Ann Thorac Surg 2001;72:1336-1342
© 2001 The Society of Thoracic Surgeons
Original article: cardiovascular
Physiologic characteristics of canine skeletal muscle: implications for timing skeletal muscle cardiac assist devices
George V. Letsou, MDa,b,c,d,
James F. Hogan, BSEEa,b,c,d,
Charles C. Miller, III, PhDa,b,c,d,
John A. Elefteriades, MDa,b,c,d,
David Francischelli, MSa,b,c,d,
John C. Baldwin, MDa,b,c,d,
Hazim J. Safi, MDa,b,c,d
a Department of Cardiothoracic and Vascular Surgery, University of Texas Medical School-Houston, Houston, Texas, USA
b Yale Medical School, New Haven, Connecticut, USA
c Dartmouth Medical School, Dartmouth, New Hampshire, USA
d Medtronic, Inc, Minneapolis, Minnesota, USA
Accepted for publication June 11, 2001.
Address reprints requests to Dr Letsou, University of Texas Medical School-Houston, 6410 Fannin, Suite 450, Houston, TX 77030
e-mail: george.v.letsou{at}mail.uth.tmc.edu
 |
Abstract
|
|---|
Background. Optimal clinical stimulation for skeletal muscle cardiac assist systems (such as dynamic cardiomyoplasty) is not clearly defined. The pressure-generating capacity of canine skeletal muscle ventricles (SMVs) at a variety of preloads and stimulation frequencies was examined as was time for SMVs to develop peak pressure.
Methods. SMVs were analyzed just after construction and after 3 months of electrical conditioning. Pressure generation and time to develop peak pressure were determined using a distensible mandrel.
Results. Higher preloads resulted in increased pressure generation; conditioned SMVs generated significantly less pressure than unconditioned SMVs. Increasing stimulation frequency from 20 to 50 Hz increased pressure-generating capacity; increases beyond 50 Hz did not result in further increases. Time to 90% peak pressure was least at 10 HZ and 65 Hz.
Conclusions. Higher stimulation frequencies and preloads result in a more quickly contracting muscle, which generates more pressure. Midrange stimulation frequencies of 30 Hz provide optimal muscle strength and minimize time to develop peak pressure. Initiation of contraction should begin before the time maximal pressure is desired.
|
Introduction
|
|---|
Despite thorough investigations into the basic relationships of muscle contraction and electrical stimulation, the use of skeletal muscle to create effective cardiac assist devices such as cardiomyoplasty wraps and skeletal muscle ventricles (SMVs) has been unrewarding.
The fundamental physiologic characteristics of electrically stimulated skeletal muscle have been extensively investigated [1]. Based on these studies, skeletal muscle cardiac assist devices were developed and extensively investigated by Acker and colleagues [2] and Chiu and others [3]. Such devices require a period of electrical stimulation or conditioning before maximal effectiveness, which results in decreased power generating capacity [4]. The influence of preload and stimulation frequency on pressure generation by the latissimus dorsi is less completely understood; but it is clear from several studies that higher frequencies and higher preload result in higher pressure generating capacity [5].
Despite these extensive investigations, the clinical use of skeletal muscle for cardiac assist has not been as beneficial as predicted. The recent clinical trial of dynamic cardiomyoplasty (Cardiomyoplasty-Skeletal Muscle Assist Randomized Trial [C-SMART]) documented statistically significant clinical symptomatic improvement yet failed to demonstrate objective improvements in cardiac performance such as ejection fraction [6]. The failure of this well-designed trial to document objective improvements in cardiac performance makes an investigation of clinically available stimulation factors of great interest.
