| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2004;127:812-822
© 2004 The American Association for Thoracic Surgery
Evolving technology |
a Department of Cardiovascular Surgery, University of Freiburg, Freiburg, Germany
b Institute of Medical Biometry, University of Freiburg, Freiburg, Germany
c Department of Cardiovascular Surgery, University of Lübeck, Lübeck, Germany
d Department of Cardiovascular Surgery, The University of Erlangen-Nürnberg, Erlangen-Nürnberg, Germany
Read at the Eighty-third Annual Meeting of The American Association for Thoracic Surgery, Boston, Mass, May 4-7, 2003.
Received for publication May 2, 2003; revisions received August 26, 2003; revisions received September 9, 2003; accepted for publication November 11, 2003.
* Address for reprints: Michael P. Siegenthaler, MD, Department of Cardiovascular Surgery, University of Freiburg, Hugstetterstrasse 55, 79106 Freiburg , Germany
siegenth{at}ch11.ukl.uni-freiburg.de
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionDiscussionReferences |
|---|
METHODS: Impella patients were compared with 198 patients treated with an intraoperative intra-aortic balloon pump between January 2000 and December 2002. Three risk scores were used: the Hausmann score, the Texas Heart Institute score, and the Cleveland intensive care unit score. Between September 2001 and March 2003, 24 patients were treated with the Impella Recover for low-output syndrome. Before device insertion, 21 could not be separated from cardiopulmonary bypass, and 3 had postoperative hemodynamic instability despite high-dose catecholamines. Sixteen were treated with the Impella and intra-aortic balloon pump and 8 with the Impella alone (no intra-aortic balloon pump because of peripheral vascular disease or because deemed unnecessary).
RESULTS: No technical problems with device insertion occurred. Pump flow was 3.3 ± 0.7 L/min at 28,000 ± 4500 RPM. Support time was 61 ± 56 hours (range, 7-228 hours). Four devices required repositioning. One device failed (leaking purge line) and was removed. Hemolysis was minimal (lactate dehydrogenase levels of 540 ± 260 U/dL for Impella survivors). Mortality for Impella patients was 54% (13/24), similar to that for high-risk intra-aortic balloon pump patients (Hausmann score 
2 [57%], intensive care unit score 2 [51%], Texas Heart Institute score 0.75 [55%], and cardiac index 
2.3 [45%]). Cardiac output data were available in 19 Impella patients. Impella patients able to increase their cardiac output to 1 L/min or more above the pump flow of the Impella Recover had a 10% (1/10) mortality, versus 88% (8/9) in patients with a residual cardiac function of 1 L/min or less (P = .001). Comparison of high-risk intra-aortic balloon pump patients with Impella patients with residual cardiac function of 1 L/min or more showed a significant reduction in mortality, regardless of the high-risk definition used. Residual cardiac function was the strongest predictor of survival in Impella patients.
CONCLUSIONS: The Impella Recover device provides 3 to 4 L/min flow. It improves survival in patients with low-output syndrome if the heart is able to pump 1 L/min or more above device flow.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionDiscussionReferences |
|---|
Impella patients
All 24 Impella devices were inserted between September 2001 and March 2003. The device was used according to the attending surgeon's judgment, with no standard protocol for use of the device. Most patients were unable to be separated from cardiopulmonary bypass. All charts were available for review. Electronic flow data were available in 13 Impella patients. In 8 patients, the ICU flow sheet had hourly device flow rates recorded. Both cardiac output data and device flow rates were present in 19 patients. All postoperative events, including surgical and device-related complications, were recorded. Estimated residual cardiac function in patients with the Impella device was defined as follows:
 |
Impella Recover left ventricular assist device
The Impella Recover device is a microaxial left ventricular assist device with a diameter of 6.4 mm at the body of the pump and a 7.3-mm diameter at the level of the outflow opening (Figure 1, A). A small electric motor is built into the device, and the thin (2.8-mm) cable leading to the device contains the electrical power supply, which is connected to an external control unit as well as a purge line connected to a purge perfusor, through which heparin (concentration of 1000-2500 IU in 50 mL of glucose 40% solution) is flushed continuously with a pressure of 300 to 700 mm Hg. The heparin/glucose solution is delivered at the level of the pump, and the patient does not need systemic anticoagulation in the early postoperative phase. Once the bleeding has subsided, it is recommended to heparinize the patient to an activated clotting time of 160 seconds or an activated partial thromboplastin time of 50 to 60 seconds. The glucose 40% solution is added to increase the viscosity of the purge-line solution to achieve the required purge-line pressure without giving large amounts of volume to the patient. The usual amount of volume given is approximately 5 mL/h, equivalent to a heparin dose of 250 U/h. The device can be set at 10 different speed levels, up to 33,000 RPM. At this speed, depending on the afterload of the device, a pump performance of approximately 3.5 to 5.0 L/min is possible, according to the manufacturer (Impella Cardiosystems AG). A memory chip built into the controller console allows pump flow and pump speed data collection. The different speed levels allow for gradual weaning to the lowest setting (10,000 RPM), at which the net device flow is 0, because the device overcomes only the mild aortic insufficiency it induces.
