Percutaneous venoarterial extracorporeal membrane oxygenation for emergency mechanical circulatory support

RESUf!XiTATlON

Resuscitation 33 (1996) 29 -34

Percutaneous venoarterial extracorporeal membrane oxygenation for emergency mechanical circulatory support P. Maif’,*, C. Hoermann”, M. Moertl”, J. Bonattib, C. Falbesonerb, D. Balogh” ,‘Depurfment oj Anarsthesiu and lntensice Care Medicine,

University c$ Innsbruck School of’ Medicine. Anichstras,rr 3.7 4 -6020 Innsbruck. Austria hDepartment of Surgery, University of Innsbruck School of Medicine. Anichstras.re 35, A-6020 Innsbruck. .4~vtriu

Received 6 September 1995; revised 26 January 1996: accepted 14 March 1996

Abstract In this retrospective study we report our initial experience with percutaneous venoarterial extracorporeal membrane oxygenation in the emergency treatment of intractable cardiogenic shock or pulseless electrical activity. Between January 1994 and July 1995, percutaneous venoarterial extracorporeal membrane oxygenation was attempted in seven patients (pulse&ear,electrical activity. five patients; cardiogenic shock, two patients). In two of the seven patients, eRorts at arterial cannulatian resulted in cannula perforation at the level of the iliac artery. In the remaining five patients, percutaneous venoarterial extracorporeal membrane oxygenation could be established and was maintained for 3-84 h. Major bleeding remained a common complication during extracorporeal membrane oxygenation despite the use of heparin-coated bypass circuits and was responsible for death during extracorporeal membrane oxygenation in one patient. The remaining four patients could be weaned from mechanical circulatory support within 24 h, two after surgical interventions (resection of right atria1 tumor, heart transplantation), one after thrombolytic therapy. In one patient, cardiac function recovered spontaneously after 6 h on venoarterial extracorporeal membrane oxygenation. Three patients were discharged from hospital, two of them made a full recovery, one sustained severe hypoxic brain injury. A few patients with intractable cardiogenic shock or pulseless electrical activity can be resuscitated with the help of emergency percutaneous venoarterial extracorporeal membrane oxygenation. Emergency venoarterial extraeorporeal membrane oxygenation is associated with a high rate of complications and its use should therefore be limited to selected patients with a rapidly correctable underlying cardiopulmonary pathology (anatomic, metabolic or hypothermic) who do not respond to conventional advanced cardiac life support. Keywords:

Cardiocirculatory arrest; Shock; Membrane oxygenation; Resuscitation; Circulatory support; Emergency

1. Introduction Several authors reported the successful use of mechanical circulatory support in the resuscitation of patients with intractable cardiogenic shock or car-

AhbretGarions: CA, cardiocirculatory arrest; CABG, coronary artery bypass grafting; ECMO, extracorporeal membrane oxygenation; PEA, pulseless electrical activity; PTCA, percutaneous transluminal coronary angioplasty. * Corresponding author. Tel: + 43 512 5042400; fax: + 43 512 5042450

diocirculatory arrest (CA) [l-3]. However, mechanical circulatory support never gained widespread acceptance in the treatment of cardiac emergencies [l]. The main reasons were the obvious complexity of the technique, the necessity for emergency surgery (either thoracotomy or femoral vessel preparation) and the need for anticoagulation. Recent advances in cardiopulmonary bypass technology (heparin-coated extracorporeal circuits, percutaneous cannulation technioues) may overcome some of these limitations and have renewed interest in emergency mechanical circulatory support [4]. In this retrospective study, we want to report our

0300-9572~96;$17.00 G 1996 Elsevier Science Ireland Ltd. All rights reserved P/I s0300-‘)572(‘)~~0o98c)-6

P. Mair

30

et al. /Resuscitation

33 (1996) 29-34

Table 1 Patients’ characteristics and underlying cardiac pathology in seven patients scheduled for percutaneous emergency venoarterial ECMO Patient

Hemodynamics

Suspected diagnosis at the moment of cannulation

Underlying disease

1. F (58 years) 2. M (48 years) 3. M (37 years) 4. F (53 years) 5. M (70 years) 6. M (53 years) 7. M (38 years)

Shock PEA PEA PEA PEA PEA Shock

Right ventricular inflow obstruction Acute myocardial infarction CA during PTCA Mitral valve prothesis dysfunction Pulmonary embolism Pulmonary embolism Acute right heart failure

Cavernous hemangioma with intramural bleeding Coronary artery disease, CABG 9 years ago Myocardial infarction, left main dissection Thrombosis of mitral valve prothesis Multiple trauma Multiple trauma, fat embolism syndrome Rejection after heart transplantation

M, male; F, female.

initial experience with percutaneous venoarterial extracorporeal membrane oxygenation (ECMO) in the emergency treatment of cardiogenic shock and CA.