We used the experimental model of canine skeletal muscle ventricles (SMVs) positioned subcutaneously outside the chest (which are easily constructed and accessible to physiologic measurements) for an investigation of factors clinically relevant to skeletal muscle cardiac assist devices. The effects of preload and clinically available electrical stimulation frequencies on unconditioned and chronically stimulated conditioned SMV pressure generation was analyzed. We built on our experience using isolated canine latissimus dorsi to identify the latency of skeletal muscle ventricle contraction after nerve stimulation to measure the time to 90% peak tension, an important factor in optimal timing of cardiac assist devices [7]. The magnitude and time course of pressure generation by SMVs after pulse-train stimulation was characterized.
|
Material and methods
|
|---|
Surgical technique
Mongrel dogs weighing 25 to 35 kg were anesthetized using intravenous pentobarbital (20 mg/kg) and halothane (0.5% to 1.5%) as inhalational anesthetic. A volume ventilator (Harvard Apparatus, Inc, Millis, MA) was used to administer inhalational anesthetic and oxygen. After induction of anesthesia, animals were placed in supine position. A unilateral incision was made ventral to the latissimus dorsi. The latissimus was dissected away from the body wall and the thoracodorsal nerve isolated high in the axilla. A 360° nerve cuff electrode (Medtronic, Inc, Minneapolis, MN) was applied loosely about the thoracodorsal nerve such that the diameter of the nerve cuff was 2 to 3 times the diameter of the nerve itself as shown in Figure 1. The latissimus was then wrapped about a silastic mandrel. The silastic mandrel consisted of a cylindrical chamber connected to a spherical chamber by a collar; a port for fluid instillation and pressure measurement was attached to the collar. A sewing cuff placed about the top of the cylindrical portion of the mandrel was used for muscular attachment (Fig 2). Latissimus dorsi was then wrapped about the cylindrical portion of the cuff in two or three layers. Each muscle layer was fastened to the layer above it; stitches were used to attach the top of the muscle wrap to the silastic collar and to close the bottom portion of the muscle wrap over the cylinder. The wrap was fashioned in such a way that when the cylinder was filled with fluid the muscle was under slight tension.
View larger version (15K):
[in this window]
[in a new window]
|
Fig 1. Diagram of 360° nerve cuff electrode wrapped loosely about the thoracodorsal nerve in such a way that its diameter is 2 to 3 times that of the nerve trunk.
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig 2. Silicone elastomer mandrel used for creation of canine skeletal muscle ventricles. Note the cylindrical portion, the spherical portion, and the Dacron sewing cuff.
|
|
The newly formed SMV was analyzed for pressure generating capacity at different preloads and frequencies and for latency of contraction. These analyses are described below. The nerve cuff electrode was then attached to a pulse-train stimulator. Either an Itrel 7421 or a SP 1005 (Medtronic, Inc, Minneapolis, MN) was used for chronic stimulation. All animals were then allowed to recover from the first operation. At 1 to 2 weeks (when the surgical wound was completely healed), the stimulator was activated and SMV conditioning was begun. Contraction rate was initially 26 times per minute and increased progressively to 72 times per minute. Stimulation frequency was 10 Hz for the first 4 weeks and increased to 25 Hz for the following 8 weeks. All conditioning measurements were as outlined in Table 1. The amplitude of the pulse train was maintained at two to three times the visual threshold of SMV contraction (the visual threshold was determined weekly by increasing the pulse amplitude until SMV contraction was first observed).
After 3 months of conditioning, all animals were restudied. Under general endotracheal anesthesia using pentobarbital and halothane as described above, each animal had the SMV dissected away from the chest wall. This dissection was easier than would be anticipated (contraction of the SMV may minimize scarring) and no skeletal muscle ventricles were injured during this dissection. Repeat pressure-generating studies and latency studies were carried out.
Measurement of physiologic variables
Measurements took place at both the first and second surgeries (ie, just after creation of the SMV and after 3 months of electrical conditioning). The amount of pressure generated by each SMV was determined by filling the distensible silastic mandrel to pressures of 10, 30, 80, or 130 mm Hg. The mandrel was accessed through its sidearm port, and pressure was measured through the same port. After the mandrel was filled to the desired pressure (ie, preload), electrical stimulation was carried out at successively increasing frequencies of 2, 10, 20, 30, 50, and 65 Hz (the clinically available frequencies). Other stimulation measurements included a voltage amplitude of 2 times visual contraction threshold, pulse width of 150 µs, and pulse-train duration of 0.6 seconds. These produced SMV contractions of sufficient duration to permit measurement. The pressure generated by the SMV was recorded both in real time and by electrical-mechanical transducer on hard copy. At each preload, a plot of pressure development versus time was produced (Fig 3). Thus, at each frequency and preload a characteristic and unique pressure versus time plot was created. At low frequencies (ie, 2 Hz), the peak of the pressure curve generated was taken as the peak pressure and at higher frequencies (ie, 65 Hz), the plateau of the curve was taken as peak pressure.