|
Operative technique
A 10-mm vascular graft was sewn to the ascending aorta (Figure 1, B), and the device was inserted retrograde across the aortic valve from the ascending aorta into the left ventricle (Figure 1, C). Positioning of the device occurred via a pressure-differential curve on the device console, as well as with transesophageal echocardiographic control (Figure 1, C). The vascular graft with the device was brought out of the skin either though the apex of the sternotomy incision or a small stab wound near the upper aspect of the sternotomy. The sternum was left open with either a primary skin closure or closure of the incision with an Epigard patch (Biovision GmbH, Ilmenau, Germany). Once the patient was weaned from ventricular support, the device was pulled out of the Dacron graft, which was then either ligated flush with the ascending aorta or was oversewn, followed by standard closure of the sternotomy.
Risk scores
Table 1 shows the risk scores described by Baldwin and colleagues3 and Hausmann and colleagues4 for all IABP patients and the risk score of Higgins and associates5 for patients who have had cardiac operations. The 3 risk scores were applied to all IABP and Impella patients. The risk scores were validated on our patient cohort and were subsequently used to identify high-risk patients.
|
Icu admission score
This complex risk score includes multiple preoperative, intraoperative, and postoperative parameters that are predictive of morbidity and mortality (Table 1). Patients with more than 1 missing data point were excluded from the analysis unless they had already accumulated a score of 21 or more, which put them in the highest risk category regardless of missing data. Data of the ICU admission score were available for analysis in 158 IABP patients and in all 24 Impella patients.
Hausmann score
This score is mainly based on patients' hemodynamic status (Table 1). Patients with more than 1 possible score point missing were excluded from the analysis. Because no left atrial pressure was measured at our institutions, we substituted the pulmonary capillary wedge pressure for the left atrial pressure used in the original publication. Patients with a score of 3 or higher were included in the analysis, because they were already in the highest-risk group for mortality, regardless of missing data. Data were available for analysis in 137 IABP patients and 20 Impella patients.
Statistical analysis
Data are presented as mean ± SD for quantitative variables and as absolute and relative frequencies for qualitative variables. Group comparisons were performed with Student t tests and Wilcoxon rank sum tests for quantitative data, as appropriate, and with the Fisher exact test for qualitative data. Logistic regression analysis was used to calculate odds ratios for clinical covariates with respect to survival. All significance tests were 2 sided. Data analysis was performed with SAS software (SAS Institute, Inc, Cary, NC).
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionDiscussionReferences |
|---|
|
Heparin was fully reversed after cardiopulmonary bypass in 19 patients, and 5 had a partial reversal of their heparin. Most patients were given low-dose heparin through the Impella purge perfusor in the early postoperative period until the chest tube drainage had slowed down, and they were subsequently heparinized. Five surgical revisions for bleeding were necessary in 4 Impella patients (5/24; 21%), and no surgical source of the bleeding could be found in 3. Anticoagulation and bleeding data of the Impella patients are shown in Figure 2. The incidence of surgical revision for bleeding in Impella patients was similar to that in high-risk IABP patients, with a reoperation rate for bleeding of 8% to 14%, depending on the high-risk definition used (such as a THI score 0.75, a Hausmann score 2, an ICU admission score 21, an epinephrine dose 0.14 µg · kg-1 · min-1, or a cardiac index at 6 hours of 2.3 L · min-1 · m-2).
|
Device flow data
The average pump flow was 3.3 ± 0.7 L/min in the first 12 hours. Figure 3 shows pump flow and device rotations per minute for all patients. To evaluate maximal pump performance, all patients with the device set at a speed of 27,000 RPM or more were analyzed, resulting in a maximal pump flow of 3.5 ± 0.5 L/min. Average support time was 61 ± 56 hours (range, 7-228 hours): 60 ± 37 hours (range, 22-123 hours) in survivors and 62 ± 69 hours (range, 7-228 hours) in nonsurvivors. Sixteen patients were weaned off the device, and successful weaning with a survival time of longer than 72 hours was observed in 15 patients (63% of all Impella patients). Weaning practice was variable. Average weaning time was 19 ± 13 hours (range, 2-42 hours), with a tendency toward longer weaning times as experience with the device grew. Weaning usually consisted of gradually reducing the device flow as tolerated. The lowest pump setting with a net flow of 0 was tested only for a few minutes at a time, and no patient remained on this minimal setting for a prolonged period of time.
|
|
|
|
The 4 patients who did not survive their hospital stay but were able to be successfully weaned from the device had a residual cardiac output between 0 and 1 L/min. No patient with a negative number of the residual cardiac function in the early postoperative phase could be weaned off the device.
