Outcomes of heart transplantation in children with previously palliated hypoplastic left heart syndrome

Abstract

OBJECTIVES

Paediatric heart transplantation in children who fail multistage palliation for hypoplastic left heart syndrome is associated with challenges related to immune, clinical or anatomic risk factors. We review current outcomes and risk factors for survival following heart transplantation in this challenging patient population.

METHODS

The United Network for Organ Sharing transplantation database was merged with Paediatric Health Information System database to identify children who received heart transplantation following prior palliation for hypoplastic left heart syndrome. Multivariable Cox analysis of outcomes and factors affecting survival was performed.

RESULTS

Our cohort included 849 children between 2009 and 2021. The median age was 1044 days (interquartile range 108–3535), and the median weight was 13 kg (interquartile range 7–26). Overall survival at 10 years following heart transplantation was 71%, with most of the death being perioperative. On multivariable analysis, risk factors for survival included Black race (hazard ratio = 1.630, P = 0.0253), blood type other than B (hazard ratio = 2.564, P = 0.0052) and male donor gender (hazard ratio = 1.367, P = 0.0483). Recipient age, the use of ventricular assist device or extracorporeal membrane oxygenation were not significantly associated with survival. Twenty-four patients underwent retransplantation, and 10-year freedom from retransplantation was 98%. Rejection before hospital discharge and within 1 year from transplantation was 20% and 24%, respectively, with infants having lower rejection rates.

CONCLUSIONS

Compared with existing literature, the number of children with prior hypoplastic left heart syndrome palliation who receive heart transplantation has increased in the current era. Survival following transplantation in this patient population is acceptable. Most of the death is perioperative. Efforts to properly support these patients before transplantation might decrease early mortality and improve overall survival.

Open in new tabDownload slide

INTRODUCTION

Children born with single ventricle physiology, such as hypoplastic left heart syndrome (HLHS), are increasingly surviving the multistage palliation strategy towards Fontan completion [1]. This can be attributed to a multitude of factors, such as continued advancements in surgical strategies, perioperative care and outpatient management. Despite increasing survival, a number of palliated HLHS patients fail this multistage palliation strategy with 15–20% requiring listing for paediatric heart transplant (HT), up to 30% of which are after Fontan completion [1–3]. As this group of patients is rapidly growing, and perhaps the most complex cohort presenting for HT, it is important to identify risk factors that contribute to survival [2–7].

Reported outcomes of HT in palliated single ventricle children in general and those with HLHS in particular varied in the literature. While all studies showed that congenital heart disease was associated with lower survival compared with dilated cardiomyopathy, some studies showed that single ventricle and HLHS were associated with lower survival among patients with congenital heart disease, while others showed that survival in this group was equivalent to those with other forms of congenital heart disease [6, 8, 9]. The HLHS patient listed for HT poses several challenges due to prior surgeries, immunologic exposure and a variety of comorbidities [1]. The complexity of surgical intervention (redo surgery, collateral load, need for pulmonary artery and aortic reconstruction and longer graft ischaemic times), higher incidence of sensitization and poor clinical condition (protein losing enteropathy, liver and renal dysfunction, malnutrition and deconditioning) all complicate the care of the child failing HLHS palliation prior and during HT [5].

Most prior studies have been limited by single institution cohort size that restricted their power to assess risk factors associated with survival. In the current study, we sought to identify and analyze a large cohort of previously palliated HLHS children who underwent HT by merging the Paediatric Health Information System (PHIS) and the United Network for Organ Sharing (UNOS) databases. We examined factors affecting major morbidities and mortality following HT on outcomes.

METHODS

Statement of Institutional Review Board and ethics board approval

Institutional Review Board Approval was obtained: IRB 14.0343 on 8/17/2022. Due to the use of de-identified data, this study did not qualify as human studies research and therefore was determined to be not regulated as human subjects research by the University of Louisville Institutional Review Board and consent as waived. All ethical standards were followed in compliance with ISHLT.

