Listing and Transplanting Adults With Congenital Heart DiseaseClinical Perspective
Background—An increasing number of patients with congenital heart disease (CHD) are reaching adulthood and may require heart transplantation. The survival of these patients after listing and transplantation has not been evaluated.
Methods and Results—A total of 41 849 patients (aged >18 years) were listed for primary transplantation during 1995–2009. Patients with a history of CHD (n=1035; 2.5%) were compared with those with other causes (non-CHD group) (n=40 814; 97.5%); 26 055 (62.3%) reached transplantation and were subdivided into those with (reoperation group; n=10 484; 40.2%) and without (nonreoperation group; n=15 571; 59.8%) a previous sternotomy. Survival on the waiting list was similar between groups, but mechanical ventricular assistance was not associated with superior survival to transplantation among CHD patients. CHD patients were more likely to have body mass index <18.5 at transplantation (P<0.0001), were younger, and had fewer comorbidities. Early mortality among patients with CHD was high (reoperation, 18.9% versus 9.6%; P<0.0001; nonreoperation, 16.6% versus 6.3%; P<0.0001), but by 10 years, overall survival was equivalent (53.8% versus 53.6%). Analysis was limited by the lack of specific information regarding the CHD diagnosis in most patients.
Conclusions—Adults with CHD have high 30-day mortality but better late survival after heart transplantation. Mechanical circulatory assistance does not improve waiting list survival in these patients. This may be due to a combination of highly complex reoperative surgery and often poor preoperative systemic health.
Forty or 50 years ago, few patients with congenital heart disease (CHD) could be expected to live past early childhood.1 Since that time, remarkable advancements in operative techniques and perioperative care mean that as many as 85% of children with CHD will reach adulthood.2,3 Adult patients with CHD, especially those with complex single-ventricle repairs, provide significant and increasingly frequent challenges to medical care providers.4,5
Clinical Perspective on p 767
Although medical therapy provides an option for the treatment of these patients, an increasing proportion of them are being referred to transplantation centers, and an accurate understanding of their outcomes after transplantation is essential to optimizing their outcomes and the allocation of limited donor allografts. Most reports of transplantation in the adult population have been limited to small, single-institution series6,–,10; here, we report the first analysis of a large population of listings and transplantations for adult CHD.
The United Network for Organ Sharing provided deidentified patient-level, encrypted center-specific data from the Thoracic Registry (data source No. 122609–1) of the United Network for Organ Sharing Standard Transplant and Research Dataset. Use of these data is consistent with the regulations of our universities' institutional review boards.
Between 1995 and 2009, 41 849 adult patients (aged ≥19 years) were listed for primary cardiac transplantation. They were divided into 2 groups: those with CHD as the primary cause of heart failure (CHD group; n=1035; 2.5%) and those with other causes (non-CHD group; n=40 814; 97.5%); outcomes after listing were stratified by these groups. In addition to the diagnostic classification provided within the data set, text fields were reviewed to better identify the cause of heart failure, the presence of previous operations, and the type of mechanical circulatory support (MCS). Transplantation centers were divided into 2 groups: those with a stand-alone program at an independent children's hospital and all other programs.
Clinical status at and outcomes after transplantation were examined in the subset of patients (n=26 055; 62.3%) who received an allograft. Because a high proportion of CHD patients had a previous cardiac operation, and reoperation is independently associated with high posttransplantation mortality,11 we separately analyzed posttransplantation outcomes for patients with (reoperation group; n=10 484; 40.2%) and without (nonreoperation group; n=15 571; 59.8%) a previous sternotomy. Four groups were compared: CHD/nonreoperation (n=193; 1.2%) versus non-CHD/nonreoperation (n=15 378; 98.8%) and CHD/reoperation (n=365; 1.8%) versus non-CHD/reoperation (n=10 119; 98.2%).
Statistical analysis was performed with the use of SAS 9.2 for Windows (SAS Institute, Cary, NC). For outcomes after listing, the primary outcome measures were 2-month mortality on the waiting list and long-term survival after listing. Competing risks analysis of outcomes on the waiting list was used in a manner reported previously12; outcomes considered were as follows: alive and waiting, transplanted, removed, died.