In 6 patients, cardiogenic shock had been present for 4 or more hours before Impella insertion. All of them died. Causes of death in the Impella patients included multisystem organ failure in 10, ventricular arrhythmia in 2, and an intracranial hemorrhage in 1 patient that occurred 18 days after removal of the device.
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionDiscussionReferences |
|---|
In the second subgroup of Impella patients, we were able to identify patients who will not benefit from the Impella Recover device. Patients with a residual cardiac function of 1 L/min or less had an 88% chance of death. This observation is likely due to 2 factors. First, the Impella Recover device provides insufficient support in the presence of virtually absent myocardial function; second, patients with more extensive myocardial damage have less potential for myocardial recovery. The likely failure of therapy and the high probability of death in patients with a residual cardiac function of 1 L/min or less were predictable in 6 of 7 available cardiac function measurements at 2 hours after implantation. This seems important, because before leaving the operating room the decision to implant a conventional assist device can be made. Even without waiting 2 hours, these patients can probably easily be identified if they do not have any pulsatile output when CPB flow decreases to less than 2 to 2.5 L/min despite adequate filling. Such patients, to have a realistic chance of survival, should receive a conventional assist device that is able to replace the entire left heart function. A significant early increase in residual cardiac function between 2 and 12 hours leading to survival was observed in only 1 patient (Figure 5). One patient had a significant decline in cardiac function between 6 and 12 hours after surgery, when he approached the cutoff point of a residual cardiac function of 1 L/min; this ultimately led to his death (Figure 5). If such a course were observed clinically in the future, a switch to a different device could be considered.
The surgical complication rate in Impella patients was similar to that in high-risk patients treated with IABP alone, which was lower than other experiences with left ventricular assist support.7,8 The observed device-related complications were minimal, with 1 device failure and 4 device repositions, all performed in the ICU. No observations of familiar problems associated with axial flow technology were made, including hemolysis and thromboembolic events, and no wound- or device-related infections were observed.
It is interesting to note that the only risk score with predictive power in Impella patients was the Hausmann score (P = .051), despite a relatively low risk discrimination, as illustrated in Figure 4, C. This is most likely due to its focus on patient hemodynamics, which seem to outweigh most preoperative risk factors.
Our data do not support the indiscriminate use of this device, because no evidence could be shown that the use of the Impella Recover device in all patients with severe LOS can reduce mortality compared with high-risk IABP patients. It has to be re-emphasized that the Impella patients were substantially more ill, because they could not be separated from cardiopulmonary bypass or were hemodynamically unstable despite high doses of catecholamine support. Hence, the comparable overall mortality in all Impella patients and high-risk IABP patients gives only indirect evidence for a potential patient benefit with indiscriminate use of this device. Patients who become device dependent for postcardiotomy LOS have a mortality rate of 75% to 80%.8-10 Even in a recently published revised screening scale for left ventricular assist device recipients, the presence of postcardiotomy LOS was highly predictive of mortality.11
Even when compared with all IABP patients, Impella patients with a residual cardiac function of 1 L/min or more seemed to have a survival benefit. Plotting their risk scores on the risk score survival curves of all IABP patients (including high- and low-risk patients) showed a higher survival than predicted, at the border of the 95% confidence interval (Figure 4).
The stratification of the Impella patients into those with a residual cardiac function of 1 L/min or more and those with virtually absent myocardial function (residual heart function 1 L/min) allowed us to identify the patient cohort with a dramatic survival benefit. This observation seems to hold true despite the retrospective nature of this study, as well as the lack of a standard protocol for use of the Impella Recover device in this series. The analysis of this data mandates a standardized protocol for implantation of the Impella Recover device in patients with postcardiotomy LOS. The indications for the use of this device should be expanded for IABP patients with a high predicted mortality, and this device should no longer be used in patients with virtually absent myocardial function, because it is a true assist device that is unable to replace the entire left ventricular function.
