time are depicted in

Fig. 7. Maximum systolic pressure was taken as that produced by an 85 Hz burst against an afterload of 200 mm Hg and a preload of 60 mm Hg. On day zero (30 minutes after initiation of work against the mock

The Journal of

7 3 8 Acker et al.

Thoracic and Cardiovascular Surgery

Fig. 6. Continuous systolic pressure (Aj and corresponding flows (Bj produced by the SMV of each dog over 2 weeks of continuous pumping (afterload 80 mm Hg, preload 40 to 50 mm Hg).

circulation), the maximum systolic pressure generated was 243 ± 11 mm Hg; at 8 days it was 205 ± 19 mm Hg, and at 2 weeks 214± 8 mm Hg (Fig. 7, A). Maximum flow was taken as that produced by an 85 Hz burst against an afterload of 60 mm Hg and a preload of 50 mm Hg. On day zero (30 minutes after initiation of work against the mock circulation), the maximum flow was 697 ± 32 ml/min, at 8 days 474 ± 87 mljmin, and at 2 weeks 369 ± 86 mljmin (Fig. 7, B). Both the continuous and maximal systolic pressures produced by dogs 1 and 6 for longer periods of time are shown in Fig. 8. There was essentially no decrease in either the continuous or maximum pressures, produced

over 5 and 9 weeks, respectively, by the SMVs of these two animals. The relationship of flow to changes in preload and afterload at a burst frequency of 25 Hz on day 7 are shown in Fig. 9. Significant flow was generated at preloads as low as 10 mm Hg. The relationships of stroke work and ejection fraction to afterload and burst frequency at a constant preload pressure, after 1 week of continuous pumping, are depicted in Figs. 10 and 11. The systolic pressure that was produced by the SMVs increased as the afterload against which it was pumping was increased. In contrast to pressure, the flow

Volume 92 Number 4 October 1986

Autologous biologic pump motor 7 3 9

Fig. 7. Maximum systolic pressures (A) maximum flows (B) that could be produced by the SMV of each dog over 2 weeks of continuous pumping. A, Afterload 200 mm Hg; preload 60 mm Hg. B, Afterload 60 mm Hg; preload 50 mm Hg.

decreased as the afterload increased. Therefore, stroke work changed little over the range of afterloads studied, but did increase with burst frequency (Fig. 10). The amount of stroke work produced remained remarkably stable over 1 week's time. Like flow, the ejection fraction of SMVs varied inversely with afterload and directly with burst frequency. Except at afterloads of 200 rom Hg, the ejection fraction ranged between 50% and 90%, depending on the burst frequency and afterload (Fig. 11). Only the ventricular cavities were examined during implantation of the mock circulation device (second

operation). A very smooth fibrous lining covered the inner ventricular surface at this time. At termination, extensive collateralization was noted grossly and the SMVs had formed numerous new collaterals with the subcutaneous tissue and chest wall adjacent to them. The SMVs retained their initial shape although their cavities appeared smaller. The muscle in both the inner and outer layers appeared healthy. In all cases the muscle was still contracting when the dog was sacrificed. There was a buildup of connective tissue around the Teflon sewing ring, which caused varying degrees of outflow obstruction in some ventricles (Table

The Journal of Thoracic and Cardiovascular Surgery

7 4 0 Acker et al.

260 240 220

.:

200

~ I

E

180

,5

160

~

140

~

~

120

.~

100

1,-

80

.

p-------o.

..........

//

+--.....c.:'/

<, '0"...,;

LJ

---+.................. +_------+-...........

.....+,

0"/

',\ \,\

0-

B

~

+-----+ Dog #6 Max. Press. o------{] Dog # I Max. Press.

60 40

+--+ Dog #

20

0----0 14

0

21

28

42

35

6 Cant. Press

Dog # I Cont. Press 49

56

63

Days of Continuous Pumping

Fig. 8. Continuous and maximum systolic pressures produced by SMVs I and 6 over 5 and 9 weeks of continuous pumping, respectively. Afterload 80 mm Hg; preload 40 to 50 mm Hg.

600

500

400

C'

E

<, 0

~

300

• e 0

200

100

0 10

20

30

40

50

60

preload (mmHg)

o Afterload 60 mmHg

+ Aftertoad 80mmHg

o Afterload 100 mmHg

Fig. 9. Flow produced versus preload and afterload after I week of continuous pumping (25 Hz burst frequency).

