The recommended steps for treating postcardiotomy

Transcription

1 Feasibility Study of a Temporary Percutaneous Left Ventricular Assist Device in Cardiac Surgery Antonis A. Pitsis, MD, FETCS, Aikaterini N. Visouli, MD, Daniel Burkhoff, MD, PhD, Petros Dardas, MD, FESC, Nikolaos Mezilis, MD, FESC, Georgios Bougioukas, MD, and Gerasimos Filippatos, MD, FESC St. Luke s Hospital, Thessaloniki, Greece; Columbia University, New York, New York; Democritus University of Thrace, Alexandroupolis, and University of Athens, Athens, Greece Background. The aim of this study is to evaluate a percutaneous left ventricular assist device (Tandem Heart pvad; Cardiac Assist, Pittsburgh, Pennsylvania) in the postcardiotomy setting. Methods. Between August 2001 and August 2004, 11 high-risk male patients who had undergone heart failure surgery or surgical revascularization were supported by the TandemHeart postcardiotomy. The major indication for pvad insertion was failure to wean from cardiopulmonary bypass. Three different techniques were employed for cannulation: the closed percutaneous technique, the open transeptal technique with percutaneous cannulas insertion, and direct central cannulation. Results. The mean duration of support was 88 hours. The mean pump flow was 3.09 L/min. The weaning rate was 72.72%. Survival to discharge and at 1 and 4 years was 54.54%, 45.45%, and 36.36%, respectively. The main complication was pericardial bleeding, noted mainly in patients receiving antiplatelet treatment preoperatively. Conclusions. The TandemHeart appears to be safe for temporary support after cardiotomy. It is a versatile device allowing different techniques of insertion. Device application yielded high weaning rate and satisfactory early and long-term survival. (Ann Thorac Surg 2007;84:1993 9) 2007 by The Society of Thoracic Surgeons The recommended steps for treating postcardiotomy cardiogenic shock are early and aggressive medical treatment with inotropic agents, intra-aortic balloon pumping (IABP), and in patients refractory to these measures, mechanical circulatory support (MCS) [1, 2]. Weaning from cardiopulmonary bypass (CPB) with use of IABP is well established, although evidence for the efficacy of this approach is hard to produce [2, 3]. Cardiac output increases by 10% to 20%, mainly owing to reduced afterload and increased coronary perfusion, but IABP does not directly pump blood and does not significantly redistribute blood flow [3, 4], suggesting that the degree of hemodynamic compromise that can be treated with this device is relatively limited. Numerous devices, which generally require major surgery for insertion and removal, have been used with similar success for postcardiotomy MCS, none being inclusive of all the characteristics of an ideal device [4]. Devices for postcardiotomy support should allow easy implantation [4] and removal, and versatility of cannulation. Devices that allow peripheral cannulation include left ventricular assist devices (LVADs) with extracorporeal centrifugal pumps, intracorporeal miniaturized axial flow pumps, and the extracorporeal membrane oxygenation (ECMO) that provides cardiopulmonary support [4 6]. The TandemHeart percutaneous ventricular assist device (Tandem Heart pvad; Cardiac Assist, Pittsburgh, Accepted for publication June 15, Address correspondence to Dr Pitsis, St. Luke s Hospital, Panorama, Thessaloniki, 55236, Greece; apitsis@otenet.gr. Pennsylvania) allows percutaneous placement of inflow and outflow cannulas [7 9], with a technique resembling the one introduced by Dennis and used clinically by Edmunds and coworkers [10]. The TandemHeart includes a dual chambered, paracorporeal, continuousflow centrifugal pump. The upper pump housing provides inflow and outflow ports. The lower housing provides communication with the controller and an integral infusion line to provide local cooling and anticoagulation [8]. The device is Conformité Européenne marked. It has been most extensively used for short-term left ventricular (LV) support during high-risk percutaneous coronary interventions and for treatment of periinfarct cardiogenic shock [8, 9]. The goals of this feasibility study were to evaluate the safety and efficacy of the TandemHeart for postcardiotomy MCS. Patients and Methods Between August 2001 and August 2004, 14 patients (0.61% of cardiac cases) were supported by LVADs. Eleven patients were supported by the TandemHeart. We retrospectively reviewed the data of these 11 high-risk male patients. The Institutional Ethics Committee approved this study, and consent was obtained from the patients. Seven patients had chronic heart failure with severely dilated hearts due to ischemic cardiomyopathy or valvular heart disease. They mainly underwent heart failure surgery including surgical ventricular remodeling. Acute 2007 by The Society of Thoracic Surgeons /07/$32.00 Published by Elsevier Inc doi: /j.athoracsur