Patients and methods

Study design

This is a retrospective analysis of multicentre data from a merged cohort of patients between 2 large databases, PHIS and UNOS. UNOS is a private, non-profit organization that manages the nation’s organ transplant system. Information collected includes detailed pre- and post-transplant demographic and clinical information for donor and recipient. PHIS is an administrative database containing inpatient data from 43 paediatric hospitals nationwide. Information is assigned a de-identified medical record number and can be tracked across multiple hospitalizations. Specifically, PHIS contains diagnosis and procedure codes by International Classification of Diseases (ICD) codes, and billed resource utilization data. While all pertinent characteristics of donors and recipients, and all clinical follow-up was obtained from the UNOS database, the purpose of the merger with the PHIS administrative database is to identify children with previously palliated HLHS, and information that is not available in the UNOS database. The use of administrative databases is naturally complicated by many disadvantages and biases. However, limitation of the utility of the PHIS database to the proper identification of HLHS patients likely minimizes these disadvantages.

Patient cohort

The initial patient cohort was identified through the PHIS database and included all patients under 18 years of age with an ICD-9 and ICD-10 codes for HT between 2009 and 2021 as well as a diagnosis of HLHS using the ICD-9 and ICD-10 revision codes. This cohort was then matched with the patients available in the UNOS database by transplant centre and transplant date. A secondary match using age and gender was performed to obtain the final study cohort (Fig. 1). Our initial goal was to divide our patient cohort based on the HLHS palliation stage (post-Norwood, post-Glenn and post-Fontan). Nonetheless, we appreciated that the PHIS database, while effective in identifying HLHS patients, was not uniformly effective in identifying the complete prior palliation history. As a result, last known palliation stage was not entered as a variable in our study.

Variables

Patient demographic data, pretransplant data, pretransplant support and outcomes were analyzed. Demographic data included age, gender, weight, body mass index and race. Pre-transplant data included blood group, mean pulmonary artery pressure, mechanical ventilation, creatinine, dialysis requirement, bilirubin, UNOS listing status, panel reactive antibodies, ABO incompatibility and transplant list waiting duration. Pre-transplant support included mechanical ventilation, extracorporeal membrane oxygenation (ECMO), ventricular assist device (VAD) and inotrope use. Donor characteristics, such as age, gender, body mass index, weight, donor/recipient weight ratio and organ ischaemic time, were gathered as well. Our primary outcomes included post-transplant survival, with freedom from retransplantation and rate of rejection being secondary outcomes. Risk factors for survival were analyzed. The follow-up information on survival is clearly recorded in the database. After the merging process, there were <1% patients with any missing data, which was of the random occurrence. We excluded these patients from the final analysis. The cohort is representative of patients with all the available data (Fig. 1).

Statistical analysis

Non-parametric univariate statistical methods were initially used to evaluate baseline characteristics. Continuous variables were analyzed using Kruskal–Wallis test, and the categorical variables were analyzed using the Chi-squared test or Fisher’s exact test. Kaplan–Meier survival curves were computed, and log-rank test was used to evaluate differences in survival. The Kaplan–Meier curves truncated at 11 years, and patients were censored at that point. Of note, we used the cumulative hazard function to generate the comparative outcomes curves for post-transplant mortality and retransplantation where patients who died or received retransplantation were treated as events and patients still alive were censored. However, we did not perform any competing risk analysis because of the low number of re-transplantations. A multivariable Cox regression model was generated for the study to identify factors associated with post-transplant survival. The variables included in the multivariable model are listed in Supplementary Material, Table S1. All potentially significant clinical, surgical and demographic variables were entered in the multivariable model. One exception was dialysis that was excluded, as it had significant collinearity with creatinine. All the data are presented as median with interquartile range (IQR) or N (%). The statistical analysis was done using SAS 9.4 software (SAS Inc., Cary, NC) at 95% confidence level.