I). In some SMVs (Table I) the Teflon anulus had either partially or completely separated so that outflow was obstructed. Histologic examination of samples from SMVs 1 and 6 indicated some fiber splitting and regeneration plus thickening of perimysial connective tissue septa, which were continuous with the newly formed connective tissue lining of the cavity. Sections stained histochemically for myofibillar adenosinetriphosphatase showed that the stimulated muscles had been transformed to a uniformly type I (fatigue-resistant) composition and had a dense

capillary blood supply. Sections stained histochemically for the demonstration of NADH* tetrazolium reductase showed a dense reaction product in all fibers, with clumping of the stain in the periphery of many fibers. Discussion These mock circulation experiments provided a means of evaluating SMV function chronically, in the conscious animal, and with full control of the physiologic *Nicotinamide-adenine dinucleotide, reduced.

Volume 92 Number 4 October 1986

Autologous biologic pump motor 7 4 1

1.8 1.5 1.4 1.3 1.2 1.1

r
Q '.~ e'~ • c

~.

f ~

:t 00

s:

~

0.9 0.8 0.7 0.8 0.5 0.4 0.3 0.2 0.1 0 80

80

IZZl

lSSI

25 Hz Bum

100

Aftorlood (mmHg) 4J Hz Bum

200 ~

85 Hz Bum

Fig. 10. Stroke work versus afterload and burst frequency on day 7 (preload 50 mm Hg).

100 90 80 70

II

C

•e • 0.

80 50 40

30 20 10 0 80

60

IZZl

25 Hz Bum

lSSI

100

Afterlood (mmHg) 43 Hz Bum

200 ~

85 Hz Bum

Fig. 11. Ejection fraction versus afterload and burst frequency on day 7 (preload 50 mm Hg).

load. This model allowed us to answer the basic question of whether an SMV was capable of this type of repetitive work in an experiment uncomplicated by potential difficulties, such as blood surface interactions and complex synchronous burst pacing, that would have been introduced by connecting it to the systemic circulation chronically. A comparison of the pressure and flow produced against the systemic and mock circulations (Fig. 5) by the same SMV at almost the same time illustrates the similarity between the two circuits. The cannister bladder system is similar to the Starling heart-lung preparation." According to the Frank-Starling relationships, cardiac muscle is like skeletal muscle

in that the energy of contraction is a function of the length of the muscle fibers before contraction. As the fiber length increases, the energy of contraction also increases up to an optimal length. Any further lengthening will cause the energy of contraction to decrease." In an elegant acute experiment, Spotnitz, Merker, and Malm" demonstrated in canine rectus muscle pouches that the same geometric and physical laws relate intraventricular pressure and skeletal muscle force during both systole and diastole. Stroke volume is a function of the extent of ventricular shortening at any level of diastolic length and contractility. 10 As is true with the heart, the SMV has an

The Journal of

742

Acker et al.

inverse relationship between afterload and stroke volume." We also found that the stroke work (stroke volume X pressure) of the SMV remained relatively unchanged over the ranges of afterloads studied. The stroke volume was high at low afterloads and low at high afterloads, which resulted in little change in total stroke work. As with the heart, stroke work did decline sharply when, at very high afterloads, little or no flow was generated. II Contractility reflects the intensity of the active state of the muscle. Unlike the heart, which is an electrical and mechanical syncytium, skeletal muscle is modulated by the number and rate at which fibers are activated." It has been demonstrated by Dewar and associates,I as well as by our laboratory," that single electrical stimulus, resulting in a single muscle twitch, does not normally generate sufficient force to augment cardiac function. However, rapid repetitive stimuli delivered before the muscle fiber completes its relaxation results in mechanical summation (until fusion occurs), which thereby causes the muscle to generate substantial force." The burst stimulation frequency of the SMV governs the cumulative duration of the active state of the skeletal muscle and produces an effect similar to the contractility of the heart. Increasing the burst frequency of the SMV produces more work. The ejection fraction of the SMV compares favorably with the ejection fraction of the left ventricle. Ejection. fraction, like stroke volume, is inversely related to afterload. Over the past several years, our laboratory has been studying ways in which skeletal muscle can actively assist the heart.v":" The general concept, however, was first introduced by Leriche and Fontaine" in 1933. Kantrowitz and Mckinnon" in 1959 attempted to use a skeletal muscle pump experimentally for cardiac assistance. Other investigators have demonstrated short-term hemodynamic benefits of skeletal muscle for cardiac assistance. However, muscle fatigue inevitably occurred from minutes to hours after the muscle was asked to assume the unrelenting demands of the myocardium.!' Past investigators have usually conducted their studies of SMVs on acute preparations immediately after their construction. 1·5 In many cases, ischemia from the acute ligation of collateral blood vessels necessary to construct the SMVs probably occurred and thereby contributed to the rapid fatigue. Using radioactive microspheres, we have demonstrated in previous experiments that in unconditioned muscle, division of collateral blood supply results in a drastic reduction in muscle blood flow during exercise. Incorporating a delay of 3