After transplantation, primary outcome measures were postoperative mortality (defined as death within 30 days or before hospital discharge) and long-term survival among patients surviving 1 year after transplantation. Univariate (P<0.05, 2-sided) and multivariable analyses were performed; variables significant (P<0.05) in univariate analysis were candidates for inclusion in multivariable models. Models for postoperative mortality and 2-month mortality on the waiting list were estimated by logistic regression (backward selection, P<0.20 to stay); models for long-term survival were estimated by Cox proportional hazards regression (backward selection, P<0.20). For multivariable analyses, missing data were handled with the technique of multiple imputation (PROC MI; imputations=10).13
Continuous variables were compared with the Student t test and 1-way ANOVA with the Bonferroni method, in which comparisons were made between >2 groups. Categorical variables were compared by χ2 test. Some continuous variables, including body surface area, weight, age, hemodynamic variables (cardiac index, pulmonary capillary wedge pressure, pulmonary vascular resistance [PVR]), laboratory variables (creatinine, albumin), and transplantation procedure variables (ischemic time, donor transplant volume) were both used as continuous variables and categorized into subgroups. Subgroups were defined on the basis of previously published stratification schemes when available14,15; the most predictive method was used.
Kaplan-Meier analysis and Cox regression were used for time-to-event analysis; the assumption of hazards proportionality was tested by introducing terms of interaction with log(time) in the Cox model. As a sensitivity analysis to assess the effect of modeling structure, we repeated the final models in a hierarchical structure (PROC GLIMMIX), including a random transplantation center effect. Odds ratio estimates obtained from hierarchical models were virtually identical to those obtained from binary logistic regression. The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
After increasing from 1.3% to 4.1%, the percentage of listings for CHD has stabilized at 2.4% to 2.9% in the past 3 years (Figure 1). Baseline characteristics for all patients listed for cardiac transplantation are shown in Table 1.
Among the non-CHD patients, the cause of heart failure was predominantly dilated (17 837; 43.7%) or ischemic (20 170; 49.4%) cardiomyopathy. The cause of CHD was not further defined in most cases (909; 98.2%), preventing any further analysis on the basis of the type of CHD. Of the 1035 patients with CHD, 104 (10.1%) were listed at children's hospitals; the remainder were at adult or combined programs.
Outcomes on the Waiting List
Competing risks analysis of outcomes on the waiting list for heart transplantation is illustrated in Figure 2. Fewer of the CHD patients were transplanted (54.2% versus 63.1%; odds ratio, 0.59; 95% confidence interval [CI], 0.61 to 0.78). Two-month mortality on the waiting list was similar between groups: CHD, 8.6%; non-CHD, 7.7% (P=0.3).
In univariate analysis, among patients without CHD, MCS was associated with higher 2-month mortality (odds ratio, 2.34; 95% CI, 2.12 to 2.57), which was primarily the result of high mortality on extracorporeal membrane oxygenation (61.0%; odds ratio, 16.7; 11.2 to 24.8) and with intra-aortic balloon pump use (36.1%; odds ratio, 7.1; 6.3 to 8.0); patients with ventricular assist devices (VADs) had lower mortality than those without support (5.0%; odds ratio, 0.5; 0.4 to 0.6). In contrast, patients with CHD had higher mortality on MCS (44.7%; odds ratio, 13.5; 6.6 to 27.5) independent of the type (intra-aortic balloon pump, 66.7%; odds ratio, 28.1; 8.2 to 96.5; extracorporeal membrane oxygenation, 83.3%; odds ratio, 66.7; 7.7 to 580.8; VADs, 30.4%; odds ratio, 5.9; 2.3 to 15.1).
Within the CHD group, male patients (odds ratio, 2.8; 95% CI, 1.5 to 5.4), those with albumin <3.5 (9.8; 4.2 to 23.1), and those hospitalized outside of the intensive care unit (3.3; 1.6 to 6.9) had particularly high mortality. Other univariate predictors of early waiting list mortality were similar across groups (data not shown).
In multivariable analysis, the non-CHD but not the CHD patients had decreased waiting list mortality on VAD support (Table 2).CHD patients spent more time on the waiting list (median, 119 days versus 152 days; P<0.001, Mann-Whitney test), and a higher percentage of time was spent as status 1/1A/1B (46% versus 54%; P<0.0001).
