Initial Experience With Miniature Axial Flow Ventricular Assist Devices for Postcardiotomy Heart Failure

Transcription

1 CARDIOVASCULAR Initial Experience With Miniature Axial Flow Ventricular Assist Devices for Postcardiotomy Heart Failure Michael J. Jurmann, MD, Henryk Siniawski, MD, Michael Erb, MD, Thorsten Drews, MD, and Roland Hetzer, MD, PhD Department of Cardiothoracic and Vascular Surgery, Deutsches Herzzentrum Berlin, Berlin, Germany Background. The recently introduced Impella Recover (Impella CardioSystems AG, Aachen, Germany) microaxial flow left and right ventricular assist devices (LVAD/ RVAD) were evaluated as they provide circulatory support in the setting of postcardiotomy heart failure refractory to high-dose inotropic and intra-aortic balloon pump (IABP) support. Methods. Between May 2002 and November 2002, the Recover LVAD was implanted in six patients (64 11 years) with acute left heart failure following coronary artery bypass procedures. Preoperative left ventricular (LV) ejection fraction was compromised (28% 12%, 12% to 45%). Three patients presented with unstable circulation or cardiogenic shock following acute myocardial infarction, with a predicted mortality rate of 44% 11% (EuroSCORE). Intraoperatively, severe heart failure was present with a more than 70% mortality rate predicted by the IABP score. The Recover RVAD and LVAD were combined to provide biventricular assist device (BVAD) support in one case of post-transplant graft failure. Results. The Recover LVAD delivered blood flows of up to 5 L/min. A moderate degree of hemolysis and a reduction in platelet count were noted. Four patients were weaned from the LVAD after hours, two of whom remain long-term survivors. Full recovery of graft function allowed weaning of the patient from BVAD support after six days. Conclusions. The initial experience with the Impella Recover VADs proved the new systems to be advantageous regarding the ease of implantation and device removal, low anticoagulation requirements, and advanced weaning features. In cases of severe heart failure, survival was improved by using LVADs when compared to that predicted by solely continuing IABP and drug support. (Ann Thorac Surg 2004;77:1642 7) 2004 by The Society of Thoracic Surgeons Providing circulatory support in the setting of postcardiotomy heart failure or acute heart failure remains challenging. Given the inability to separate a patient from cardiopulmonary bypass during the operation using conventional measures (high-dose inotropic support, use of the intra-aortic balloon pump [IABP]), the expected mortality rate continues to be exceptionally high (40% to 80%) [1 3]. Some progress has been made in this regard over the past years as the following factors have been clinically proven to promote survival in this situation: fast decision making towards implantation of ventricular assist devices (VADs) [2, 4, 5] to limit cardiopulmonary bypass time, the utilization of standardized clinical protocols on VAD implantation [6, 7], accurate operative hemostasis, early VAD system change in favor of implantable blood pumps [8], and eventually, early selection of patients with lack of recovery of native heart function for bridge-to-transplant procedures [9]. In addition, the availability of short-term VADs, which (by virtue of their design) are inexpensive, allow for simple and fast Accepted for publication Oct 2, Address reprint requests to Dr Jurmann, Deutsches Herzzentrum Berlin, Augustenburger Platz 1, Berlin, Germany; jurmann@dhzb.de. surgical implantation and removal procedures, and provide broader variations in blood flow to enhance the weaning strategies, is desirable to improve on the results. Only recently, miniaturized axial flow left and right ventricular assist devices (LVADs/RVADs) have been introduced into clinical practice. This communication summarizes our initial clinical experience with the utilization of the Impella Recover VADs (Impella CardioSystems AG, Aachen, Germany) in the setting of postcardiotomy acute heart failure. Material and Methods The Impella Recover LVAD (Fig 1) is a small (6.4 mm in diameter) intracardiac axial flow ventricular assist device. At 32,500 rpm, the Recover LVAD will provide between 4 and 5 L/min of continuous blood flow against physiologic afterloads [10]. The Recover LVAD bears a flexible inlet cannula (55 or 65 mm in length), which is advanced into the LV cavity in a retrograde fashion via a prosthetic graft sutured to the ascending aorta, thereby passing the aortic valve (Fig 2A). Although the correct intraventricular position and proper function of the Recover LVAD might be verified from the differential pressure trace as pro by The Society of Thoracic Surgeons /04/$30.