Thoracic and Cardiovascular Surgery

weeks between muscle collateral ligation and exercise permits substantial recovery of the exercise-induced increase in muscle blood flow necessary to prevent muscle ischemia.25 Skeletal muscle has the ability to adapt to different patterns of use, although the extent of this capacity was not fully appreciated until the introduction of the chronic stimulation technique by Salmons and Vrbova." Since then Salmons," Pette,30,JI and others have shown that chronic low-frequency electrical stimulation of skeletal muscle results in the acquisition by that muscle of fatigue-resistant properties. In this transformed muscle, there are increases in the capillary density, the activity of oxidative enzymes, and the mitochondrial volume. Thus experiments conducted for the most part in rabbits have shown that chronic electrical stimulation results in a muscle that has many of the properties of myocardium. In past experiments we have demonstrated that a similar transformation occurs in canine muscle in response to electrical stimulation at rates as low as that of the average heart rate- 14·18, 24, 25 and that the fatigueresistant characteristics of electrically conditioned muscle remain stable over 1 year of continuous stimulation." We J3 have also demonstrated that chronic electrical burst stimulation of the type and frequencies required for meaningful work are well tolerated by the muscle and also result in a transformed fatigue-resistant muscle. In acute studies we" have connected the electrically preconditioned SMVs to the systemic circulation, where they have functioned as diastolic counterpulsators, generating flows of 20% of the cardiac output for 8 hours. Using radioactive microspheres, we" have additionally demonstrated that neither the inner nor outer layers become ischemic while pumping continuously against the systemic circulation. In an acute experiment in which similar SMVs were connected to the systemic circulation as a diastolic counterpulsator, SMV stroke work was 650 ± 234 ergs X IOJ. The stroke work of the left and right ventricles was 1,948 ± 890 ergs X 103 and 219 ± 131 ergs X 103, respectively." In contrast, after 1 week of continuous pumping against the mock circulation, SMV stroke work was 606 ± 275 ergs X IOJ. Increasing the burst frequency to 85 Hz more than doubled the stroke work. For example, the SMV of dog 3 after 1 week of continuous pumping was generating a continuous stroke work of 1,190 ergs X 103, flows of greater than 400 ml/rnin, and systolic pressures of 143 mm Hg. When the burst frequency was increased to 85 Hz, the pressure obtained (preload 40 mm Hg; afterload 80 mm Hg) was