MCS Among CHD Patients
CHD patients were less likely to have MCS at listing (Table 1) or transplantation (7.7% versus 16.4%; P<0.0001). VAD use at listing has increased from 1995 to 2005 for non-CHD patients from 6.1% to 18.2% (P<0.0001) but not for CHD patients (1995: 6.4% to 4.7%; P=0.8). Among non-CHD patients, 28.9% of VADs were Heartmate II nonpulsatile devices compared with only 17.1% among CHD patients (P<0.0001). VAD use at listing in CHD patients did not differ between children's hospitals and all other hospitals (3.9% versus 4.6%; P=0.8).
Characteristics at Transplantation
Recipient and Donor Characteristics
Between 1995 and 2009, a total of 27 361 primary transplants were performed (26 055 on patients listed 1995–2009; the remainder were listed earlier than 1995 and were excluded from further analysis). The proportion of transplanted patients with CHD over time is given in Figure 1. CHD patients were younger and smaller at transplantation but had fewer comorbidities; VADs were used rarely (Table 3). Donor characteristics were similar between groups.
In both the reoperation (odds ratio, 2.3; 95% CI, 1.6 to 3.2) and nonreoperation groups (1.8; 1.5 to 2.3), CHD patients were more likely to have allograft ischemic times >4 hours, although the distance between donor and recipient hospitals was similar (reoperation, 145 versus 145 miles; nonreoperation, 164 versus 153 miles; P=NS).
A total of 109 centers performed transplants on adult CHD patients over this time (median number of transplants, 3; range, 1 to 46); of these, 101 were centers that also performed pediatric heart transplants during the same time period (median number of pediatric transplants, 8; range, 1 to 274). These centers performed a median of 197 transplants over the 15-year study period (range, 12 to 1263). Of the 109 centers, 91 were adult hospitals, and the remaining 18 were primarily children's hospitals.
Short-Term Posttransplantation Outcomes
Postoperative mortality was higher among CHD patients in both the reoperation (18.9% versus 9.6%; odds ratio, 2.2; 95% CI, 1.7 to 2.9) and nonreoperation (16.6% versus 6.3%; 2.9; 2.0 to 4.3) groups. Univariate analysis among CHD patients identified the standard risk factors for postoperative mortality after heart transplantation (Table I in the online-only Data Supplement).
Multivariable analysis among all patients demonstrated that CHD was an independent predictor of high mortality (odds ratio, 3.6; 95% CI, 2.8 to 4.7) (Table II in the online-only Data Supplement). Multivariable predictors of postoperative mortality were similar across groups; however, centers performing >5 pediatric transplants had higher mortality in the CHD groups and lower mortality in the non-CHD groups (Table 4).
Posttransplantation complications occurred at higher frequency in the CHD patients, independent of reoperation status. CHD patients required reoperations for cardiac complications at higher rates (CHD/nonreoperation, 21.9% versus non-CHD/nonreoperation, 8.4%; odds ratio, 3.0; 95% CI, 2.0 to 4.5; CHD/reoperation, 20.4% versus non-CHD/reoperation, 13.7%; 1.6; 1.2 to 2.2). They also had a high incidence of posttransplantation dialysis (CHD/nonreoperation, 15.3% versus non-CHD/nonreoperation, 6.8%; 2.5; 1.6 to 3.8; CHD/reoperation, 24.7% versus non-CHD/reoperation, 10.8%; 2.7; 2.1 to 3.5). The need for permanent pacemaker, incidence of stroke, and postoperative length of stay did not differ between groups.
Poor long-term survival in the CHD group was primarily the result of early mortality (Figure 3). At 10 years, survival between CHD and non-CHD patients was equivalent (52.8% versus 53.6%; P=NS). Cox proportional hazard regression among patients surviving the first year is shown in Table 5 and Table II in the online-only Data Supplement. Among all patients, CHD was associated with a decrease in long-term mortality (hazard ratio, 0.73; 95% CI, 0.56 to 0.94). Predictors of long-term mortality were similar between groups, including the improved long-term survival associated with centers performing >5 pediatric transplants annually (non-CHD, 0.88; 0.81 to 0.95; CHD, 0.61; 0.32 to 1.16).