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg JURMANN ET AL 2004;77: MINIATURE AXIAL FLOW VADS Abbreviations and Acronyms BVAD biventricular assist device CABG coronary artery bypass grafting IABP intraaortic balloon pump ITA internal thoracic artery LVAD left ventricular assist device LVEF left ventricular ejection fraction RVAD right ventricular assist device TEE transesophageal echocardiography SvO 2 mixed venous oxygen saturation VAD ventricular assist device 1643 CARDIOVASCULAR vided by the external control console alone, transesophageal echocardiography (TEE) guidance is recommended during implantation of the LVAD (Fig 2B). Driven by an electrical motor and by means of a small impeller (both of which reside within the blood stream of the ascending aorta) the blood is drained directly from the LV cavity to be expelled into the ascending aorta. The impeller housing is continuously flushed with a hyperosmolaric solution, also containing heparin (2500 ie or 5000 ie/50 ml), to maintain the pressure within a predefined range ( purger ). A 3-mm driveline has to be brought out through the skin in the supraclavicular, jugular, or parasternal regions and connects the LVAD to an external mobile control console and the purger. The Recover RVAD (Fig 3) is a small paracardiac pump bearing a short inlet cage, which is inserted into the right atrium. The outlet portion is a ring-enforced graft prosthesis, which is anastomosed to the pulmonary artery. An electrical motor drives a small impeller to deliver 5 to 6 L/min of continuous blood flow. The thin driveline has to be brought out through either the jugular or parasternal region to be connected to the external mobile control console and the purger unit. The Recover LVAD and RVAD might be used in combination to provide biventricular assist device (BVAD) circulatory support. The Fig 2. (A) Situs of an Impella Recover intracardiac LVAD after implantation in a patient following triple aortocoronary saphenous vein and left ITA bypass grafting. The blood pump was advanced into the LV in a retrograde fashion via a prosthetic graft, which had been anastomosed to the ascending (Asc.) aorta. The patient s head is oriented to the right. (B) Intracardiac position of the Recover LVAD as demonstrated by intraoperative transesophageal echocardiography. (ITA internal thoracic artery; LV left ventricle; LVAD left ventricular assist device.) Fig 1. Impella Recover left ventricular assist device. The tip of the blood pump bears the inlet cannula, above which the impeller and the electrical motor are situated. The thin driveline is exterritorized and connected to the external control console and the purger. Recover VADs provide enhanced blood flow capabilities; at first, blood flows below 2 L/min can be safely used for extended periods of time and second, a dynamic pump stop feature is available where the VAD produces no effective forward flow. During dynamic pump stop mode the rotational speed of the VAD (and the resultant forward blood flow) is lowered to an extent that is anticipated to only compensate for the backward flow of blood through the VAD itself and across the aortic valve. Systemic heparinization aiming at an activated partial thromboplastin time above 55 seconds was originally recommended [11]. However, broader clinical application of the VADs showed that the amount of heparin provided by the purger solution (see above) did allow for safe function of the VADs with an even lower activated partial thromboplastin time. The Recover VADs have

3 CARDIOVASCULAR 1644 JURMANN ET AL Ann Thorac Surg MINIATURE AXIAL FLOW VADS 2004;77: Fig 3. Combined use of the Impella Recover LVAD and RVAD in a patient with biventricular graft failure after cardiac transplantation. The intraoperative situs illustrates the remarkably compact implants of the smallest BVAD available at this time. The patient s head is oriented to the left. (BVAD biventricular assist device; LVAD left VAD; RVAD right VAD.) been designed for short-term use (7 days). The study was approved by our local institutional human research committee (application no.: 64/2001), with the latest submissions being approved on June 21, The Recover LVAD was implanted in six patients (five males, one female) with postoperative left heart failure following coronary artery bypass grafting (CABG) between May 2002 and November 2002 (Table 1). The patients mean age was years, ranging from 45 to 77 years; 50% of the patients were above 70 years of age. On average, the preoperative left ventricular function as assessed by left ventricular ejection fraction (LVEF) was Table 1. Preoperative and Intraoperative Parameters Related to Cardiac Function and Circulatory Status Pre-OP Pre-LVAD Age [years] (48 77) LVEF [%] (12 45) 17 4 (10 25) Dopamine/Dobutamine (n 3) (n 6) Epinephrine (n 1) (n 6) Enoximone (n 6) EuroSCORE logistic [%] (4 54) IABP score [n] (3 5) Parameters indicating cardiac function, circulatory status, and risk assessment scores of six patients before the operation ( Pre-OP ) and at the time they were considered for LVAD