| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 1995;109:1127-1137
© 1995 Mosby, Inc.
CARDIOPULMONARY BYPASS, |
West Lafayette and Marion, Ind., and Houston, Tex.
Supported by a grant from the Trask Fund, Purdue University, West Lafayette, Ind.
Received for publication April 1, 1994. Accepted for publication August 4, 1994. Address for reprints: L. A. Geddes, ME, PhD, FACC, Purdue University, The William A. Hillenbrand Center for Biomedical Engineering, 1293 A. A. Potter Engineering Center, West Lafayette, IN 47907-1293.
Abstract
The pumping and left ventricular unloading capabilities of the left ventricular, ventricular synchronous skeletal-muscle ventricle were determined in nine anesthetized dogs ranging in weight from 20.7 to 31.8 kg. The ventricular synchronous skeletal-muscle ventricle consists of the left rectus abdominis muscle wrapped around a 4-mil-thick polyethylene pouch (wrapped volume 80 to 100 ml) connected to the left ventricular apex with no valve and to the aorta via a prosthetic heart valve. The rectus muscle is timed to contract tetanically and relax during left ventricular ejection. This arrangement provides a high precontraction pressure for the rectus muscle and a high muscle capillary blood flow during skeletal muscle relaxation. The timing signal for initiation of the train of stimulating pulses (40/sec) was derived from the ventricular electrogram. The delay for the stimulus train determines the preload for the rectus muscle and along with the stimulus train duration determines ventricular synchronous skeletal-muscle ventricle stroke volume, which was measured by electric impedance. With unconditioned rectus muscles (70 to 120 gm) and with a pumping ratio of 1:3, ventricular synchronous skeletal-muscle ventricle stroke volume averaged 26.1 ml, which provided an average output of 876 ml/min. The normalized ventricular synchronous skeletal-muscle ventricle output was 35.6 ml/min per kilogram of body weight. In a typical resting dog (and man), the normalized cardiac output is 70 ml/min per kilogram. Therefore the ventricular synchronous skeletal-muscle ventricle is capable of pumping 52% of the cardiac output (with a pumping ratio of 1:3). The optimum train delay from the apex of the ventricular electrogram ranged from 10 to 100 msec. The left ventricular ejection period averaged 309 msec, and this determines the time available for the rectus muscle to contract and relax. Evidence for unloading the left ventricle is shown by the reduced left ventricular diastolic pressure and stroke volume for the postassisted beats. (J THORAC CARDIOVASC SURG 1995;109:1127-37)
There are two critical requirements for the optimum use of electrically stimulated skeletal muscle, wrapped around a blood-filled pouch, to assist the circulation. First, a precontraction pressure on the order of aortic pressure is needed to obtain a forceful contraction. 
1,2 Second, a high muscle capillary blood flow is necessary to avoid muscle fatigue. 3,4 The ventricular synchronous skeletal-muscleventricle (VS-SMV) described herein satisfies both requirements. 5 The VS-SMV consists of a muscle-wrapped pouch connected directly to the apex of the left ventricle (LV) with no valve and to the aorta via a prosthetic heart valve. Therefore LV and VS-SMV pouch pressure are always the same: low during ventricular diastole (ensuring a high muscle capillary and myocardial blood flow) and high during ventricular systole when the skeletal muscle is timed to contract, thereby ensuring a high preload and forceful contraction. The objective of this report is to demonstrate the importance of timing to obtain LV unloading and circulatory assistance provided by the VS-SMV in a short-term animal model with the use of unconditioned rectus abdominis muscles.
Theory of operation
The VS-SMV is applied as shown in Fig. 1. There is no valve between the LV and the pouch, which is wrapped with the left rectus abdominis muscle. At the outlet of the pouch is a prosthetic heart valve at the entrance of the conduit that leads to the aorta. The method used to attain the dual goal of a low pouch diastolic pressure, followed by a high precontraction pressure, is based on precise timing of the tetanic contraction of the rectus muscle relative to LV systole. When the desired pouch precontraction pressure is reached, the rectus muscle is contracted tetanically by an appropriately delayed train of stimuli after the R wave of the electrocardiogram (ECG). The pulse train duration determines the duration of muscle contraction and is chosen so that the skeletal muscle is completely relaxed at the end of the LV ejection period. The pulse train delay determines the precontraction pressure (preload) and is selected so that the skeletal muscle starts to contract at the beginning of LV ejection.