In summary, the Impella Recover device is a low-cost microaxial assist device that pumps 3 to 4 L/min and is associated with a low complication rate. This study showed a survival benefit in mortally ill patients with postcardiotomy LOS if the patient's heart is able to contribute at least 1 L/min to the cardiac output in the immediate postoperative period.
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionDiscussionReferences |
|---|
There has been significant progress in the development of axial flow pumps for temporary assist, and you can see in the background [slide] the initial Hemopump for peripheral access, then in the middle the one for central access, and finally the percutaneous one. These pumps had a drive shaft, and I think one of the major improvements with the current Impella device is that the drive shaft has disappeared; the second major difference is that there are sensors now available that give us some information about pump positioning and flow.
I have 3 questions. You have already alluded to the bleeding issue. I would like to know if there was a problem with platelet consumption with this pump and if any of the deaths were related to bleeding.
The second question is about thrombosis. There have been emboli, I think, central emboli, but have the explanted pumps been analyzed, and, if yes, were there any deposits?
Finally, what is your weaning protocol, or, in other words, what would be the mean pump speed (RPM) required for 0 flow at a mean arterial pressure of about 60 mm Hg?
Dr Siegenthaler. Thank you very much for those kind comments and questions. Four Impella patients needed to go back to the operating room because of bleeding. All required several transfusions. In 3 of them, no surgical source of bleeding was evident. The question of whether bleeding was responsible for the patients' deaths is difficult to answer. Certainly, no patient died directly related to exsanguinating hemorrhage. Also, the incidence of bleeding was no different in Impella patients than in high-risk IABP patients. The cause of death in most of the 13 nonsurvivors with the Impella pump was multisystem organ failure. Transfusions certainly can be a contributing factor to such a syndrome and therefore also could have been a factor leading to adverse outcomes. We observed no platelet consumption.
As to thromboembolic events, we know from the carotid literature that surgeons do not have the same horizon to detect such events as neurologists. I can tell you with certainty only that there were no obvious strokes. One patient had a major intracranial hemorrhage from which he died 18 days after successful removal of the device, and we cannot rule out completely that he had bleeding into a small and previously asymptomatic embolic area of his brain.
We looked at all the pumps when we explanted them, and we could not see any residual fibrin or thrombi in the pumps. But this information has to be viewed with caution, as there was no systematic analysis of these pumps.
The pump is suitable for a slow weaning process, as it can be set at 10 different speed levels, up to 33,000 RPM. At the lowest pump setting, the pump just overcomes the mild aortic insufficiency that it causes. This allows assessment of the patient's native heart function at the bedside. Gradual weaning has been more successful than our initial procedure, in which we evaluated the ventricular performance with echocardiography at full or only slightly reduced device speed. If satisfactory contractility was found, we went to the operating room and tried to turn off the pump. This approach was sometimes not well tolerated. We now try to observe the patient in the ICU at a low pump setting after a slow weaning process before proceeding to the operating room to explant the device.
Dr O. Howard Frazier (Houston, Tex). This pump, of course, is an extension and refinement of the Hemopump we first used in 1988. The Hemopump was powered by a cable, and its most effective use was in the postcardiotomy patient. This was aborted because the investors felt that there was not a big enough market, and I have always felt that this was the 1 technology that would be so important to directly unload the ventricle without requiring heparinization or anticoagulation, which we did not use in the immediate postoperative period with the Hemopump. Because the Hemopump had a cable, we sewed the graft on to keep the cable from fracturing.
Now, this pump is powered with an electric wire, and it is much more effective from that standpoint. Do you think you could get rid of the graft altogether, since those grafts are not inexpensive and simply use a purse-string suture? The other question is, was heparinization continued in the immediate postoperative period?
Dr Siegenthaler. Thank you very much for those questions. Vascular grafts are expensive indeed; we pay about 500 Euro for such a 10-mm Vascutek graft at our institution.
Nobody at our institutions has tried to insert the device without a vascular graft, just by using a purse-string suture with a tourniquet. I could imagine that in a patient with a soft aorta this might be actually possible, similarly to how we do it for thoracic intra-aortic balloon pumps. But I think it would be risky to avoid the use of a graft in the presence of a diseased aorta.