implantation intraoperatively ( Pre- LVAD ). The large variation in EuroSCORE [12] predicted hospital mortality rate is caused by the diversity of the patients preoperative status. Immediately before LVAD insertion, an IABP score of 3 to 5 would predict a 70% to 100% 30-day mortality rate [13]. For further explantations see text. The majority of the data are presented as the mean one standard deviation and the range. IABP intra-aortic balloon pump; LVAD left ventricular assist device; LVEF left ventricular ejection fraction. compromised in all patients (LVEF 28% 12%, range 12% to 45%). Three out of six patients presented with unstable circulatory status or profound cardiogenic shock, secondary to acute myocardial infarction, and underwent emergency CABG (Table 1). Two of these shock patients were intubated; one required cardiopulmonary resuscitation, defibrillation, very high-dose catecholamine support, and insertion of an IABP a few hours before surgery. The remainder of the patient population consisted of two elective cases and one urgent case. One patient had a severely reduced LVEF of 12% before the operation for elective CABG and LV aneurysmectomy. Two other patients developed intraoperative left heart failure during CABG, in one case it being a redo procedure. The preoperative logistic EuroSCORE [12] predicted an average 28% 20% mortality rate for the whole group (n 6). However, the EuroSCOREpredicted mortality showed a large variability (4% to 54%). If the patients are separated into those being operated on in stable condition (n 3, logistic Euro- SCORE 13% 12%) and those requiring emergency CABG procedures with instable circulatory patterns (n 3, logistic EuroSCORE 44% 11%), major differences in predicted mortality become obvious. The patients were considered for LVAD implantation if the function of their native heart was inadequate to establish stable circulatory patterns on attempted discontinuation from cardiopulmonary bypass or up to 1 hour thereafter despite high-dose inotropic (epinephrine 0.5 g kg 1 min 1 ) and IABP support (Table 1). At this time the intraoperative IABP score was also calculated (2 score points: epinephrine 0.5 g kg 1 min 1 ; 1 score point each: S v O 2 60%, diuresis 100 ml/h, and left atrial pressure 15 mm Hg) [13]. Between 3 and 5 score points were assigned to the patients in this series (Table 1), corresponding to a predicted 30-day mortality rate of 70% and 100% [13]. Initially, the LVAD blood flow was maximized to provide full unloading of the LV. With echocardiographic evidence of an increase in LVEF and a decrease in left ventricular end-diastolic dimensions, minimized catecholamine support, and stable circulatory patterns (cardiac index 2.2 l min 1 m 2 ), weaning from LVAD support was initiated. During the weaning phase, the blood flow in the LVAD was lowered 0.5 to 0.8 L/min every 8 to 12 hours, until LVAD blood flows of 1.0 to 1.5 L/min were reached. Native heart function was also evaluated during pump-off trials (dynamic pump stop). Once persistent improvement of LV function could be documented, the patient was referred to the operating room, the chest was reopened, the LVAD was withdrawn from the LV, and the graft prosthesis was shortened and its stump was oversewn. One 50-year old male with dilative cardiomyopathy underwent an elective heart transplantation. Intraoperatively, biventricular graft failure developed; the patient was maintained on cardiopulmonary bypass for an extended period of time and an IABP was inserted. After prolonged surgical reperfusion and several unsuccessful attempts to separate the patient from cardiopulmonary

4 Ann Thorac Surg JURMANN ET AL 2004;77: MINIATURE AXIAL FLOW VADS 1645 CARDIOVASCULAR Fig 4. Course of LVEF in six patients before, during, and after coronary artery bypass grafting and temporary circulatory support with the Recover LVAD. Four patients were weaned from the LVAD and two remain long-term survivors. (LVAD left ventricular assist device; LVEF left ventricular ejection fraction; preop preoperative; postop postoperative.) Fig 5. Serial average values for plasma free hemoglobin (hb) [g/100 ml] in six patients undergoing coronary artery bypass grafting and temporary support with the Recover LVAD. Data are presented as the mean one standard deviation. (LVAD left ventricular assist device; op day of operation; pod postoperative day.) bypass, echocardiography revealed persistent moderately impaired systolic, but severely impaired diastolic, biventricular graft function. A Recover LVAD and an RVAD were implanted to provide BVAD support (Fig 3) and the patient could be weaned from cardiopulmonary bypass immediately. Results The degree of circulatory support provided by the Recover LVAD was found to be adequate in clinical terms. If the mixed venous oxygen saturation (S v O 2 ) is regarded as a measure of tissue perfusion, the average S v O 2 was maintained well