https://orcid.org/0000-0002-2654-8117
Abstract
This study aims to explore characteristics and clinical outcomes of patients with congenital heart disease (CHD) in the European Registry for Patients with Mechanical Circulatory Support (EUROMACS).
This is a retrospective study of EUROMACS participants receiving MCS as bridge-to-transplant, possible bridge-to-transplant, or rescue therapy/bridge-to-recovery from 2011 to 2023 (n = 5340). Adult and paediatric cohorts were analysed separately. The primary outcome was mortality on MCS; secondary outcomes included recovery, transplant and complications including bleeding, cerebrovascular events, and sepsis.
Among adult patients, mortality at 1-year was 33.3% among the CHD cohort vs 22.1% in the non-CHD cohort. Adult CHD patients had higher hazards of mortality within the first year after MCS implantation [hazard ratios 1.98, 95% confidence interval (CI) 1.35–2.91, P < 0.001] and bleeding events (subdistribution hazard ratios 2.10, 95% CI 1.40–3.16, P < 0.001) compared with non-CHD patients. Both associations remained significant after accounting for multiple mediators. Among paediatric patients, mortality at 1 year was 22.1% in the CHD cohort vs 17.3% in the non-CHD cohort (hazard ratios 1.39, 95% CI 0.83–2.32, P = 0.213).
Adult and paediatric patients with CHD on MCS have higher adverse event risk compared with non-CHD MCS patients, though children did not have greater risk of mortality. As the number of CHD patients requiring advanced heart failure management continues to grow, these findings can enhance informed decision-making.
Registry name: EUROMACS.
INTRODUCTION
With the significant improvement in survival of patients with congenital heart disease (CHD) to adolescence and adulthood [1], CHD is an increasingly frequent indication for advanced heart failure therapies [2]. There is a paucity of data regarding the characteristics and outcomes of mechanical circulatory support (MCS) in this population, and how their outcomes compare to patients without CHD [3]. Furthermore, there is currently little information on risk factors that may be associated with worse prognosis amongst the CHD population requiring MCS.
These questions are clinically relevant as CHD patients are known to have longer heart transplant waiting times, despite being younger and having less comorbidities than the non-CHD population [4]. These factors suggest that despite being seemingly good candidates for this treatment, there may be a gap in service delivery. The relative underutilization of MCS as a bridge-to-transplant in the CHD patient cohort may relate in part to the lack of data regarding its use and outcomes [5].
At present, the only available large-scale study specifically exploring MCS in CHD is derived from the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) and its corresponding paediatric cohort, the Paediatric Interagency Registry for Mechanical Circulatory Support, both based in the United States [6, 7]. In this study, paediatric CHD patients had greater mortality in the first year following MCS implantation. However, outcomes of CHD patients might differ outside this specific cohort and the United States. It is already recognized that trends in outcomes such as infection and neurological event rates differ between INTERMACS and its European counterpart, the European Registry for Patients with Mechanical Circulatory Support (EUROMACS) [8]. In addition to this, the characteristics of cardiac donors and donor organ acceptance trends are recognized to be different between the 2 continents [9] with donors tending to be older and with more cardiovascular risk factors in Europe, but higher rates of obesity and abnormal ejection fraction in the United States where different organ utilization rate is observed.
In this study, we aim to describe the clinical characteristics and outcomes of MCS in the population with CHD in EUROMACS. First, we will describe the clinical characteristics and types of congenital lesions in the CHD cohort requiring MCS. We will then compare baseline, clinical, admission and device characteristics between CHD patients and non-CHD patients. Next, we will compare the clinical outcomes between CHD and non-CHD cohorts, in order to establish whether patients with CHD have similar or worse prognosis compared to those without. Finally, we aim to identify predictors of adverse outcomes specifically in the CHD population. Separate analyses will be performed in the paediatric and adult cohorts.