Volume 92 Number 4 October 1986

168 nun Hg, the flow 700 ml/min, and the stroke work 2,253 ergs x 103• After an initial decrease, the continuous and maximal pressure produced by the SMVs stabilized and showed no indication of decreasing further. However, flow did gradually decrease although it was still significant at 2 weeks. This decrease in flow accounted for the decrease seen in stroke work. The decreased stroke volume or flow may have been caused by the gradual decrease in cavity size and compliance of the SMVs. The partially permeable polyurethane bladder may have contributed to the gradual decrease in ventricular compliance and resting volume by allowing the cavity of the SMV to gradually shrink as fluid was absorbed slowly from the bladder, although those small amounts of fluid were replaced every day or two. An SMV connected to the systemic circulation, where filling pressure is maintained, may not undergo cavity shrinkage. A different design of the SMV utilizing a larger ventricular volume might create a ventricle of higher compliance. This would make possible greater stroke volumes and higher flows. This possibility is under present investigation. Another possible cause for the decrease in stroke work over time was the chronic preload setting of 40 to 50 mm Hg, which forced the ventricle to function on the downward slope of its ventricular function curve. This high preload pressure, and resultant overstretching of the muscle, may have been responsible for the histologic abnormalities and the buildup of connective tissue around the outlet of some of the SMVs. Such problems were not encountered in previous experiments, in which muscle was stimulated chronically in situ, and may also have resulted from the pattern and conditions of stimulation or preconditioning used in this experiment, which have yet to be optimized. The mock circulation system itself is a rather large amount of foreign material that no doubt also contributed to the buildup of connective tissue in this area. In the human heart the muscle force that must be developed in the cardiac wall to produce systemic arterial pressure is about 500 gm/cm'. This is well below the 2 kg/ern' of force typical for skeletal muscle." This study confirms that appropriately constructed latissimus dorsi SMVs are capable of work approaching the same magnitude as the left ventricle and certainly greater than that of the right ventricle. It further suggests that an appropriately constructed SMV can produce meaningful amounts of work at relatively low preloads (10 nun Hg), a fact that may allow for direct coupling with the heart. Neilson and associates" have pointed out that SMVs used as assist devices might not necessarily be

Autologous biologic pump motor 7 4 3

required to assume the entire function of the left ventricle but rather only a fraction of the cardiac output. In these experiments we have demonstrated for the first time that ventricles constructed from skeletal muscle are capable of doing significant and useful continuous work in a living animal for weeks at a time. Muscle fatigue still looms as a potential problem but one that should be solvable. Skeletal muscle, like the heart, is an engine that transforms energy from food through chemical reactions into useful work. From this study a chronic nonfatiguing biologic pump motor constructed from skeletal muscle appears to be achievable. This concept holds promise for long-term augmentation or replacement of failing cardiac ventricles. We gratefully acknowledge the assistance of Drs. Aida Khalafalla and Arthur Coury, Senior Scientists, Medtronic Corporation, Minneapolis, Minnesota with the burst stimulator and the polyurethane bladders. We would also like to thank Fred DiMeo for his dedicatedcare of the animals used in this study. REFERENCES Dewar ML, Drinkwater DC, Wittnich C, Chiu RC: Synchronously stimulated skeletal muscle graft for myocardial repair. J THORAC CARDIOVASC SURG 87:325-331, 1984 2 Kusba E, Stule W, Sawatoni S, Jaron D, Freed P,

Kantrowitz A: A diaphragmatic graft for augmenting left ventricular function. Trans Am Soc Artif Intern Organs 19:251-257,1973 3 Spotnitz HM, Merker L, Maim JR: Applied physiology of

the canine rectus abdominis. Force-length curves correlated with functional characteristic of a rectus powered "ventricle." Potentialfor cardiac assistance. Trans Am Soc Artif Intern Organs 20:747-756, 1974 4 Vachon BR, Kunov H, Zingg W: Mechanicalproperties of diaphragm muscle in dogs. Med BioI Eng Comput 13:252260, 1975

5 VonRecum A, Stirke JP, Hamada 0, Baba H, Kantrowitz A: Long-term stimulation of a diaphragm musclepouch. J Surg Res 23:422-427, 1977 6 Mannion JD, Hammond R, Stephenson LW: Canine latissimus dorsi hydraulic pouches. Potential for left ventricular assistance. J THORAC CARDIOVASC SURG 91:534544, 1986 7 National Academy of Sciences: Guide for the Care and Use of Laboratory Animals. NIH Publications No. 80-23, revised 1978 8 Patterson SW, Piper H, Starling EH: The regulation of the heart beat. J Physiol (London) 48:465, 1914

9 Milnor WR: The heart as a pump, Medical Physiology, Vol 2, V Mountcastle, ed., St. Louis, 1980, The C. V. Mosby Company, p 922

The Journal of Thoracic and Cardiovascular

7 4 4 Acker et al.