above 63% in all patients and at all times (including the weaning phase). Full unloading of the LV can be achieved with this device as demonstrated by intraoperative TEE, initial reduction of left atrial pressures down to 2 to 4 mm Hg in some cases, and the presence of a nonpulsatile arterial blood pressure trace. Intra-aortic balloon pump support was continued in all cases. Recovery of LV function was observed in five out of six patients supported by the Recover LVAD leading to removal of the LVAD after hours in four patients (Fig 4). In these four, the IABP was removed three days later ( hours). Two patients died on LVAD support; the fifth patient with demonstrated improvement of LV function could not be weaned from the LVAD because he had incurred septicemia and multiorgan failure early on during the treatment. The remaining patient showed no evidence of myocardial recovery and died later from multiorgan failure and visceral ischemia. Because of early development of multiorgan failure and septicemia, the latter two patients were not candidates for an extension of the therapy; ie, change to implantable LVADs or cardiac transplantation. Two of the weaned patients continue to remain long-term survivors, one of them now for more than 1 year. The two remaining patients who had been weaned from the LVAD died later of multiorgan failure secondary to septicemia. Moderate damage to the cellular elements of blood was observed during LVAD support. The presence of hemolysis was noted, with initial values of 18 4 g/100 ml settling around 10 to 15 g/100 ml on average after 5 to 6 days, which is coincident with the lowering of the rotational speed of the LVAD during the weaning phase at that time (Fig 5). A decrease in platelet count was a uniform observation (Fig 6). Cardiac rhythm disturbances rarely occur during utilization of the Recover Fig 6. Average daily platelet count [10 12 /L] of six patients receiving coronary artery bypass grafting and temporary support with the Recover LVAD. Data are presented as the mean one standard deviation. (LVAD left ventricular assist device; op day of operation; pod postoperative day; preop preoperative.)

5 CARDIOVASCULAR 1646 JURMANN ET AL Ann Thorac Surg MINIATURE AXIAL FLOW VADS 2004;77: LVAD as a consequence of direct mechanical interaction between the inlet cannula and the endocardium. A hypovolemic state with inadequate filling of the LV seems to predispose the occurrence of this phenomenon. Defibrillation because of ventricular fibrillation became necessary once in this series. In the one case with Recover BVAD support, the blood flow rates were maintained at 3 L/min in the RVAD and 5 L/min in the LVAD. Over the next 5 days full recovery of graft function was observed (LVEF 60%, RVEF 50%) during BVAD support, allowing removal of the Recover BVAD after 143 hours. Graft function remained stable thereafter; however, the patient showed signs of inadequate recovery of cerebral function. An intracranial hemorrhage was demonstrated by cranial computer tomographic scan and the patient died 5 days later. Comment The present results of circulatory support in the setting of acute heart failure or postcardiotomy heart failure need substantial improvement. First, the medical and surgical management of such patients has to address the factors which have been clinically proven to promote survival as cited above. However, the patients with acute postoperative heart failure are different from those, for example, receiving circulatory support for bridge-to-transplant procedures. Particularly, older age, prolonged cardiopulmonary bypass time, acute myocardial infarction, but also presence of shock, coagulation disorders, and a variable degree of right heart failure obviously subject these patients to early development of multiorgan failure and an inflammatory state or septicemia despite adequate circulatory support. Therefore, to some extent we might not expect the results of circulatory support for postcardiotomy heart failure to match those in the bridge-to-transplant patient population. Another source of improvement is the development of VADs that are suited to the situation of acute heart failure. Ventricular assist devices to be used in this situation should enable timely and simple implantation and explantation techniques without compromise of native heart function by the surgical procedure itself. They should be inexpensive as they may only be used as intermediate VADs in some cases. Such VADs should feature simple device control algorithms and provide a broad range of blood flows to enhance our present weaning strategies. By virtue of their design, the Impella Recover VADs correspond to most of these requirements; comparatively simple surgical maneuvers are necessary during implantation and explantation. They provide enhanced blood flow capabilities blood flows below 2 L/min can be safely used for extended periods