Outcomes of Children Bridged to Heart Transplantation With Ventricular Assist Devices
A Multi-Institutional Study
Background— Current ventricular assist devices (VADs) in the United States are designed primarily for adult use. Data on VADs as a bridge to transplantation in children are limited.
Methods and Results— A multi-institutional, prospectively maintained database of outcomes in children after listing for heart transplantation (n=2375) was used to analyze outcomes of VAD patients (n=99, 4%) listed between January 1993 and December 2003. Median age at VAD implantation was 13.3 years (range, 2 days to 17.9 years); diagnoses were cardiomyopathy (78%) and congenital heart disease (22%). Mean duration of support was 57 days (range, 1 to 465 days). Seventy-three percent were supported with a long-term device, with 39% requiring biventricular support. Seventy-seven patients (77%) survived to transplantation, 5 patients were successfully weaned from support and recovered, and 17 patients (17%) died on support. In the recent era (2000 to 2003), successful bridge to transplantation with VAD was achieved in 86% of patients. Peak hazard for death while waiting was the first 2 weeks after VAD placement. Risk factors for death while awaiting a transplant included earlier era of implantation (P=0.05), female gender (P=0.02), and congenital disease diagnosis (P=0.05). There was no difference in 5-year survival after transplantation for patients on VAD at time of transplantation as compared with those not requiring VAD.
Conclusions— VAD support in children successfully bridged 77% of patients to transplantation, with posttransplantation outcomes comparable to those not requiring VAD. These encouraging results emphasize the need to further understand patient selection and to delineate the impact of VAD technology for children.
Received July 25, 2005; revision received February 21, 2006; accepted February 24, 2006.
Mechanical circulatory support as a bridge to heart transplantation in children has been used routinely in many centers.1–5 Extracorporeal membrane oxygenation (ECMO) has traditionally been the mainstay of this therapy but is limited to support only in the acute setting.6–8 Waiting times for transplants often exceed the time that patients can be supported successfully with ECMO. Ventricular assist devices (VADs) in adults have been successful as a long-term bridge to transplantation and as destination therapy.9–11 There is substantially less experience with VADs in children; existing reports are limited to relatively small numbers of patients.
This report uses the Pediatric Heart Transplant Study (PHTS) database, which contains data from 23 pediatric heart transplantation centers in North America. The aims of this study were (1) to describe the clinical course and adverse events in pediatric patients requiring VADs as a bridge to transplantation, (2) to define risk factors for death while waiting on support, and (3) to compare posttransplantation survival between patients who did and did not receive pretransplantation VAD support.
Editorial p 2266
Clinical Perspective p 2319
Patient Selection and Data Collection
Data were analyzed from the PHTS database, which is a prospectively maintained database of patients <18 years of age listed for heart transplantation at 23 North American centers. Data collection and management have been described previously12 (see Appendix in the online Data Supplement for participating centers). All centers had Institutional Review Board approval for data collection through the PHTS. All approvals are kept on file at the Data Analysis and Coordinating Center. Between January 1, 1993, and December 31, 2003, 2375 patients were enrolled in PHTS. Of this cohort, 99 children underwent implantation of VAD before transplantation. The decision for VAD implantation was made at the discretion of the primary team on the basis of individual institutional clinical guidelines.
Supplemental data collection was undertaken retrospectively of the 99 VAD patients to delineate complications and clinical course of the VAD therapy. Of the 18 institutions with at least 1 patient on VAD, 100% submitted supplemental data collection forms. Duplicate data were cross-referenced with PHTS forms, and any discrepancies were resolved.
The VAD study group (n=99) was compared with the remaining 2276 patients listed for transplantation during the same time era from all 23 institutions in the PHTS. Patients were divided into status 1 and status 2 groups on the basis of United Network of Organ Sharing urgency status at the time of listing. The status 1 group included all patients listed as status 1 before 1999 and as status 1A or 1B after 1999. Preliminary data collection separated patients who were on VAD at the time of listing from patients who underwent VAD implantation after listing for transplantation. Preliminary analysis revealed that patient characteristics and outcomes were not distinct between these groups. They were therefore combined into a single group of patients, and standard Kaplan-Meier and parametric analyses were used for survival analysis. Patients supported by ECMO alone (without conversion to VAD support) were not included in the VAD group. Analysis was done with the ECMO group included within the non-VAD status 1 group and repeated with the ECMO patients excluded completely. Competing-outcomes methods were used to analyze outcome after implantation.13 Multivariate analysis in the hazard-function domain was used to identify risk factors for death.14
The authors had full access to the data and take full responsibility for their integrity. All authors have read and agree to the manuscript as written.
