Biventricular Assist Devices as a Bridge to Heart Transplantation in Small Children
Background— Experience with the use of biventricular assist device (BiVAD) support to bridge small children to heart transplantation is limited.
Methods and Results— We used BIVAD support (Berlin EXCOR) in 9 pediatric heart transplant candidates from 4/05 to 7/07. The median patient age was 1.7 years (12 days to 17 years). The median patient weight was 9.4 kg (3 to 38 kg). All children were supported with multiple intravenous inotropes±mechanical ventilation (6) or ECMO (3) before BiVAD implantation. All had significant right ventricular dysfunction. The median pulmonary vascular resistance index (Rpi) was 6.0 WU/m2. Eight patients were successfully bridged to heart transplantation after a median duration of BiVAD support of 35 days (1 to 77 days). One death occurred after 10 days of support from perioperative renal failure in a 3 kg infant. Five patients required at least 1 blood pump change. One patient had a driveline infection requiring treatment. There were no acute neurological complications, no thromboembolic events, and no bleeding complications. In 2 patients with Rpi >10 WU/m2 unresponsive to pulmonary vasodilator therapy, Rpi dropped to 1.4 and 4.6 WU/m2, after 33 and 41 days of support, respectively. All 8 survivors underwent successful heart transplantation. Of 5 patients supported >30 days, 3 developed an extremely elevated (>90%) panel reactive antibody by ELISA that was not confirmed by other methods; none had a positive donor-specific retrospective crossmatch. There was 1 episode of rejection (with hemodynamic compromise) in the 8 transplanted patients. Rpi was normal (<3 WU/m2) without pulmonary vasodilators in all patients within 3 months after transplant. There have been no deaths after transplant with a median follow-up of 19 months.
Conclusions— BiVAD support can effectively be used in small children as a bridge to heart transplantation and can be accomplished with low mortality and morbidity. BiVAD support may offer an additional means to reverse extremely elevated pulmonary vascular resistance. Surveillance for HLA antibody sensitization during BiVAD support may be complicated by the development of non-HLA antibodies which may not reflect true HLA presensitization.
Though the treatment of children with acquired and congenital heart disease has advanced significantly in recent years, a subset of patients will experience refractory myocardial failure, necessitating cardiac transplantation. In an effort to keep the patient alive until a suitable organ becomes available, mechanical circulatory support may be necessary. In adults, many ventricular assist devices (VADs) have been developed to clinical maturity that can function as an effective bridge to transplantation. Some pulsatile paracorporeal VADs are implantable in larger sized adolescent patients.1–3 In reality, nonpulsatile devices such as extracorporeal membrane oxygenation (ECMO) and centrifugal pumps have been the mainstay of pediatric circulatory support technology, particularly when biventricular support is required.4,5 These technologies reliably provide immediate cardiopulmonary support, however their long-term use is associated with significant potential risks.6 Also, the intricate circuits mandate patient immobilization, which has a detrimental impact on rehabilitation. Although small pulsatile devices suitable for ventricular support in children have been available in Europe for some time, such miniaturized pumps have not been approved for use in North America. Recently, the Berlin Heart EXCOR device became available for use in North America on a compassionate use basis. We describe our experience with biventricular assist device (BiVAD) support using the Berlin Heart pump in 9 pediatric heart transplant candidates.
Materials and Methods
We conducted a retrospective, nonrandomized review of all patients who underwent insertion of a Berlin Heart EXCOR VAD at Saint Louis Children’s Hospital. The study spans from April 2005 until July 2007. Demographic and clinical outcome data, including adverse events and information regarding pump performance and device malfunction, were collected prospectively on all VAD recipients at the time of device implantation. Data points are summarized in tabular format and are expressed as a mean or median value with observed extremes where appropriate. The Washington University Human Research Protection Office approved this scientific review. The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.
Berlin Heart EXCOR Ventricular Assist Device
All children in the study underwent placement of Berlin Heart EXCOR VADs (Berlin Heart AG Berlin, Germany) on a compassionate-use basis. The Berlin Heart consists of a paracorporeal, pneumatically driven polyurethane blood pump (10 to 60 mL) with a multilayer flexible membrane that separates the blood and air chambers. Silicon cannulae connect the blood pumps to the patient and triple-leaflet polyurethane valves prevent blood reflux. All blood-contacting surfaces are heparin-coated. The pump is driven by a pulsatile electropneumatic system. The drive unit (IKUS 2000) has a triple operational control and pneumatic system, with synchronous and asynchronous operating modes.
