| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2002;73:556-562
© 2002 The Society of Thoracic Surgeons
a The Michael E. DeBakey Department of Surgery, Baylor College of Medicine, Houston, Texas, USA
* Address reprint requests to Dr Anstadt, Cardiothoracic Surgery, BA 4300, 112 15th St, Medical College of Georgia, Augusta, GA 30912-4040, USA
e-mail: manstadt{at}mail.mcg.edu
Presented at the Forty-seventh Annual Meeting of the Southern Thoracic Surgical Association, Marco Island, FL, Nov 9–11, 2000.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentDiscussionReferences |
|---|
Methods. Bovines (n = 10) underwent thoracotomy and were instrumented for continuous hemodynamic monitoring. Cardiac failure was induced by ß1-blockade to achieve a cardiac index of < 1.5 l/min/m2 for 1 hour. Heart rate was maintained at 100 bpm by atrioventricular sequential pacing. Synchronous DMVA support was then applied for 3 hours.
Results. Eight animals achieved significant reductions in cardiac index and mean arterial pressures (35%* and 43%* control, respectively; *p < 0.05). DMVA restored cardiac index to baseline and significantly increased arterial pressures (p < 0.05; DMVA versus cardiac failure). Pulmonary flow and mean pulmonary artery pressures were similar to baseline during DMVA (p = NS). Pathologic exam did not demonstrate evidence of significant device trauma.
Conclusions. DMVA support can effectively restore pump performance of the acutely failing heart. Synchronization may be inherent to the stimulus of cardiac compression. These data further substantiate DMVA’s potential as an adjunct to the field of circulatory support.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionReferences |
|---|
Direct mechanical ventricular actuation (DMVA) augments myocardial pump performance but avoids blood contact (Fig 1). DMVA employs a pneumatically regulated, vacuum attached, flexible membrane for transferring rhythmic systolic and diastolic forces. Applied to the fibrillating or non-beating heart, DMVA maintains physiologically normal hemodynamics without significant myocardial trauma [1–10]. Recent clinical experience has demonstrated that DMVA is effective for assisting the failing heart [11]. These results have renewed interest in non-blood contacting mechanical support in the treatment of severe heart failure.
|
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionReferences |
|---|
Induction and maintenance of cardiac failure
Cardiac failure (CHF) was simulated pharmacologically using ß1-blockade. A loading dose of esmolol hydrochloride (7.5 mg/kg IV) was followed by a continuous infusion (400 to 600 µg/kg/min). Sequential atrioventricular (AV) pacing was begun at 100 bpm and the esmolol infusion titrated to achieve a cardiac index (CI) of less than 1.5 l/min/m2. Cardiac failure was maintained for 1 hour before initiating DMVA support.
DMVA support protocol
A DMVA assist cup fabricated from silicone rubber was used for biventricular support. The same device was used in all experiments. Attachment was provided by a vacuum delivered via the housing’s apical port. Mechanical actuation was maintained at a cycle rate of 100/min. The solenoid valve initiating mechanical systole was electronically synchronized to a ventricular pacing stimulus. Pneumatic drive settings were adjusted within the following ranges as previously described [3, 11]: systolic duration 55% to 60%; systolic forces +135 to +155 mmHg; diastolic forces -115 to -135 mmHg; and vacuum attachment -70 to -110 mmHg. These settings along with in-line damping valves were adjusted to achieve complete ventricular compression and expansion as previously described [10]. The chest was closed and the drive lines exited through separate intercostal incisions. During DMVA, esmolol was decreased to avoid progressive systemic hypotension. The dose was adjusted to keep mean arterial pressures greater than or equal to 60 mm Hg requiring its discontinuation by 90 minutes of support. DMVA was continued for a total of 3 hours and the chest reopened for device removal. Hearts were immediately extirpated and perfusion fixed with Z-fix (Anatech, Ltd, Battle Creek, MI) buffered formaldehyde.
Tissue analysis
Formalin fixed hearts underwent complete gross and histologic examination by a Veterinarian Pathologist specializing in cardiovascular disease who was blinded to the experimental protocol. Hearts were sectioned at 1-cm intervals from apex to base. Gross findings were noted and representative sections from the left ventricle (LV), RV, and septum examined by light microscopy. Analysis for apoptotic chromatin changes was performed as previously described [15]. Specimens were labeled with an oligonucleotide probe that binds double-stranded DNA breaks, counterstained with an immunofluorescent dye, and examined by fluorescence microscopy.
