Extracorporeal membrane oxygenation

An ECMO circuit is available and rapidly deployable in many cardiac centers and in some large intensive-care units. The use of ECMO to provide cardiac circulatory support, respiratory support, or both offers a simple way of maintaining the circulation in children. ECMO is frequently the first choice of support for patients with intracardiac defects, concomitant respiratory failure and unknown potential for recovery from shock-induced organ dysfunction, because the technique represents a flexible emergency rescue system. Additional left atrial venting or balloon atrial septostomy allows complete cardiac decompression. Despite survival of 35–60%, ECMO offers only short-term support, with increasing onset of complications and fatal outcome beyond the second or third week.^{3, 7, 8} Although sufficient cardiac output with unloading of the poorly contracting heart can be established, ECMO also has potential negative side effects, such as hemorrhage, infection and cerebral events.⁹ This approach should generally not be considered if the underlying cause of cardiac failure is uncorrectable, although cases of bridge to heart transplantation with ECMO have been reported.²⁷

After corrective surgery, if immediate weaning from bypass fails, ECMO is a reliable tool for the first few days and offers the option of performing further diagnostic procedures in the catheter laboratory under hemodynamic protection. Further interventions may also be performed and, if myocardial function does not recover within the first week, ECMO can be replaced by a pulsatile VAD before complications arise.^{16, 28} A novel form of extracorporeal support is so-called rescue ECMO, or extracorporeal cardiopulmonary resuscitation, during cardiac arrest. In these circumstances, bypass circuits and equipment must be instituted within a very short space of time. In one study, the overall survival in almost 60 patients placed on ECMO during resuscitation was 40%.²⁹

Ongoing ECMO can be used in hemodynamically stabilized children during transfer by plane or ground transport to a cardiac center that can provide full surgical, long-term VAD support, and transplant options. Foley et al.²⁸ described their experience with 100 patients who underwent ECMO during transport, although these were mostly adults. Despite several complications, all patients survived the transport and 78% survived to hospital discharge.

Centrifugal pumps

The currently available centrifugal pump VADs are BioMedicus Biopump, (Medtronic, Minneapolis, MN), CentriMag^® (Levitronix, Zürich, Switzerland), RotaFlow^® (Jostra, Hirrlingen, Germany) and Capiox^® (Terumo, Ann Arbor, MI). This type of pump has been available since 1989 to support neonates and older children with postoperative cardiac failure but competent lung function. Based on vortex technology, these VAD systems use turbine spins of 10,000–20,000 rpm to create a flow of 5–6 l/min and have generally been applied for temporary assistance of stunned myocardium of the left ventricle.^{10, 11, 12, 13} The VAD system is designed for isolated left ventricular or right ventricular support, but with two pumps connected in series it can also be modified to provide biventricular support. This approach is, however, rather complex and demanding in terms of pump regulation.

The use of centrifugal pump systems has some advantages over other systems: no oxygenator is needed, they have low priming volume and low requirements for heparin and hemolysis, adequate left ventricular decompression can be achieved, transport is easy and costs are low. Adequate pulmonary function is, however, a requirement, and in most pediatric cases the patient's chest must remain open, as with ECMO. Such patients have to be constantly sedated and mechanically ventilated. Most patients will need parenteral nutrition and periodic infusions of fresh plasma and platelets. Duncan et al.³⁰ reported the outcomes of 29 pediatric patients supported with BioMedicus VADs. Survival was 71% among patients with anomalous origin of the left coronary artery arising from the pulmonary artery or with cardiomyopathy. In those who underwent heart transplantation after support, survival was 50% and fewer instances of neurologic complications and blood trauma were seen than in the ECMO group reported in the same paper.

Other common complications are recurrent bleeding, thrombosis and infections; complications frequently increase in number and severity after 2–3 weeks of centrifugal pump support. Thuys et al.¹⁰ reported their experience with centrifugal pump support in 34 children who had congenital heart defects and who weighed 6 kg or less. This team was able to wean 64% of the patients from the VAD and overall survival to discharge was 31%.

Pneumatic pulsatile ventricular assist devices

Only the Berlin Heart Excor^® and the Medos HIA pulsatile systems have so far proved successful in children of all ages.^{20, 21, 22, 23, 24, 25, 26} Both systems are extracorporeal and consist of a pneumatic compressor-operated diaphragm pump with valves. Both devices have transparent polyurethane pump housings that allow early thrombus detection, and the external position of the ventricles enables fast and safe pump changes if required. For the Medos HIA, results from only small series of children have been reported.^{21, 24, 25} More results are available for the Berlin Heart Excor^®.^{31, 32, 33} We will, therefore, discuss this device rather than the Medos HIA.