The operations were performed using cardiopulmonary bypass with mild hypothermia on a beating heart. Short periods of fibrillatory or cardioplegic arrest were used only to close intracardiac shunts. Inflow cannulation for the left ventricular assist device (LVAD) was completed via the left ventricular apex in 8 children and via the left atrium in 1 child with a small, nonapex forming, left ventricle. Left-sided outflow cannulation was to the ascending aorta. Right atrial to main pulmonary artery cannulation was used for the right-sided device in 8 children and right ventricle to pulmonary artery cannulation was used in 1 child with L-TGA anatomy. The Berlin Heart assist device was run in a synchronous operating mode in 7 children and asynchronously in 2 patients.
Postoperative anticoagulation was started with intravenous heparin 24 hours after admission to the intensive care unit. The goal for the partial thromboplastin time was 40 to 51 seconds for the first 72 hours and 52 to 80 seconds thereafter. The chronic anticoagulation regimen was initiated on postoperative day 3. In the first 3 patients in the series, this consisted of oral warfarin to keep the international normalized ratio between 3 to 3.5. For the latter 6 patients in the series, a triple drug regimen was used, consisting of low molecular weight heparin (initiated at 1 mg/kg/dose subq b.i.d.), aspirin (5 mg/kg/dose p.o. b.i.d.), and dipyridamole (1 mg/kg/dose p.o. q.i.d.). The goal antifactor XA level was between 0.75 and 1.0.
Patient characteristics and demographics are summarized in Table 1. The study included 7 girls and 2 boys. The median age of the patients was 1.7 years (12 days to 17 years). The median patient weight was 9.4 kg (3 to 38 kg). The etiology of heart failure was cardiomyopathy in 7 children and complex congenital heart disease in 2. Of the latter 2 patients, one patient had congenitally corrected transposition of the great vessels with a small ventricular septal defect, status post epicardial pacemaker placement, and one had a forme fruste variant of hypoplastic left heart syndrome, with no previous intervention.
Preimplantation variables are summarized in Table 2. All children were supported with multiple intravenous inotropes±mechanical ventilation (6 patients) or ECMO (3 patients) before BiVAD implantation. All had severe systemic ventricular dysfunction. At least moderate right ventricular dysfunction, as defined by a 2-dimensional echocardiographic ejection fraction less than 40%,7 was present in all patients. The median pre-VAD pulmonary vascular resistance index (Rpi) was 6.0 WU/m2.
Management During Support
Time courses for ventilatory support, intensive care unit stay, and duration of VAD support are summarized in Table 3. The median ventilatory requirement post VAD insertion was 4 days (1 to 22 days). Only 3 of 8 survivors required mechanical ventilation longer than 1 week. The median length of stay in the intensive care unit was 11 days (1 to 65 days). Chronic care for all but 1 patient was managed on a regular cardiac ward. The median duration of VAD support was 35 days (1 to 77 days).
The primary end points included survival to and after heart transplantation. Eight of the 9 patients (89%) who received VADs survived the period of circulatory support. A 3-kg baby died who required immediate ECMO after birth; the child had significant pre-VAD renal insufficiency which progressed to anuric renal failure post Berlin Heart insertion. Despite the institution of hemodialysis, the child never thrived and, at the request of the parents, support was withdrawn on postoperative day 10.
Adverse events during VAD support are indicated in Table 4. Three patients had 4 infectious complications. One child suffered a blood borne infection, while 2 experienced respiratory infections. One patient suffered a driveline infection, requiring treatment with local wound care and intravenous antibiotics. There was one episode of postoperative renal insufficiency not requiring dialysis. There were 12 total pump changes secondary to fibrin deposition or thrombus in 5 patients; 4 patients required ≥2 pump changes. One patient suffered a hemidiaphragm paralysis necessitating plication. There was an acute mechanical failure of the IKUS driver in one instance, necessitating manual pumping until the substitute driver was attached. There were no complications related to postoperative bleeding. There were no acute neurological complications and no thromboembolic events. There was no significant detectable hemolysis, as measured by serial hemoglobin and plasma free hemoglobin assays. The chest was closed primarily in all patients.
Overall, 8 of 9 patients were successfully bridged to heart transplant. Posttransplant survival has been 100%, with a median follow-up of 19 months. The clinical course of each patient is summarized in Table 5.
Human leukocyte antigen (HLA) sensitization, defined as a dithiothreitol-treated T-cell panel-reactive antibody (PRA) titer greater than 10% immediately before transplantation, occurred in 3 patients in the series. All were supported longer than 30 days and all developed an extremely elevated (>90%) PRA by ELISA that was not confirmed by Luminex; none had a positive donor-specific T or B cell retrospective crossmatch. There has been 1 episode of rejection (with hemodynamic compromise) in the 8 transplanted patients.