2 1994 PITSIS ET AL Ann Thorac Surg FEASIBILITY OF TEMPORARY PERCUTANEOUS LVAD 2007;84: Table 1. Baseline Characteristics and Indications for Percutaneous Left Ventricular Assist Device (pvad) Support Mean SD (Range) Patients With Preexisting CHF Patients With Acute Ischemia Number Age (years) (48 87) (48 71) (61 87) Body surface area (m 2 ) ( ) ( ) ( ) EuroSCORE (6 17) (6 13) (6 17) Preoperative LVEF (%) (10 55) (10 25) (30 55) Evolving myocardial infarction 3 of 11 0 of 7 (0%) 3 of 4 (75%) Redo surgery 3 of 11 (27.3%) 2 of 7 (28.6%) 1 of 4 (25%) Emergency operation 6 of 11 (54.5%) 3 of 7 (48.9%) 3 of 4 (75%) IABP before LVAD 5 of 11 (45.5%) 1 of 7 (14.3%) 4 of 4 (100%) Surgery Isolated CABG 4 patients 0 of 7 4 of 4 Emergency redo MVR 1 patient 1 of 7 0 of 4 Surgical ventricular remodeling 6 patients 6 of 7 0 of 4 Indication for pvad Failure to wean from CPB 9 of 11 (81.81%) 6 of 7 (85.7%) 3 of 4 (75.0%) Preoperative decompensation 1 of 11 (9.1%) 1 of 7 (14.3%) 0 of 4 (0%) Postoperative cardiac arrest 1 of 11 (9.1%) 0 of 7 (0%) 1 of 4 (25%) CABG coronary artery bypass graft surgery; CPB cardiopulmonary bypass; EuroSCORE European System for Cardiac Operative Risk Evaluation; IABP intra-aorticballoon pump; LVAD left ventricular assist device; LVEF left ventricular ejection fraction; MVR mitral valve replacement. heart failure occurred as decompensation of preexisting end-stage chronic heart failure. Four patients underwent coronary artery bypass graft surgery for coronary artery disease and had acute de novo heart failure due to perioperative coronary syndromes. Indications and Criteria for TandemHeart Insertion The main indication for TandemHeart insertion was failure to wean from CPB (9 patients). Other indications were preoperative acute decompensation of end-stage chronic heart failure (1 patient) and postcardiotomy refractory cardiac arrest (1 patient). High doses of at least two catecholamines including epinephrine were administered to all 11 patients before initiation of mechanical assistance. In 5 patients (all 4 patients with acute ischemia and 1 patient with chronic heart failure), the TandemHeart was inserted after IABP failed. In 6 patients with chronic heart failure, the TandemHeart was inserted without previous IABP. In these patients with chronically dilated and hypokinetic hearts, criteria for MCS were failure to wean from CPB on first attempt despite maximum inotropic support, and minimal intrinsic pumping activity and marked LV dilatation not only on CPB withdrawal but even with CPB flows less than 1.5 L/min. The patients baseline characteristics and indications for MCS are shown in Table 1. Cannulation Among the 7 patients with preexisting chronic heart failure, the device was inserted in 1 patient preoperatively with the standard percutaneous technique (Fig 1c) as reported by Thiele and colleagues [8]. Two outflow arterial cannulas 15F were used. In the remaining 6 patients with preexisting chronic heart failure, an open transeptal insertion technique was employed intraoperatively. On CPB, the right atrial cannulation was converted to direct bicaval cannulation. Through the opening of the right atrial appendage, a pursestring suture was placed to the fossa ovalis. The long transeptal inflow cannula was inserted percutaneously through the right femoral vein, advanced inside the right atrium, guided under direct vision into the left atrium through a stab wound to the fossa ovalis, and secured in place by the pursestring suture (Fig 1). Outflow was by percutaneous bilateral femoral artery cannulation with two 15F Medtronic biomedicus cannulas in 4 patients or unilateral with a single 17F cannula in 2 patients (Fig 2). Direct cannulation of the ascending aorta and the left atrium through the right superior pulmonary vein (Fig 3) was employed in all 4 patients with acute ischemia, owing to peripheral