2. Materials

and methods

2.1. Patients

Percutaneous venoarterial ECMO for emergency mechanical circulatory support was attempted in seven patients from January 1994 to July 1995 at the University Hospital Innsbruck (five men, two women, mean age 51 years, range 37-70 years). The decision to call the ECMO team and discuss the use of venoarterial ECMO was made by the physician in charge of the patient and was based solely on clinical necessities.Two patients were in intractable cardiogenic shock with a systolic blood pressure below 50 mmHg despite highdose adrenaline therapy and in both patients hemodynamics had rapidly deteriorated within the previous hour. Five patients had already arrested at the moment of cannulation. All five had pulseless electrical activity (PEA) and underwent cardiopulmonary resuscitation (CPR) according to the guidelines of the European Resuscitation Council [5]. All five patients had arrested after cardiogenic shock resistant to maximum medical therapy (including intra-aortic balloon-pumping in one patient) and the physicians in charge of the patient expected conventional advanced cardiac life support (ACLS) not to be successful due to the underlying pathology. Duration of CPR before venoarterial ECMO ranged from 20 to 60 min. Six of the seven patients were cannulated in the intensive care unit, one in the catheterisation laboratory. Patient characteristics and underlying cardiac pathology are shown in Table 1. 2.2. Extracorporeal

circuit and cannula insertion

The extracorporeal circuit consisted of a centrifugal pump (Biopump, Biomedicus, Eden Prairie, MN), a hollow fiber oxygenator (Maxima, Medtronic Blood

System, Anaheim, CA) with an integrated heat exchanger (Stoeckert, Germany). No reservoir was included in the system. The whole extracorporeal system (pump head, oxygenator, tubing, cannulas) was heparin coated (Carmeda Bioactive Surface, Carmeda AB, Sweden). The system was mounted on a mobile cart. The time necessary to prime the system (albumin, Ringer’s lactate) was about 15 min. Cannulation of the femoral vessels was attempted percutaneously in all patients using a modified Seldinger technique. Cannulation was performed either by an anesthesiologist or a cardiac surgeon. After guidewire placement, a 17 or 19 French cannula was inserted for arterial and 21 or 23 French cannula for venous vascular access,after step-wise dilatation of the femoral vessels (Percutaneous Cannula Set, Biomedicus, Eden Prairie, MN). If repeated attempts at percutaneous cannulation failed, surgical preparation of the femoral vessels was performed and the percutaneous cannulas were inserted under direct visualisation. Surgical preparation of the femoral vesselswas performed by cardiac surgeons only. The tip of the arterial cannula was positioned in the common iliac artery or the distal abdominal aorta, the tip of the venous cannula at the junction of the inferior vena cava and the right atrium. Bypass flow was initiated at flow rates of 2-2.4 l/min per m2 body surface area, while tissue perfusion was monitored by blood pH, lactate levels, urine output and mixed venous oxygen saturation (if available). During mechanical circulatory support, patients were anesthesized with a continuous infusion of midazolam and fentanyl. Transesophageal echocardiography was used in six of the seven patients during cannulation (to monitor guidewire and cannula position), during ECMO (to monitor unloading of the heart and detect left ventricular distension) and during weaning from mechanical support (to monitor ventricular function). Before cannula insertion, all patients received a heparin bolus dose of 40-70 U/kg body weight. During ECMO, a small dose of heparin was given continuously to keep the activated clotting time between 150 and 180 s (Hemochrom 401, International Technidyne, Edison,

P. Mair et al. / Resuscitation 33 (1996) 29 34

Table 2 Outcome and complications of five patients treated with venoarterial ECMO Patient

Duration of ECMO (h)

1

3

3

18

5

7

6

6

7

84

Complications

Interventions during ECMO

Outcome

Surgical reconstruction of the femoral vessels Major bleeding, left ventricular distension Major bleeding

Resection of right atria1 hemangioma Unsuccessful attempt at revascularisation, heart transplantation Thrombolysis

Long-term burvivot

Major bleeding, upper body hypoxemia Major bleeding

Change to venovenous ECMO Thoracotomy ____~--

WI). In the event of any serious bleeding complication, heparin was stopped. Decannulation was done by simply removing the percutaneous cannulas and applying manual compression for at least 20 min.