Data analysis
Hemodynamic and arterial gas measurements were averaged over the three experimental states: baseline, CHF, and DMVA support. Results were expressed as the mean ± the standard deviation. Variables were then analyzed for distribution. Comparisons were made between the three experimental states using Student’s t and Wilcoxon Rank tests. Differences were considered statistically significant at alpha levels less than 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionReferences |
|---|
ß-Blockade reduced mean CI by more than 50% (p < 0.05; Fig 2). The pharmacologically induced CHF state resulted in similar reductions in both left (Fig 1) and right (Fig 3) ventricular output. Decreases in cardiac output were paralleled by significant reductions in mean systemic arterial pressures (Fig 4). However, pulmonary artery pressures (p = NS; Fig 5) did not change substantially during the hour of untreated CHF.
|
|
|
|
An electronic shortage created pacer malfunction during DMVA support in 3 animals as the pacer output was directly coupled to the DMVA drive chip. In these experiments, epicardial pacing wires contacted and directed energy pulses toward the output chip resulting in its overheating and breakdown. The pacer chip then oscillated instead of pulsing with the DMVA system. Two such instances resulted in ventricular fibrillation during the latter hour of DMVA support. In a third experiment, sinus rhythm was unaltered and DMVA support was continued in an asynchronous mode. Systolic compressions in the latter circumstance were noted to stimulate QRS complexes, and hemodynamics were unaltered compared with synchronized assist. In the other two instances of pacer malfunction, support during fibrillation resulted in no identifiable hemodynamic alternations compared with synchronous assist of the failing heart. Hemodynamics in the animal supported for 3 hours following electrically induced VF were generally similar to the other eight experiments. At DMVA termination, electrical cardioversion was successful in 2 of these animals. The other was refractory to multiple cardioversion attempts. The remaining 6 animals exhibited marked sinus bradycardia and hypotension following device removal. The brief period of post-support hemodynamics observed prior to cardiac extirpation were generally similar to the CHF state consistent with persistent ß-blockade.
Cardiac examination demonstrated fibrin patches over the epicardium. Subepicardial hematomas were noted in the LV apex and the RV outflow tract corresponding with apical port and outer rim contact regions of the assist cup, respectively. Histologic exam demonstrated subepicardial reactive changes and superficial hemorrhage in these same areas. Scattered areas of focal myocyte necrosis were noted throughout the myocardium occurring most frequently in the LV subendocardium. Double-stranded DNA breaks, identified by immunofluorescence staining, paralleled light microscopy findings, being most prevalent in the LV subendocardium. In contrast, the animal supported for 3 hours without antecedent CHF exhibited a marked paucity in apoptotic changes.
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionReferences |
|---|
A number of non-blood contacting circulatory support methods have been investigated for their clinical potential. Most approaches were conceived with the primary aim of improving systolic function. Dynamic cardiomyoplasty (DCM) utilizes a conditioned muscle wrap for ventricular compression. Time required for muscle conditioning limits the utility of DCM support. The muscle wrap can augment cardiac output with significant reductions in MVO2. Reduced MVO2 is most likely a consequence of secondary changes in end-diastolic volume (the "girdling" effect) which might also attenuate myocardial remodeling [18]. Recently, two mechanical devices have come under investigation for providing more effective systolic augmentation or DCC [12, 13]. Laboratory studies with these devices are limited to short-term applications but demonstrate significant improvements in cardiac output when applied to the acutely failing heart [13]. The potential traumatic effects of these devices have yet to be determined. Furthermore, lack of diastolic assist, cumbersome designs, and unresolved installation techniques may limit their efficacy.
The ideal cardiac compression device should provide systolic and diastolic augmentation, rapid application, and cause minimal cardiac trauma. DMVA is the only described device that meets these challenging goals. Application requires only minutes as demonstrated in both laboratory [3] and clinical trials [3, 6, 11]. Device sizing can be estimated by measuring ventricular dimensions using a chest radiograph or echocardiogram. A precise fit is not critical to effective support as demonstrated in the present study where the same device was used in all animals. Despite natural variations in heart size, this approach was also utilized in previous canine studies illustrating an average fit is sufficient for short-term support [4, 7]. Vacuum provides a nontraumatic means for attaching the contoured housing to the ventricular epicardium leaving the atria unopposed. A silicone rubber diaphragm dynamically conforms to the myocardial wall as compressive and dilating forces augment systolic–diastolic function. Improvements in systolic function are thereby complemented by accelerated diastolic filling [19]. The process can be continued indefinitely without significant myocardial trauma [6–9, 11]. Laboratory studies have confirmed that the fibrillating heart can be completely supported for hours to days [1–3, 5, 7–10] while clinical trials in the failing heart have resulted in long-term survival after 1 week and no significant trauma after up to 3 months of support [6, 11].