View larger version (17K):
[in this window]
[in a new window]
|
Fig 3. Representative training of skeletal muscle ventricle pressure development plotted against time. The time to 50% peak pressure and 90% peak pressure is determined from these plots as indicated.
|
|
The latency of SMV contraction after nerve stimulation was calculated using the hard copy plot of the pressure generated versus time. Time to 90% peak pressure was determined as shown in Figure 3. The determination of time to 90% peak pressure was made using a prolonged pulse train duration of 1.6 seconds (rather than 0.6 seconds used elsewhere in the study) to allow better definition of the maximal pressure plateau.
All animals received humane care in compliance with "Principles of Laboratory Animal Care" and "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences and published by the National Institutes of Health (National Institutes of Health Publication No 85-23, revised 1985). The Yale University Animal Care and Use Committee approved and regulated the use of all animals.
Statistics
Data were analyzed by repeated measures mixed-model analysis of covariance using the SAS mixed procedure version 6.11 (Litell and colleagues; 1996 SAS System for Mixed Models, SAS Institute, Cary, NC). Subject was modeled as a random effect with nesting in the examination period (unconditioned versus conditioned). Preload was modeled as a continuous variable and frequency was modeled as a quadratic. An autoregressive error structure was specified for the model. Because some data were missing owing to wound infection in 2 dogs, type III sums of squares were used to test the fixed effects and least means squares were computed as effect estimates. The null hypothesis was rejected at p less than 0.05. The Figures (Figs 4, 5, 7A and B, and 8A and B) showing the effects of preload, stimulation frequency, and conditioning are two- and three-dimensional spline plots of the least squares means.
View larger version (23K):
[in this window]
[in a new window]
|
Fig 4. Three-dimensional representation of the influence of preload and stimulation frequency on pressure generation by skeletal muscle ventricles in the acute setting (just after surgical construction). Preload is on the x-axis, stimulation frequency is on the y-axis, and pressure generated is on the z-axis.
|
|
|
Results
|
|---|
Ten animals were used in this study and all completed the protocol. Wound infections developed in 2 and they were discarded from the analysis of conditioned animals.
Analysis of stimulation frequency revealed it to be an important factor in the amount of the pressure generated by the SMV in the acute setting. As can be seen from Figures 4 and 7, pressure-generating capacity was greatest of a frequency of 47.2 Hz.
Unexpectedly, the pressure generated at each preload, defined as the peak pressure generated in the SMV minus the preload, did not increase significantly with increasing preload (Fig 4). In other words, the amount of pressure generated by the SMV did not increase significantly with increasing preload.
The time to 90% peak pressure is maximized at a frequency of 20 Hz in the acute setting. Muscle stimulated at a frequency of 20 Hz takes longer to develop peak pressure than at all other frequencies tested (Fig 5). At 20 Hz, the latissimus dorsi/SMV takes an average of 0.20 seconds to develop 90% of peak pressure; at 65 Hz it takes only an average of 0.11 seconds (Fig 6). In part these findings are explained by little pressure being generated at the lower frequencies and preloads and in part by fusion of muscular contraction not occurring at the lower frequencies. This is represented graphically in Figure 6, which shows six different pressure versus time plots for SMV stimulation at the varying frequencies. All these tracings were at the same preload of 80 mm Hg. Clearly, lower frequencies such as 2 Hz result in the development of small amounts of pressure; thus the time taken to develop these very low pressures is short. In addition, fusion of individual muscle fiber contractions into a sustained contraction of the entire muscle does not occur at low frequencies such as 2 and 10 Hz; the desired prolonged plateau of pressure generation does not occur either. Figure 7A and B plot pressure-generating capacity against stimulation frequency and preload, respectively, acutely and after 3 months of conditioning using the stimulation protocol outlined in Table 1. Though pressure generated is still maximized at a frequency of 46.5 Hz, conditioned SMVs generate significantly less pressure than unconditioned SMVs at all stimulation frequencies. Similarly, conditioned SMVs generate significantly less pressure at all preloads as well.