Improvements in congenital cardiac surgery, especially neonatal corrective surgery, have resulted in remarkable gains in life expectancy among children born with CHD.1,–,3 However, some CHD patients (especially those with morphological right systemic ventricle) may develop end-stage heart failure.8,10,16,–,19 Although the long-term course of these patients remains undefined, as many as 10% to 20% of CHD patients will eventually require cardiac transplantation.17
Because of the small proportion of transplants performed for adults with CHD, accurate data with which to guide decisions regarding listing and transplantation have been lacking. Early publications were often single-institution studies with small sample sizes10,18; more recent multi-institutional studies have been published but have suffered from methodological problems: statistical issues (for example, handling of missing variables in large data sets), inclusion of a small group of adults within a larger group of children in the same analysis, and inclusion of noncontemporary transplants performed >22 years ago.20,21
More importantly, patients cannot survive after transplantation without survival to transplantation, so that without data to guide listing and pretransplantation management, we cannot optimize the long-term survival of adults with CHD and heart failure. Accordingly, this represents the first large study to specifically and rigorously evaluate both waiting list and posttransplantation outcomes within this population.
Listing for Transplantation
Over the past 15 years, patients with CHD have formed an increasing percentage of cardiac transplants (Figure 1), but, more recently, the percentage of listings for CHD has stabilized at ≈2.5%. This plateau was not identified in studies with older data sets.21 In the context of the growing population of adults with CHD,2 this is a surprising finding. We hypothesize that it may be related to transplantation centers recognizing the higher early mortality associated with transplants for CHD and narrowing the criteria under which they will list adult CHD patients. We hope that data presented here and elsewhere may help to identify specific patients likely or unlikely to benefit from transplantation and assist in the development of data-driven criteria for listing.
There has been concern that the need for additional donor tissue for allograft implantation in CHD recipients might lead to increased waiting times because of the resultant exclusion of donors simultaneously donating lungs. We found that CHD patients spend longer on the waiting list despite a higher percentage of time as status 1/1A/1B. Other, nonanatomical considerations may contribute to the longer waiting time, such as the desire to await the “perfect” donor in younger patients.14,21 However, the fact that fewer patients in the CHD group reach transplantation suggests that decisions regarding the timing of listing, management of waiting list candidates, and donor selection all need improvement.
Several findings suggest that listing criteria for patients with CHD may not be properly defined and that listing may occur too late. First, although VAD support was associated with a higher likelihood of survival on the waiting list for patients without CHD (hazard ratio, 0.30), this association was not seen among patients with CHD. Mechanical ventricular assistance may not address several pathological processes leading to transplantation in adults with CHD (protein-losing enteropathy, for example). In addition, complex anatomy (including arteriovenous malformations, collateral flow, and shunts) may preclude effective support and ventricular decompression in many patients. However, anatomic contraindications are not always present. The absence of an increase in VAD use over time and the less frequent use of the more contemporary continuous-flow devices may indicate a lack of familiarity with VAD use among the surgeons and cardiologists who specialize in congenital disorders compared with their colleagues specializing in adults, who treat a much larger population of patients with heart failure. Improvements in survival on the waiting list require improved education and training in VAD implantation as well as further analysis to identify CHD diagnoses amenable to MCS.
Elevated waiting list mortality in the absence of traditional risk factors (and particularly among the often younger, status 2 CHD patients with few comorbidities) suggests that traditional listing criteria may result in late listings. It is likely that we need to list patients with CHD earlier, before the development of end-organ dysfunction necessitating mechanical support or deterioration of health and nutritional status.
CHD and non-CHD patients had similar levels of clinical support required at transplantation. Within the reoperation group, CHD patients were much less likely to have mechanical circulatory assistance at transplantation, likely representing both underutilization of and fewer appropriate candidates for MCS.