DMVA has been predominantly evaluated in the non-beating, fibrillating heart [1–3, 5, 9, 10, 19, 20]. In this state, ventricular filling is entirely passive. DMVA application restores ventricular dynamics with active diastolic function and improved ventricular filling [19]. These effects, along with other demonstrated attributes, may significantly benefit the failing heart. Alternative devices recently proposed for DCC are limited to systolic augmentation and have undetermined effects on the myocardium. Therefore, the present study evaluated DMVA’s effectiveness for assisting the failing heart.
The bovine model was selected because of its accepted validity in the field of investigative circulatory support. Cardiac failure was modeled using ß-blockade to avoid direct myocardial damage. However, profound cardiac shock resulted in global myocardial ischemia as evidenced by histologic findings. Both apoptotic changes and focal myocyte necrosis were diffuse and more marked in the subendocardium suggestive of global ischemia. These findings are in contrast with prior laboratory [3, 7, 8] and clinical [3, 6, 11] studies using DMVA in which such pathology was scarce to absent. The heart supported during VF without antecedent failure in the present study also exhibited a paucity of such pathologic findings, corroborating prior investigations. On the other hand, the presence of subepicardial contusions has been previously characterized in earlier studies and appears to be of minimal functional consequence.
Cardiac dysfunction induced prior to DMVA support was clearly severe. The mean cardiac index was less than 1 L/min/m2 for 1 hour and was accompanied by significant hypotension. Clearly, this was sufficiently representative of states requiring mechanical support in the clinical setting. Ventricular pacing overcame the chronotropic effects of ß1-blockade during the CHF state. DMVA support was subsequently synchronized with pacing stimuli to ensure appropriate device timing. Prior experimentation has documented drive-line lag times preceding myocardial compression [21]. These calculated delays were used to set the interval between pacing and systolic drive system activation. Mechanical assistance thereby occurred in synchrony with the cardiac cycle. The augmentation in pump performance was sufficient to restore baseline flow demonstrating DMVA’s ability to support the failing heart. The degree of augmentation exceeded that reported using direct cardiac compression techniques [12, 13]. These comparatively better results may be explained, in part, by the added feature of diastolic augmentation.
Both left and right ventricular function were augmented in an equally effective fashion as exhibited by comparisons in aortic and pulmonary flow. One might anticipate DMVA would over-pump the lower pressured right ventricle. However, the discrepancies between right and left ventricular flows were consistent when comparing either baseline, CHF and DMVA support. These persistent differences between RV and LV flow measurements may have been accountable to physiologic shunts, coronary flows being excluded from LV flow measurements, and underlying differences between the probes. Slight elevations in pulmonary artery pressures were most likely due to increases in RV dp/dt during systolic augmentation [19]. These small hemodynamic changes did not affect pulmonary function as indicated by stable ventilatory dynamics and blood gas exchange. These findings were consistent with investigations using DMVA for support of the fibrillating heart.
Two animals fibrillated during DMVA support in the present study due to a pacemaker malfunction caused by epicardial lead contact and a secondary short-circuit. There were no apparent changes in measured hemodynamic variables following fibrillation. Furthermore, hemodynamics in the animal intentionally supported during VF were similar to the other 8 receiving synchronous support. These observations indicate that DMVA support augments the fibrillating and failing heart to similar degrees. One animal received DMVA without an accompanying electrical pacing stimulus. In this experiment, electrocardiographic signals demonstrated a QRS following each mechanical compression. Cardiac compression served as its own pacing stimulus. There were no discernable hemodynamic differences comparing DMVA support with or without electrical pacing. These findings reflect prior laboratory [21] and clinical experiences [7, 11] using nonsynchronized DMVA support of the failing heart. Cardiac compression generally provides a means of "over-drive" pacing, and device synchronization may thereby be inherent to the method. Attention is probably best focused on appropriate rate and drive-line dynamics to achieve optimal hemodynamic results [3, 11]. However, further studies are required to substantiate these inferences.
Cardiac dysfunction was evident at the termination of the experiment. Post-support cardiac function was not assessed. Instead, hearts were immediately extirpated for rapid fixation and subsequent microscopic examination. Cardiac function following support was adversely impacted by the experimental preparation. ß-Blockade persisted as indicated by the marked sinus bradycardia. Clearly, the esmolol doses required to achieve severe cardiac failure overwhelmed this drug’s normal systemic breakdown by plasma esterases. Pharmokinetics of this agent have not yet been characterized in the bovine species. The severe hypotension associated with this model led to global myocardial ischemia and attendant cardiac dysfunction. Although DMVA might also adversely impact the heart, extensive investigations using this method for hours to days have not indicated any such detrimental effects. Laboratory results have shown DMVA can acutely resuscitate and restore cardiac function when applied to the ischemically injured heart [2, 3, 5]. Even the functional integrity of saphenous vein bypass grafts is preserved following hours of DMVA support [9]. Clinical applications have included successful bridge to transplantation (days), post-cardiotomy support (days), long-term recovery from severe myocarditis (1 week), and prolonged circulatory support (3 months) without evidence of device-related dysfunction or trauma [6, 11]. Importantly, significant traumatic effects have been noted after cardiac support with devices constructed of less compliant, polyurethane diaphragms [20]. Design characteristics, use of silicone rubber diaphragms and appropriate drive mechanics are felt to be paramount to effective, atraumatic support.