of time and a dynamic pump stop feature allows safe evaluation of native heart function with the VAD creating no net forward blood flow. Most pulsatile flow VADs used for the treatment of acute heart failure, as the ABIOMED BVS 5000i (ABIOMED Inc., Danvers, MA, USA) or the Berlin Heart Excor system (Berlin Heart AG, Berlin, Germany), will not allow blood flows below 2 L/min to be maintained for extended periods of time. While the decision for VAD explantation using these pulsatile VADs carries a substantial risk, it is anticipated that the improved flow capabilities of the Recover VADs will increase the safety in predicting sustained recovery of native heart function and the timing of VAD explantation. Our initial experience, at this time, is too limited to prove these assumptions. The Recover VADs might be used without additional systemic heparinization besides the heparin administered as part of the purger solution. As postoperative bleeding continues to be a substantial source of morbidity and mortality in such patients [7], the low anticoagulation requirements are clearly advantageous in the setting of postcardiotomy circulatory support. The observed damage to the cellular elements of blood, such as hemolysis and a decrease in platelet count, might be attributed to the compact design of the LVAD and the sheer stresses imposed by the high rotational speed of the impeller. These data (Figs 5 and 6) are biased to some extent since previous shock and prolonged cardiopulmonary bypass times will lead to similar observations. While hemolysis, as assessed by plasma free hemoglobin values between 10 to 15 g/100 ml, on average might also be observed with other VADs in this situation, we found it helpful to administer packed platelets immediately after implantation of the LVAD and eventually by the time of device removal. Utilizing the advanced blood flow capabilities of the Recover VADs and a slow weaning protocol in this series, the LVAD weaning rate was 66% with a 33% discharge rate. With the obvious limitation of the small sample size, these results are on a par with, but do not improve on, previously published material [1 5]. As an alternative, these results might therefore be compared to established risk scores assessing the risk of mortality on an individual basis. The EuroSCORE is now well established as a score reflecting the operative risk in cardiac surgery on a large scale, and also in the North American patient population (The Society of Thoracic Surgeons [STS] database) [14]. The average predicted 30-day mortality rate by Euro- SCORE for the whole series was far below the actual mortality rate (Table 1). Regarding only the patients with preoperative unstable circulatory patterns (n 3), the predicted logistic EuroSCORE mortality rate (44% 11%; range, 32% to 54%) was much closer to the actual situation. More important,, the actual mortality rate was lower than that expected if treatment had been continued solely with IABP and drug support (as calculated by the intraoperative IABP score [13]) and without LVAD implantation. The IABP score has been developed to stratify the actual conditions of intraoperative IABP support versus outcome. Within the data set forming the basis for the development of the IABP score, patients with a score of 3 corresponded to 70%, a score of 4 to 92%, and a score of 5 to 100% predicted 30-day mortality rate [13]. This led us to the conclusion earlier, that patients need to be considered for VAD implantation with any IABP score higher than 0 (corresponding to a 30-day mortality rate of 14%). In this series, a mean score of (between

6 Ann Thorac Surg JURMANN ET AL 2004;77: MINIATURE AXIAL FLOW VADS 3 and 5 score points) was calculated translating into 70% to 92% expected mortality rates. Therefore, the utilization of the Recover LVADs provided some survival benefit when compared to mortality rate predicted by the IABP score [13] in this initial series of patients with severe heart failure. Another algorithm based on drug use and hemodynamic criteria to predict the necessity of VAD implantation for patients with intraoperative heart failure, was introduced earlier by Samuels [6]. In his formula [6], the cut-off dosage of epinephrine was substantially lower (at 25% to 30% of that used in our score [13]), but his algorithm accounted for administration of milrinone, dopamine, and dobutamine in addition. Post-transplant graft failure (depending on the underlying pathophysiology) might be managed by either univentricular [15] or biventricular circulatory support to aim at either recovery of graft function or early cardiac retransplantation [16, 17]. Previously, cardiac retransplantation appeared to be the more promising clinical approach [15, 16]. However, with improvements in circulatory support technology and patient management, and given the current shortage of donor hearts, more recently bridge to recovery of