A versatile intracorporeal ventricular assist device based on the thoratec VAD system.

BACKGROUND: As patients are supported for longer durations with paracorporeal Thoratec left ventricular and biventricular assist devices (longest durations: 515 and 457 days, respectively), there is a need for implantable options. METHODS: We are developing a small, simple, and versatile intracorporeal ventricular assist device (IVAD) for left, right, or biventricular support as an alternative to the large, implantable, pulsatile left ventricular assist device (LVAD) systems available today. The new device is based on the Thoratec paracorporeal VAD that has been used in more than 1,400 patients weighing from 17 to 144 kg and for durations exceeding 1 year including patient discharge (using the portable driver). RESULTS: The IVAD has the same blood flow path and Thoralon polyurethane blood pumping sac as the paracorporeal VAD, but the housing is a smooth contoured, polished titanium alloy. The IVAD has a new sensor to detect when the pump is full and empty, and is controlled with the Thoratec TLC-II portable VAD driver, which is a small, briefcase-sized, battery-powered, pneumatic control unit. A small flexible (9 mm OD) percutaneous pneumatic driveline for each VAD is tunneled out of the body from the LVAD or right VAD in a pre- or intraperitoneal position. Small size and simplicity are the major advantages of the new device. The IVAD weight (339 g) and implanted volume (252 mL) are approximately one-half that of the current implantable pulsatile electromechanical LVAD systems. CONCLUSIONS: The small size of the IVAD should not only allow support of a large range of patient sizes and body habitus, but also provide options for implantable left, right, or biventricular support. By implanting only the mechanically simple blood pump, the more complex control unit is external, where it can be serviced and replaced without surgery. The IVAD with the portable driver will be a viable alternative to large implanted electromechanical systems and should address a larger segment of the physically diverse patient population.

Characterization and work optimization of skeletal muscle as a VAD power source.

Powering a ventricular assist device (VAD) with skeletal muscle in a linear configuration will require the understanding of basic muscle mechanics and efficient use of available power. Accordingly, a mathematical model incorporating aspects of the Hill equation has been developed. This model relates whole muscle length, force, velocity, and time during cyclic contraction to investigate coupling with a hydraulically actuated VAD. Parameters of the model have been determined from in vivo isometric and isotonic measurements of electrically stimulated pig latissimus dorsi with the humerus insertion reattached to a hydraulic loading system. The in vivo results show an exponential passive force-length relationship and active isometric forces increasing from 2 to 8 kgf over a 5 cm change in length. The maximum shortening velocity extrapolated from isotonic data in 85 cm/sec. With the experimentally determined parameters, the model system of differential equations was optimized computationally. Predicted maximum cycle work and corresponding muscle force are nonlinear functions of contraction duration; an increase in duration yields little improvement in work output for longer contraction times. The model helps clarify VAD system design parameters for optimal muscle coupling; for example, the model predicts that operating at maximum instantaneous power does not optimize stroke work.

In vivo measurements of skeletal muscle in a linear configuration powering a hydraulically actuated VAD.

Linear contracting skeletal muscle can provide more power and physiologic efficiency for cardiac assistance than muscle wrapping configurations. In this study, the insertion of the porcine latissimus dorsi muscle was removed from the humerus and reattached to a muscle powered ventricular assist device (MVAD), consisting of a mechanical to hydraulic piston energy convertor coupled to a Thoratec VAD. Effects of muscle preload stretch and thoracodorsal nerve stimulation parameters on in vivo unconditioned muscle work and MVAD stroke volume were studied. Stroke work increased linearly with muscle preload, and the slope of this relationship (Mprsw) provided an index of muscle "contractility," similar to the preload-recruitable stroke work relationship for the heart. With 5 V, 220 microseconds stimulation pulses over a 200 msec contraction period at 60 bpm, the Mprsw increased with stimulation frequencies from 0.055 J/cm at 30 Hz to 0.149 J/cm at 60 Hz, and to 0.212 J/cm at 90 Hz. Stroke work up to 1 J was achieved during muscle shortening of 2.5 cm with forces up to 6 kgf and energy convertor pressure of 112 psi (approximately 760 kPa). This produced an ejected MVAD stroke volume of 40 ml into a systolic pressure of 92 mmHg on a mock loop, at a filling pressure of 10 mmHg. The MVAD is designed as an alternative to cardiac transplantation, to provide completely implantable circulatory support free from batteries and other power conditioning hardware required with electromechanical systems.

In vivo studies of an implantable energy convertor for skeletal muscle powered cardiac assist.