In summary, DMVA is a unique non-blood contacting device that can effectively restore performance of the failing, asystolic, or fibrillating heart. It differs from other cardiac compression devices with respect to design specifications, silicone rubber construction, vacuum attachment, and diastolic actuation properties. Compared to conventional assist devices, DMVA is relatively easy to install, versatile, and avoids blood contact. There is compelling evidence that DMVA could fill a critical void in the field of circulatory support. Its potential utility as a resuscitative device might be best viewed as an adjunct to other existing methods. The role for DMVA and related concepts as a means for long-term support of the failing heart also has promise.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionReferences |
|---|
If you had patients who were not beta-blocked and AV paced and perhaps had much more rapid heart rates, would it actually be possible to synchronize the pump? I suppose the main question is, how important is it that you synchronize it?
I have questions related to your intention of use of this device. My understanding is that some of the problems clinically have been the appropriate fixation of the device so there is not movement, and if you really were to use this, for example, as a bridge to transplant, could you actually get a patient out of bed and walk him and have it not become dislodged?
Finally, I am sure you are aware that there are several of these external devices now being developed by others with various features. What is the advantage of this device over some of the others in development?
It is a very nice paper and I appreciate having the opportunity to discuss it.
DR SCHULTE-EISTRUP: Well, this is a very elaborate question.
To start off with, we have not had any problems with the fixation of this device due to the continuous suction that we apply to the apex of the heart. We had problems when we pumped some herniation on the right ventricular outflow tract, where we have seen some epicardial hemorrhages as well, but overall the heart was not ejected when it was pumped and while cardiac output was reapplied.
The unique feature that we have in this device, compared to the other ones that are out right now and are developed, is the diastolic augmentation where we do not just apply systolic pressures and basically squeeze the heart in order to pump it; we also apply a negative force, so we augment the diastole as well. When we did a series of these cows, where we just skipped the diastolic actuation, we saw that we could achieve better cardiac output if we also had an augmented diastolic phase.
Concerning the rapid heart rates that you were asking about, that is true—we beta-blockaded the heart and we paced it rapidly to exclude the negative chronotropic effect by the beta-blockade, but we are not that far yet that we can say that the heart itself is leading and the pacemaker is sensing and we adapt to it. This is a technical problem which has not been solved so far.
Concerning the significance of synchronization, if you do not mind, I would like to pass this question on to Dr Anstadt, who is also in the audience and has been the primary PI of this, and who would probably like to answer that himself.
DR ANSTADT: Thank you, Dr Pennington, for your questions. In reference to your question about synchronization, work has been done in the past by my father using DMVA synchronized to the beating heart. These investigations demonstrated that, when synchronized to the QRS signal, DMVA can provide effective mechanical compression prior to the onset of ventricular contraction. In addition, there has always been an impression by myself and others that mechanical compression functions as a pacing stimulus and, thereby, initiates myocardial contraction. So your question regarding the need for pacing or synchronization is quite relevant.
In the clinical setting, we have seen patients with poor ventricular function, such as those discussed by Dr Schulte-Eistrup, where rapid heart rates were essentially ignored. The device was applied with what was felt to be the best physiologic rate based on cardiac size. As the heart became stronger, we found that adjusting the rate to more closely match the native rhythm resulted in better hemodynamics. This seems to, in part, support the phenomenon that the mechanical stimulus leads to myocardial contraction. I do not think we really know the answer as to whether or not synchronous support is essential. Clear, arrythmias, or rapid heart rates, seen in patients with severe end-stage failure, are best managed by applying DMVA with the actuation rate based on cardiac size, which would decrease with increasing heart size.
And then, finally, your question about other devices—I would just echo comments already made about the concept of diastolic augmentation. I think the major difference between the current devices that are now being developed by Abiomed, which was recently presented at the STS Circulatory Support Meeting, as well as other related technologies, is that they focus on systolic actuation. There are no efforts, to my knowledge, on diastolic augmentation, which I think is critical in end-stage failure and, of course, the fibrillating or asystolic heart. The best data published on a device that participates only in cardiac compression comes from Burkoff’s group at Columbia University. Their results indicate that there are negative effects on diastolic filling. I would be concerned that this may limit the effectiveness of such approaches.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