PATIENTS AND METHODS
Ethical statement
Participation in the EUROMACS Registry was approved by the institutional committees of all participating centres. All subjects gave informed consent for inclusion in the EUROMACS Registry and utilization of anonymized data for subsequent studies. This study was approved by the EUROMACS scientific review committee (EUROMACS study number 73) and was granted national ethical approval (IRAS project ID 328981, approved 11 July 2023) as well as approval from institutional ethical committee (R&D approval reference P03601).
Study design
EUROMACS is a multinational prospective registry of the European Association for Cardio-Thoracic Surgery. Consenting patients across 70 institutions in 21 countries who received MCS were prospectively recruited since 2011 and their relevant clinical, echocardiographic, haemodynamic and laboratory parameters collected. Outcomes were ascertained on an ongoing basis and recorded for participants. Detailed descriptions of the data collection protocol have been previously reported [10–12]. Data for the study was extracted from the EUROMACS in September 2023. Details regarding initial quality control of the data are provided in Supplementary Material, Methods.
Patient selection
The patient selection flowchart for the study is depicted in Fig. 1. From the EUROMACS cohort (n = 7923), patients were excluded if they were duplicate patient records (n = 308), those with a missing primary diagnosis (n = 456), and those who did not have any follow-up data (n = 131). Patients were also excluded if they had received MCS as destination therapy (n = 1688, none of whom had a primary diagnosis of CHD). If the device strategy was missing, it was assumed to be potential bridge-to-transplant or recovery; these patients were therefore not excluded from the study.
Patient selection flowchart. EUROMACS: European Registry for Patients with Mechanical Circulatory Support.
After identification of eligible patients, patients were grouped into those with a diagnosis of CHD (n = 152) and those without CHD (n = 5188). Finally, the patients were divided between adult (aged ≥ 18, n = 4836, among whom CHD n = 75 and non-CHD n = 4761; mean age 51.4 years [± 11.6]) and paediatric (aged < 18, n = 504, among whom CHD n = 77 and non-CHD n = 427; mean age 9.04 years [± 5.17]).
Congenital heart disease definition
For the purpose of this study, a primary diagnosis of CHD was determined if the patient:
Had a named primary or secondary diagnosis of CHD; unless the primary diagnosis related to a major pathology and the secondary diagnosis only described a minor congenital lesion (e.g. hypertrophic cardiomyopathy as primary diagnosis and atrial septal defect as secondary diagnosis).
Had clear indication of having undergone a surgical procedure which indicates an underlying diagnosis of CHD (e.g. Norwood procedure).
Study outcomes
The primary outcome of this study was mortality. Secondary outcomes included recovery, transplant, and serious adverse events including bleeding, cerebrovascular events, and sepsis. Further details regarding definition of outcomes are provided in Supplementary Material, Methods.
Statistical analysis
Statistical analysis was carried out using R version 4.2.2 (2022–10-31). All statistical analyses were performed separately for adult (aged ≥ 18) and paediatric (aged < 18) cohorts.
The baseline, clinical, admission and device characteristics of study participants by CHD status are summarized using descriptive statistics. Continuous variables are reported as means ± standard deviations, and differences in these variables by CHD status tested using t-tests if their distribution appeared normal in histograms; otherwise, they were described as median and interquartile range and compared using Mann–Whitney U tests. Categorical or binary variables are reported as numbers and percentages, and compared by CHD status using χ2 tests for trend.
Time-to-event analyses
For the primary outcome, Cox proportional hazards models were used to evaluate the association between CHD and mortality. Two sets of models were fitted: firstly, survival was compared between each individual non-CHD primary diagnosis with CHD (i.e. dilated cardiomyopathy versus CHD, coronary artery disease versus CHD, etc.); and secondly, mortality among patients with CHD was compared with all other patients when considered as a single non-CHD group (i.e. CHD versus all non-CHD). For the secondary outcomes, patients with CHD were compared with all non-CHD patients only. For these analyses, Fine and Gray models were used to account for competing risks of mortality.