3. Results 3.1. Outcome

In two patients with prolonged CA (patients 2 and 4, Table 1). venous but no arterial cannulas could be inserted percutaneously. Therefore, surgical preparation of the femoral artery was performed and arterial cannulas were inserted under direct visualisation (without using a guidewire in one patient). Adequate extracorporeal blood flow could not be achieved after cannula insertion and resuscitation efforts were stopped. Autopsy showed severe arteriosclerotic vascular disease and perforation of the arterial cannula into the retroperitoneum at the level of the iliac artery in both patients. In five patients, percutaneous cannulation of the femoral vesselswas successfully completed and venoarterial ECMO was established with extracorporeal blood flow rates between 2.0 and 2.2 l/min per m2 body surface area. Mean arterial blood pressure was equal to or above 55 mmHg in all five patients and available parameters of tissue perfusion (blood pH, lactate, urine output, mixed venous oxygen saturation) improved rapidly. Once stable extracorporeal blood flow had been established, patients were evaluated for a correctable underlying cardiopulmonary pathology. Two patients (patient 1, right ventricular inflow obstruction; patient 3, left main coronary artery dissection) were immediately transferred to the operating theatre. Two patients were transferred to the radiology department for contrast-enhanced high-resolution computed tomography scanning of the thorax to confirm suspected central

Long-term survivor Long-term survivor (hypoxic cerebral injury) Weaned. brain death after 20 h Died on ECMO

pulmonary embolism. Central pulmonary embolism could be confirmed in one patient (patient 5, thrombolysis) and could be excluded in the other patient (patient 6, fat embolism syndrome). One patient (patient 7, rejection after heart transplantation) was scheduled for retransplantation on a special urgency request. Patient 7 underwent thoracotomy while on mechanical circulatory support for intrathoracic bleeding after tube thoracostomy. Patients remained on femorofemoral venoarterial ECMO for between 3 and 84 h. Four of the tive patients were weaned from mechanical circulatory support within 24 h, one died during ECMO (patient 7, Tables 1 and 2, cerebral and intrapulmonary bleeding while awaiting cardiac retransplantation). In one patient, cardiac function recovered spontaneously (patient 6, fat embolism syndrome, Tables 1 and 2) and the patient was weaned from circuIatory support, but remained on venovenous ECMO for respiratory support. Extracorporeal life support was removed in this patient 20 h later when brain death was confirmed. One patient (patient 5, pulmonary embolism, Tables 1 and 2) could be weaned from venoarterial ECMO after thrombolysis, two patients could be weaned after surgical intervention (patient 3, remained on venoarterial ECMO after failed emergency coronary artery bypass grafting and underwent successful cardiac transplantation; patient 1, resection of right atria1 hemangioma; Tables 1 and 2). Three patients were long-term survivors. TWO of them (patient 1 and 3, Tables 1 and 2) made a full recovery, patient 5 (pulmonary embolism, thrombolysis, Tables 1 and 2) was discharged from hospital with severe hypoxic cerebral injury. 3.2. Complications

Beside vascular injury during cannulation, we observed several other complications directly related to the use of ECMO in our patients: major bleeding ( > 10 units of blood within 6 h) in four patients, surgical reconstruction of femoral artery after decannulation in