RESULTS

Patient characteristics

Eight hundred forty-nine patients were matched between the PHIS and UNOS databases. Our cohort included 298 infants <1 year of age. The median age was 1044 days (IQR 108–3535), and the median weight was 13 kg (IQR 7–26). There were 535 males (63%). The race of patients was White (n = 569, 67%), Black (n = 102, 12%) and other (n = 178, 21%). Listing status was 1A in 82% of patients, 1B in 14% and 2 in 2%. Infants were more likely to be listed 1A (98%). Blood type was A in 40%, B in 7%, AB in 12% and O in 41%. ABO-incompatible donors were used in 92 patients, including 80 infants (27% of all infants) and 12 between 1 and 2 years of age. Median panel reactive antibodies were 0 (IQR 0–39).

Median creatinine at time of listing was 0.4 mg/dl (IQR 0.3–0.5), and 2% of children were receiving dialysis at time of transplantation. Median bilirubin was 0.6 mg/dl (IQR 0.4–1.1).

The support at the time of transplantation included mechanical ventilation (n = 144, 17%, including 37% of infants), inotropes (n = 467, 55%), ECMO (n = 34, 4%) and VAD (n = 85, 10%).

Median donor age was 3 years (IQR 1–10), and median weight was 16 kg (IQR 10–38) with a median donor/recipient weight ratio of 1.2 (IQR 1.0–1.6). The gender of the donor was male in 57% of cases. Median ischaemic time was 3.9 hours (IQR 3.3–4.6). The median waitlist duration was 70 days (IQR 28–148; Table 1).

Outcomes and risk factors after heart transplant in hypoplastic left heart syndrome

Overall survival at 10 years following HT was 71% (Fig. 2). Risk factors associated with survival following HT were sought on multivariable Cox regression analysis (Supplementary Material, Table S1). On multivariable analysis, risk factors for survival included Black race [hazard ratio (HR) = 1.630, P = 0.0253; Fig. 3a], blood type other than B (HR = 2.564, P = 0.0052; Fig. 3b) and male donor gender (HR = 1.367, P = 0.0483).

On univariate Kaplan–Meier analysis, while 1-year survival was less in infants <1 year of age (81%) than that of older children (1–5 years of age 96%, ≥6 years of age 91%), 10-year survival was similar (Fig. 4a). Moreover, on multivariable analysis, recipient’s age was not found to be a significant factor for survival (HR = 0.946, P = 0.2562; Fig. 4b). Also, ECMO use was not found to be a significant factor for survival (HR = 1.175, P = 0.6653; Supplementary Material, Fig. S1a). Interestingly, evaluation of survival following HT on Kaplan-Meir showed that there were some discrepancies in survival following HT when ECMO was utilized based on patient’s age, with the difference in survival being more pronounced in patients 1–5 years of age requiring ECMO (Supplementary Material, Fig. S1b).

During follow-up, 24 patients received retransplantation with 10-year freedom from retransplantation of 98% (Fig. 5). Five years following HT, 2% of patients had received retransplantation, 28% had died without retransplantation, while 70% were alive and free from retransplantation.

Rejection at discharge for the entire cohort was 20% and was lower in infants compared with older patients (12% vs 24%, P < 0.01). Rejection within the first-year post-transplant for the entire cohort was 24% and was also lower in infants compared with older patients (16% vs 29%, P < 0.01), although the difference seemed mainly due to the lower rejection rate before discharge.

DISCUSSION

This merge between the UNOS and PHIS databases have provided us with a large cohort of children with HLHS who received HT at various stages in the multi-stage surgical palliation strategy. Our data show increasing number of children receiving HT after failed palliation for HLHS during our study period [2, 10]. In this increasing option for failed palliation, 1-year survival was less in infants compared with older patients, although the 10-year survival was similar, suggesting that age and potentially palliation stage might affect early mortality but not overall survival. This finding of higher early mortality might be related to the poor clinical status of the patient and difficulty in providing mechanical circulatory support that would allow better rehabilitation of their body and organs before transplantation, with potentially better early survival. Efforts to address this issue, including innovations in the support of these infants, might prove to be beneficial.