The pediatric version of the Berlin Heart Excor^® VAD is mounted with trileaflet polyurethane valves and is available with pump sizes of 10 ml, 25 ml, 30 ml, 50 ml, 60 ml and 80 ml. The 10 ml pumps are suitable for neonates and infants with body weight of up to 9 kg (body surface area 0.43 m²), and the 25 ml and 30 ml pumps can be used in children up to the age of 7 years (weight 30 kg and body surface area of about 0.95 m²); adult sized pumps can be implanted in older children (Figure 1).

Figure 1 Adult (left) and pediatric (right) Berlin Heart Excor^® (Berlin Heart AG, Berlin, Germany) ventricular assist devices.

The adult pump has a stroke volume of 80 ml and tilting disk valves. The pediatric-sized pump is suitable for children with a body weight of 3–9 kg. This pump has a stroke volume of 10 ml and polyurethane trileaflet valves. Polyurethane valves are now available in pumps of all sizes.

Full figure and legend (12K)Figures & Tables indexDownload Power Point slide (255K)

Special silicone cannulae connect the blood pumps to the body (Figure 2). These cannulae are anastomosed to the right atrium and pulmonary artery for right ventricular support, and to the apex of the left ventricle or, more rarely, to the left atrium and the ascending aorta for left ventricular support. In the past 5 years we have used apical cannulation whenever possible, because we have found the unloading of the left ventricle to be much more efficient than with left atrial cannulation. Good unloading of the left ventricle immediately reduces the afterload of the right ventricle by 20–25 mmHg and together with medical right-heart support (including nitric oxide), in our experience, can prevent the need for implantation of an additional VAD in the right ventricle in most children. To distinguish children in whom an additional VAD might be required, transesophageal echocardiography is performed in the operating room in all patients and right-heart filling pressure is monitored.

Figure 2 Different types and sizes of ventricular assist device cannulae.

(A) Arterial cannulae. (B) Atrial cannulae. (C) Left ventricular apical cannulae.

Full figure and legend (13K)Figures & Tables indexDownload Power Point slide (322K)

Various cannula designs and diameters exist to match patients' sizes and anatomies. For pediatric patients, internal diameters of 4.8 mm, 6.0 mm and 9.0 mm are available. All Berlin Heart Excor^® cannulae are designed to exit the body through the upper abdominal wall. A Dacron^® velour cover in the middle portion of the cannula stimulates rapid ingrowth of tissue as a biological barrier to ascending infections (Figure 3).

Figure 3 Standard configuration of Berlin Heart Excor^® (Berlin Heart AG, Berlin, Germany) biventricular support.

The preference is to insert the cannulae into the apex of the left ventricle, which unloads the left ventricle most completely and in many cases makes the addition of a right ventricular assist device unnecessary.

Full figure and legend (17K)Figures & Tables indexDownload Power Point slide (254K)

The pumps are driven by a pulsatile electropneumatic system and all blood-contacting surfaces are heparin-coated, which protects effectively against thrombosis for at least several weeks. The in-hospital driving unit (Ikus, Berlin Heart, Germany) operates pumps of any size in univentricular or biventricular configurations. For pumps of 50 ml or larger, a mobile system allows the patient to move around freely and to go home.

Results of our institution's experience with the pediatric Berlin Heart Excor^® device between January 1990 and October 2005 are shown in Table 2.23 Overall survival to heart transplantation or discharge home was 60%. In our discussion here, however, we exclude this learning phase and focus on the results of the past 5 years, during which 30 children were supported with Berlin Heart Excor^® VADs. Eight children were successfully weaned from the device, 16 underwent heart transplantation, and 6 died during support. No children died after weaning or transplantation, and the survival rate to discharge home was 80% (Table 3). Of the 21 children treated for either cardiomyopathy or fulminant myocarditis, 90% survived and were discharged home. Within the past 5 years, the smallest Berlin Heart Excor^®, which has only 10 ml stroke volume, has become a reliable and acceptable longer-term support device for neonates and small infants, down to a body surface of 0.2 m² (Figure 4). Since 1999, survival among our infants (age 0–12 months) has increased to 78% (Table 3).²⁰ Three children underwent ECMO before being switched to Berlin Heart Excor^®, for 5, 2 and 13 days, respectively. Currently, we make the decision to switch from ECMO to Berlin Heart Excor^® between days 5 and 8 of support because of the unpredictable length of time required until recovery or transplantation and the expected increase in ECMO-related complications with time.