vascular disease (3 patients) or to urgent need for initiation of support in an arrested postcardiotomy patient. The sternum was closed in all patients, with the exception of 1 patient in whom only the skin was closed. Pump Set-Up and Patient Management The pump speed was initially set between 5,900 rpm and the maximum speed of 7,500 rpm aiming to a cardiac index greater than 2.2 L/min measured by thermodilution, a mean systolic arterial pressure greater than 70 mm Hg, and a mixed venous oxygen saturation greater than 70%. After weaning from CPB, heparin was partially reversed with protamine. Continuous heparin infusion through the pump incorporated in the TandemHeart

3 Ann Thorac Surg PITSIS ET AL 2007;84: FEASIBILITY OF TEMPORARY PERCUTANEOUS LVAD 1995 Fig 1. Open transeptal technique for device insertion. (1a) The percutaneously inserted inflow cannula is guided to the right atrium. (1b) The inflow cannula has been inserted to the left atrium through a stab wound to the fossa ovalis. (1c) The end result of both pure percutaneous cannulation and the open transeptal technique is the same. system was started intraoperatively or early postoperatively to maintain an activated clotting time of 180 s to 200 s. In 3 of 4 patients with acute ischemia, IABP was maintained during support; in 1 patient with severe peripheral atherosclerosis, the balloon inserted through the ascending aorta was removed to facilitate insertion of the TandemHeart outflow cannula. In the 7 patients with decompensated chronic heart failure, IABP was not used during support, as the LV offloading provided by the device and pulsatility provided by residual LV function were adequate. Progressively decreasing doses of inotropes were administered during support, and weaning was generally attempted after the first 24 hours. When the patients were hemodynamically stable, without excessive inotropic and vasoactive drugs (such as epinephrine or norepinephrine) and with good vital organ function, weaning was considered. Preload, afterload, cardiac rhythm, and ventricular function were optimized. Low to moderate doses of mirlinone, dobutamine, or dopamine were administered during the weaning procedure, which lasted between 2 and 12 hours. The pump flow was gradually decreased as tolerated, to a flow of 0.5 L/min (for short Fig 2. A patient with percutaneous insertion of cannulas during support. Fig 3. Direct cannulation.

4 1996 PITSIS ET AL Ann Thorac Surg FEASIBILITY OF TEMPORARY PERCUTANEOUS LVAD 2007;84: periods with activated clotting time 220 s). If under these settings the cardiac index remained above 2.2 L/min, without ventricular dilatation and with adequate ventricular contraction, weaning was decided. Weaning from the ventilator started during support in hemodynamically stable patients, and most patients were extubated shortly after device removal. Device Removal In the 7 patients with percutaneous inflow and outflow cannulas insertion, the device was removed by simple withdrawal of the cannulas, followed by manual compression to achieve hemostasis at the percutaneous insertion sites. When direct cannulation was employed the device was removed in theater. Definitions and Measurements Cardiac output during support is defined as the right ventricular output measured by thermodilution. Residual cardiac output is defined as the cardiac output during support minus the device flow measured by the device flowmeter. Residual cardiac output in the setting of partial support LVAD represents the LV output through the aortic valve, and it is attributed to residual intrinsic function of the assisted left ventricle. Residual cardiac output equals cardiac output (thermodilution) minus LVAD flow [11]. Hemodynamic measurements immediately before initiation of support are not available because most patients were on CPB (n 9) and 1 patient was in cardiac arrest. Hemodynamic and pump measurements obtained within the first 2 hours from initiation of support (or reinitiation of support in the patient with preoperative insertion) are defined as initial measurements. Measurements made at a mean of hours from initiation of support are defined as measurements during support. Comparisons of these measurements were performed with the paired t test. Results The mean initial pump flow, pump speed, cardiac output (thermodilution), cardiac index, and residual cardiac output are shown in