With improvements in surgical technique and postoperative management, most patients with HLHS can undergo staged palliation, with HT reserved for the highest risk patients [11, 12]. Thus, it is unsurprising that historically out of all CHD patients, post-transplant survival has been inferior in the palliated HLHS patient [2, 7, 13–20]. Despite this, studies focusing on transplant outcomes and risk stratification between palliated HLHS, and other CHD patients are limited. A recent encouraging study compared survival after HT in palliated HLHS vs non-single ventricle CHD between 2016 and 2021 and found equivalent post-transplant outcomes at 1st and 5th year (87% vs 88%, 76% vs 79%) [13]. The present study adds to this limited body of evidence by further examining HT survival in this challenging group of patients. Our data similarly indicate early, and late survival is improving in children with palliated HLHS undergoing HT, including infants.

In addition to the diagnosis of HLHS and history of prior cardiac surgery alone, risk factors for survival in prior reports have also included lower weight, mechanical ventilation or dialysis at time of transplant, or ECMO as a bridge to transplant [21, 22]. We did not identify inotropic or mechanical ventilatory support at transplant to significantly impact long-term survival outcomes across the 3 groups. In the aforementioned study by Greenberg et al., rates of organ dysfunction were similar between palliated HLHS and non-CHD groups before transplant, with requirement for mechanical ventilation and ECMO actually lower in the HLHS group, suggesting that till now, children with HLHS refereed to transplantation are still well selected [13].

During prior surgeries and while waiting, the HLHS population often undergoes multiple blood transfusions before transplant, leading to increased allosensitization, which may affect mortality if transplant is delayed, and ABO-incompatible heart transplantation is not an option [9]. This time could lead to worsening organ dysfunction and increased operative risk. In our study, Groups B and C had longer waitlist times and unsurprisingly limited ABO incompatible transplants as compared to Group A, possibly accounting for the relatively low 10-year survival compared with 1 year in the 2 groups.

Mortality associated with long transplant waitlist times, particularly for patients of O blood type, has improved with the advent of ABO incompatible transplant [18–20]. While a multitude of studies have proven equivalent outcomes between ABO compatible and incompatible HTs in infants, this is unfortunately not an option in older patients [20, 21]. We identified blood type (other than B) as a risk factor for survival across the entire cohort. This finding is hard to explain. Blood B patients had a slightly longer wait time, however, whether that affected outcomes remain speculative. In prior adult studies, type O has been associated with worse outcomes for all HT recipients in respect to both wait list mortality and post HT [22]. While we did not further analyze blood type as a risk factor by age group due to sample size limitations, further study may provide more insight into the significance of these findings for risk stratification.

We also identified Black race as a risk factor for survival after HT, consistent with existing literature. In the overall cohort, patients of black race had lower survival after HT compared with White race, most significant at 10 years post-HT. Paediatric HT recipients of Black race are known to be at higher risk of death and graft loss, an association thought to be mediated by a higher risk of rejection [23, 24]. Black children have been reported to be at higher risk of rejection, rejection accompanied by haemodynamic compromise and late rejection after HT, consistent with our data [25]. This risk may be due to genetic, immunologic and socioeconomic factors [23–27]. Some of these factors may include donor-recipient human leucocyte antigen mismatch, higher anti–human leucocyte antigen antibodies, rapid metabolism of calcineurin inhibitor immune suppression, proinflammatory state and variation in gene expression.