In the population examined by Patel et al, 21 CHD patients were more likely to have a higher PVR (>4 Wood units); in this more contemporary population, we did not find a similar difference. We also did not find a statistically significant impact of elevated PVR on posttransplantation outcomes. PVR is a problematic variable in CHD transplantation patients, many of whom have single-ventricle physiology.20 Neither the United Network for Organ Sharing database nor most single-institution studies18,22 have information about the catheter locations used to measure pretransplantation and posttransplantation pulmonary pressures or about the presence of aortopulmonary collaterals. Further work is needed to clarify the following: (1) whether accurate estimates of PVR can be obtained in patients with complex CHD and (2) what levels of PVR elevation might preclude successful transplantation.
Several other findings of our study are notable. First, patients with CHD were more likely to have a body mass index <18.5 at transplantation. This may occur as a sequela of poor nutritional status or as a consequence of the neurohormonal alterations of severe heart failure (cardiac cachexia).23,24 Although body mass index as a metric may result in misclassification of individuals, it remains a commonly used surrogate for body habitus, and extremes of recipient body mass index predict poor outcomes after transplantation.15 Improvements in the care of CHD patients are necessary to optimize the preoperative nutritional and cardiac functional status.
Consistent with studies of posttransplantation mortality, as well as those analyzing CHD patients, postoperative mortality was significantly higher among patients with CHD (odds ratio, 3.6).10,25,26 Unfortunately, the United Network for Organ Sharing data set does not contain information enabling identification of subgroups expected to have higher early mortality. Lamour and colleagues20 have shown that in a combined population (121 adults, 367 children), there is a wide variability in congenital diagnoses resulting in transplantation, and patients with more complex disease, such as a previous Fontan procedure, have higher early mortality. Other single-institution and multi-institution studies (although dominated by pediatric CHD patients) show that diagnosis matters (for example, failed Fontan procedures) and that the specific transplantation procedure required (including reoperative sternotomy, pulmonary artery reconstruction, and transplantation to a single lung) may affect outcomes.10,20,27 Accurate delineation of postoperative risks requires improved data collection in this population.
Interestingly, although later transplantation year is usually associated with improved outcomes,28 this was not the case among the CHD patients. We speculate that this is the result of an increasing number of transplantations for complex CHD. Higher mortality associated with increased complexity may hide the improved outcomes over time usually seen in transplantation.10 Among the factors likely resulting in increasing complexity are the following: (1) The risks of long-term pulmonary insufficiency and right ventricular failure in patients with repaired tetralogy of Fallot have been increasingly recognized and treated; (2) Fontan procedures have become increasingly more common over the last 20 years; and (3) increasing experience may lead to an increasing comfort with transplantation in even the most complex patients.27
We expected that an increasing volume of pediatric transplantations would correlate with improved outcomes after adult transplantation both because adults undergoing surgery for CHD may fare better when operated on by CHD surgeons29,30 and because transplantation center volume has been associated with improved outcomes.31 Surprisingly, although non-CHD patients with a previous sternotomy had lower mortality at centers performing >5 pediatric transplants per year (likely because of the correlation between pediatric and adult transplantation volume [r=0.8, P<0.0001]), the opposite effect was seen among CHD patients, for whom these centers had higher early mortality.
We theorize that the increased mortality seen within the CHD group at these high-volume centers reflects a higher proportion of transplants for complex CHD with high mortality.27 Our results suggest that such centers should proceed with caution, recognizing the significant potential for high mortality rates. Alternatively, adult transplant recipients (even those with CHD) may be better managed at adult than at children's hospitals. Although studies suggest that adults with CHD have better outcomes with experienced CHD surgeons, the appropriate facility has not been determined,29,30 and the postoperative management of transplant recipients is less influenced by their specific preoperative diagnosis than by factors such as age and size.
Finally, and most encouragingly, among early survivors, long-term survival for adult CHD transplant recipients was better than that among non-CHD patients, resulting in equivalent overall survival at 10 years. This is consistent with previous studies and with the International Society for Heart and Lung Transplantation registry.10,20,21,26 This likely reflects the younger age and lack of extracardiac comorbidities within the CHD population; those who survive the technically challenging allograft implantation will have improved long-term outcomes.27 This encouraging late survival suggests that increasing experience, with improvements in donor selection, operative technique, and perioperative care, may result in a transplant population with excellent survival.