View larger version (23K):
[in this window]
[in a new window]
|
Fig 5. Three-dimensional representation of the influence of preload and stimulation frequency on time to 90% peak pressure by skeletal muscle ventricles in the acute setting (just after construction). Preload is on the x-axis, stimulation frequency is on the y-axis, and time to 90% peak pressure is on the z-axis.
|
|
View larger version (42K):
[in this window]
[in a new window]
|
Fig 6. Representative pressure generated versus time tracings at 6 different stimulation frequencies. Preload is 80 mm Hg in all. Note the lack of muscle fusion and, therefore, lower pressure generation at stimulation frequencies of 2 Hz and 10 Hz.
|
|
View larger version (14K):
[in this window]
[in a new window]
|
Fig 7. Skeletal muscle ventricle pressure generation plotted against preload and stimulation frequency acutely and after 3 months of chronic electrical stimulation. Note the diminution in pressure-generating capacity after chronic stimulation. Curves in both (A) and (B) are statistically different from each other at p < 0.0001 for conditioning effect. Preload is 80 mm Hg.
|
|
Figure 8A and B demonstrates the relationship of time to 90% peak pressure to both stimulation frequency and preload. A conditioned SMV requires an average of 0.32 seconds at 50 Hz and 80 mm Hg preload to develop 90% of peak pressure and only 0.28 seconds at 65 Hz. The average conditioned SMV required only 0.23 seconds to reach 90% of peak pressure at a preload of 10 mm Hg and 50 Hz compared with 0.33 seconds at 130 mm Hg and 50 Hz. Conditioned SMVs require significantly longer to develop peak pressure at all stimulation frequencies and preloads than unconditioned SMVs.
View larger version (15K):
[in this window]
[in a new window]
|
Fig 8. Time to 90% peak pressure plotted against preload and stimulation frequency acutely and after 3 months of chronic electrical stimulation (conditioning). Note the prolongation of time to 90% peak pressure after conditioning. Curves in both (A) and (B) are statistically different from each other at p < 0.0001 for conditioning effect. Preload is 80 mm Hg.
|
|
|
Comment
|
|---|
The recent closing of the North American multicenter dynamic cardiomyoplasty trial with equivocal results makes further investigation of skeletal muscle cardiac assist and latissimus dorsi of interest [6]. This randomized trail was stopped prematurely because of difficulties with patient recruitment but preliminary data indicated statistically significant improvement in patient symptoms and NYHA class. However, the anticipated improvements in objective measures of cardiac function, such as ejection fraction, did not materialize. Thus further basic investigations are necessary to explain this result despite many years of careful experimental work. Our study characterizes canine latissimus dorsi function over time and changes with chronic electrical stimulation (conditioning). As would be expected from previous experimental studies and clinical experience with diaphragm pacing (the only currently successful long-term application of skeletal muscle pacing), conditioned latissimus dorsi develops different response characteristics to electrical stimulation than unconditioned latissimus dorsi [8–10]. Chronically stimulated, conditioned skeletal muscle can be made fatigue resistant but requires more time to contract than unconditioned skeletal muscle.
Stimulation frequency is known to be an important factor in strength and quickness of latissimus dorsi and SMV contraction [11]. Available muscle stimulators permit electrical stimulation at frequency from 2 to 65 Hz. The present study shows that lower frequencies (10 to 20 Hz) result in smaller amplitude contractions but peak muscle tension is achieved rapidly. Midrange stimuli (ie, 20 to 35 Hz) result in strong amplitude contractions, which take longer to develop maximal tension. High-frequency stimulation (65 Hz) results in high pressure generation and shorter time to peak pressure. Thus, frequency of stimulation is an important factor in determining both the amount of pressure generated and the latency of contraction after stimulation (time to peak pressure). Although the relationship between stimulation frequency and strength of contraction is well known, this study offers practical information on clinical stimulation and timing of skeletal muscle cardiac assist devices. As skeletal muscle cardiac assist devices become conditioned, optimal stimulation determination is important, changes with time, and also changes with preload. Therefore, retiming of clinical skeletal cardiac assist devices (using echocardiography) and choice of appropriate stimulation based on preload is mandatory; such a requirement was absent from the recently closed dynamic cardiomyoplasty trial.