Table 2. On comparison of initial hemodynamics and hemodynamics during support of the 9 patients who survived beyond the first 6 hours of support, there was a significantly decreased diastolic systemic arterial pressure during support (p 0.002), reflecting increased pulsatility due to LV contribution to cardiac output, suggesting less severe ventricular compromise in those patients (Table 3). The mean durations of support, mechanical ventilation, intensive care unit stay, and postoperative hospital stay are shown in Table 4. Prolonged mechanical ventilation is attributed mainly to persisting or recurrent heart failure. Eight of 11 patients (72.72%) were weaned from the device. Six of 11 patients survived to discharge (54.54% of treated patients, 75% of weaned patients; Table 2). Among hospital survivors, 83.33% remained alive at 1 Table 2. Pump Measurements and Cardiac Output Direct Central Cannulation (Acute Ischemia) Percutaneous Cannulas Insertion (CHF) All Patients Mean SD, (Median, Range) Initial pump flow (L/min) (2.9, ) (2.5, ) (3.45, ) Initial pump speed (rpm) (6,500, 5,900 7,500) (6,400, 5,900 6,800) (7,500, 6,000 7,500) Initial total cardiac output (thermodilution) (L/min) (5.5, ) (5.6, ) (4.55, ) Initial total cardiac index (L/min/m 2 ) (2.9, ) (2.9, ) (2.41, ) Initial residual cardiac output (L/min) (2.8, ) (3, ) (0.9, ) CHF congestive heart failure.

5 Ann Thorac Surg PITSIS ET AL 2007;84: FEASIBILITY OF TEMPORARY PERCUTANEOUS LVAD 1997 Table 3. Initial and During Support Hemodynamic Variables Means (mm Hg) Initial (n 9) During Support (n 9) Paired t Test Systolic arterial NS, p 0.49 BP mm Hg Diastolic p arterial BP mm Hg Mean systemic NS, p 0.78 BP mm Hg Systolic PAP NS, p 0.54 mm Hg Diastolic PAP NS, p 0.96 mm Hg Mean PAP mm NS, p 0.38 Hg PCWP mm Hg NS, p CVP mm Hg NS, p 0.61 BP blood pressure; CVP central venous pressure; NS not significant; PAP pulmonary artery pressure; PCWP pulmonary capillary wedge pressure. year and 66.66% remained alive at 4 years. Among survivors in 4 years, 1 patient was in New York Heart Association class I and 3 patients were in class II. All patients reported that they were independent in daily activities and considered the treatment worthwhile. Causes of Death All 3 patients who were not weaned from the device were taken to theater with evolving acute myocardial infarction and died of biventricular failure (Table 5). Their left ventricle never recovered; right-sided heart failure became evident (within 0.5 hour to 8 days after initiation of support) and was rapidly followed by death or multiple organ failure. Table 4. Weaning, Survival to Discharge, and Duration of Treatments Variable Value Able to be weaned 8 of 11 (72.72%) Survival to discharge 6 of 11 (54.54%) Survival to discharge 2 in 5 (40%) with previous IABP depending on IABP 4 in 6 (66.66%) with early support before LVAD support Duration of support (41, ) (hours) Duration of mechanical (7.5, ) ventilation (days) Duration of ICU stay (16.5, ) (days) Duration of hospital (26.5, ) stay (days) IABP intra-aortic balloon pump; ICU intensive care unit; L- VAD left ventricular assist device. Table 5. Causes of Death Patients not weaned Patients weaned Patient Characteristics Acute de novo heart failure due to preoperative acute myocardial infarction Decompensation of preexisting congestive heart failure Cause of Death Biventricular failure (n 3) Recurrence of left heart failure (n 2) Two patients with preexisting chronic heart failure were weaned but did not survive to discharge owing to recurrent heart failure. These patients died of low cardiac output syndrome refractory to inotropes and IABP or cardiorespiratory arrest on grounds of LV failure, unstable sternum, and severe respiratory infection. Adverse Events The most common adverse event during support was pericardial bleeding with similar incidence in central and femoral cannulation (Table 6). Five patients (45.45%) required reexploration for bleeding complications. In 1 patient, bleeding was attributed to a small vessel, and in the remaining patients, to bleeding diathesis. Diffuse bleeding requiring reexploration was significantly higher among patients receiving aspirin or clopidogrel preoperatively ( 2 test, p 0.033). Among the 7 patients with percutaneous cannulas insertion, 1 patient (14.28%) with unilateral arterial cannulation presented with late bleeding from the left common