Figure 4 A Berlin Heart Excor^® (Berlin Heart AG, Berlin, Germany) left ventricular assist device with 10 ml stroke volume, implanted in a 5-month-old infant.

The child can be extubated, mobilized and regularly fed.

Full figure and legend (16K)Figures & Tables indexDownload Power Point slide (348K)

Table 2 Results in our institution for the pediatric Berlin Heart Excor^® ventricular assist device^a between January 1990 and October 2005.

Full tableFigures & Tables indexDownload Power Point slide (250K)

Table 3 Results in our institution for the pediatric Berlin Heart Excor^® ventricular assist device^a between October 2000 and October 2005 regarding diagnosis and age.

Full tableFigures & Tables indexDownload Power Point slide (243K)

We have learned several lessons from our experience so far. The selection of patients remains a major factor influencing outcome and survival. In our early experience, VADs were mostly implanted as a last resort in extremely ill patients who had required resuscitation and whose prognosis was dismal. As the system proved to be technically safe and reliable, earlier implantation could be justified, and this has contributed to notable improvements in outcome.

Children who are in cardiogenic shock and show critically compromised organ perfusion are candidates for mechanical support. Some of these children might experience sudden deterioration of hemodynamic function. The indicators for critical low-cardiac-output syndrome are shown in Box 1. First, all medical treatment options, including afterload reduction combined with catecholamines, diuretics, a phosphodiesterase-III inhibitor, fluid and transfusion management and mechanical ventilation, should be exhausted. If all medication fails, there is a need for a mechanical device.

The Berlin Heart Excor^® system enables long-term circulatory support, which offers time to restore organ function, allows extubation, mobilization and neurological examination and increases the chance of transplantation. One experience we have found particularly important is myocardial recovery in children with fulminant myocarditis and in some cases of cardiomyopathy where the devices could be explanted; in most of these cases ventricular function has remained stable during long-term follow-up.³⁴

A factor that has played a part in our achieving satisfactory results in the past 5 years has been a change in our anticoagulation and monitoring protocols. In the earlier years we gave only heparin continuously and measured the activated clotting time. Currently, we give heparin in doses dependent on thrombelastography and activated partial thromboplastin time, and add antiaggregation and antiadhesion drugs (oral aspirin and dipyridamol) according to the platelet aggregation tests.

Other factors leading to improved survival in our most recent cohort have been earliest possible extubation, enteral nutrition and mobilization. All kinds of minor infections—even viral febrile infections that would be normal in healthy children every few weeks—are dangerous. Endogenous reactions against infection always activate the coagulation cascade because of rising fibrinogen levels, and other factors. During an infection, clotting must be anticipated and the frequency of checking the pumps and targets for activated partial thromboplastin times should be increased.

If echocardiographic signs of minimal recovery of myocardial function are detected, we administer angiotensin-converting-enzyme inhibitors and -blockers to reduce afterload and to influence the neurohumoral system. Our goal is to reach sinus rhythm, an age-dependent normal heart rate, and blood pressures in the lower range of age-dependent normal levels. Once these goals are achieved, the pumping frequency is reduced and the pump is stopped for a few minutes, with additional anticoagulation and three to four manual pump inflations per minute to prevent clotting under echocardiographic monitoring.

Weaning is considered possible if the ejection fraction is greater than 50% and the left ventricular diameter is within the 97th percentile. The weaning procedure is performed in the operating room with pharmacologic support (milrinone, low-dose catecholamines and afterload reduction). Transesophageal echocardiography and optimum anesthetic management are essential. In infants, cardiopulmonary bypass is used for explantation of the apical cannula. Improvements in the management of patients and in the technical equipment of the Berlin Heart Excor^® device mean that children now have even better chances for recovery or successful heart transplantation than adults.²⁰

Other pulsatile devices

Several other pulsatile devices developed for the adult population are used in school-age children: the HeartMate^® I (Thoratec, Pleasanton, CA), Toyobo (National Cardiovascular Center Tokyo, Japan), Abiomed^® BVS 5000 (Abiomed Inc., Delaware, MA) and Novacor^® (Baxter Healthcare Corporation, Irvine, CA). Results for application of these devices in adolescents and older children have been encouraging.^{17, 35, 36, 37, 38} The major disadvantage of these devices is that they are suitable only for patients with body surface area larger than 1.2 m² and require pump flows of more than 2 l/min.