Pulmonary Vascular Resistance
In 2 patients with pulmonary vascular resistance (Rpi) >10 WU/m2 unresponsive to pulmonary vasodilator therapy, Rpi dropped to 1.4 and 4.6 WU/m2, after 33 and 41 days of support, respectively. Pulmonary vascular resistance was normal (<3 WU/m2) without pulmonary vasodilators in all patients within 3 months after transplant.
Pediatric Mechanical Support
Since the 1980s, cardiac transplantation has been the most effective long-term therapy for children with intractable heart failure. However, it is not unusual for a child listed as a status 1A heart transplant candidate to wait several months before an organ becomes available. Options for mechanical support in children include miniaturized intraaortic balloon pumps, ECMO, centrifugal pumps, and, more recently, both pulsatile VADs and axial flow devices.5,8–12 ECMO remains the most common form of mechanical support available and is the best option for acute decompensation. ECMO provides total cardiopulmonary support, is relatively rapid, and allows the flexibility of peripheral and central cannulation. ECMO pumps, however, achieve nonpulsatile flow and the circuit is complex. The incidence of medium and long-term bleeding and infectious complications is exceedingly high and neurological impairment with extended use is also common.6 ECMO also restricts patient mobility, impairing physical rehabilitation.
Ventricular assist devices have potential advantages over ECMO as a mechanical bridge. Pulsatile pumping results in better tissue perfusion and specifically provides better recruitment of the microcirculation of the brain, lungs, and kidneys during extracorporeal circulation. In addition to improving the patient’s hemodynamic status and reversing end-organ dysfunction, VADs can be partially or fully implanted and allow for physical rehabilitation to improve the patient’s overall condition and likelihood for successful transplantation. The first successful pediatric bridge to heart transplantation using a pulsatile paracorporeal LVAD was reported in 1991 in an 8-year-old boy.13 Despite a substantial worldwide experience with pediatric specific devices, centers in the United States have relied on adult-sized devices because of a lack of US Food and Drug Administration approved miniaturized VADs.14 Specific concerns regarding using “oversized” devices have been documented.15 Pumping large stroke volumes into a small aorta can perpetuate systolic hypertension and subsequent intracranial hemorrhage, stasis in the device can cause thromboembolic complications, and placement of multiple adult size cannulae in a limited pericardial space can be technically challenging.
Our institutional experience with BiVAD support using the Berlin Heart system began in April 2004 for a child after cardiac arrest and emergent ECMO cannulation. The overall 89% survival rate in this series during circulatory support is favorable compared to other trials. Studies examining pediatric bridge to transplant experiences have reported bridging success rates between 51% to 78% and 1-year survival posttransplant from 62% to 88%.2,16–19 Our posttransplant survival after VAD support has been excellent, with 100% of children alive to date.
Timing of device implantation is of critical importance. However, classic guidelines for VAD implantation have not always been successful. Though risk factor summation scores systems have been calculated, these models are largely unsubstantiated when applied to children.20 The indications for pediatric VAD usage are evolving. Our strategy for children with significant biventricular failure is planned BiVAD implantation. Though implantation of VADs too early exposes the patient to unnecessary surgery and potential device-related morbidity, it is mandatory to attempt to institute mechanical assist device therapy before the onset of any end-organ dysfunction. Studies in adult VAD recipients have demonstrated the greatest benefit from mechanical support in those patients who do not exhibit signs of secondary organ malfunction, supporting the strategy of early intervention.21 The importance of patient selection and timing of device implantation was also illustrated in a pediatric study from the German Heart Center.16
When comparing adverse events in this study with other reports in the literature, certain complications occurred with a much lower frequency, including postoperative bleeding, neurological events, and thromboembolism. We did experience a high incidence of fibrin or thrombin deposition in the blood pumps requiring pump changes. In the literature, neurological events vary considerably, from 6% to 45%. Some of these problems have been attributed to technical issues, such as left atrial cannulation, which is a known risk factor for stroke.17,19 Both the Berlin Heart EXCOR and Medos HIA blood pumps are manufactured in multiple sizes, obviating large stroke volumes in small children. The use of polyurethane valves instead of mechanical valves may be less thrombogenic, and transparent blood chambers allow for visual control of filling and emptying and transillumination detection of thrombotic deposits. Nevertheless, neurological event rates using these devices range from 11% to 45%, using a combination of left atrial and apical inflow cannulation and varied anticoagulation schemes.10,17 Pump exchanges owing to thrombus formation are not infrequent in Berlin Heart recipients. Recent data may support the use of a modified anticoagulation protocol, similar to that employed in the latter patients in our series.