femoral artery due to cannula dislocation (resolved by device removal), and 1 patient with bilateral arterial cannulation required fasciotomy owing to compartment syndrome of the right leg. Documented respiratory infection presented during support in 3 patients but its relationship to the device per Table 6. Adverse Events During Percutaneous Ventricular Assist Device (pvad) Support Adverse Event Pericardial bleeding Respiratory tract infection Leg ischemia Femoral artery bleeding (minor dislocation of outflow cannula) Percentage of Patients 45.4% Reopening How Resolved 27.27% Intravenous antibiotic treatment 14.3% of patients Fasciotomy who underwent percutaneous insertion 14.3% of patients who underwent percutaneous insertion Removal of the pvad on the 5 th day of support

6 1998 PITSIS ET AL Ann Thorac Surg FEASIBILITY OF TEMPORARY PERCUTANEOUS LVAD 2007;84: Table 7. Blood Product Transfusions se was not proven. All device cannula cultures after removal were negative. No thromboembolic or cerebrovascular events or device failures were noted. Excluding the patient in whom multiple organ failure followed the right-sided heart failure, there were no hepatic or renal complications based on clinical and laboratory data. Creatinine remained below 1.4 mg/dl, liver enzymes remained below triple the higher normal value. The mean lactate dehydrogenase was U/L. The platelet count increased after the third day of support in all 5 patients supported longer than 3 days. On transthoracic echocardiogram within 24 hours from device removal, among the 7 patients with transeptal inflow cannula insertion, no atrial septal defect was detected in 6 patients and a small, clinically insignificant atrial septal defect was detected in 1 patient. No atrial septal defect was detected on serial echocardiography. Blood product transfusions are shown in Table 7. Comment Mean / SD (Units) Range (Units) Median (Units) Packed red blood cells 5.8 / Fresh frozen plasma 6.72 / The generalized tendency toward minimally invasiveness has influenced the design and technology of VADs [12]. High mortality rates are reported early after weaning and device removal in postcardiotomy support [4, 5]. Avoiding surgical trauma during this critical period is desirable. TandemHeart is a versatile VAD allowing not only percutaneous insertion but also intraoperative open transeptal or direct cannulation. Although not purely percutaneous, the open transeptal technique involves percutaneous cannulas insertion and allows easy removal, avoiding surgical trauma. We performed percutaneous cannulas insertion to avoid resternotomy for device removal, prolonged mechanical ventilation, increased risk of infection, and hemodynamic and respiratory derangement. We performed direct central cannulation in patients with severe peripheral vascular disease or when support was urgently required for resuscitation in a patient with open sternum. No interference of the inflow cannula with the left ventricle with either cannulation technique had advantages since it was used in patients with acute myocardial infarction or after LV surgical remodeling. Although not maximal or directly comparable, in our patients, the TandemHeart flow with transthoracic cannulation was higher than that with percutaneous femoral cannulation. The flow with the transthoracically inserted TandemHeart (mean L/min, maximum 6.0 L/min) is comparable with that achieved by transthoracically inserted temporary VADs with other centrifugal or pulsatile pumps, microaxial flow pumps, and ECMO [4 6, 11, 13]. Although maximum flow of 4.5 L/min has been reported with the percutaneously inserted Tandem- Heart [14], in our patients, the maximum flow with percutaneous insertion was 3.1 L/min at a submaximal speed of 6,200 rpm (the mean flow being ). This flow is less than that achieved by ECMO and the Hemopump (Medtronic, Minneapolis, MN) inserted with open femoral cannulation and higher than that of the percutaneously inserted Hemopump [6, 13]. The incidence of pericardial bleeding is typical of that reported with commonly used extracorporeal centrifugal pumps and ECMO [4, 5], but it was higher that that reported for Impella (Cardiosystems AG, Aachen, Germany) [11], Hemopump [6], and temporary pulsatile VADs [4]. The incidence of bleeding was similar with percutaneous and direct cannulation, and it was higher among patients receiving antiplatelet