Of the 2375 patients enrolled in PHTS, 99 patients (4%) had VAD implantation. Patient characteristics of the VAD group are summarized in Table 1. There were 32 girls and 67 boys, with a median age at implantation of 13.3 years (range, 2 days to 17.9 years) and median weight of 56 kg (range, 3 to 150 kg). Origin of heart disease included 21 patients with congenital heart defects (single ventricle, n=5), L-transposition of the great arteries (n=5), critical aortic stenosis (n=3), and other 2-ventricle postcardiotomy congenital disease (n=8). The median age at listing of the congenital group was 13.5 years and of the cardiomyopathy group was 13 years. All patients met criteria for status 1 listing before VAD implantation, including 34 patients on inotropes alone, 41 requiring ventilatory support in addition to inotropes, and 26 bridged to VAD implantation from another form of mechanical support. These other types of mechanical support included intra-aortic balloon pump in 11 patients, ECMO in 10 patients, and transition directly from cardiopulmonary bypass in 3 patients. Two patients required a second VAD as a result of failure of the primary device.
Frequency of VAD use over the decade of this study is illustrated in Figure 1. Patients with VAD at listing and patients with VAD at transplantation increased over time; in the most recent era, 7.2% of pediatric patients who had undergone transplantation were supported with VADs at the time of transplantation. Over time, the device type changed also, with long-term devices used more commonly than short-term devices in the current era (30 long-term devices of 47 [64%] in the early era, 40 of 49 [82%] in the current era; P=0.05). Of the 18 PHTS centers implanting VADs, 7 centers placed 1 or 2 VADs during the study period, 9 centers implanted 4 to 8 VADs, and 2 centers placed 13 and 20 VADs, respectively.
Of the VAD patients, 59 were supported by left ventricular support alone, whereas 37 patients required biventricular support (Table 2). The distribution of body surface area and age in each category illustrates that the long-term pulsatile systems were used in larger adolescent patients, whereas centrifugal VAD was the predominant form of support in smaller children (Table 3). Two patients were supported with a combination of a pulsatile device for left ventricular support and a centrifugal pump for right ventricular support. Twenty-six patients were supported with short-term devices (BVS 5000, Abiomed Inc, Danvers, Mass, or Bio-Pump, Medtronic, Minneapolis, Minn); 70 patients were supported with long-term devices. Ten patients were supported with ECMO acutely and “double”-bridged to transplantation with VAD implantation after ECMO resuscitation.
Survival to Transplantation
Seventy-seven patients were successfully bridged to heart transplantation, with an additional 5 patients explanted because of myocardial recovery. The probability of successful bridge to transplantation (censored at transplantation) after VAD implantation was 85%, 80%, and 76% at 1, 3, and 6 months, respectively (Figure 2). The hazard function (instantaneous risk) was highest immediately after implantation. Competing outcomes by era (Figure 3) show that at 6 months after transplantation in the current era (2000 to 2003), 86% of patients had undergone transplantation, 8% had died, and 6% were still listed. There were 17 deaths while awaiting transplantation among the 99 VAD patients (17%). Causes of death after VAD implantation in the 17 nonsurvivors were stroke in 11 (65%), infection and sepsis in 3 (17%), multisystem organ failure in 2 (12%), and arrhythmia in 1 (5%).
Of the 10 patients bridged to VAD implantation from ECMO, 9 had undergone successful transplantations. There was no difference in the median duration of support based on the intensity of support required before implantation (inotropes versus ventilator with or without mechanical support, 38 days [range, 1 to 343 days] versus 34 days [range, 1 to 465 days]; P=0.39). However, the need for mechanical ventilation and/or mechanical support before VAD support tended toward a higher pretransplantation mortality rate than that of patients supported by inotropes alone (25% versus 9%; P=0.07).