BiVADs Versus LVAD
Isolated LVAD insertion may be sufficient for bridging some patients to transplantation, even in the face of significant preoperative right ventricular dysfunction. Such therapy allows for maximal unloading of the left ventricle, which can appreciably reduce the afterload of the right ventricle. Combined with pharmacological right heart support, the need for additional mechanical assistance can be more limited.
All of the children in our series had echocardiographic evidence of significant pre-VAD right ventricular dysfunction. Though MRI would have more accurately estimated right ventricular ejection, the critical nature of the patients precluded remote hospital transport. Rather than attempting to pharmacologically support the right ventricle, our tact was to empirically use biventricular assistance in these patients. Other than a slightly prolonged operative time, we have observed few disadvantages to the addition of right ventricular mechanical support. Technical limitations imposed by adult devices are easily overcome by pediatric sized hardware. BiVAD support significantly simplifies postoperative management. Elimination of RV dysfunction completely may aid in postoperative bleeding, secondary to improved liver function, may limit renal dysfunction, and may permit earlier extubation, hence promoting a quicker recovery.
Pulmonary Vascular Resistance
In 2 patients with severely elevated pulmonary vascular resistance unresponsive to pulmonary vasodilator therapy, Rpi normalized after prolonged mechanical support. Pulmonary vascular resistance was normal without pulmonary vasodilators in all patients soon after transplant. Continuous unloading of the left ventricle, thereby lowering left atrial pressure, is the theoretical basis for VAD implantation in these patients as a mechanism of improving their pulmonary hypertension. This theory has been substantiated in adult patients.22
Perioperative blood transfusions, device-related infections, and the interaction between the device surface and the patient’s immune system are recognized mechanisms for high rates of sensitization in VAD patients. The association of preformed antidonor antibodies with hyperacute rejection and the persistence of the antidonor immune response secondary to immunologic memory make allosensitization a relative contraindication to transplantation. Sensitized untreated VAD recipients have been noted to have a prolongation of waiting time to transplantation, an increased risk of acute rejection, and a trend of increased 30-day mortality.23 In our report, 33% of patients showed evidence of sensitization by ELISA, not verified by Luminex methodology. These children all had negative retrospective donor cross-match results. Only 1 patient has had a hemodynamically significant rejection episode postoperatively. It is important to recognize that surveillance for HLA antibody sensitization during VAD support may be complicated by the development of non-HLA antibodies, which may not reflect true HLA presensitization.
This series was limited by its retrospective, nonrandomized nature. As our experience with the Berlin Heart device increased, the selection criteria for implantation became more aggressive. Timing of pump insertion was also dependent on obtaining institutional IRB and FDA approval, as well as device availability by the manufacturer; all of these factors varied on a case by case basis. This lack of standardization may have had an impact on outcomes.
Insertion of pulsatile paracorporeal VADs has been validated as an effective strategy to bridge patients with refractory myocardial dysfunction to heart transplantation. Our experience demonstrates the feasibility of BiVAD support in small children with low mortality and reasonable morbidity. This experience emphasizes the importance of continued development and refinement of mechanical ventricular assist devices in the pediatric population.
D.T.B is a consultant for AGA Medical Corporation. C.E.C. is a consultant for Blue Cross/Blue Shield.
Presented at the American Heart Association Scientific Sessions, November 4–7, 2007, Orlando, Fla.
Haddad F, Hunt SA, Rosenthal DN, Murphy DJ. Right Ventricular Function in Cardiovascular Disease, Part I: Anatomy, Physiology, Aging, and Functional Assessment of the Right Ventricle. Circulation. 2008; 117: 1436–1448.
Baldwin JT, Borovetz HS, Duncan BW, Gartner MJ, Jarvik RK, Weiss WJ, Hoke TR. The National Heart, Lung, and Blood Institute Pediatric Circulatory Support Program. Circulation. 2006; 113: 147–155.
Warnecke H, Berdjis F, Hennig E, Warnecke H, Berdjis F, Hennig E, Lange P, Schmitt D, Hummel M, Hetzer R. Mechanical left ventricular support as a bridge to cardiac transplantation in childhood. Eur J Cardiothorac Surg. 1991; 5: 330–333.
Burlington DB. FDA regulation of medical devices. FDA perspective. Ann Thorac Surg. 1996; 6: 482–484.
Deng MC, Loebe M, El-Banayosy A, Gronda E, Jansen PG, Vigano M, Wieselthaler GM, Reichart B, Vitali E, Pavie A, Mesana T, Loisance DY, Wheeldon DR, Portner PM. Mechanical circulatory support for advanced heart failure: effect of patient selection on outcome. Circulation. 2001; 103: 231–237.