Continuous-flow ventricular assist devices

Two continuous flow rotary VADs that use axial flow or centrifugal flow—the INCOR^® VAD (Berlin Heart) and the DeBakey VAD^®—have been introduced into routine clinical care. Some further devices are still being subjected to clinical trials: Jarvik 2000^® (Jarvik, New York, NY), HeartMate^® II, Duraheart^® (Terumo Kabushiki Kaisha Corporation, Shibuya-ku, Japan), Ventrassist (Ventracor, Chatswood, NSW, Australia) and CorAid^®, (Cleveland Clinic, Cleveland, OH).^{39, 40}

The first axial flow pump to be introduced into clinical practice for intermediate to long-term treatment of end-stage heart failure in adults was the DeBakey VAD^®.⁴¹ This device is powered via a small, flexible, percutaneous cable. The inlet cannula is positioned in the left ventricular apex and the outlet cannula is connected to the ascending aorta. The primary difference between this pump and the earlier generation of pulsatile devices is its small size (86 mm long, 25 mm wide) and low weight (95 g). The comparable INCOR^® device is 146 mm long and 30 mm wide and weighs 200 g. In Europe, the DeBakey VAD^® has been awarded the CE Mark for both bridging to transplantation and destination therapy. MicroMed modified the adult pump to fit children and in 2004 received FDA humanitarian device exemption status, enabling implantation of the DeBakey VAD^® Child in children aged 5–16 years awaiting heart transplantation.¹⁹

Cardiomyopathy

Children in whom all intensive pharmacologic treatment fails should be supported with a VAD suitable for long-term implantation, since heart transplantation is the primary goal. The waiting time to transplantation is unpredictable and many children die while they are on the waiting list if they do not receive mechanical support.²² Children, like adults, however, infrequently achieve complete cardiac recovery through VAD support, allowing pump explantation and the prospect of long-term clinical stability.^{42, 43}

Fulminant myocarditis

Children who have acute viral myocarditis comprise the group most likely to benefit from VAD use. These patients are healthy until the onset of fulminant myocarditis, and long-term circulatory support with a pulsatile pneumatic device is an effective method for bridging until cardiac recovery.

Duncan et al.⁴⁴ reported ECMO or centrifugal-pump treatment of 15 children with fulminant myocarditis, in whom 80% survival was achieved. In this study, the median support time was only 6 days, and five children underwent early heart transplantation. Mechanical support for a longer time with a VAD such as the Berlin Heart Excor^® might have led to complete myocardial recovery, which we have seen after 10–21 days, making pump explantation possible. If myocardial function does not improve with VAD use after several weeks, transplantation is a possible therapy, to which the VAD will serve as a bridge.³⁴

Cardiac surgery

Despite improvements in surgical techniques and in the management of cardiopulmonary bypass and myocardial protection, myocardial dysfunction can occur after surgery for complex congenital heart disease, resulting in failure in the left, right or both ventricles. Of children undergoing open heart surgery, 2–5% need mechanical circulatory assistance. The decision whether to use mechanical support should be made very soon after surgery, and the indication should be based on poor hemodynamic profile combined with metabolic acidosis, oliguria, rising lactate level, poor perfusion despite high-dose inotropic treatment, afterload reduction and, dependent on the case, despite an optimum combination of pharmacologic agents. Before a child is placed on mechanical circulatory support, any residual cardiac defect should be ruled out and good homeostasis is required. In this situation, selection of the optimum device is important. If full recovery within 2 weeks can be expected, and the patient's lung function is not compromised, a centrifugal pump circuit could be the best choice.^{11, 12} Alternatively, in cases of concomitant respiratory failure, ECMO should be considered, and for some complex cases a pulsatile VAD is the best option.^{38, 39}

Chronic stages of congenital heart disease

Children in chronic stages of congenital heart disease can develop ventricular failure at some time after cardiac surgery and become candidates for transplantation when all other surgical and medical options have been exhausted. As in children with cardiomyopathy, pneumatic pulsatile devices for long-term support should be implanted when transplantation is the only realistic goal. Since children in this group have complex anatomic disorders and have often had several previous operations, surgical placement of the cannulae might be more demanding than in the cardiomyopathy group.