The overall mean length of support was 57 days (median, 25 days; range, 1 to 465 days). The adverse events of pediatric patients after VAD implantation are summarized in Table 4. Major adverse events during VAD support included reoperation for bleeding in 35 patients, infection in 35, stroke in 19, and hemolysis in 16. There was no significant difference in rate of adverse events based on age or body surface area.
Because outcomes are different in adults between short-term and long-term devices, these subgroups were analyzed separately (Table 4). As expected, mean length of support differed between the 2 groups (9.4 versus 70 days). Clinical course of the patients on long-term support devices revealed 84% extubated and 73% successfully ambulating, whereas the vast majority of the short-term device patients were not. A comparison of primary outcomes shows that the 70 patients with long-term devices were significantly more likely to successfully bridge to transplantation than were the 26 patients with short-term devices (P=0.003). In addition, there was a statistically significant difference in the adverse event of stroke, with 9 of 26 patients (35%) supported with short-term devices experiencing a cerebrovascular event as compared with 9 of 70 patients (13%) with a long-term device (P=0.02). The rate of infection was significantly higher in the long-term device patients than in the short-term device patients (29 of 70 [41%] versus 3 of 26 [12%]; P=0.004). There were no significant differences in the adverse events of reoperation for bleeding, hemolysis, or thrombosis between the device types, nor was there a difference in cause of death between groups.
Comparison With Non-VAD Listed Patients
Compared with the other 1619 patients listed as status 1 in the same era who did not require VAD support, survival after listing was comparable between patients with VAD and non-VAD patients (P=0.4; Figure 4). After exclusion of the 174 patients on ECMO at the time of listing from the status 1 group, there was still no difference in survival to transplantation between the VAD group and the status 1 group who did not require any mechanical support (P=0.63).
Risk Factors for Death While Waiting in VAD Patients
The univariate analysis was significant for several variables, including younger age (P=0.006; Figure 5A), congenital diagnosis (P=0.005; Figure 5B), smaller body surface area (P=0.01), earlier era (P=0.01), gender (P=0.05), and device type (P=0.0001) (all of which were used in the multivariate model). Examples of these univariate Kaplan-Meier curves are depicted in Figure 5, showing the unadjusted effect of age and congenital diagnosis on survival to transplantation after implantation. Of note, there was no difference in survival between patients needing LVAD alone and those supported with biventricular support (P=0.9). In the multivariate model (Table 5), congenital heart disease (P=0.05), female gender (P=0.02), and earlier year of implantation (P=0.02) were the independent risk factors for death while waiting.
To examine the effect of VAD implantation on posttransplantation survival, the Kaplan-Meier posttransplantation survival curves were analyzed, comparing the patients who were successfully bridged to transplantation with VAD (n=76) with those other patients who had undergone transplantation as a status 1 (n=1144) and those patients who had undergone transplantation as status 2 (n=350). There was no statistically significant difference (P=0.8) in posttransplantation mortality between these groups (Figure 6).
The use of long-term mechanical circulatory support in children has increased substantially over the past decade as waiting times for scarce pediatric heart allografts have increased and understanding of recovery of cardiac function has improved. Despite using VAD technology developed for adults, this study demonstrates that VADs may be used as bridge to transplantation therapy in appropriately sized children with the expectation of a successful outcome in the majority of patients. In the current era, 86% of pediatric patients who received a VAD were successfully bridged to transplantation. On closer analysis, it is apparent that successful application of these devices is disappointing in the subgroups of patient with congenital heart disease and in the smaller, younger patients who rarely are large enough for most long-term assist devices. Perhaps most important, the present analysis is the first to demonstrate that posttransplantation survival for pediatric patients who were bridged with VAD implant is similar to that of status 1 patients who did not require such support.