|
6 was the first to point out that there are two types of skeletal muscle: (1) fast twitch (pale) and (2) slow twitch (red), with the latency between the stimulus and maximum force of contraction being longer for slow-twitch (fatigue-resistant) muscle. However, when a muscle is contracted by nerve electrodes there are additional delays. First, there is a propagation time for the nerve action potential to reach the myoneural junctions. Second, there is a myoneural junction delay. Third, there is a propagation time for the muscle action potential. Fourth, there is a latent period between muscle excitation and the development of force, this latter time being known as the latent period, which is typically 40 to 100 msec for slow-twitch muscles and 5 to 10 msec for fast-twitch muscles. 7 Of all of these delays, the latent period is the longest. Not only is there a latency between the onset of the stimulus train and the development of muscle force, there is also a relaxation latency, that is, the time between termination of the stimulus train and complete muscle relaxation. In preliminary dog studies with muscle-wrapped pouches in which the motor nerves were stimulated, we found that the contraction latency is on the order of 70 msec and the relaxation latency is about 100 msec. These two latencies have an important bearing on selection of the onset and termination of the VS-SMV stimulus train so that the skeletal muscle contracts only during the LV ejection period and is relaxed by the end of the ejection period.
As illustrated in Fig 2, A, for the skeletal muscle to start contracting at the onset of LV ejection, which occurs about 100 msec after the peak of the first ventricular excitation wave (Q or R) in the dog, the stimulus train must be initiated 100 - 70 (30) msec (delay) after the Q or R wave, whichever is earliest. Likewise, for the skeletal muscle to be fully relaxed at the end of LV ejection, which occurs about 400 msec after the Q or R wave (whichever is first), the stimulus train must be terminated at 400 - 100 (300) msec. Therefore the maximum duration of the stimulus train, under these circumstances, is 300 - 30 (270) msec, as shown in Fig. 2, A.
|
From the foregoing it is clear that the VS-SMV stroke volume depends on the train delay, which determines the preload on the skeletal muscle, and on the train duration, which controls the duration of muscle contraction. Reducing train duration reduces VS-SMV stroke volume and increasing train duration increases VS-SMV stroke volume; but there is a long-duration limit for train duration, namely one that permits complete relaxation of the skeletal muscle by the end of the LV ejection period to avoid restriction of LV filling.
When any LV unloading technique is evaluated, it is necessary to recognize that the Frank-Starling law is bidirectional, that is, increasing preload increases the force of contraction and decreasing preload decreases the force of contraction. Fig. 3 (inset) shows LV pressure of an isolated coronary-perfused canine heart as the preload is increased with each beat. The graph of force of contraction versus resting tension identifies two regions: (1) the Frank-Starling region (0-B) and (2) a failure region (beyond B). Choosing an operating point A, increasing the preload increases the force of contraction. However, if the preload is decreased, the force of contraction decreases, that is, unloading decreases the force of contraction as shown. Therefore, when an LV-unloading device is evaluated, the force of contraction of the postassisted beat will be less if the LV precontraction pressure is lower. Consequently the stroke volume of the postassisted (unloaded) beat will be less than that for the preassisted beat.
|
All studies described herein were done in accordance with the Guide for Care and Use of Laboratory Animals, published by the U.S. Department of Health, Education and Welfare (DHHS Publication No. NIH 85-23, revised 1985). In addition, this study was approved by the Purdue University Animal Care and Use Committee.
Nine mongrel dogs ranging in weight from 20.7 to 31.8 kg were anesthetized with intravenous thiopental (10 mg/kg) and a cuffed endotracheal tube was inserted. Anesthesia was maintained with a mixture of isoflurane and oxygen. Lead II ECG and femoral artery pressure were monitored continuously.
Surgical procedure
A midline skin incision was made from the ventral midthorax to just inferior to the umbilicus to provide exposure of the left rectus abdominis muscle, which was dissected free from its attachments. The muscle was then transected just below the umbilicus and its aponeurotic origin divided. The anterior portion of the muscle is supplied by the inferior intercostal nerves and care was taken to preserve the five or six branches that supply the muscle belly. The vascular pedicle was preserved and the muscle remained attached to the body by only the intercostal nerves and the epigastric blood supply.