32

P. Muir et al. /Resuscitation

one patient, arterial hypoxemia of the upper half of the body during venoarterial ECMO in one patient and inadequate left ventricular drainage and left ventricular distension in one patient (Table 2). We did not observe clot formation within the extracorporeal circuit or leg ischemia in any of our patients. Complications directly related to the use of ECMO were potentially lethal in three of the patients (patients 2 and 4: cannula perforation; patient 7: cerebral and intrapulmonary bleeding; Tables 1 and 2). The major complication not directly related to the use of ECMO was severe hypoxic brain injury after prolonged resuscitation, responsible for death in one patient (patient 6, Table 2) and disability in another patient (patient 5, Table 2). 4. Discussion 4.1. Outcome Two of the seven patients were long-term survivors and have,resumed normal activity. We relate survival in these two patients only to the use of venoarterial ECMO. Considering the underlying pathology (left main coronary artery dissection, progressive right ventricular inflow obstruction), we believe that conventional ACLS would have failed in both patients. In the literature, resuscitation from intractable shock or circulatory arrest using emergency mechanical circulatory support is credited with survival rates between 4% and 64% [2,3,6-lo]. Survival rates strongly depend on patient selection and the underlying disease. Higher survival rates were reported in patients arresting in the catheter laboratory [1,3,6,7], in patients with pulmonary embolism [ 1,2] or postcardiotomy patients [3,11]. Very low survival rates were reported for patients resuscitated in the emergency department [8,9] or patients with major trauma [1,2]. Results from our small series of patients are in accordance with these previous data and our study confirms that long-term survivors of emergency mechanical circulatory support will be those patients with a rapidly correctable underlying cardiopulmonary pathology (anatomic, metabolic or hypothermic) weaned from extracorporeal life support early [l]. 4.2. Complications A major complication of percutaneous emergency venoarterial ECMO observed in this small series of patients was perforation by the arterial cannula at the level of the iliac artery. Two patients in prolonged circulatory arrest were affected. Collapse of the iliac artery associated with prolonged CA and arteriosclerotic vascular disease - proved at autopsy in both patients - were obvious predisposing factors. Perfora-

33 (1996) 29-34

tion occurred only after surgical preparation of the femoral artery and prolonged efforts to advance the long percutaneous cannula. In one patient, perforation was most likely caused already by the guidewire, in the other patient no guidewire was used. Consequently, we have changed our technique of arterial cannulation for emergency femorofemoral ECMO. Dilatators or cannulas are inserted into the femoral artery only after correct position of the guidewire in the thoracic aorta has been assured by transesophageal echocardiography. If attempts to place the tip of the guidewire in the thoracic aorta or attempts to advance dilatators or cannulas over the guidewire are not successful immediately, we change to an open surgical cannulation technique early. With surgical preparation of the femoral artery, the long percutaneous cannulas are no longer necessary and shorter standard cardiopulmonary bypass cannulas can be used. This approach should avoid vascular injury and perforation of an arteriosclerotic, small-sized or collapsed iliac artery [12]. In patients who might become candidates for venoarterial ECMO, we place single lumen central venous catheters in the femoral vessels very early. If mechanical circulatory support becomes necessary, the guidewires of the percutaneous cannulation set can be advanced immediately through the central venous catheters. Thereby, we hope to reduce hasty and forceful efforts at femoral vessel cannulation during prolonged CPR efforts. Severe vascular injury is an uncommon complication of percutaneous ECMO in the controlled environment of the catheter laboratory, also in patients cannulated during CA [6,7,13]. In the catheter laboratory, cannulas are advanced with the help of fluoroscopy only after angiography has yielded good information on both the site and the severity of arteriosclerotic vascular disease. Inability to cannulate the femoral vesselsfor emergency mechanical circulatory support and vascular injury are more common outside the catheter laboratory [3], although other authors have reported no vascular complications at all [8]. Vascular perforation with percutaneous arterial cannulas in a substantial number of patients treated with emergency ECMO has not been reported so far. The extracorporeal circuit used can be primed by a trained individual within about 15 min. Therefore, the duration of cannulation predominately limits the use of percutaneous ECMO in emergency situations. The time needed for cannula placement could not be recorded exactly in this retrospective study. However, in almost one third of our patients, cannula insertion was timeconsuming and associated with lethal complications. Major bleeding has remained a common reason for morbidity and mortality in our patients, despite the use of heparin-coated extracorporeal systems. Platelet dysfunction due to blood cell trauma, activation of the coagulation cascade,hyperfibrinolysis and low-dose an-

