Author + information
- Received February 25, 2016
- Revision received May 22, 2016
- Accepted May 24, 2016
- Published online August 30, 2016.
- Laith I. Alshawabkeh, MD, MSca,b,∗ (, )
- Nan Hu, MSc,
- Knute D. Carter, PhDc,
- Alexander R. Opotowsky, MD, MPHa,
- KellyAnn Light-McGroary, MDb,
- Joseph E. Cavanaugh, PhDc and
- Heather L. Bartlett, MDd
- aBoston Adult Congenital Heart Program (BACH) at the Department of Cardiology, Boston Children's Hospital, and Department of Medicine, Division of Cardiology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts
- bDivision of Cardiovascular Medicine, Department of Internal Medicine, Carver College of Medicine, Iowa City, Iowa
- cDepartment of Biostatistics, College of Public Health, University of Iowa, Iowa City, Iowa
- dDepartment of Pediatrics and Medicine, School of Medicine and Public Health, University of Wisconsin, Madison, Wisconsin
- ↵∗Reprint requests and correspondence:
Dr. Laith I. Alshawabkeh, Boston Adult Congenital Heart Program (BACH) at the Department of Cardiology, Boston Children’s Hospital, and the Department of Medicine, Division of Cardiology, Brigham and Women’s Hospital, Harvard Medical School, 300 Longwood Avenue, BCH 3215, Boston, Massachusetts 02115.
Background Heart failure represents a common end-stage syndrome for many adults with congenital heart disease (ACHD). These patients, however, have been excluded from most heart transplantation research. It is not known how current criteria, derived from non-ACHD populations, used to determine priority at the time of transplant listing, impact the outcomes for ACHD patients listed for heart transplantation.
Objectives The goal of this study was to investigate outcomes of ACHD in comparison to non-ACHD patients while listed for heart transplantation.
Methods We conducted a retrospective study using the Scientific Registry of Transplant Recipients on patients ≥18 years of age listed in the United States between 1999 and 2014. The probability of mortality or delisting due to clinical worsening was estimated using cumulative incidence functions, where transplantation was a competing event.
Results Among 1,290 ACHD and 38,557 non-ACHD patients listed, 237 ACHD and 6,377 non-ACHD patients died or were delisted due to clinical worsening. Death or delisting for clinical worsening was more likely for ACHD patients initially listed as status 1A (24% ACHD vs. 17% non-ACHD after 180 days; p < 0.001). There were no significant differences between ACHD and non-ACHD patients listed as status 1B or 2. In multivariable analysis, factors associated with death or delisting due to clinical worsening within 1 year in ACHD included: estimated glomerular filtration rate <60 ml/min/1.73 m2 (hazard ratio [HR]: 1.4; 95% confidence interval [CI]: 1.0 to 1.9; p = 0.043); albumin <3.2 g/dl (HR: 2.0; 95% CI: 1.3 to 2.9; p <0.001); and hospitalization at the time of listing, whether in the intensive care unit (HR: 2.3; 95% CI: 1.6 to 3.5; p < 0.001) or not (HR: 1.9; 95% CI: 1.2 to 3.0; p = 0.006) relative to outpatients.
Conclusions Wait-list mortality or delisting due to worsening clinical status is disproportionately common for ACHD patients listed as status 1A. An allocation system that takes into account the distinctive aspects of ACHD patients may help better care for this growing population.
Advances in surgical and medical care have led to a dramatic increase in survivorship of children born with congenital heart disease (CHD) (1,2). As a consequence, most patients living with CHD are now adults (3). This growing population of adults with congenital heart disease (ACHD) is faced with an increasing burden of chronic diseases, including heart failure (HF) (4). The magnitude of this issue is reflected in the almost doubling of hospitalizations for ACHD with HF between 1998 and 2005 (5). As mortality in survivors of CHD shifts into adulthood, HF is emerging as a leading cause of death (6).