While blood type and race are more static risk factors for survival, encouragingly, ECMO use at time of transplant has become less common [13]. However, patients that do require ECMO at time of HT have inferior survival outcomes, consistent with our study findings [28, 29]. Our findings suggest the necessity of ECMO to be most common in the infant age group (Group A); however, 10-year survival was the lowest in B and C who required ECMO. Therefore, although patients in Group A were more likely to require ECMO, it did not appear to significantly change their long-term outcomes as compared to children >1 year of age (Groups B and C). In some cases, a feasible alternative to ECMO support is a VAD. Patients that have been bridged to transplant with a VAD have outcomes after HT comparable to those who did not need a VAD [30]. A study compared patients <18 years of age between 1993 and 2015 supported with either ECMO or VAD to HT [30]. Of the 5145 patients with CHD or cardiomyopathy that underwent HT, 19.3% were supported by a VAD, although <1% were neonates and overall survival at 2 and 20 years between the ECMO and VAD patients were similar [27]. Only a small percentage of our cohort was supported with a VAD after transplant with an overall similar survival outcome as compared to no VAD support. This modality has historically been used to support children with cardiomyopathy but advances in implantation techniques to accommodate the complex single ventricle make it an increasingly accessible option for the HLHS patient.

This study shows an increasing number of patients with HLHS undergoing HT in the current era consistent with predictions that this population will continue to grow in numbers of annual paediatric HTs [7, 12]. Everitt et al. identified 69 HLHS patients with prior surgery that underwent transplant between 1993 and 2008. Another study examined outcomes of HT after Norwood, Glenn and Fontan palliation identifying 118 patients who underwent HT between 1993 and 2012 [10]. Greenberg et al. identified 360 palliated HLHS that underwent transplant between 2016 and 2021 through the UNOS database [13]. Although unable to make a direct comparison between our cohort of 849 patients between 2009 and 2021 from 2 combined databases and these prior reports, the overall trend would indicate an increase in HT after failed palliation for HLHS. This underscores the high-risk nature of this population that despite advancements in surgical palliation management, many advance to early heart failure. Continued efforts to identify risk factors for survival and clinical optimization before HT are imperative.

Limitations

This study should be interpreted within the context of its limitations. As with any retrospective database study, there are limitations to the data collected from the registries. Potential sources of bias include missing data, selection bias and confounding variables (such as centre variation). The study is limited to patients who underwent HT and may not be generalizable to all HLHS patients. Due to the nature of the databases, there was a lack of listing information, inconsistency of detailed surgical approach and inconsistency of panel-reactive antibody reporting. Additionally, as we have discussed before, the PHIS dataset does not specify the stage of palliation performed before transplantation; thus, sub-categorizations depending upon palliation stage were not possible.

SUMMARY

The patient with HLHS who has undergone previous palliative surgery and ultimately fails single ventricle palliation is a growing population that presents a multitude of challenges. Ultimately, HT is a valid strategy to salvage children who fail palliation with single ventricle physiology. Patients who require HT after palliation for HLHS have acceptable survival outcomes in the current era.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

FUNDING

Institutional funding was provided for this research. There is no external source of funding to report.

Conflict of interest: none declared.

DATA AVAILABILITY

The data examined for this article are owned by a third party, UNOS and PHIS, and is accessible for participating institutions.

Author contributions

Johnna Sizemore: Conceptualization; Methodology; Writing—original draft; Writing—review & editing. Jamie Furlong-Dillard: Supervision; Writing—review & editing. Sarah Wilkens: Writing—review & editing. Deborah Kozik: Writing—review & editing. Shriprasad Deshpande: Writing—review & editing. Jaimin Trivedi: Writing—review & editing. Bahaaldin Alsoufi: Conceptualization; Methodology; Supervision; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Martin M. Kostolny, Supreet P Marathe and Joachim O.M. Photiadis for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• ECMO

  Extracorporeal membrane oxygenation

 
• HLHS

  Hypoplastic left heart syndrome

 
• HR

  Hazard ratio

 
• HT

  Heart transplant

 
• ICD

  International Classification of Diseases

 
• IQR

  Interquartile range

 
• PHIS

  Paediatric Health Information Systems

 
• UNOS

  United Network for Organ Sharing

 
• VAD

  Ventricular assist device