The relationship between latency of contraction and stimulation frequency is clinically important [7]. Unlike mechanical cardiac assist devices, a period of time is necessary for skeletal muscle cardiac assist devices to develop peak tension. Minimizing this latency is important in obtaining good clinical results with cardiomyoplasty or experimental cardiac augmentation with skeletal muscle ventricle. For example, at a heart rate of 60 beats per minute there is 1 second between cardiac contractions; to obtain optimal augmentation by either SMV assist or dynamic cardiomyoplasty, the latissimus dorsi must be stimulated at an appropriate time in the cardiac cycle. As latency has been shown to be significant in this study, on the order of 0.25 seconds for canine SMVs to achieve 90% of peak pressure, electrical stimulation of the latissimus/SMV must begin in mid systole for peak SMV contraction in early diastole. In dynamic cardiomyoplasty, stimulation must begin in late cardiac diastole for peak cardiomyoplasty wrap contraction during cardiac systole. If the native heart rate doubles to 120 beats per minute, stimulation must be reassessed. In previous cardiomyoplasty studies, the importance of early stimulation has been underappreciated and some groups have even delayed initiation of stimulation until after appearance of the QRS complex on electrocardiogram, possibly compromising their results [12]. The importance of early stimulation was also underappreciated in the recent multicenter randomized trial of dynamic cardiomyoplasty, in which initiation of skeletal muscle contraction was recommended with mitral valve closure and appearance of the QRS complex on electrocardiogram.
Frequency of stimulation is also important in producing muscular fusion, an important component of effective muscle contraction. Lower frequencies of 2 Hz do not result in muscular fusion whereas higher frequencies do produce fusion.
This study confirms the well-known importance of increasing the preload of the SMV and stretching the latissimus dorsi appropriately to increase its strength of contraction [13]. Somewhat unexpectedly, increasing preload did not result in significant increases in pressure generation. Thus, a preload of 80 or 120 mm Hg does not result in significantly more SMV pressure generation than a preload of 2 or 10 mm Hg. This finding is consistent with the findings by Stephenson’s group at preloads of up to 60 mm Hg [5, 13]. However, Gealow and colleagues [14] examined the effect of preload during more than 3 months of stimulation and found a significant increase in pressure generation with a preload of 80 mm Hg compared with 20 mm Hg; our study does not address chronic stimulation at different preloads nor does it examine whether increases in preload cause proportional increases in resting muscle length (both are important issues). Thus, preload may still be an important factor in obtaining optimal performance from skeletal muscle cardiac assist devices.
Our study confirms that chronic repetitive electrical stimulation, conditioning, produces important and highly significant decrements in muscle strength. This decrement in function has been elegantly described at the cellular level. Complete conversion to type I fatigue-resistant fibers occurs with appropriate electrical stimulation, as do attendant changes in ATPase (as well as other mitochondrial enzymes) [15]. Our study shows the decrement in SMV function with such electrical conditioning is in the range of 50%. Unconditioned SMVs stimulated at 40 Hz can be expected to generate 110 mm Hg pressure whereas conditioned SMVs produce little more than half that (55 mm Hg). This is probably one explanation for the lack of improvement in ejection fraction documented in the recently closed trail of dynamic cardiomyoplasty. Our stimulation protocol was chosen to mimic that used in clinical cardiomyoplasty and is similar to that used in chronic electrical diaphragm stimulation. It might be criticized for being overly aggressive. However, it only serves as a basis for further studies of optimal training of protocols, which should be designed with an eye toward maximizing strength and minimizing fatigue. Such investigations should use a range of stimulation frequencies and stimulation patterns. Studies using the mandrel to simulate the pressure wave generated by cardiac contraction and aortic filling would also produce valuable information concerning muscle function in vivo.
Canine latissimus dorsi stimulation and conditioning protocols play an important role in skeletal muscle cardiac assist devices and dynamic cardiomyoplasty. Stimulation frequency has important effects on strength of muscle contraction and latency of contraction. Midrange stimulation frequencies of 40Hz produce strong contractions and minimize latency. Preload may not be as important as optimal stimulation frequency and timing. The production of a suitable conditioned muscle is of extreme importance in obtaining excellent long-term results with skeletal muscle cardiac assist devices. As shown in this study, decrements in muscle strength of approximately 50% can be expected with currently used patterns of chronic electrical stimulation.