A device that harnesses the mechanical energy of skeletal muscle contracting in a linear configuration has been implanted in goats. This energy convertor transforms muscle work to hydraulic energy that could drive a variety of cardiac assist devices. The device is mounted with a rib clamp and plate affixed to the sternum by cortical bone screws. A transcutaneous hydraulic line carries a silicon based working fluid to an external system that controls the muscle load. In 60 to 70 kg goats, the latissimus dorsi insertion was reattached to the energy convertor. A Telectronics myostimulator with intramuscular electrodes stimulated the latissimus dorsi. In acute implants, hydraulic pressures in excess of 150 psi were obtained. Chronic implantation of the device allowed system evaluation in the conscious unanesthetized animal. Two weeks after implant, hydraulic pressures in excess of 200 psi were obtained and energy transferred to the external loading system exceeded 1 J per contraction. Six weeks after implant, the device continued to cycle freely. These initial results are very promising and suggest an implantable energy convertor is feasible. Development of an energy convertor is an important step toward tether-free skeletal muscle powered cardiac assist devices.

Mechanical advantage of skeletal muscle as a cardiac assist power source.

The insertion of unconditioned latissimus dorsi muscle at the humerus was reattached in anesthetized goats to the first stage piston of a two stage mechanical-to-hydraulic energy convertor for a skeletal muscle powered ventricular assist device. To study a range of forces, pistons of different cross sectional areas were evaluated during thoracodorsal nerve stimulation. Each energy convertor piston was coupled hydraulically to an actuator on a Thoratec VAD (Thoratec Laboratories, Berkeley, CA) in a mock circulatory loop. Maximum force (70.1 +/- 10.8 N) was greatest for the largest piston, and stroke length (4.0 +/- 0.7 cm) was greatest for the smallest piston. However, maximum stroke work (1.2 +/- 0.5 J) and muscle powered VAD ejected stroke volume (45 +/- 17 ml) were greatest for the middle size piston. These results are consistent with a biomechanical model of whole muscle contraction that predicts that there is an optimum force that produces maximum cycle work. Thus, with a two stage energy convertor, by changing the ratio of the cross sectional areas of the energy convertor and muscle powered VAD actuator pistons, the effective mechanical advantage for the muscle can be optimized to produce more work output and muscle powered VAD flow. Skeletal muscle powered devices using such an energy convertor could provide completely implantable circulatory support free from batteries and other power conditioning hardware required with electromechanical systems.

Passive characteristics of conditioned skeletal muscle for ventricular assistance.

Systems to drive a ventricular assist device (VAD) with power from skeletal muscle have been proposed and are under development. During VAD filling, these systems must counter passive muscle force to control the precontraction length and optimize power output. To determine how muscle conditioning with electrical stimulation alters basic biomechanical characteristics and influences available power, goat latissimus dorsi were evaluated in vivo after an 8 week training protocol with an implanted myostimulator. Conditioned muscles displayed increased passive stiffness. After conditioning, the slope of the exponential passive force-length relation, at a passive force of 10 N, significantly increased from 5.1 to 7.6 N/cm (p = 0.003). Similarly, for a passive force of 10 N, the length relative to the zero developed force length decreased from 5.5 to 4.2 cm (p < 0.014). The linear relationship between slope (dF/dL) and force (F) also demonstrated a significant intercept shift. The latter relationship is independent of absolute length. Consistent with other studies, conditioning also resulted in fatigue resistance, fiber type transformation, and reductions in maximum developed force and shortening velocity. In the context of available power for cardiac assist, the results demonstrate that the influence of passive characteristics is accentuated after conditioning and has a substantial effect on available power.

Development of an intracorporeal Thoratec ventricular assist device for univentricular or biventricular support.

There is a need for a small, simple, and versatile intracorporeal ventricular assist device (IVAD) as an alternative to the large implantable electromechanical LVAD systems in current use. Because the basic design of the Thoratec paracorporeal VAD has been demonstrated in over 1,000 patients, weighing from 17 to 144 kg, and for durations up to 515 days including patient discharge (by using the portable driver), we are developing a new intracorporeal version of our VAD. This IVAD has a smooth contoured, polished titanium housing, and maintains the same blood flow path and Thoralon polyurethane blood pumping sac as the paracorporeal VAD. The IVAD is controlled with the Thoratec TLC-II Portable VAD Driver, which is a small briefcase sized, battery powered, pneumatic control unit. Intracorporeal LVADs and/or RVADs are implanted in a preperitoneal position, with a single small (9 mm OD) percutaneous pneumatic driveline for each VAD. The major advantages of the new IVAD design are size and simplicity. The IVAD weight (339 g) and implanted volume (252 ml) are substantially smaller than current implantable electromechanical LVAD systems. Only the small blood pump is implanted, leaving the more complex control unit external, where it can be serviced and replaced. The versatile design is intended for left and/or right heart support in large or small patients. The IVAD in combination with the TLC-II portable driver will be a viable and attractive alternative to large, implanted electromechanical systems.

Chronic implantation of a skeletal muscle energy convertor for cardiac assist devices: a preliminary report.