10 Parmley WW, Talbot L: Heart as a pump, Handbook of Physiology, Vol I, Section 2, RM Berne, ed., Bethesda, Md., 1979, American Physiological Society, pp 440-442 II Braunwald E, Ross J: Control of cardiac performance, Handbook of Physiology, Vol I, Section 2, RM Berne, ed., Bethesda, Md., 1979, American Physiological Society, pp 538-539 12 Johnson E: Force-interval relationship of cardiac muscle, Handbook of Physiology, Vol I, Section 2, RM Berne, ed., Bethesda, Md., 1979, American Physiological Society, p. 475 13 Carson FD, Wilkie DR: Muscle Physiology, Englewood, N. J., 1974, Prentice-Hall, p 33 14 Armenti FR, Bitto T, Macoviak JA, Kelly AM, Chase CT, Hoffman BK, Rubenstein NA, St. John-Sutton M, Edmunds LH, Stephenson LW: Transformation of skeletal muscle for cardiac replacement. Surg Forum 35:258260, 1984 15 Bitto T, Mannion JD, Hammond R, Macoviak JA, Rashkind WJ, Edmunds LH Jr, Stephenson LW: Pectoralis and rectus abdominis for potential correction of congenital heart defects, Pediatric Cardiology, Proceedings of the Second World Congress of Pediatric Cardiology, New York, 1986, Springer-Verlag, pp 609-612 16 Bitto T, Mannion J, Hammond R, Cox J, Yamashita J, Duckett SW, Salmons S, Stephenson LW: Preparation of fatigue-resistant diaphragmatic muscle grafts for myocardial replacement, Progress in Artificial Organs, Y Nose, C Kjellstrand, P Ivanovich, eds., Cleveland, 1986, ISAO Press, pp 48-52 17 Bitto T, Armenti F, Hoffman RK, Rubinstein NA, Stephenson LW: Time course of transformation of dog diaphragm muscle with continuous low frequency stimulation at 10 Hz and 2 Hz, Proceedings of Second Vienna Muscle Symposium, 1985, pp 175-179 18 Hoffman BK, Gambke B, Stephenson LW, Rubenstein NA: Myosin transitions in chronic stimulation do not involve embryonic isoenzymes. Muscle Nerve 8:796-805, 1985 19 Macoviak JA, Stephenson LW, Spielman S, Greenspan A, Likoff M, St. John-Sutton M, Reichek N, Rashkind WJ, Edmunds LH Jr: Electrophysiological and mechanical characteristics of diaphragmatic autograft used to enlarge the right ventricle. Surg Forum 31:270-271, 1980 20 Macoviak JA, Stephenson LW, Spielman S, Greenspan A, Likoff M, St. John-Sutton M, Rieichek N, Rashkind WJ, Edmunds LH Jr: Replacement of ventricular myocardium with diaphragmatic skeletal muscle. Acute studies. J THORAC CARDlOYASC SURG 81:519-527, 1981 21 Macoviak JA, Stephenson LW, Alavi A, Kelly AM, Edmunds LH Jr: Effects of electrical stimulation on diaphragmatic muscle used to enlarge right ventricle. Surgery 90:271-277,1981 22 Macoviak JA, Stephenson LW, Kelly AM, Likoff MJ, Reichek N, Edmunds LH Jr: Partial replacement of the right ventricle with a synchronously contracting diaphrag-