The rectus muscles (70 to 120 gm) were wrapped spirally around the empty cylindrical pouch (Fig 4, A) made of 4-mil-thick polyethylene tubing. The muscle was trimmed so that a single layer encompassed the pouch. The unwrapped volume was typically 200 ml; the muscle-wrapped volume was 80 to 100 ml. Ribbed polytetrafluoroethylene (PTFE)* grafts (15 mm) were used as conduits to the pouch and a St. Jude Medical bileaflet valve (St. Jude Medical, Inc., St. Paul, Minn.) (19 mm) was placed in an adapter at the outlet of the pouch facing the aorta. The length of the conduit from the LV to the pouch was made as short as possible to ensure rapid filling. A 12 mm diameter plastic stent was sewn to one end of a PTFE conduit.
|
Aortic flow velocity was measured with an ultrasonic perivascular flow probe (Transonics Systems, Inc, Ithaca, N.Y.) placed at the root of the aorta. LV pressure was measured with a catheter passed therein via the left atrial appendage. This method was used to avoid compromising aortic valve function.
Pouch volume measurement
Stroke volume of the VS-SMV was measured by the change in electric impedance (Z) appearing between hollow sleeve electrodes 1 and 2 in Fig. 1, within the ends of the pouch (Fig. 4, A), in the manner previously described by us. 8,9 In brief, a direct-coupled, 50 kHz impedance recorder measured the pouch impedance and its increase with ejection. Calibration was achieved by crossclamping the two PTFE inlet and outlet conduits and blood was withdrawn in 5 ml increments from the pouch with a syringe while the pouch impedance was recorded. A calibration curve relating ohms change (
Z) to the volume of blood withdrawn was then constructed for each animal; Fig. 4, B, illustrates the calibration curve for each animal.
Stimulator
The trains of tetanic stimuli were delivered to the five or six rectus abdominis motor nerves via a stimulus isolation unit driven by two Grass stimulators (Grass Instruments, Quincy, Mass.). The motor-nerve electrodes were connected to the negative pole of the stimulus isolation unit. The positive pole was connected to the indifferent electrode, which was placed caudal to the rectus muscle to direct the return-path stimulating current away from the sensing electrodes on the right ventricle. The second stimulator generated each train of tetanic stimuli (40/sec, each pulse being 0.1 msec). The first stimulator gated the second stimulator on and off and thereby controlled the stimulus train duration, which determined the duration of muscle contraction. The first stimulator was activated by an electronic circuit that was triggered by the first deflection in the ventricular electrogram (Q or R) detected by two electrodes on the base of the right ventricle. The circuit allowed control of the delay between the R wave and the onset of the train of stimuli. The circuit also permitted delivery of a train of stimuli for each ventricular electrogram wave or a submultiple thereof, that is, 1:2 to 1:10.
Train delay studies
The stimulus train delay (d) determines the preload on the skeletal muscle and therefore the VS-SMV stroke volume. To ensure accurate measurement of LV stroke volume and LV plus VS-SMV stroke volume, the outlet (valve) end of the conduit was crossclamped. Therefore emptying of the VS-SMV was via the LV. With the use of a train duration of 150 msec, the augmentation in aortic root flow velocity was determined for delays ranging from 10 to 100 msec with a pumping ratio of 1:5. In all of the animals, the times between the apex of the first ventricular wave (Q or R) to the beginning and end of ejection were measured by the aortic root flow-velocity recording.
Train duration studies
After the optimum delay was determined, studies were done to determine the effect of stimulus train duration (TD) on VS-SMV stroke volume measured by the blood-calibrated pouch-impedance record.
Pumping capability studies
In a previous study, 10 we showed that on a per gram basis, unconditioned skeletal muscle can deliver about one third of the power of cardiac muscle. Therefore, for the pumping studies, we selected a pumping ratio of 1:3. Under this condition, the VS-SMV was contracted with each third heart beat and the volume pumped was the VS-SMV stroke volume (measured by pouch impedance), multiplied by the heart rate divided by three. Various train delays (d) and train durations (TD) were used.