treatment preoperatively. Delayed sternal closure could have been an option, but not without the disadvantages of restrainement to prevent cannulas dislodgment, sedation [5], prolonged mechanical ventilation, increased risk of infection, and reoperation for closure. Thromboelastography could guide the strategy of closure and blood product administration. Heparin withdrawal in the early postoperative hours could minimize the need of reexploration. The incidence of vascular complications is comparable with that reported with peripherally inserted ECMO [4, 15]. Anastomosis of a vascular graft to the common femoral artery to serve as a conduit for the arterial cannula could help in avoiding leg ischemia and cannula dislocation [4]. Despite advantages of easy device removal in percutaneous cannulation, owing to flow limitation and vascular complications, we would not recommend this approach in patients with no residual LV function, peripheral atherosclerosis, and small femoral arteries. Evaluation of results is closely related to the indications for MCS. Although the recommendation for IABP before LVAD support in acute heart failure is characterized as class I, level of evidence C [2], Baskett and colleagues [3] consider that there is class IIa, level of evidence C, for postoperative IABP in coronary artery bypass graft surgery patients, as some authors suggest [13]; or apply in practice [4, 5, 11] earlier initiation of mechanical support when the shock is less prevalent without delay after failure of IABP. The high incidence of complications after VAD implantation is directly related to prolonged attempted weaning periods from CPB [5]. Early MCS is considered justifiable if the risk for the patients is not increased by the adverse events related to the device itself [6]. Two randomized studies comparing TandemHeart support to IABP in patients with cardiogenic shock showed improved hemodynamic parameters with the TandemHeart and similar survival benefit [9, 16]. Intra-aortic balloon pumping to treat low cardiac output rather than ischemia has been associated with poorer prognosis [3]. In all patients with acute ischemia and in 1

7 Ann Thorac Surg PITSIS ET AL 2007;84: FEASIBILITY OF TEMPORARY PERCUTANEOUS LVAD 1999 patient with decompensated chronic heart failure, IABP preceded TandemHeart support. All these patients would have probably died as they were not responsive to all means of conventional treatment, but TandemHeart support led to survival of 2 of 5 patients (40%). In 6 patients with decompensated chronic heart failure, the TandemHeart was inserted without previous IABP. These patients would have probably died without some kind of mechanical support as they were not responsive to maximal conventional medical treatment. Although we can not produce evidence of superiority of Tandem- Heart in comparison with IABP (and perhaps IABP or another LVAD would have led to the same or better results), it was TandemHeart assistance that led to survival of 4 of 6 patients (66.6%). Weaning and survival to discharge with the Tandem- Heart was comparable with that reported with generally accepted centrifugal pumps, ECMO, pulsatile temporary VADs, and microaxial flow pumps in similar settings. Hospital discharge rates reported for patients receiving VADs postcardiotomy range between 20% and 60%, higher survival being observed when temporary VADs are used for bridging procedures [1, 4 6, 11, 13, 15, 17 19]. All patients who died on support had preoperative acute myocardial infarction and died of biventricular failure. The extent of myocardial injury has been underestimated in these patients who might have benefited by biventricular support or ECMO. Withdrawal of the inflow cannula to the right atrium and incorporation of an oxygenator to the circuit would allow switch to ECMO. Nevertheless, given that extensive myocardial damage is present in the majority of patients not surviving after postcardiotomy MCS, it has been considered unlikely that a specific VAD or mode of support could improve outcome [15]. In failure of cardiac recovery, early conversion to long term support could offer a life-saving alternative [15]. Mehta and Pae [14] recently reported conversion of the TandemHeart to long-term support. The cause of death in patients weaned from the device was recurrence of chronic left-sided heart failure. In these patients, heart failure