Surgery

23

24

25

26

27

28

29

30

31 32

33

34

35

36

37

matic skeletal muscle autograft, Proceedings of the Third Meeting of the International Society for Artificial Organs vol 5, 1981, pp 550-555. Macoviak JA, Stephenson LW, Armenti F, Kelley AM, Alavi A, Mackler T, Cox J, Palatianos GM, Edmunds LH Jr: Electrical conditioning of in situ skeletal muscle for replacement of myocardium. J Surg Res 32:429-439, 1982 Mannion JD, Bitto T, Hammond R, Rubenstein NA, Stephenson LW: Histochemical and fatigue characteristics of conditioned canine latissimus dorsi muscle. Circ Res 58:298-304, 1986 Mannion JD, Velchik M, Alavi A, Stephenson LW: Blood flow in conditioned and unconditioned latissimus dorsi muscle (abstr), Second Vienna Muscle Symposium, 1985, p 28 Leriche R, Fontaine R: Essai experimental de traitment de certains infarctus du myocarde et de l'aneuvrisme du coeur par une greffe de muscle strie. Bill Soc Nat Chir 59:229, 1933 Kantrowitz A, McKinnon WMP: The experimental use of the diaphragm as an auxiliary myocardium. Surg Forum 9:266-268, 1959 Salmons S, Vrbova G: The influence of activity on some contractile characteristics of mammalian fast and slow muscles. J Physiol (London) 210:535-549, 1969 Salmons S, Henriksson 1: The adaptive response of skeletal muscle to increased use. Muscle Nerve 4:94-105, 1981 Pette DL, Vrbova G: Neural control of phenotype expression in mammalian muscle fibers. Muscle Nerve 8:676689, 1985 Pette D: Activity-induced fast to slow-transitions in mammalian muscle. Med Sci Sports Exerc 16:517-528,1984 Acker MA, Mannion JD, Brown WE, Salmons S, Henriksson J, Bitto T, Gale DR, Hammond R, Stephenson LW: Canine diaphragm muscle after one year of continuous electrical stimulation at 2 Hz and 4 Hz. Its potential as a myocardial substitute. J Appl Physiol (in press) Mannion JD, Acker MA, Khalafalla AS, Salmons S, Stephenson LW: Electrical pulse trains to activate latissimus dorsi muscle chronically for potential cardiac augmentation (abstr). J Am Coli Cardiol 7:10, 1986 Mannion JD, Acker MA, Hammond RL, Stephenson LW: Circulatory assistance with canine skeletal muscle ventricles. Acute studies. Surg Forum (in press) Mannion JD, Velchik MA, Acker MA, Hammond R, Alava A, Stephenson LW: Transmural blood flow of multilayered latissimus dorsi pouches during circulatory assistance. Trans Am Soc Artif Intern Organs 15:14, 1986 Mommaerts WFHM: Heart muscle, Circulation of the Blood: Men and Ideas, AP Fishman, DW Richard, ed., Bethesda, Md., 1964, American Physiological Society, p 152 Neilson IR, Brister SJ, Khalafalla AS, Chiu RC: Left

Volume 92 Number 4 October 1986

ventricular assistance in dogs using a skeletal muscle powered device for diastolic augmentation. Heart Transplant 4:343-347, 1985

Discussion DR. RA Y CHU-JENG CHIU Montreal, Quebec, Canada

I congratulate Dr. Acker and his colleagues for their continued contribution in demonstrating that fatigue resistance can be induced in the skeletal muscle. This study is another important step in showing that such a muscle has the potential to be used to assist the circulation for a medium to long-term period. In our own laboratory at McGill, we have taken the approach of connecting such a muscle-powered pouch to the circulation from the outset, to assure that the assist provided is hemodynamically relevant and significant. We also have used a canine latissimus dorsi muscle transformed into an almost pure type 1 fatigue-resistant muscle, as described by Dr. Acker. Then we connect a simple balloon device to the thoracic aorta and wrap the balloon by the transformed muscle. Using a synchronized pulse-train stimulator, which we developed for this particular purpose, we can summate the contractile period and force of the muscle, time them to compress and empty the balloon during diastole, and achieve hemodynamically significant counterpulsation. This totally implantable extra-aortic counterpulsation device is undergoing long-term study at present. However, to avoid stagnation of blood in such a balloon, we think our second generation device being developed will be more suited for patients. The muscle is relaxed during systole but made to contract during diastole to squeeze the "hydraulic" balloon, which transmits energy to the blood chamber and returns the blood to the circulation to achieve diastolic augmentation. The continuous flow through the blood chamber prevents stagnation. In perspective, we are calling this approach "biomechanical cardiac assist," which may supplement the current purely mechanical or purely biological (i.e., transplantation) approach in managing patients with intractable heart failure. Dr. Acker's work seems to confirm that such a goal will be attainable. I wish to ask Dr. Acker how he envisions his SMV to be connected to the circulation in the future. DR. JOHN A. MACOVIAK Stanford, Calif

Like some things, conditioned skeletal muscle gets better with time. These authors have unmasked the inherent ability of skeletal muscle to be stimulated fatiguelessly at rates similar to those of cardiac contractions. Pioneer investigators led by Kantrowitz in 1959 were limited by muscle fatigue in their attempts to augment ventricular function. Recognition of work by Salmons and other muscle plastologists justified our reexploration of this field. In the past 5 years since these concepts were first joined and

Autologous biologic pump motor

745

advanced, many things have evolved. Dr. Carpentier and Dr. Magovern have both reinforced left ventricular repairs in two separate patients. As these and most previous authors have targeted left ventricular failure, at Stanford we have undertaken an approach to the right ventricle. Right ventricular free wall myoventriculoplasty and neoventricle myografts can provide total pulmonary blood flow at programmable transpulmonary pressures, intended for use in entities currently treatable only by organ transplantation. The pressures producible by this entity, the neoventricle, can provide transpulmonary blood flow at pressures of up to 125 mm Hg under conditions of elevated pulmonary artery pressures or left ventricular dysfunction, features that currently restrict the Fontan operation. 1 would ask the authors how they foresee conditioning skeletal muscle in patients with intractable left ventricular failure. In other words, will skeletal muscle in those particularly ill patients be able to be conditioned? Also, as Dr. Chiu asked, if the particular entity that they are proposing is placed in the circulation, as our neoventricle has been, what do they foresee doing about the nature of the muscle surface thrombogenicity problem, which must be resolved before permanent placement of these entities in circula tion? D.R. HENRY M. SPOTNITZ New York, N. Y.

Drs. Maim, Merker, and I had some experience in this area about 10 years ago. We investigated factors relating force in a linear muscle model, the rectus abdominis, to pressure in a hydraulic chamber formed from the same muscle. We found that the observed "systolic" pressure correlated well with predictions based on simple geometry and force generated in situ. The mean isovolumic pressure, which exceeded 400 mm Hg, was more than adequate for circulatory assistance, even with minimal stimulation. End-diastolic pressure was high, however, and volume performance was disappointing unless the muscle was tetanized. Conditioning with direct rather than neural pacing proved of little value. Although we were able to demonstrate both counterpulsation and left heart bypass in our experiments, fatigue proved too hard to beat and led us to abandon our efforts until new ideas or instrumentation, similar to the results presented by Dr. Acker, could provide new insights. The present work is important not only because it presents a practical application of chronic conditioning, but also because of a geometric principle overlooked in prior work. The value of tetany in our preparation lay primarily in increasing stroke volume rather than in increasing pressure performance. One of the great values of the thick-walled structure of the normal left ventricle is that it allows high volume efficiency with less than 10% shortening of sarcomeres at the epicardial surface and 15% shortening at midwall, providing ejection fractions in excess of 50%. The use of a thick-walled model in the present study may be an innovation of equal importance to chronic conditioning for volume performance.

7 4 6 Acker et al.

My question relates to precisely what increment in performance can be provided by conditioning. Reference to athletics suggests that while conditioning is important, it is unlikely alone to make an Olympic champion out of an ordinary citizen, or even an ordinary cardiac surgeon. However, this would appear to be necessary to produce effective hemodynamic performance from skeletal muscle. Therefore, can we hope to achieve a right or left ventricle totally and chronically powered by skeletal muscle alone, or should we look to more modest goals? DR. ACKER (Closing) Dr. Chiu has been one of the early investigators in this field. We are presently investigating various techniques for connecting SMVs to both the right and left circulations. Dr. Mannion has previously reported one method of using the SMV as a diastolic counterpulsator." Dr. Macoviak was the first, with Dr. Stephenson, to point out the importance of muscle conditioning in this field of research." It is unlikely that SMVs will be applicable for

The Journal of Thoracic and Cardiovascular Surgery

patients in acute cardiogenic shock. We see its primary role in those individuals with end-stage chronic left ventricular failure. The problem of thrombogenicity must be addressed now that we have shown that skeletal muscle can do continuous useful work for prolonged periods. The knowledge gained in the field of blood surface interactions, the artificial heart, and heart valves should help in solving this problem. Dr. Spotnitz is a pioneer in this field. The benefit of conditioning is to avoid muscle fatigue. Electrical conditioning alone does not increase the strength of the muscle. Conditioned SMVs, however, are capable of the work required to augment the failing heart. The stroke work of one of our SMVs, after 7 days of continuous pumping, was 1,200 ergs X 10J. This compares favorably with canine left ventricular stroke work of 1,900 ergs X 103 and right ventricular stroke work of 300 X IOJ• Therefore, we believe that a nonfatiguing, conditioned SMV is quite capable of the work necessary to augment a failing ventricle.