graft function by circulatory support devices remains the more realistic option. The latter approach has proven to be clinically successful in the immediate past, in our experience and that of others (as, for example, illustrated by the one case of Recover BVAD support reported here). In this case full recovery of graft function was observed to occur within 5 days of circulatory support, which was maintained to full extent after removal of the BVAD. Since this patient had been maintained on cardiopulmonary bypass for a substantial period of time and considering the moderate anticoagulation requirements of the Recover BVAD, it can be speculated that the fatal event of intracranial hemorrhage can most likely be attributed to the original operative procedure. Combining the Impella Recover LVAD with the RVAD creates a BVAD with remarkably compact implants (Fig 3). However, careful observation of device performance is necessary since there is no interaction between the LVAD and the RVAD with regard to flow control. The patient s pathophysiology allowed us to maintain the RVAD flow at about 1.5 L/min lower than the LVAD flow, a setting which provides a safety margin with regard to the risk of a sudden pulmonary volume overload in case of an acute drop in LVAD flow. The Impella Recover VADs are now being used in seven European countries, with more than 120 clinical implantations having been performed to date [11]. The broader clinical application will show whether the patients will benefit from the improved handling and weaning features of the Recover VADs. A variant of the Recover LVAD, designed for percutaneous application, was found to have similar blood flow capabilities during clinical tests and will be available for clinical use in the near future. The authors wish to thank Anne Gale, Medical Editor, Deutsches Herzzentrum Berlin, Berlin, Germany, and Beate Jurmann, MD, Berlin, Germany, for their editorial assistance in preparing the manuscript. References Pae WE Jr, Miller CA, Mathews Y, Pierce WS. Ventricular assist devices for postcardiotomy cardiogenic shock. A combined registry experience. J Thorac Cardiovasc Surg 1992; 104: Noon GP, Lafuente JA, Irwin S. Acute and temporary ventricular support with BioMedicus centrifugal pump. Ann Thorac Surg 1999;68: Farrar DJ. The Thoratec ventricular assist device: a paracorporeal blood pump for treating acute and chronic heart failure. Semin Thorac Cardiovasc Surg 2000;12: Parascandola SA, Pae WE Jr, Davis PK, Miller CA, Pierce WS, Waldhausen JA. Determinants of survival in patients with ventricular assist devices. ASAIO Trans 1988;34: Dreyfuss GD. Hemopump 31, the sternotomy Hemopump: clinical experience. Ann Thorac Surg 1996;61: Samuels LE, Kaufman MS, Thomas MP, Holmes EC, Brockman SK, Wechsler AS. Pharmacological criteria for ventricular assist device insertion following postcardiotomy shock. J Card Surg 1999;14: Samuels LE, Holmes EC, Thomas MP, et al. Management of acute cardiac failure with mechanical assist: experience with the ABIOMED BVS Ann Thorac Surg 2001;71:S DeRose JJ Jr, Umana JP, Argenziano M, et al. Improved results for postcardiotomy cardiogenic shock with the use of implantable left ventricular assist devices. Ann Thorac Surg 1997;64: Helman DN, Morales DL, Edwards NM, et al. Left ventricular assist device bridge-to-transplant network improves survival after failed cardiotomy. Ann Thorac Surg 1999;68: Siess T, Reul H, Rau G. Concept, realization and first in vitro testing of an intraarterial microaxial blood pump. Artificial Organs 1995;19: Impella Cardiosystems AG. Klinische Ergebnisse. Available at Accessed on June 28, Roques F, Michel P, Goldstone AR, Nashef SA. The logistic EuroSCORE. Eur Heart J 2003;24: Hausmann H, Potapov EV, Koster A, et al. Prognosis after the implantation of an intra-aortic balloon pump in cardiac surgery calculated with a new score. Circulation 2003; 106(Suppl 1):I Nashef A, Roques F, Hammill BG, et al. Validation of the European system for cardiac operative risk evaluation (EuroSCORE) in North American surgery. Eur J Cardio-thorac Surg 2002;22: Jurmann MJ, Wahlers T, Coppola R, Fieguth HG, Haverich A. Early graft failure following cardiac transplantation: management by extracorporeal circulatory assist and retransplantation. J Heart Transplant 1989;8: Jurmann MJ, Haverich A, Schäfers HJ, Wahlers T, Cremer J, Borst HG. Early graft failure after heart transplantation: circulatory assist versus retransplantation. In: Akutsu T, ed. Artificial Heart 3. Tokyo, Japan: Springer-Verlag, 1991: Jurmann MJ, Schäfers HJ, Haverich A, Demertzis S, Wahlers T, Borst HG. Mechanical circulatory support after heart transplantation: prognosis of right versus biventricular failure. In: Körner MM, Posival H, Körfer R, eds. Thoracic organ transplantation: routine as a challenge. Excerpta Medica International Congress Series Amsterdam: Elsevier Science, 1994: CARDIOVASCULAR