ticoagulation (commonly used during mechanical circulatory support even with heparin-coated systems) cause a typical coagulopathy during venoarterial ECMO [14]. Together with patient-related factors commonly observed in our study population (hepatic dysfunction, surgery or trauma, thrombolytic therapy) this coagulopathy predisposes patients for hemorrhagic complications. Results with mechanical support of the failing heart using centrifugal pumps are best if the duration of support can be kept short [ 1,151.Integrated in a venoarterial ECMO system, centrifugal pumps are generally not recommended for support longer than 24 h because of the rate of complications associated with long-term use [15]. Only one of our patients (patient 7, graft failure after heart transplantation) remained on venoarterial ECMO for more than 24 h, because a paracorporeal ventricular assist device for long-term use (Thoratec VAD) was not readily available. This patient had a smooth and uncomplicated course until he died from sudden and unexpected cerebral and intrapulmonary hemorrhage after 84 h. Death was most likely directly related to the coagulopathy caused by the long-term use of venoarterial ECMO. Although Aranki [ 121reported successful use of venoarterial ECMO up to 129 h, our data confirm previous suggestions that patients not weaned from venoarterial ECMO within 24 to 48 h should be switched to a more sophisticated device for long-term support. A major problem of circulatory support with venoarterial ECMO systems is incomplete unloading of the left ventricle [ 1,4,16]. The cause for left ventricular volume load and left ventricular distension during venoarterial ECMO is blood flow from thebesian and bronchial veins as well as minimal aortic valve insufficiency 14,161.Incomplete left ventricular unloading and left ventricular distension was not a common problem in our patients. All seven patients had PEA with some degree of residual mechanical ventricular activity during ECMO support. Although residual ventricular activity was insufficient to create a palpable pulse in five of our patients, transesophageal echocardiography and direct visualisation after sternotomy proved that it was sufficient to unload the left ventricle [17]. Only one of our patients (patient 3, Table 2) had episodes of ventricular fibrillation, which caused left ventricular distension as demonstrated by transesophageal echocardiography. Defibrillation and restoration of a perfusing rhythm resulted in sufficient unloading of the left ventricle. Restoration of a perfusing rhythm facilitates left ventricular unloading. even if systemic circulation is completely maintained by extracorporeal pump flow. In a patient with severely impaired pulmonary but preserved cardiac function. femorofemoral venoarterial ECMO can disrupt the usually uniform oxygen saturation within the arterial circulation [18]. The upper parts

of the body (coronary and cerebral vessels) receive poorly oxygenated blood from the lung ejected by the native heart, while the lower parts of the body receive well-oxygenated blood from the extracorporeal circuit [18]. Cardiac but not pulmonary function recovered in one of our patients during venoarterial ECMO (patient 6, Tables 1 and 2). In this patient, arterial blood gas analysis obtained from the radial artery demonstrated severe hypoxemia (oxygen saturation of only 76’S). Blood gas analysis in the arterial line of the extracorporeal circuit and pulse oximetry of the legs demonstrated 100% oxygen saturation. Hypoxemia in the upper half of the body of this patient could be improved only after changing from venoarterial to venovenous ECMO. 4.3. Echocardiogruphic

monitorin,q

In our hands, transesophageal echocardiography proved a most useful instrument to monitor patients resuscitated with venoarterial ECMO. During cannulation, transesophageal echocardiography allowed to identify whether the femoral vessel cannulated was an arterial or venous vessel (guidewire in the aorta-arterial vessel, guidewire in the right atrium-venous vessel) and later on to control guidewire and cannula position. During ECMO, trdnsesophageal echocardiography was an effective and readily available technique to monitor venous drainage and left ventricular unloading. During weaning from ECMO, transesophageal echocardiography gave valuable information about ventricular function and volume status. The value of transesophageal echocardiographic monitoring during mechanical circulatory support has been demonstrated before for postcardiotomy patients and patients bridged lo cardiac transplantation [19]. In conclusion, a few patients with intractable cardiogenie shock or PEA can be resuscitated with emergency percutaneous venoarterial ECMO. However, emergency venoarterial ECMO is associated with a high rate of complications and its use should be limited to selected patients with a rapidly correctable underlying cardiopulmonary pathology (anatomic, metabolic or hypothermic), who do not respond to conventional A(‘LS.