This is a significant improvement when compared with studies of outcome of ECMO as a bridge to transplantation that demonstrate 47% to 57% survival to transplantation.1–3 Previously, smaller single-center studies4 and device registry data5,15,16 reported an &60% successful bridge to transplantation in pediatric patients supported with VAD. This is the first study that clearly demonstrates an era effect for the success of VAD implantation in the pediatric population. In general, the earliest experience involved the most critically ill children who often were near death at the time of VAD implantation. Improved outcomes in the most recent era may be influenced by the centers’ increasing experience with the surgical techniques, timing, and postoperative care; the use of more long-term devices over time; and refinements in patient selection. In fact, these refinements appear to have resulted in improved outcomes over time despite the increasing use of VADs in smaller and more complex patients.17–19 Further study is warranted to optimize criteria for patient20 and device selection.
With the decrease in the early multisystem failure after implantation, the adverse events associated with the long-term support have become the most critical obstacle. A major source of morbidity after pediatric VAD implantation continues to be neurological complications,5,15 with a 20% incidence of stroke in this study, a large proportion of which (11 of 19) were fatal. This rate was statistically different between short-term devices (35%) and those intended for long-term support (13%; P=0.02). A better understanding of the developmental changes in the coagulation cascade after implantation might facilitate a reduction in the high rate of bleeding and thromboembolic events during VAD support in children. In addition, research efforts into pediatric VAD technology must address biocompatibility issues to mitigate the risks arising from blood–surface interactions in these devices that may be unique to both lower flows and the pediatric patient. The introduction of pediatric-specific devices in the past year has brought this issue to the forefront.
The adverse event profile and hazard function for death from this study are remarkably similar to those reported from the larger International Society for Heart and Lung Transplantation mechanical circulatory support device database.11 Outcomes of long-term versus short-term devices are also similar to the adult data. The significant increase in mortality of adult patients11 bridged to transplantation with biventricular support versus left ventricular support alone, however, was not found in our study. Although the numbers are smaller, there was no survival difference for children supported with either left ventricular or biventricular support. One possible explanation for this comes from an examination of the underlying pathophysiology that necessitates mechanical circulatory support in children as compared with adults. The progression to right ventricular (biventricular) failure as a result of coronary artery disease or valvular disease in adults may suggest advanced and chronic disease that may be unresponsive to treatment with or without mechanical circulatory support. However, right ventricular failure and biventricular failure in children may be due to more acute primary pathophysiologies that occur without the advanced systemic disease.
In adults on long-term devices, normalization of cardiac output provided by long-term and implantable VADs allows a period of improvement in end-organ function that, combined with patient ambulation and cardiac rehabilitation, optimizes patient condition before transplantation. The present study demonstrates that most pediatric patients bridged to transplantation with long-term VAD implantation were extubated and ambulating. This represents a substantial advantage over both ECMO support and the short-term devices with which ambulation and extubation are generally not possible during support.
The present study is limited by the retrospective nature of the supplementary data collection and the relatively small number of patients with VAD implantation. It also must be noted that the inclusion criteria for the PHTS include listing for transplantation. Thus, it is possible that centers may have had post–VAD implantation deaths in patients who were not listed for transplantation and therefore not enrolled in PHTS. A survey of all enrolling centers identified only 7 additional patients who received a VAD but were not listed within the PHTS during the study period. We acknowledge that death after implantation may be underestimated in this report and may limit the generalizability of this data set. In addition, we acknowledge that only a randomized clinical trial can fully address the issue of efficacy and survival benefit with this technology and that this will be difficult because of small numbers.
Risk factors for death after pediatric VAD implantation in this study included era of implantation, diagnosis of congenital heart disease, and gender. It should be noted that although device type and age of patient were highly significant univariate factors, both lost significance once entered into the multivariate model. Device type and era were closely linked to one another. Other preoperative markers of end-organ function, anticoagulation protocols, and antibiotics prophylaxis protocols will be collected in the prospective registry (Intermacs), which should provide additional information of the role of VADs in pediatric circulatory support.
In conclusion, VAD support provides successful bridge to transplantation in 77% of listed patients and 86% in the current era. Posttransplantation survival is comparable to non-VAD patients. These encouraging results emphasize the need to further understand patient selection and to further delineate the impact of VAD technology and its associated adverse events within the pediatric end-stage heart failure population.