The presentation of HF in ACHD patients is diverse and often atypical, owing to the heterogeneity of the underlying cardiac defects and subsequent surgical and transcatheter interventions. Consistent definition, identification, and prognostic evaluation of HF in this population is challenging (4,7). Furthermore, landmark clinical trials in HF have excluded patients with CHD, resulting in uncertainty in evaluating mortality benefit from medical or device therapies (4,8,9). ACHD patients who undergo transplantation experience higher mortality in the immediate postoperative period, but those who remain alive enjoy significantly better long-term survival compared with non-ACHD patients (10–12).
Because the criteria for determining priority status for listing for heart transplantation are derived from evidence in the non-ACHD population, these criteria could misclassify prognosis for the ACHD population and may therefore not reflect the relative benefit of transplant versus alternative therapies. Previous studies suggest similar overall wait-list mortality for ACHD and non-ACHD patients listed for heart transplantation despite a higher likelihood of transplantation in the non-ACHD population (13,14). Untested clinical factors potentially germane to this question may account for the incongruity, such as divergent initial listing status and incomplete characterization of those delisted due to worsening clinical status precluding transplant. The 1-year probability of survival after delisting due to worsening is 26% for all patients; those with CHD had a 40% greater likelihood of death after delisting compared with non-ACHD patients, which translates into approximately 16% survival at 1 year after delisting (15).
Using a national database, we investigated outcomes of ACHD compared with non-ACHD patients while listed for heart transplantation, as well as correlates of mortality or delisting due to worsening of clinical status in ACHD patients.
The Scientific Registry for Transplant Recipients is a multicenter cohort database that continuously receives data from the Organ Procurement and Transplantation Network. This network gathers information on all patients listed for solid organ transplantation in the United States via reports from the transplanting centers (16). The present study included 1,649 ACHD and 60,667 non-ACHD patients ≥18 years of age listed for first orthotopic heart transplantation between April 1, 1986, and June 2, 2014. Due to observed significant improvement in survival over time, the analysis was restricted to the current era, defined as initial listing after January 19, 1999, on the basis of when a 3-tiered (1A, 1B, and 2) priority listing status was introduced. Patients listed for multiple organs or for repeat transplantation were excluded from this analysis.
The Institutional Review Board of the University of Iowa approved the study. The data reported were supplied by the Minneapolis Medical Research Foundation as the contractor for the Scientific Registry for Transplant Recipients.
Clinical endpoints and variable definitions
Our primary hypothesis was that ACHD patients listed for heart transplantation experience higher mortality or delisting due to worsening of clinical status than patients without CHD. In addition, we aimed to assess predictors of death or worsening clinical status for ACHD patients while listed for heart transplantation. Patients were followed up from the time of listing until 1 of the following outcomes occurred: death, transplant, delisting due to worsening, or delisting due to improvement. Patients were considered to experience the primary outcome if they died or were delisted due to worsening of clinical status. Listing status was defined as the initial listing status, irrespective of status changes during the observation period.
Race/ethnicity categories were reported by the transplanting center as black or African American, white, multiracial, American Indian or Alaska Native, Asian, Hispanic/Latino, Native Hawaiian, or other Pacific Islander. For this study, race was dichotomized as white versus nonwhite. Estimated glomerular filtration rate (eGFR) was calculated by using the Modification of Diet in Renal Disease formula (17). Activities of daily living included walking around the home, getting out of bed or a chair, eating, dressing, bathing or showering, or using the toilet. Mechanical circulatory support (MCS) included an intra-aortic balloon pump, ventricular assist device, total artificial heart, and extracorporeal membrane oxygenation. Pulmonary vascular resistance (in Wood units) was calculated as: mean pulmonary artery pressure − pulmonary capillary wedge pressure/cardiac output.