Timing of contraction is extremely important for skeletal muscle cardiac assist systems. Initiation of muscle stimulation before mitral valve closure and appearance of the QRS electrocardiographically is implicated as an important factor in improving the performance of such systems. Future studies need to identify optimal stimulation protocols producing maximal pressure generation, minimal time to peak pressure, and minimal decrements in pressure generation over time.
|
Footnotes
|
|---|
Dr Francischelli discloses that he has a financial relationship with Medtronic, Inc.
|
References
|
|---|
-
Jarvis J.C. Power production and working capability of rabbit tibialis ante muscles after chronic electrical stimulation at 10 Hz. J Physiol 1993;470:157-169.[Abstract/Free Full Text]
-
Acker M.A., Hammond R.L., Mannion J.D., et al. Skeletal muscle as the potential power source for a cardiovascular pump: assessment in vivo. Science 1987;263:324-327.
-
Chiu R.C.-J., Walsh G.L., Dewar M.L., et al. Implantable extra-aortic balloon assist powered by transformed fatigue-resistant skeletal muscle. J Thorac Cardiovasc Surg 1987;89:1032-1040.
-
Salmons S., Jarvis J.D. The working capacity of skeletal muscle transformed for use in a cardiac assist role. In: Chiu R.C.-J., Bourgeois I., eds. Transformed muscle for cardiac asist and repair. Mount Kisco, NY: Futura Publishing, 1990:105-115.
-
Acker M.A., Hammond R.L., Mannion J.D., et al. An autologous biologic pump motor. J Thorac Cardiovasc Surg 1986;92:733-746.[Abstract]
-
Young J., Kirklin J. Cardiomyoplasty-Skeletal Muscle Assist Randomized Trial (C-SMART): 6 month results. Circulation 1999;100:514.
-
Letsou G.V., Zarif A., Smith A., et al. Latency of skeletal muscle contraction after pulse train stimulation: an important factor in correct timing of skeletal muscle cardiac assist devices. J Thorac Res 1994;57:672-676.
-
Glenn W.W.L., Hogan J.F., Loke S.S.O., Ciesielski T.E., Phelps M.L., Rowedder R. Ventilatory support by spacing of the conditioned diaphragm in quadriplegia. N Engl J Med 1984;310:1150-1155.[Medline]
-
Kochamba G., Chiu R.-J. The physiologic characteristics of transformed muscle for cardiac assist. ASIAO J 1987;10:404-407.
-
Van der Veen F.H., Dassen W.R.M., Havenith M.G., et al. In vivo contraction analyses as a measure of canine latissimus device muscle adaption following chronic electrical stimulation. In: Chiu R.-J., Bourgeois I., eds. Transformed muscle for cardiac assist and repair. Mount Kisco, NY: Futura Publishing, 1990:105-115.
-
Hohenhaus E, Anderson WA, Pochettino A, et al. Skeletal muscle for cardiac assist. In: Cardiac surgery: state of the art reviews. Philadelphia: Hanley & Belfus, 1990:677–701.
-
Schreuder J.J., van der Veen F.H., van der Velde E.T., et al. Beat-to-beat analysis of left ventricular pressure-volume relation and stroke volume by conductance catheter and aortic modelflow in cardiomyoplasty patients. Circulation 1995;91:2010-2017.[Abstract/Free Full Text]
-
Bridges C.R., Brown W.E., Hammond R.L., et al. Skeletal muscle ventricles: improved performance at physiologic preloads. Surgery 1989;106:275-282.[Medline]
-
Gealow K.K., Solien E.E., Bianlo R.W., Chiu R.C.-J., Shumway S.J. Conformational adaption of muscle: implications in cardiomyoplasty and skeletal muscle ventricles. Ann Thorac Surg 1993;56:520-526.[Abstract/Free Full Text]
-
Salmons S., Henriksson J. The adaptive response of skeletal muscle to increased use. Muscle Nerve 1981;4:94-105.[Medline]
Related Article
-
Invited commentary
- George J. Magovern, Jr and Dennis R. Trumble
Ann. Thorac. Surg. 2001 72: 1343.
[Extract]
[Full Text]
[PDF]
Top
Abstract
Introduction