References [I] Aufiero TX, Pae WE. Extracorporeal cardiopulmonary support I‘or resuscitation and invasive cardiology outside the operating suite. In: Gravlee GP, Davis RF, editors. Cardiopulmonary bypass, principals and practise. Baltimore. MD: Williams and Wilkins, 1993: 682 692. 121 Mattox KL, Beal AC. Resuscitation of the moribund patient using portable cardiopulmonary bypass. Ann Thorac Surg 1976: 27: 4’6 443.

34

P. Mair et al. / Resuscitation 33 (1996) 29-34

[3] Reichmann RT, Joyo CI, Dembitsky WP, Griffith LD, Adamson RM, Daily PO et al. Improved patient survival after cardiac arrest using a cardiopulmonary support system. Ann Thorac Surg 1990; 49: 101-105. [4] Safar P, Abramson NS, Angelos M, Cantadore R, Leonov Y, Levine R et al. Emergency cardiopulmonary bypass for resuscitation from prolonged cardiac arrest. Am J Emerg Med 1990; 8: 55-61. [5] Chamberlain D, Bossaert L, Carli P, Edgren E, Ekstrom L, Hapnes S et al. Guidelines for advanced life support. Resuscitation 1992; 24: 111-121. [6] Mooney RM, Arom KV, Joyce LD, Fishman Mooney J, Goldenberg IF, Von Rueden TJ et al. Emergency cardiopulmonary bypass support in patients with cardiac arrest. J Thorac Cardiovast Surg 1991; 101: 450-454. [7] Rees MR, Browne T, Sivanathan UM, Whittaker S, Hick D, Verma SP et al. Cardiac resuscitation with percutaneous cardiopulmonary support. Lancet 1992; 340: 513-514. [8] Hoed1 W, Sterz F, Zeiner A, Frossard M, Laggner AN. Percutaneous cardiopulmonary bypass for CPR-ACLS resistant cardiac arrest in patients. Resuscitation 1994; 28: S4 (abstract). [Y] Hartz R, Locicero J, Sanders J, Frederiksen J, Michaelis L. Portable bypass does not improve survival in cardiac arrest patients. J Am Co11Cardiol 1989; 13: 121A (abstract). [lo] Overlie RA, Reichmann RT, Smith SC. Emergency use of portable cardiopulmonary bypass in patients with cardiac arrest. J Am Co11Cardiol 1989; 13: 160A (abstract). [l I] de1Nido PJ, Dalton HJ, Thompson AE, Siewers RD. Extracorporeal membrane oxygenator rescue in children during cardiac arrest after cardiac surgery. Circulation 1992; 86 (Suppl II): 300-304.

[12] Aranki SF, Adams DH, Rizzo RJ, Couper GS, DeCamp MM, Fitzgerald DJ et al. Femoral veno-arterial extracorporeal life support with minimal or no heparin. Ann Thorac Surg 1993; 56: 149-155. [13] Shawl FA, Domanski MJ, Wish MH, Davis M. Percutaneous cardiopulmonary bypass support in the catheterization laboratory: technique and complications. Am Heart J 1990; 120: 195203. [14] Pearson GA, Underwood MJ, Firmin RK. Bleeding in extracorporeal membrane oxygenation. In: Lewis T, Graham TR, editors. Mechanical circulatory support. London: Edward Arnold, 1995: 299-305. [15] Pennington DG, Swartz MT. Mechanical circulatory support: patient and device selection. In: Lewis T, Graham TR, editors. Mechanical circulatory support. London: Edward Arnold, 1995: 159-168. [16] Ward KE, Tuggle DW, Gessouroun MR, Overholdt ED, Mantor PC. Transseptal decompression of the left heart during ECMO for severe myocarditis. Ann Thorac Surg 1995; 59: 749-151. [17] Bocka JJ, Overton DT, Hauser A, Oak R. Electromechanical dissociation in human beings: an echocardiographic evaluation. Ann Emerg Med 1988; 17: 450-452. [18] Snider MT, High KM, Campbell DB, Williams DR. Extracorporeal membrane oxygenation. In: Henseley FA, Martin DE, editors. The practice of cardiac anesthesia. Boston, MA: Little Brown, 1990: 662-709. [lY] Brack M, Olsen JD, PedersenWR. Transesophageal echocardiography in patients with mechanical circulatory assistance. Ann Thorac Surg 1991; 52: 1306-1309.