Summary statistics were reported for quantitative variables as medians (interquartile ranges) and for qualitative variables as percentages; quantitative variables were additionally presented as categorical when appropriate. Variables that were missing in >70% of the sample, including bilirubin levels, were excluded. The probability of an event was estimated by using cumulative incidence functions, in which transplantation was regarded as a competing event to the outcomes of interest (death or delisting due to worsening). Univariate relationships between the tested variables and the primary outcome were assessed with Gray’s test for 1 year of follow-up, accounting for transplantation as a competing event. Variables that were significant at p < 0.10 in either of the 2 groups were included in multivariable analysis for both groups based on Cox proportional hazards regression censored at 1 year of follow-up; some group comparisons (outcomes stratified according to listing status) were restricted to outcomes occurring within 180 days because few ACHD patients who were listed at status 1A survived longer. The multivariable analyses used competing risk models as described by Fine and Gray (18).
After examining the group of patients with missing information, the following variables were determined to have informative missing values: hypertension, coronary artery disease, previous cardiac surgery, and peak oxygen consumption. Missing values for these variables were coded by using a categorical indicator, and the patients were retained in the full multivariable model but were excluded from subsequent models that involved multiple imputation. Logistic imputation was used for the binary variables diabetes mellitus, eGFR <60 ml/min/1.73 m2, difficulty with activities of daily living, antiarrhythmic medication use, presence of a defibrillator, pulmonary capillary wedge pressure >18 mm Hg, and pulmonary vascular resistance >4 Wood units; regression imputation was used for albumin and body mass index, which were then categorized; and discriminant function imputation was used for the medical condition variable (3 levels: in intensive care unit [ICU], in the hospital but not in the ICU, and ambulatory/outpatient) as well as for MCS. Ten imputation datasets were generated, and the results obtained from each dataset were combined by using the multiple imputation procedures provided within the SAS statistical system (19). No substantial collinearity was detected between the tested variables. The final model was further assessed after conditioning on status 1B and 2 patients only.
Analyses were performed by using SAS version 9.4 (SAS Institute, Inc., Cary, North Carolina) and R version 3.2.0. (R Foundation for Statistical Computing, Vienna, Austria).
There were 1,290 ACHD and 38,557 non-ACHD patients listed after January 19, 1999 (Table 1). The median age was 34.5 years (interquartile range: 25.1 to 44.7 years) in ACHD and 55.6 years (interquartile range: 46.8 to 61.8 years) in non-ACHD patients. Most patients were white (82.9% and 70.4%, ACHD and non-ACHD, respectively) and male (61.5% and 75.9%, ACHD and non-ACHD). Most ACHD patients had undergone previous cardiac surgery (87.3% compared with 38.6% for non-ACHD patients). Compared with other candidates, ACHD patients had a low prevalence of coronary artery disease (5.5% vs. 27.2%). Only 5.8% of ACHD patients had a ventricular assist device at the time of listing, compared with 21.3% for non-ACHD.
ACHD patients were less likely to be initially listed at the highest priority status of 1A (11.8% vs. 21.7% for non-ACHD) and more likely to be initially listed at the lowest priority status 2 (59.4% vs. 40.9% for non-ACHD; p < 0.0001).
Death or delisting due to worsening
The distribution of outcomes for 1,257 ACHD and 37,248 non-ACHD patients is shown in Figure 1, stratified according to listing status. As of June 2, 2014, a total of 176 ACHD and 3,982 non-ACHD patients remained active on the transplant list.
Because few ACHD patients listed as status 1A remained listed beyond 180 days, follow-up was restricted to this time frame for both groups; results of cumulative incidence functions are presented in Figure 2. The probability of the primary outcome at 180 days for patients initially listed as status 1A was 23.9% for ACHD patients compared with 16.7% for non-ACHD patients (p = 0.011). For the ACHD and non-ACHD populations, respectively, 180-day probability of the primary outcome was 12.1% versus 9.5% for those listed as status 1B (p = 0.118) and 6.8% versus 5.9% for those listed as status 2 (p = 0.385). The 180-day probability of death alone in the status 1A patients was 14.6% for ACHD versus 11.9% for non-ACHD (p = 0.274). Because of differences in distribution of the initial listing status, the overall incidence of the primary outcome was similar between the ACHD and non-ACHD groups (13.2% vs. 12.1%; p = 0.293).
Most ACHD and non-ACHD patients on the wait-list underwent transplantation. However, transplantation was less frequent for listed ACHD than for non-ACHD patients, irrespective of the listing status (62.2% and 67.4% respectively; p < 0.0001) (Figure 1).
Univariate factors associated with death or delisting due to worsening within 1 year are shown in Table 2. Some predictors of adverse outcomes in non-ACHD patients were not associated with these outcomes in ACHD, including: age, body mass index, coronary artery disease, use of antiarrhythmic medication, having a defibrillator, pulmonary capillary wedge pressure >18 mm Hg, and blood type. Conversely, sex was associated with the primary outcome in the ACHD population but not in the non-ACHD population.
ACHD versus non-ACHD
All the predictors in the univariate analysis were included in the multivariable model, except for race and cardiac index (which were not significant in either group in the univariate analysis). The multivariable factors associated with the primary outcome are shown in Figure 3 and Online Table 1. No significant interaction was found between ACHD patients versus non-ACHD patients.
Hypertension, coronary artery disease, previous cardiac surgery, and peak oxygen consumption seemed to have informative missing values in at least 1 of the groups (i.e., patients who had missing data for these variables were more or less likely to experience the primary outcome compared with those who had valid data). Imputation was not performed on these missing variables, and they were not incorporated in subsequent models.
Table 3 summarizes the multivariable predictors of death or delisting due to worsening. Among the overall group of ACHD patients listed at any status, independent predictors of 1-year death or delisting due to worsening included eGFR <60 ml/min/1.73 m2, albumin level <3.2 g/dl, mechanical ventilation, being in the ICU, and hospitalized but not in the ICU.
Because it would be clinically useful to assess the factors associated with death and delisting due to worsening for patients initially assigned a lower priority status (which would have been based on variables collected at the time of listing and clinical judgment), we fit the model for patients initially listed at status 1B and status 2. The predictors were not notably different, except that mechanical ventilation was no longer a relevant predictor. Multivariable predictors indicate that some patients with high-risk features for adverse outcomes at the time of listing were initially listed at a low priority status.
The present study of patients listed for heart transplantation in the United States identified several important findings that add to the results of previous studies. ACHD patients who were listed as status 1A were more likely to die or became too sick to undergo transplantation compared with non-ACHD patients. The clinical variables associated with mortality or worsening were generally similar between ACHD and non-ACHD patients, with several notable exceptions, including previous cardiac surgery; in contrast to congenital patients, non-ACHD patients potentially fared worse if they had undergone previous cardiac surgery. In addition, new predictors of death or worsening for ACHD patients were identified.
Substantial improvements in anatomic palliation and acute and chronic medical care for patients with CHD now allow the majority to survive to adulthood; HF is a common late consequence (20). Although progress in the management of HF in adults who do not have CHD has led to improved survival and quality of life, a lack of information remains about the presentation, effective therapies, and predictors of outcomes of HF in ACHD. Consequently, our understanding of optimal care for these patients is limited to reasoning from basic principles and extrapolation from evidence in other populations. As a result, HF is the most common cause of death in adults with CHD (Central Illustration) (6).
Most ACHD patients are initially listed at the lowest priority status, in stark contrast to adults listed for heart transplantation who have HF related to other causes. Recent data suggest that 67% of transplants performed in the United States are for patients listed as status 1A, whereas status 2 accounts for only 5% (21). Although the initial status listing is subject to adjustment, there are a number of reasons why an initial listing at lower status puts ACHD patients at a disadvantage. For example, cumulative time listed at a certain status is a key determinant of priority within that status; because more ACHD patients are initially listed as status 2, the majority will have to be upgraded before accruing time at a higher status with the attendant increase in priority. Non-ACHD patients have the added advantage of routine candidacy for left ventricular assist devices that are known to extend their life; non-ACHD patients may additionally benefit from the use of a Swan-Ganz catheter. The presence of such interventions often constitutes a rationale for listing at higher priority status. Such interventions are less frequently used for ACHD patients for a variety of reasons; for example, patients with Fontan circulation are generally not candidates for current mechanical support options and are usually unlikely to benefit clinically from hemodynamic monitoring or inotrope support. Non-ACHD patients are thus able to accrue more time at a higher priority status on the transplantation list (22). The disparity in initial listing status and outcomes for patients listed at higher priority between the ACHD population and the non-ACHD population is troubling, and it highlights the need for additional evidence to appropriately classify risk and benefit when listing ACHD patients for heart transplantation.
ACHD patients who were initially listed as status 1A had the highest probability of mortality or delisting due to worsening of all the patient groups studied, with substantially worse outcomes than non-ACHD patients listed at equivalent priority. This finding seems at first glance to contrast with previous reports, which indicated similar wait-list mortality for ACHD and non-ACHD groups (13,14). Our observation is actually consistent with those previous findings for ACHD versus non-ACHD groups listed for heart transplantation overall without consideration of disparities in initial listing status. Our study highlights differences in mortality or delisting for worsening specifically in patients initially listed as status 1A. A larger proportion of ACHD patients are listed as status 2, and this finding obscures the difference in outcomes between the groups of equivalent priority. Although this difference in outcomes could be related to factors associated with lower probability of undergoing transplantation, late referral for heart transplantation in ACHD patients likely contributes. This scenario is supported by the observation that the probability of death or delisting due to worsening in ACHD is markedly higher than for non-ACHD patients in the immediate period after initial listing as status 1A. This finding may be, in part, related to larger issues in care delivery for this population; more than one-quarter of adults with severe forms of CHD who are at high risk of developing complications report important gaps in their care (23). Lapses in care for this population-in-flux probably contribute to presentation in the late stages of HF, and even death before presentation (24,25). Improved awareness of the clinical needs for this population and studies of early markers for HF and disease progression are urgently needed (26).
We found that renal disease (i.e., eGFR <60 ml/min/1.73 m2) was independently associated with the probability of death or delisting due to worsening in ACHD patients listed for heart transplantation; interestingly, given how robustly renal function predicts outcomes in diverse populations, a previous study did not find eGFR to be an independent predictor of outcomes (13). This finding may simply be a matter of inadequate power; the previous study used a more stringent cutoff (creatinine clearance <40 ml/min); it also included records from before 1999, which suffered from more missing data. Because of these issues, only 28 patients with CHD were classified as having low eGFR in that study, compared with 506 patients with eGFR <60 ml/min/1.73 m2 in the present analysis. CHD can also be associated with a chronic hepatic insult in many patients, which may lead to a higher probability of poor outcomes if advanced HF ensues (27). This result should encourage attention to both renal and hepatic function before deciding to list the patients for transplantation, and it corroborates expert opinion in managing these challenging patients by listing them earlier in their disease course (28). Furthermore, when considering the clinical data to support requests for exception to achieve a higher listing status for ACHD patients, recurrent or persistent hepatic and renal dysfunction should be considered. Conversely, more data are required to understand how to predict the type and degree of multiorgan dysfunction that best portends unacceptable post-transplant outcomes; overabundant caution would also be detrimental to ACHD patients.
We observed no differences in the outcomes for ACHD patients who had MCS at the time of listing. Because of anatomic difficulties and differences in expertise with MCS implantation, ACHD patients do not undergo MCS support as frequently as non-ACHD patients (22). We hypothesize that careful timing, thoughtful device selection, and anatomic adaptation in MCS implantation in ACHD patients might improve the survival benefit. However, interpretation is limited because MCS implantation after listing was not examined.
The main strength of the present study is its comprehensive inclusion of all patients listed for heart transplantation in the United States with complete ascertainment of vital status. A secondary strength is that missing data were carefully examined and addressed. When appropriate, multiple imputation was used to accommodate missing values so that the patient records could be retained in the analyses. However, we recognize several limitations. First, the underlying CHD diagnosis was not collected in this dataset and, as such, the heterogeneity in the ACHD population could not be accounted for. Regrettably, there is no dataset currently available that collects this information at a national level in the United States. Second, there was a proportion of missing data for certain variables, which could introduce bias. A careful examination of the vital status of patients with missing information was performed. Nonetheless, we had to exclude a number of variables that could be important predictors. Third, delisting due to worsening was included in the primary outcome despite its subjective nature and unknown eventual outcome. Patients delisted due to worsening are at very high risk of death, and this claim is especially true for ACHD patients (15). Finally, the current database cannot provide insight on the decision-making process regarding candidacy for transplant. A better understanding of current impediments to transplantation listing in this group of patients is critically needed. Further data are vital to determine how best to calibrate our thinking on which subjects constitute suitable ACHD transplant candidates to improve a process that currently, of necessity, proceeds without the benefit of evidence.
When listed for heart transplantation, ACHD patients are more frequently listed as lowest priority. ACHD patients listed at the highest priority status are more likely to die or be delisted due to clinical worsening, suggesting they are listed too late in the disease process. Renal dysfunction and liver disease portend worse outcomes in ACHD. These findings highlight a gap in understanding of prognostic evaluation for ACHD patients who experience HF; thus, epidemiological studies are needed to define patterns of HF in ACHD patients. These findings should be considered within the newly proposed Organ Procurement and Transplantation Network heart organ allocation algorithm (29).
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: ACHD patients listed for heart transplantation are at risk of adverse outcomes, particularly when certain comorbidities such as renal dysfunction are present.
TRANSLATIONAL OUTLOOK: Larger observational studies of patients with CHD and heart failure may identify additional markers associated with adverse outcomes and prompt reconsideration of priorities for organ allocation.
The authors thank Michael Landzberg, MD, for providing critical revisions of the manuscript.
For a supplemental table, please see the online version of this article.
This research was made possible with support from the Research For The Kids Foundation. Interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy of or interpretation by the Scientific Registry for Transplant Recipients or the U.S. government. Dr. Light-McGroary has served as a site principal investigator for a clinical trial by Bayer Pharmaceuticals. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- adult congenital heart disease
- congenital heart disease
- estimated glomerular filtration rate
- heart failure
- intensive care unit
- mechanical circulatory support
- Received February 25, 2016.
- Revision received May 22, 2016.
- Accepted May 24, 2016.
- American College of Cardiology Foundation
- Marelli A.J.,
- Mackie A.S.,
- Ionescu-Ittu R.,
- Rahme E.,
- Pilote L.
- Khairy P.,
- Ionescu-Ittu R.,
- Mackie A.S.,
- Abrahamowicz M.,
- Pilote L.,
- Marelli A.J.
- Marelli A.J.,
- Ionescu-Ittu R.,
- Mackie A.S.,
- Guo L.,
- Dendukuri N.,
- Kaouache M.
- Stout K.K.,
- Broberg C.S.,
- Book W.M.,
- et al.
- Opotowsky A.R.,
- Siddiqi O.K.,
- Webb G.D.
- Verheugt C.L.,
- Uiterwaal C.S.,
- van der Velde E.T.,
- et al.
- McLarry J.,
- Broberg C.,
- Opotowsky A.R.,
- Burchill L.J.
- Warnes C.A.,
- Williams R.G.,
- Bashore T.M.,
- et al.
- Ross H.J.,
- Law Y.,
- Book W.M.,
- et al.
- Davies R.R.,
- Russo M.J.,
- Yang J.,
- Quaegebeur J.M.,
- Mosca R.S.,
- Chen J.M.
- Opotowsky A.R.,
- Landzberg M.J.
- Gurvitz M.,
- Valente A.M.,
- Broberg C.,
- et al.
- Mylotte D.,
- Pilote L.,
- Ionescu-Ittu R.,
- et al.