RESULTS
Fig. 5 is a record of the electrogram, LV/pouch pressure, pouch impedance (volume), and aortic root flow velocity from a 24 kg dog. To make this record the outlet (valve) end of the pouch was crossclamped so that all of the output passed through the flowmeter at the root of the aorta. On the left, the LV is unassisted and the stroke volume was 22.7 ml; the heart rate was 96 beats/min; therefore the control cardiac output was 96 x 22.7 = 2179 ml/min. In the center of the record the VS-SMV was activated with 200 msec trains of stimuli, delayed 75 msec from the R wave and activated with a pumping ratio of 1:3. The SV for the assisted beat (1) was 36.7 ml, 16 ml for the second beat (2), and 23.4 ml for the third beat (3). Therefore, during the period of cardiac assistance, the total blood flow was 32 (36.7 +16 + 23.4) = 2435 ml/min, which is 12% more than the control value. However, during the period of assistance, the VS-SMV stroke volume was 37 ml, as shown by the impedance record; therefore the VS-SMV pumped 32 x 37 = 1184 ml/min, which is 48.6% of the total blood flow during the cardiac assistance period.
|
The work performed is 
PdV, where P and V are pressure and volume. If, in Fig. 5, the PdV for the preassisted beat is taken as 100%, the work done by the LV for the first postassisted beat (2) is 64% of that done by the LV before assistance, that is, the work unloading for this beat was 34%.
As discussed in the Theory section, two factors affect stroke volume: (1) the train delay and (2) the train duration The train delay sets the precontraction load for the rectus muscle wrapped around the pouch. Fig. 6 illustrates the augmentation in aortic flow velocity (with the outlet end of the VS-SMV crossclamped) versus the train delay. Note that the augmentation in aortic root flow velocity (about 160%) was maximum for a train delay from 20 to 80 msec for this 31.8 kg animal. The optimum train delays for the other animals are shown in Table I. The average time from the apex of the first ventricular excitation (Q or R) to the onset of LV ejection was 82 msec (52 to 127 msec) and the average LV ejection period was 309 msec (229 to 400 msec).
|
|
|
|
Table II presents the VS-SMV stroke volume and output (in milliliters per minute) along with the prevailing heart rate and the train duration used. Also shown are the dog weights. The average VS-SMV stroke volume was 26.1 ml and the average VS-SMV output was 876 ml/min. The normalized resting cardiac output for the dog (and man) is 70 ml/min per kilogram; therefore from Table II the VS-SMV was capable of providing from 12.6% to 106% of the normalized cardiac output with a pumping ratio of 1:3. The average normalized VS-SMV output was 35.6 ml/min per kilogram, that is, 52% of the normal resting cardiac output of 70 ml/min per kilogram.
|
There are many reports on the use of electrically stimulated skeletal muscle to pump blood; the various techniques have been reviewed by Chiu. 11 However, there are only a few reports in which the blood pumping was via the LV-to-aorta route. LaFarge and associates 12 appear to have been the first to connect a doubly valved, air-compressed pouch between the LV apex and aorta in a long-term calf study. They reported a considerable reduction in LV work when the pouch was activated. The use of a skeletal muscle to compress a doubly valved pouch at the same site in dogs was reported by Drinkwater 13 and Neilson 14 and their colleagues. Stevens 15 and Badylak 16 and their colleagues investigated the doubly valved LV-to-aorta SMV in dogs with pharmacologically induced heart failure. In the study by Stevens and colleagues 15 , activation of the SMV during ventricular diastole improved the hypodynamic state considerably. In the study by Badylak and colleagues 16 , diastolic pumping with a ratio of 1:2 produced a blood flow equivalent to 45% of the cardiac output with normal ventricles. Lu and associates 17 also reported the use of a doubly valved LV apex-to-aorta SMV, activated during LV diastole with 1:2 ratio in a six-dog study. After a 3-week vascular delay, followed by 6 weeks of muscle conditioning, in short-term studies, the SMVs were able to pump 40% of the blood flow at 3 hours with a pumping ratio of 1:2.
The use of a doubly valved LV apex-to-aorta SMV activated during LV diastole is characterized by a high pumping capability. However, when the skeletal muscle is not contracting, the pressure in the pouch is high, and this sustained high pressure compresses the skeletal muscle capillaries and severely compromises muscle capillary blood flow, 4,5 inviting early fatigue. In the study of Badylak and colleagues 16 the lowest pouch pressure with the SMV not contracting was 39 mm Hg; in the study by Lu and associates 17 the pressure was 57 mm Hg. Both pressures are more than enough to compress the skeletal muscle capillaries. 3,4 The VS-SMV described herein eliminates this problem and provides the needed high precontraction pressure and high muscle capillary blood flow during the diastolic interval because pouch pressure is low and equal to LV end-diastolic pressure.