surgery combined with early initiation of support to allow recovery from the operative insult yielded good results (100% weaning, 71.4% survival). Longer support to allow myocardial remodeling [20] could improve survival. In conclusion, versatility of cannulation, minimal invasiveness, and low cost are advantages of the Tandem- Heart. Rapid conversions to CPB with adequate flows and to long-term support have been recently shown. The adverse events during support were typical of those reported with other devices in similar settings. Device application yielded high cumulative weaning rate and satisfactory early and midterm survival. Percutaneous cannulation requires careful patient selection and biventricular failure necessitates urgent biventricular or ECMO support. We speculate that when poor outcome is anticipated, as in cases of extensive acute ischemic injury or persisting chronic heart failure, early implementation of a strategy to long-term support might improve outcomes. The authors thank Dr Emmanuel Psaltis for designing the Figures. References 1. Kirklin/Barratt-Boyes cardiac surgery. 3rd ed. Edinburgh: Churchill Livingston, 2003: Nieminen MS, Böhm M, Cowie MR, et al. Executive summary of the guidelines on the diagnosis and treatment of acute heart failure: the task force on acute heart failure of the European Society of Cardiology. Eur Heart J 2005;26: Baskett RJF, Ghali WA, Maitland A, Hirsch GM. The intraaortic balloon pump in cardiac surgery. Ann Thorac Surg 2002;74: Moazami N, McCarthy PM. Temporary circulatory support. In: Cohn LH, Edmunds LH Jr, eds. Cardiac surgery in the adult. New York: McGraw-Hill, 2003: Noon GP, Lafuente JA, Irwin S. Acute and temporary ventricular support with biomedicus centrifugal pump. Ann Thorac Surg 1999;68: Meyns B, Sergeant P, Wouters P, et al. Mechanical support with microaxial blood pumps for postcardiotomy left ventricular failure: can outcome be predicted? J Thorac Cardiovasc Surg. 2000;120: Pitsis A, Dardas P, Mezilis P, Nikoloudakis N, Filippatos G, Burkhoff D. Temporary assist device for postcardiotomy heart failure. Ann Thorac Surg 2004;77: Thiele H, Lauer B, Hambrecht R, Boudriot E, Cohen HA, Schuler G. Reversal of cardiogenic shock by percutaneous left atrial-to-femoral arterial bypass assistance. Circulation 2001;104: Thiele H, Sick P, Boudriot E, et al. Randomized comparison of intra-aortic balloon support with a percutaneous left ventricular assist device in patients with revascularized acute myocardial infarction complicated by cardiogenic shock. Eur Heart J 2005;26: Edmunds LH Jr, Herrmann HC, DiSesa VJ, Ratcliffe MB, Bavaria JE, McCarthy DM. Left ventricular assist without thoracotomy: clinical experience with the Dennis method. Ann Thorac Surg 1994;57: Siegenthaler MP, Brehm K, Strecker T, et al. The Impella Recover microaxial left ventricular assist device reduces mortality for postcardiotomy failure: a three-center experience. J Thorac Cardiovasc Surg 2004;127: Pitsis AA, Anagnostopoulos CE. Acute heart failure: is there a role for surgery? Heart Fail Rev 2007;12: Dreyfus GD. Hemopump 31, the sternotomy Hemopump: clinical experience. Ann Thorac Surg 1996;61: Mehta S, Pae W. Conversion of the Tandemheart percutaneous ventricular assist device to long-term support. Interact Cardiovasc Thorac Surg 2006;5: Smedira NG, Blackstone EH. Postcardiotomy mechanical support: risk factors and outcomes. Ann Thorac Surg 2001; 71(Suppl): Burkhoff D, Cohen H, Brunckhorst C, O Neill WW, and the Tandemheart Investigators Group. A randomized multicenter clinical study to evaluate the safety and efficacy of the Tandem- Heart percutaneous ventricular assist device versus conventional therapy with intraaortic balloon pumping for treatment of cardiogenic shock. Am Heart J 2006;152:469.e Stevenson LW, Kormos RL, Bourge RC, et al. Mechanical cardiac support 2000: current applications and future trial design. J Am Coll Cardiol 2001;37: Joyce LD, Kiser JC, Eales F, et al. Experience with generally accepted centrifugal pumps: personal and collective experience. Ann Thorac Surg 1996;61: Körfer R, El-Banayosy A, Arusoglu L, et al. Temporary pulsatile ventricular assist devices and biventricular assist devices. Ann Thorac Surg 1999;68: Pitsis AA, Visouli AN, Vassilikos V, et al. First human implantation of a new rotary blood pump: design of the clinical feasibility study. Hellenic J Cardiol 2006;47: