Secondary Logo

Journal Logo

Biomedical Engineering

In Vivo Performance of a Muscle-Powered Drive System for Implantable Blood Pumps

Trumble, Dennis R.*; Melvin, David B.; Dean, David A.*; Magovern, James A.*

Author Information
doi: 10.1097/MAT.0b013e3181733d9e
  • Free


Circulatory support with a left ventricular assist device (VAD) has been shown to reverse the physiologic effects of severe congestive heart failure and improve patient survival and quality of life when compared with medical therapy.1 In short-term applications—as in bridge to cardiac transplantation—this approach has proven to be extremely effective. When used for permanent support, however, VAD technology has been beset with problems associated with the complexities of transcutaneous power delivery and the complications of percutaneous drivelines. An alternative means of power generation that allows a completely self-contained form of mechanical circulatory support would therefore be a major advance in the treatment of heart failure.

The use of skeletal muscle as an endogenous power source represents a promising means by which a completely implantable, tether-free cardiac assist system might be realized. Work to create a practical muscle-powered VAD has focused primarily on developing an efficient, reliable means to capture and transmit energy from electrically stimulated latissimus dorsi (LD) muscle. The impetus behind this approach stems from two essential observations: first, that skeletal muscle can, with chronic activation, be trained to express fatigue-resistant muscle fiber phenotypes2 and second, that trained LD muscle can perform steady-state work at levels compatible with long-term cardiac assistance.3

Anatomically speaking, the key to optimizing muscle energy output is to allow the LD to contract normally with its blood supply intact so that performance limitations inherent to muscle mobilization techniques can be avoided.4 We believe the most effective way to capture this energy for cardiac assist purposes is to attach the LD humeral insertion to an implanted hydraulic pump—called a muscle energy converter or MEC—which, in turn, can be used to actuate a pulsatile VAD.

Lessons learned through six successive design iterations5 have now been incorporated into a seventh-generation device (Figure 1) with significantly enhanced energy transfer and tissue interface characteristics. Here we describe the most recent improvements in MEC design and report results from the latest in a series of canine implant trials.

Figure 1.
Figure 1.:
Computer rendering of the MEC device with accumulator loading system attached. Upper housing is shown as semitransparent to reveal the inner mechanism.

Materials and Methods

Summary of Design Improvements

Four small but important design changes have recently been implemented to improve device durability and provide a more reliable means for muscle fixation. First, a semicircular channel with shelf extension was added to the actuator arm to reduce stress concentrations at the tendon/device interface. Second, the camshaft material was changed from 440C stainless steel to BioDur 316LS stainless to better protect against corrosion. Third, camshaft spacers were moved from the outer race to inside the upper housing to avoid rubbing contact with the spring-loaded lipseals. And finally, camshaft needle bearings were replaced by full-complement (cageless) bearings to reduce shaft wobble and subsequent seal wear. These modifications were aimed specifically at addressing two failure modes identified via accelerated bench testing and long-term implant studies—namely, tendon rupture due to concentration of flexion stresses, and leakage of biological fluids past the camshaft seals (described below).

In addition to the changes to the MEC itself, significant improvements have been made to the in vivo loading system that allows for better control over cycling conditions during long-term implant trials. To be specific, the pneumatic system used in prior work has been replaced by a fluid-based system that uses a sealed, air-filled bellows to provide an adjustable pressure resistance. The problem with previous pneumatic loading schemes was that gas under pressure would invariably diffuse across the walls of the pressure tubing and the septum of the vascular access port, making it impossible to maintain a set resting pressure within the pump over an extended period of time (days to weeks). Because elevated fluid pressures are far easier to maintain, we believe this approach will prove much more effective than those loading methods used to date.5–7

The metallic “accumulator” bellows is housed within a cylindrical chamber that attaches directly beneath the MEC as shown in Figure 2. The housing is fitted with a stainless steel tube and luer connector so that the fluid chamber can be accessed from outside the ribcage. Resting pressures within the pump are adjusted by injecting or withdrawing fluid via a subcutaneous vascular access port (model 72-4323, Harvard Apparatus, Holliston, MA). Accumulator pressures are measured through a y-connector that links the central fluid line to both the access port and one of two implanted pressure catheters. This arrangement allows the use of fluid-filled lines for both load adjustment and pressure measurement thereby eliminating the problem of load instability caused by gas diffusion.

Figure 2.
Figure 2.:
Cutaway view of the MEC and accumulator loading system showing arrangement of the lipseals, cam, bearings, and bellows. Note that the lower, accumulator bellows is filled with air to allow for fluid displacement from the MEC bellows into the accumulator housing.

Experimental Methods

The surgical procedures described below were performed in compliance with the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences and published by National Academy Press (ISBN 0-309-05377-3, 1996). This project was also approved by the Institutional Animal Care and Use Committee of the Allegheny-Singer Research Institute (Pittsburgh, PA).

MEC pumps fitted with hydraulic accumulators were implanted in three mixed-breed dogs (30–35 kg) to assess long-term device function. Each animal was sedated with intramuscular injections of acepromazine (0.25 mg/kg) before induction of general anesthesia with intravenous pentothal (20 mg/kg). Anesthesia was maintained with 1%–2% isoflurane delivered through an endotracheal tube. A longitudinal incision was made over the lateral border of the left LD muscle. The LD was then freed from its humeral insertion and mobilized proximally with minimal dissection, leaving the muscle origin, intercostal perforators, and thoracic neurovascular bundle intact. A four-pole spiral peripheral nerve stimulation lead (custom-built by Case Western Reserve University, Cleveland, OH) was sutured around the main trunk of the thoracodorsal nerve and the LD stimulated using an Itrel3 pulse train stimulator (Medtronic model 7425, Medtronics, Minneapolis, MN). The LD insertion was then fitted with a 16-tow, looped artificial tendon (MyoCoupler, CardioEnergetics, Inc., Cincinnati, OH) described previously8 and shown in Figure 3. A 6.5-cm length of the second rib was removed adjacent to the sternum and the MEC device placed in the gap with the actuator arm set perpendicular to the direction of LD shortening. The device was then secured to the chest wall by tying four wire sutures around the ribs and through the perforated anchor ring. Finally, the terminal loop of the artificial tendon was fixed to the actuator arm and the muscle/pump system tested acutely to confirm proper alignment and full device actuation with muscle stimulation. A lateral chest radiograph (Figure 4) was used to document device and hardware position immediately following surgery.

Figure 3.
Figure 3.:
Artificial tendon with looped termination shown after attachment to the MEC actuator arm. Each side of the loop is split into eight bundles, each comprising 2500–3000 individual 25-micron polyester fibers. These bundles are sewn into the muscle for a length of 4–6 cm and tied together in pairs to prevent pullout before stabilization via fiber ingrowth.
Figure 4.
Figure 4.:
Top: Scale drawing showing the placement of the MEC beneath the LD muscle in a 30-kg dog. Bottom: Lateral radiograph of the MEC implanted in a dog with telemetric monitoring system and muscle stimulation hardware. VAP, vascular access port.

The animals were allowed to recover for 1 week before LD stimulation whereupon muscle contractions were induced using bursts of nine pulses delivered at a rate of 33 Hz. Because it generally takes 6–8 weeks of electrical conditioning to convert skeletal muscles to mostly type-1 fatigue-resistant fibers,9,10 peak contraction rates were limited during the first 2 months of the study period to avoid muscle damage from overstimulation. The complete LD stimulation regimen is detailed in Table 1. Throughout the course of the stimulation period, housing and driveline pressures were monitored daily via telemetry (model TL11M3-D70-PCP, Data Sciences International, St. Paul, MN) and adjusted weekly to the highest level against which the LD could effect full actuator motion (2 cm stroke length).

Table 1
Table 1:
Stimulation Schedule for MEC Implant Dogs

Pressure signals were captured and processed using IOX software (version, Emka Technologies Inc., Falls Church, VA). Pressure swings within the upper housing were measured against bellows volume displacement on the bench, and these data were used to generate a calculated channel for stroke volume estimation. This volume waveform was then plotted against accumulator fluid pressure to create pressure-volume loops from which MEC stroke work was calculated. Load conditions were cycled up and down once per week to measure MEC/LD stroke work capacity and peak driveline pressure production.

At the conclusion of each implant study, the device and LD/MyoCoupler complex were removed for analysis. The dogs were allowed to recover so that possible complications resulting from device removal could be documented. Because no major postoperative problems were encountered, all three dogs were released to adoptive homes.


Two experiments were conducted concurrently using MEC6 pumps fitted with the new hydraulic loading system described above. Both experiments were stopped before their 6-month target dates due to rupture of the artificial tendon interface (Figure 5). These failures were likely caused by a combination of three factors: suture wound around the terminal cord (used to effect acute fixation of the proximal muscle); an abrupt leading edge machined into the actuator arm; and cracking of the plastic sheathing housing the terminal cord. These findings led to the first of the four modifications listed above. Bench tests were performed subsequently to confirm the efficacy of these changes before implant trials were allowed to resume.

Figure 5.
Figure 5.:
Rupture of the looped artificial tendon (right) near to where the cord exits the actuator arm channel (left). A shelf extension was later added beneath the channel openings to distribute stress and reduce the flex angle of the cord.

Examination of these two pumps following explant revealed that a small amount of serous fluid had entered the pumping chamber past at least one of the two opposing lipseals, a situation that did not appear to alter device function but could conceivably lead to pump failure over longer periods of time (months to years). Subsequent component-level inspections found signs of corrosion on the camshaft around which the lipseals operate, a reddish discoloration on the cam and bearing surfaces, and dark staining between the guidepost and bearing mount. These staining patters were originally attributed to the small amount of blood that had entered the housing. Subsequent testing, however, showed that they were likely caused by the dispersal of oxidized granules originating from the bearing mount. These findings led directly to the design improvements listed above, the first three of which were incorporated into the device before the final implant trial described here (long lead times precluded the use of cageless bearings for this study).

In the most recent animal trial, the MEC7 device functioned well for 35 days after implant and generated the highest levels of power production observed to date. Accumulator pressures (expressed as resting pressure over peak contraction pressure) ranged from a low of 181/267 mm Hg on postoperative day 7 to a high of 478/710 mm Hg on day 26. The peak driveline pressure generated in this last experiment was 1743 mm Hg, which translates to a muscle pull strength of about 60 N (13.5 lbs). Steady state power generation was measured daily and trended upward as muscle training progressed (Figure 6). Mean stroke work levels reached 478 ± 21 mJ/stroke (mean ± SD) on day 26 with a maximum value of 785 mJ being achieved on day 28 during a weekly pressure cycle test. MEC/LD stroke work was seen to vary linearly with accumulator preload pressure (Figure 7) with peak values occurring at driveline pressures near 1000 mm Hg.

Figure 6.
Figure 6.:
Plot of MEC stroke work versus implant time. Data were collected daily via radiotelemetry and averaged over five contiguous cycles (mean ± SD). The abrupt increase in pump function on postoperative day 21 was due to an increase in baseline accumulator pressure made to accommodate improved muscle function.
Figure 7.
Figure 7.:
Plot of MEC stroke work versus accumulator resting pressure taken on postoperative day 35. Load conditions were changed by injecting/withdrawing sterile water through a vascular access port. Work production was calculated in real time from accumulator and housing pressure waveforms collected via telemetry.

This final study was stopped prematurely due to loss of fluid pressure in the accumulator system caused by failure of a stainless steel weld at the port/housing interface. This anomalous manufacturing flaw is currently being addressed and is not expected to limit the duration of future implant studies.


Despite recent advances in the treatment and prevention of cardiovascular diseases, the incidence and prevalence of congestive heart failure continues to increase. It is estimated that 2.3 to 3.0 million people currently suffer from congestive heart failure in the United States, and that hundreds of thousands of Americans die each year with heart failure listed as a contributing factor.11,12 Current pharmacologic therapies result in symptomatic improvement in many patients but do not appreciably alter the natural course of the disease. Heart transplantation is a very effective treatment but is limited by a small donor pool and by the serious side effects of immunosuppressive drugs.

Clinical researchers have long maintained that a safe and effective permanent VAD would be a powerful tool for treating end-stage heart failure. This assertion was recently confirmed by clinicians participating in REMATCH, a Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure.1 This trial was conducted over 3 years from May 1998 through July 2001 in 20 cardiac transplant centers around the country. Subjects included 129 patients with end stage heart failure who had not responded to medical management of their disease and were ineligible for heart transplantation. Patients enrolled in REMATCH were randomly assigned to one of two study groups. Sixty-eight patients averaging 68 years of age received the HeartMate I, a vented electric LVAD that is fully implantable and has wearable external components. The control group comprised 61 patients (mean age: 66 years) who received optimal medical management including ACE inhibitors and monthly follow-ups. As a group and regardless of age, the HeartMate patients had a 52% chance of surviving 1 year, compared with 25% of the patients in the control group. The 2-year survival rate was 23% for the HeartMate group versus 8% for the control group. Moreover, quality of life was consistently higher in the LVAD patients at 12 months than in the control group.

Still, despite the proven efficacy of these devices, the complexities of transcutaneous power delivery and problems associated with percutaneous drivelines have severely limited this approach. The use of skeletal muscle as an endogenous power source affords a unique opportunity to bring a completely implantable, tether-free cardiac assist system to fruition. Muscle-powered devices offer an attractive alternative to current long-term support schemes by eliminating the need to transmit energy across the skin, thereby reducing hardware requirements significantly. Through this mechanism, external battery packs, power conditioning hardware, transmission coils, and internal power cells could all be replaced by natural biomechanical processes. This would significantly enhance patient quality-of-life by improving reliability and eliminating all external components. Moreover, muscle-based blood pumps would be much less expensive to implement and maintain, resulting in wider availability and reduced costs for health-care providers.

Harnessing skeletal muscle for long-term circulatory support has been a goal of medical researchers for nearly 25 years now, beginning with the first clinical cardiomyoplasty procedure performed in 1985.13 Since that time numerous methods have been explored to determine how best to capture this contractile energy and deliver it to the bloodstream. Nearly all have employed the LD as the muscle of choice owing to its large size, favorable anatomic location, and transferability. Naturally, the simplest approaches were examined first as surgeons set about wrapping the LD muscle around the heart, aorta, and various other blood-filled conduits in an effort to pump blood by direct muscular compression.14 Unfortunately, poor perfusion and suboptimal muscle mechanics15 have limited the efficacy of this approach, which has now been largely abandoned as a result.

To avoid these problems, we propose to keep the LD in its natural anatomic position in order to maintain its native blood supply and preserve normal contractile dynamics, the idea being to stimulate the muscle in situ and collect the resultant energy at the LD insertion. As expected, the main challenge here has been to engineer an implantable muscle pump that can: (a) fit the space, (b) cause minimal discomfort to the patient, and (c) efficiently capture and transmit this energy without breaking down or binding up.

We are not the first to consider this approach by any means. Indeed, the concept of powering a pump with linearly contracting muscle first appeared in the literature as early as 1964 when Kusserow and Clapp employed a quadriceps femoris muscle to drive a spring-loaded diaphragm pump.16 Numerous investigators addressed this topic in the decades that followed, yet no serious attempts to develop an implantable device were published until Sasaki in 1992.17 His system employed a flexible rod, sheath, crank, and cam to transmit muscle power to a pusher-plate pump that was tested in dogs using untrained LD muscle and a mock circulatory system. At 60 beats per minute, this device maintained 0.8–2.0 L/min for 200 minutes against an afterload of 75 mm Hg. Output power was 2.5 mW/g of muscle and system efficiency approached 50%. Later that same year, Farrar and Hill described a novel linear-pull energy converter for powering various implantable devices.18 A pair of guide shafts was added in subsequent versions to counter the torque created by the offset between the muscle and piston shaft, but friction at the piston’s o-ring seal limited energy transfer efficiency and presented inherent durability problems.19,20 Despite these limitations, however, energy transmission levels of 1.0 J per contraction were achieved in conscious goats 2 weeks after device implantation with peak isometric pressure readings exceeding 200 psi. Shortly thereafter, Takahashi et al. described a “linear-push actuator” with a bellows supported by two interlocking cylinders designed to drive a dynamic patch for ventricular assistance.21 This device worked well in acute trials using dogs but issues of device fixation, tissue adhesion, and biomaterial leaks were never addressed. Most recently, Sasaki and colleagues reported testing a “roller screw linear actuator” that transmits linear muscle power to a pusher-plate pump via a cable wound around a spool attached to a roller screw nut.22 In this case, questions regarding cable fatigue and stable tissue interfaces were left unresolved. These important pioneering studies have each served to advance the conceptual development of muscle-powered implants but none have fully addressed the key issues of biocompatibility and long-term durability. As a result, efforts to date have yet to produce a practical means by which contractile energy may be collected and transmitted in vivo to perform work within the body over an indefinite period of time.

Here we describe the latest in a series of seven prototype MECs designed to transform contractile energy into hydraulic power and report results from three recent canine implant trials. Current device improvements include critical changes to the tendon interface and seal surfaces, which are designed to reduce the risk of tendon rupture and prevent body fluid incursion into the MEC upper housing. We believe these modifications were instrumental in bringing about the most successful implant trial performed to date wherein the MEC/LD complex functioned well for over a month until a problem with the hydraulic loading system ended the experiment prematurely.

Record levels of steady-state stroke work production (478 ± 21 mJ/stroke) and driveline pressure (1743 mm Hg) were achieved after steps were taken to eliminate the tendon fixation problem experienced in the previous two animal implant trials (described above). This is a sizable improvement over levels achieved in previous MEC studies where stroke work and pressure levels topped out at 290 mJ and 430 mm Hg, respectively.5 Because normal left and right ventricular stroke work levels in dogs this size (35 kg) are roughly 700 and 150 mJ, respectively, these data suggest that MEC/LD power levels—maintained in tandem with an appropriate cardiac assist device—are sufficient to provide significant long-term circulatory support.