An efficient energy convertor capable of driving a variety of cardiac assist devices is being developed in goats. Muscle work in a linear configuration is converted to hydraulic energy and transmitted to an external test system that controls muscle loads during shortening contractions. This investigation focuses on the variation of muscle characteristics and optimal power output during muscle conditioning. The energy convertor was mounted on the rib cage, the latissimus dorsi insertion reattached to the device, and percutaneous hydraulic lines exited near the spine. Following device, stimulator, and intramuscular electrode implantation, a progressive conditioning protocol was initiated. Weekly biomechanical muscle characterization was performed in the conscious animal, with single twitch and tetanic contractions performed under isometric and isotonic conditions. The characterization data provide a measure of available power, as well as inputs, for a computer simulation that predicts optimal muscle power output and operating conditions. These ongoing implants provide insight into the available muscle power and suggest an implantable energy convertor is feasible. Development of an energy convertor is an important step toward tether free skeletal muscle powered cardiac assist. These studies will be expanded in number and duration to further investigate the effects of conditioning and identify improvements in device development.

Sustained skeletal muscle power for cardiac assist devices: implications of metabolic constraints.

A device to harness power from skeletal muscle contracting in a linear configuration is under development. This application requires a sustained level of power that is dependent upon muscle mechanics and metabolic properties. A biomechanical muscle model and a metabolic model constructed from experimental data were used to predict maximum power available in a sustainable region of loading and stimulation conditions. Latissimus dorsi (LD) of four goats were evaluated in vivo after a 10 week in situ conditioning protocol with an implanted Telectronics myostimulator. The LD insertion was reconnected to a hydraulic loading system, allowing isometric and isotonic contractions for biomechanical characterization. Metabolic utilization was measured by a thermister based myothermic technique. Brief fatigue tests of working isotonic contractions revealed stimulation conditions associated with sustained power. The results show metabolic utilization was dependent on contraction duration, rate, force, and stroke. The region of sustainable contractions was found for a range of durations of 0.1 to 0.6 sec and rates of 10 to 120 bpm. The boundary for the sustainable power region was well approximated by a constant value of metabolic utilization. A constant duty cycle (contraction to cycle duration ratio) also approximated the sustained power but differed by as much as 30% during the shorter contraction durations. The results demonstrate that a mechanical muscle model can predict maximum sustained power when the operating conditions are constrained to a sustainable range determined by a metabolic model. Furthermore, metabolic constraints influence the optimum conditions for sustained power needed in the design of skeletal muscle powered assist devices.

Evaluation of a skeletal muscle energy convertor in a chronic animal model.

A device is under development for powering cardiac assist devices with skeletal muscle contracting in a linear configuration by converting muscle work to hydraulic energy. Prototype devices are being implanted in goats to study device performance and associated muscle mechanics. Percutaneous hydraulic lines provide the means to control muscle load and evaluate muscle performance during an electrical conditioning protocol. Chronic implant durations ranged from 36 to 87 days in 7 goats. The latissimus dorsi muscle (LDM) insertion was reconnected to the device with a tendon loop. A sternal plate attached with bone screws, and a rib clamp secured the device. A new modular sternal mount design was implemented to eliminate plate loosening that complicated early implants. Extensive bone remodeling around the rib clamp was observed. The tendon attachment demonstrated sufficient initial strength; however, in five implants, efforts to repair the tendon were required. Device encapsulation was observed, but the device continued to cycle freely and no tethering adhesions to the device were found. Interactions between the capsule wall and LDM seemed to limit LDM movement in some cases. Development of a long-term animal model for energy convertor evaluations is an important step toward skeletal muscle powered cardiac assist.

Relative metabolic utilization of in situ rabbit soleus muscle: thermistor-based measurements and model.

Electrically conditioned skeletal muscle can provide the continuous power source for cardiac assistance devices. Optimization of the available sustained power from in vivo skeletal muscle requires knowledge of its metabolic utilization and constraints. A thermistor-based technique has been developed to measure temperature changes and to provide a relative estimate for metabolic utilization of in situ rabbit soleus muscle. The relative thermistor response, active tension and muscle displacement were measured during cyclic isometric and isotonic contractions across a range of muscle tensions and contraction durations. The thermistor response demonstrated linear relationships versus both contraction duration at a fixed muscle length and active tension at a fixed contraction duration (r(2)=0.90+/-0.14 and 0.70+/-0.21, respectively; means +/- s.d.). A multiple linear regression model was developed to predict normalized thermistor response, DeltaT, across a range of conditions. Significant model variables were identified using a backward stepwise regression procedure. The relationships for the in situ muscles were qualitatively similar to those reported for mammalian in vitro muscle fiber preparations. The model had the form DeltaT=C+at(c)F+bW, where the constant C, and coefficients for the contraction duration t(c) (ms), normalized active tension F and normalized net work W were C=-1.00 (P<0.001), a=5.97 (P<0.001) and b=2.12 (P<0.001).