We use no anticoagulation therapy in the early postoperative period until bleeding has subsided. Then we start heparin at 500 U/h and gradually increase it to a partial thromboplastin time of 50 to 60 seconds. The risk of pump thrombosis seems small even if patients are not systemically anticoagulated, because the pump is continuously purged with a low-dose heparin–containing solution.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. Samoukovic, T. Al-Atassi, C. Rosu, N. Giannetti, and R. Cecere Successful Treatment of Heart Failure due to Acute Transplant Rejection With the Impella LP 5.0. Ann. Thorac. Surg., July 1, 2009; 88(1): 271 - 273. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Meyns, S. Klotz, A. Simon, W. Droogne, F. Rega, B. Griffith, R. Dowling, M. J. Zucker, and D. Burkhoff Proof of Concept: Hemodynamic Response to Long-Term Partial Ventricular Support With the Synergy Pocket Micro-Pump J. Am. Coll. Cardiol., June 30, 2009; 54(1): 79 - 86. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Samoukovic, C. Rosu, N. Giannetti, and R. Cecere The Impella(R) LP 5.0 as a bridge to long-term circulatory support Interactive CardioVascular and Thoracic Surgery, June 1, 2009; 8(6): 682 - 683. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Patane, P. Centofanti, E. Zingarelli, F. Sansone, and M. La Torre Potential role of the Impella Recover left ventricular assist device in the management of postinfarct ventricular septal defect. J. Thorac. Cardiovasc. Surg., May 1, 2009; 137(5): 1288 - 1289. [Full Text] [PDF]
|
|
|
|
|
|
O. Honjo, S. L. Merklinger, J. B. Poe, A.-M. Guerguerian, A. A. Alghamdi, S. Takatani, and G. S. Van Arsdell Mechanical cavopulmonary assist maintains pulmonary and cerebral blood flow in a piglet model of a bidirectional cavopulmonary shunt with high pulmonary vascular resistance. J. Thorac. Cardiovasc. Surg., February 1, 2009; 137(2): 355 - 361. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. R. Dixon, J. P.S. Henriques, L. Mauri, K. Sjauw, A. Civitello, B. Kar, P. Loyalka, F. S. Resnic, P. Teirstein, R. Makkar, et al. A Prospective Feasibility Trial Investigating the Use of the Impella 2.5 System in Patients Undergoing High-Risk Percutaneous Coronary Intervention (The PROTECT I Trial): Initial U.S. Experience J. Am. Coll. Cardiol. Intv., February 1, 2009; 2(2): 91 - 96. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L.-G. Dahlin and B. Peterzen Impella Used for Hemostasis by Left Ventricular Unloading, in a Case With Left Ventricular Posterior Wall Rupture Ann. Thorac. Surg., April 1, 2008; 85(4): 1445 - 1447. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Aggarwal, F. Cheema, M. C. Oz, and Y. Naka Long-Term Mechanical Circulatory Support Card. Surg. Adult, January 1, 2008; 3(2008): 1609 - 1628. [Full Text]
|
|
|
|
|
|
A. A. Pitsis, A. N. Visouli, D. Burkhoff, P. Dardas, N. Mezilis, G. Bougioukas, and G. Filippatos Feasibility Study of a Temporary Percutaneous Left Ventricular Assist Device in Cardiac Surgery Ann. Thorac. Surg., December 1, 2007; 84(6): 1993 - 1999. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Markl, C. Benk, D. Klausmann, A. F. Stalder, A. Frydrychowicz, J. Hennig, and F. Beyersdorf Three-dimensional magnetic resonance flow analysis in a ventricular assist device. J. Thorac. Cardiovasc. Surg., December 1, 2007; 134(6): 1471 - 1476. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Garatti, T. Colombo, and E. Vitali Placement of the Impella Recover LD microaxial blood pump through a bioprosthesis is technically feasible J. Thorac. Cardiovasc. Surg., October 1, 2006; 132(4): 989 - 990. [Full Text] [PDF]
|
|
|
|
|
|
J. T. Strauch, U. F.W. Franke, M. Breuer, J. Wippermann, T. Wittwer, N. Madershahian, M. Kaluza, and T. Wahlers Technical feasibility of Impella Recover 100 microaxial left ventricular assist device placement after biologic aortic valve replacement (21 mm) for postcardiotomy failure J. Thorac. Cardiovasc. Surg., December 1, 2005; 130(6): 1715 - 1716. [Full Text] [PDF]
|
|
|
|
 |
|
 E. Catena, F. Milazzo, M. Merli, R. Paino, A. Garatti, T. Colombo, and E. Vitali Echocardiographic evaluation of patients receiving a new left ventricular assist device: the Impella(R) recover 100 Eur J Echocardiogr, December 1, 2004; 5(6): 430 - 437. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
) represent patients who underwent a revision for bleeding; the other patients did not undergo surgical revision. ACT, Activated clotting time; aPTT, activated partial thromboplastin time; PRBC, packed red blood cells.