We are indebted to Susan L. Myers for her dedicated statistical assistance.
Dr Blume serves as chairperson of the pediatric subcommittee on research grant NHLBI, NIH N01-HV-58198, Interagency Registry for Mechanically Assisted Circulatory Support. Dr Blume also serves without compensation as chairperson of the Pediatric Advisory Board for Micromed Technologies, Inc. Dr Duncan is the principal investigator for NHLBI contract HHSN268200448188C, entitled The Pedi-Pump: A Versatile Implantable Pediatric Ventricular Assist Device, and serves without compensation on the Micromed Technologies, Inc Pediatric Advisory Council. Dr Naftel is the director of the Data Coordinating Center for research grant NHLBI, NIH N01-HV-58198, Interagency Registry for Mechanically Assisted Circulatory Support. Dr Kirklin is the principal investigator for research grant NHLBI, NIH N01-HV-58198, Interagency Registry for Mechanically Assisted Circulatory Support. The other authors report no conflicts.
Duncan BW, Ibrahim AE, Hraska V, del Nido PJ, Laussen PC, Wessel DL, Mayer JE Jr, Bower LK, Jonas RA. Use of rapid-deployment extracorporeal membrane oxygenation for the resuscitation of pediatric patients with heart disease after cardiac arrest. J Thorac Cardiovasc Surg. 1998; 116: 305–311.
Ungerleider RM, Shen I, Yeh T, Schultz J, Butler R, Silberbach M, Giacomuzzi C, Heller E, Studenberg L, Mejak B, You J, Farrel D, McClure S, Austin EH. Routine mechanical ventricular assist following the Norwood procedure: improved neurologic outcome and excellent hospital survival. Ann Thorac Surg. 2004; 77: 18–22.
del Nido PJ, Duncan BW, Mayer JE Jr, Wessel DL, LaPierre RA, Jonas RA. Left ventricular assist device improves survival in children with left ventricular dysfunction after repair of anomalous origin of the left coronary artery from the pulmonary artery. Ann Thorac Surg. 1999; 67: 169–172.
Rose EA, Gelijns AC, Moskowitz AJ, Heitjan DF, Stevenson LW, Dembitsky W, Long JW, Ascheim DD, Tierney AR, Levitan RG, Watson JT, Meier P, Ronan NS, Shapiro PA, Lazar RM, Miller LW, Gupta L, Frazier OH, Desvigne-Nickens P, Oz MC, Poirier VL, for the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) Study Group. Long-term mechanical left ventricular assistance for end-stage heart failure. N Engl J Med. 2001; 345: 1435–1443.
Laliberte E, Cecere R, Tchervenkov C, Wan C, Bittira B, Calaritis C, Beland M, Decell M, Reyes T, Shum-Tim D. The combined use of extracorporeal life support and the Berlin Heart pulsatile pediatric ventricular assist device as a bridge to transplant in a toddler. J Extra Corpor Technol. 2004; 36: 158–161.
Reddy SL, Hasan A, Hamilton LR, Dark J, Schueler SW, Bolton DT, Haynes SR, Smith JH. Mechanical versus medical bridge to transplantation in children: what is the best timing for mechanical bridge? Eur J Cardiothorac Surg. 2004; 25: 605–609.
Currently available ventricular assist devices (VADs) in the United States are designed primarily for adult use. Data on VADs used as a bridge to heart transplantation in children are limited. The present study summarizes a 10-year experience of 99 pediatric patients from 18 centers who underwent VAD implantation as a bridge to transplantation. Seventy-seven percent of these children with end-stage heart disease were successfully bridged to transplantation with posttransplantation outcomes comparable to those not requiring VAD. Risk factors for death while awaiting a transplant on VAD included earlier era of implantation, female gender, and the presence of congenital heart disease. The greatest challenges for the future are miniaturization of devices for use in infants and small children and development of strategies to improve outcomes in those with congenital heart disease.
The online-only Data Supplement can be found at http://circ.ahajournals.org/cgi/content/full/113/19/2313/DC1.