In the shorter term, the higher perioperative mortality seen in patients with CHD may result in difficult decisions regarding listing strategy. Some have advocated using an alternate listing strategy in high-risk patients that may more equitably allocate the limited supply of donor organs. However, this has been associated with higher, although acceptable, mortality in other high-risk populations.32 Perhaps more careful donor selection may result in better long-term survival. Additional data regarding specific diagnoses and risk factors for perioperative mortality are required to improve the allocation of organs to these patients.
Although this study suffers from limitations inherent in a retrospective study of a large multi-institutional database, most notably missing data and lack of congenital diagnostic information, it supports several conclusions. First, transplantation for adults with CHD has increased over the past 15 years. Second, mechanical ventricular support is used less frequently than in the non-CHD population and does not confer the same survival advantage for those on the waiting list. Pretransplantation management and listing strategies should take this into account, and further work should be directed at improving the options for MCS in CHD patients. Third, transplantation of highly complex adult CHD patients should be undertaken only with the recognition of the potential for high early mortality, even at centers with a large transplantation experience. Fourth, although early mortality among transplant recipients with CHD is high, late survival is better than that among non-CHD patients. Fifth, risk factors for poor outcomes both on the waiting list and after transplantation differ between CHD and non-CHD patients; analyzing them together may result in confounding and a significant loss of information. Finally, and most importantly, the population of adults with CHD continues to form a “lost” population. Accurate assessment of their outcomes necessitates data sets with accurate coding of congenital diagnoses; these are not yet available on a multi-institutional scale. Until that happens, small sample sizes in single-institution studies may be our only guide to this increasingly important population.
Sources of Funding
This work was funded in part by the Health Resources and Services Administration contract 231–00-0115 and departmental funding sources. The content of this article is the responsibility of the authors alone and does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.110.960260/DC1.
- Received April 13, 2010.
- Accepted December 6, 2010.
- © 2011 American Heart Association, Inc.
- Marelli AJ,
- Mackie AS,
- Ionescu-Ittu R,
- Rahme E,
- Pilote L
- Moller JH,
- Taubert KA,
- Allen HD,
- Clark EB,
- Lauer RM
- Gatzoulis MA,
- Hechter S,
- Siu SC,
- Webb GD
- Carrel T,
- Neth J,
- Pasic M,
- Laske A,
- Jenni R,
- Maggiorini M,
- Turina M
- Hasan A,
- Au J,
- Hamilton JR,
- Hunter S,
- Hilton CJ,
- Dark JH
- McGiffin DC,
- Naftel DC,
- Kirklin JK,
- Morrow WR,
- Towbin J,
- Shaddy R,
- Alejos J,
- Rossi A
- Russo MJ,
- Davies RR,
- Sorabella RA,
- Martens TP,
- George I,
- Cheema FH,
- Mital S,
- Mosca RS,
- Chen JM
- Khairy P,
- Fernandes SM,
- Mayer JE,
- Triedman JK,
- Walsh EP,
- Lock JE,
- Landzberg MJ
- Petko M,
- Myung RJ,
- Wernovsky G,
- Cohen MI,
- Rychik J,
- Nicolson SC,
- Gaynor JW,
- Spray TL
- Davies RR,
- Russo MJ,
- Hong KN,
- Mital S,
- Mosca RS,
- Quaegebeur JM,
- Chen JM
- Hosseinpour AR,
- Cullen S,
- Tsang VT
An increasing number of patients with congenital heart disease (CHD) are reaching adulthood. These patients, often with systemic right ventricles or long-standing valvular disease, are at increased risk for the development of heart failure and may require heart transplantation. These adult patients with CHD may need to be managed differently from non-CHD patients, but few studies have specifically examined their outcomes both on the waiting list and after transplantation. A better understanding of the differences between CHD and non-CHD patients may enable improvements in the outcomes of this increasingly important population. In this study, we demonstrate that patients with CHD are different from those without CHD: They are younger and have fewer comorbidities but have higher early posttransplantation mortality. Mechanical circulatory assistance is not associated with improved survival among CHD patients on the waiting list. The increasingly common transplantation of patients with complex CHD may result in particularly high posttransplantation mortality, and centers performing these transplantations should proceed with caution. Collection of data specific to the CHD population, including accurate congenital diagnoses, is essential to better understanding and improving the outcomes with transplantation in this population.