In the present VS-SMV study, which used a pumping ratio of 1:3, the average VS-SMV output was 356 ml/min per kilogram of body weight. In the normal dog and man at rest, a typical cardiac output is 70 ml/min per kilogram. In other words, the VS-SMV, by itself, pumped 100 x 36.7/70, or 52%, of the cardiac output. However, when one views the pumping capability of the VS-SMV, it is useful to remember that the output depends on the VS-SMV stroke volume and the pumping ratio.
When the VS-SMV is not activated, the LV ejects blood through the aortic valve and the pouch-outlet valve, which can be heard to open and close. In a previous paper 5 we showed that if the output end of the pouch is closed, blood does not stagnate in the pouch. Saline solution deposited therein was washed out as a result of stretching of the pouch by the LV pressure; the pouch recoiled during diastole thus washing out the saline solution. Of course the dead-end pouch washed out more quickly when the VS-SMV was activated.
Regarding activation of the VS-SMV, it is necessary to place stimulating electrodes on all five or six motor nerves that innervate the rectus muscle. A stimulus frequency of 35 to 40/sec provides a smooth, strong tetanic contraction. Pulses of 0.1 msec allow stimulation with the least electrical energy, 18 because this duration is the chronaxie of a typical mammalian motor nerve. Timing of the stimulus train so that the onset of muscle contraction occurs after the onset of LV ejection is essential so that the skeletal muscle experiences a high preload. In the present study we found the optimum train delay from the R wave to be in the 10 to 100 msec range. In our previous dynamic cardiomyoplasty studies, 19 we found that a train delay in the range of 40 to 80 msec caused the latissimus dorsi muscle to contract just before LV ejection when the ventricles bulge maximally and provided additional preload to the latissimus dorsi muscle. However, the optimum train delay from the Q or R wave of the ventricular electrogram cannot be specified a priori for all subjects; nonetheless, the optimum delay can be found quite easily. With a pumping ratio of 1:3 and selection of a short train duration (e.g., 150 msec), measurement of VS-SMV stroke volume as a function of train delay, as shown in Fig. 6, provides the desired information. Although pouch impedance provides the best indication of VS-SMV stroke volume, aortic root blood flow velocity can be used provided the outlet of the VS-SMV is crossclamped so that all of the output is sensed by the aortic root flowmeter. After the optimum delay has been found, the train duration can be increased to increase VS-SMV stroke volume, as shown in Fig. 7.
The optimum train delay depends on the rectus contraction latency (Fig 2, A), that is, the time between the onset of the stimulus train and the onset of muscle contraction. There is very little information on the range of time for unconditioned and conditioned muscle, although it has been found that the contraction latency time increases when skeletal muscle is conditioned. For example, Letson and colleagues 20 showed that the time from the onset of the stimulus train to 50% of maximum contraction force in the unconditioned latissimus dorsi muscle was 80 msec. After conditioning, the time to 50% of maximum force increased to 140 msec. Therefore, to accommodate the increase in contraction latency, it may be necessary to use the P wave of the ECG as the timing reference and to measure VS-SMV stroke volume as a function of stimulus train delay. There are no data on the effect of conditioning on the relaxation latency.
In conclusion, from the foregoing it is clear that the VS-SMV unloads the LV, evidence for which is the reduced diastolic pressure and reduced stroke volume for the postassisted beats. The high VS-SMV stroke volume results from the high precontraction pressure, secured by selection of the optimum train delay. Even with a pumping ratio of 1:3, it has been shown that the VS-SMV by itself is capable of pumping about one half of the cardiac output. Because LV and pouch pressure are always the same, muscle capillary blood flow occurs at the same time as coronary artery blood flow, thereby ensuring good muscle perfusion.
Footnotes
From the Hillenbrand Biomedical Engineering Center,a Purdue University, West Lafayette, Ind.,Indiana Wesleyan University,b Marion, Ind., and Baylor College of Medicine,c Houston, Tex. 
Gore-Tex graft: Gore-Tex is a registered trademark of W. L. Gore & Associates, Inc., Newark, Del.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