Future Work

The next phase of this project will involve longer-term implant studies of a similar type. These experiments will run for a minimum of 6 months and will: (a) measure the steady state work capacity of fully conditioned skeletal muscle contracting in situ against a terminal load; (b) document the morphological changes that occur in skeletal muscle working under conditions of long-term circulatory support; and (c) demonstrate the stability and durability of the system as a whole. Once this is accomplished, the MEC will be used to drive a hydraulic VAD compatible with the steady-state power levels measured in vivo.


This work was supported by a grant from the National Institutes of Health (5 R01 HL059896-08).


1. Rose EA, Gelijns AC, Moskowitz AJ, et al: Long-term mechanical left ventricular assistance for end-stage heart failure. N Engl J Med 345: 1435–1443, 2001.
2. Salmons S, Jarvis JC: The working capacity of skeletal muscle transformed for use in a cardiac assist role. In: Transformed Muscle for Cardiac Assist and Repair, Futura Publishing Co., pp. 89–104, 1990.
3. Trumble DR, LaFramboise WA, Duan C, et al: Functional properties of conditioned skeletal muscle: implications for muscle-powered cardiac assist. Am J Physiol 273 (Cell Physiol 42): C588–C597, 1997.
4. Trumble DR, Magovern JA: Muscle-powered mechanical blood pumps. Science 296: 1967, 2002.
5. Trumble DR, Melvin DB, Byrne MT, et al: Improved mechanism for capturing muscle power for circulatory support. Artif Organs 29: 691–700, 2005.
6. Trumble DR, Magovern JA: A muscle-powered energy delivery system and means for chronic in vivo testing. J Appl Physiol 86: 2106–2114, 1999.
7. Trumble DR, Magovern JA: Method for measuring long-term function of muscle-powered implants via radiotelemetry. J Appl Physiol 90: 1977–1985, 2001.
8. Trumble DR, Melvin DB, Magovern JA: Method for anchoring biomechanical implants to muscle tendon and chest wall. ASAIO J 48: 62–70, 2002.
9. Acker MA, Mannion JD, Stephenson LW: Methods of transforming skeletal muscle into a fatigue-resistant state: Potential for cardiac assistance. In: Biomechanical Cardiac Assist, Futura Publishing Co., pp. 19–28, 1986.
10. Hill AB, Chiu RC-J: Dynamic cardiomyoplasty for treatment of heart failure. Clin Cardiol 12: 681–688, 1989.
11. Funk M: Epidemiology of heart failure. Crit Care Nurs Clin North Am 5: 569–73, 1993.
12. Gillum RF: Epidemiology of heart failure in the United States. Am Heart J 126: 1042–1047, 1993.
13. Carpentier A, Chachques JC. Myocardial substitution with a stimulated skeletal muscle: first successful clinical case. Lancet, 1: 1267, 1985.
14. Salmons S, Jarvis JC: Cardiac assist from skeletal muscle: a critical appraisal of the various approaches. Br Heart J 68: 333–338, 1992.
15. 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 11: 237–246, 1996.
16. Kusserow B, Clapp J: A small ventricle-like pump for prolonged perfusions: construction and initial studies, including attempts to power a pump biologically with skeletal muscle. ASAIO Trans 10: 74–78, 1964.
17. Sasaki E, Hirose H, Murakawa S, et al: A skeletal muscle actuator for an artificial heart. ASAIO J 38: 507–511, 1992.
18. Farrar D, Hill J: A new skeletal muscle linear-pull energy converter as a power source for prosthetic circulatory support devices. J Heart Lung Trans 11: 341–350, 1992.
19. Reichenbach SH, Farrar DJ, Diao E, et al: In vivo studies of an implantable energy convertor for skeletal muscle powered cardiac assist. ASAIO J 43: M668–M672, 1997.
20. Reichenbach SH, Gustafson KJ, Egrie GD, et al: Evaluation of a skeletal muscle energy convertor in a chronic animal model. ASAIO J 46: 482–485, 2000.
21. Takahashi M, Misaki T, Watanabe G, et al: Efficacy of a skeletal muscle-powered dynamic patch: Part 1. Left ventricular assistance. Ann Thorac Surg 59: 305–312, 1995.
22. Sasaki Y, Chikazawa G, Nogawa M, et al: Ex vivo evaluation of a roller screw linear muscle actuator for an implantable ventricular assist device using trained and untrained latissimus dorsi muscle. Artif Organs 23: 262–267, 1999.
Copyright © 2008 by the American Society for Artificial Internal Organs