Results are reported as hazard ratios (HR) and 95% confidence intervals (CIs) for the Cox models, and subdistribution hazard ratios (sdHR) and 95% CI for the Fine and Grey models. Potential confounders were adjusted for on the basis of the direct acyclic graph displayed in Supplementary Material, Fig. S1. For all time-to-event analyses, the proportional hazards assumption was tested using Schoenfeld’s residuals test. Where violation was noted (P < 0.05) and this was evident on ‘log-log’ [log(−log(survival)) vs log(time)] plots, the analysis was stratified by the culprit variable. If the culprit variable was the exposure, a time-stratified analysis was performed. Details of additional sensitivity analyses are outlined in Supplementary Material, Methods.
Mortality predictors among congenital heart disease patients
Finally, clinical predictors of mortality among patients with CHD requiring MCS were evaluated using multivariable Cox proportional hazards models adjusted for sex and age. Further details are outlined in Supplementary Material, Methods.
RESULTS
Baseline characteristics
A total of 5340 patients were included, among whom 152 had an underlying diagnosis of CHD, as displayed in Fig. 1. The median follow-up in the whole cohort was 1.38 years (interquartile range 0.37 to 3.25). After exclusion of censored patients, the inverse Kaplan–Meier median follow-up remained similar: 1.83 years (interquartile range 0.77 to 3.65). The specific congenital diagnoses are summarized in Supplementary Material, Table S1. The most common diagnosis among adults was transposition of the great arteries (n = 26, 34.7%), whereas in the paediatric cohort it was single ventricle circulation (n = 21, 27.3%).
Overall, patients with CHD tended to be younger (mean age 25.6 ± 20.1 vs 48.7 ± 15.5, P < 0.001), less commonly male (67.8% vs 80.1%, P < 0.001), and as would be expected, were more likely to have had previous cardiac surgery (38.2% vs 11.1%, P < 0.001) and transfusions (31.6% vs 9.4%, P < 0.001). The baseline characteristics are summarized in Supplementary Material, Table S2 for the whole cohort, Supplementary Material, Table S3 for the adult cohort and Supplementary Material, Table S4 for the paediatric cohort.
At the time of admission, CHD patients tended to have a more critical clinical status (27.0% vs 18.9% had an INTERMACS score of 1, indicating critical cardiogenic shock on admission, P for trend < 0.001), inotrope requirement (59.2% vs 46.9%, P = 0.004), mechanical ventilation (32.2% vs 15.1%, P < 0.001), and extracorporeal membrane oxygenation (ECMO) support (23.7% vs 14.9%, P = 0.004). Patients with CHD were also more likely to have had a cardiac arrest (15.1% vs 8.2%, P = 0.004), major infection (24.3% vs 11.6%, P < 0.001) and protein losing enteropathy (2.6% vs 0.1%, P < 0.001). In addition to this, CHD patients were more likely to require biventricular assist device (BiVAD) (12.5% vs 8.9%, overall P < 0.001), and significant differences also existed in devices of choice across groups.
Survival analyses
Adults
For adult patients implanted with MCS, both 1-month and 1-year mortality rates were significantly higher in CHD patients compared with non-CHD patients: 21.3% vs 12.4% (P = 0.033) and 33.3% vs 22.1% (P = 0.029) respectively. At median follow-up of 1.53 years (interquartile range 0.44–3.38), mortality rates in CHD and non-CHD patients were similar at 41.3% and 38.3% respectively (P = 0.919), as summarized in Supplementary Material, Table S5.
Unadjusted survival analysis by underlying diagnosis demonstrated significant heterogeneity in mortality across diagnosis groups (log-rank P < 0.001) (Fig. 2A). After adjustment for confounders, patients with coronary artery disease had lower hazards of mortality compared to CHD (HR 0.65, 95% CI 0.46–0.92, P = 0.016), as did those with dilated cardiomyopathy (HR 0.57 95% CI 0.40–0.80, P = 0.001) (Fig. 3A). After accounting for mediators relating to disease severity and device characteristics, neither of these associations were significant. This suggests that the association was at least in part mediated by these factors.
Cumulative mortality over the first year from the time of implantation of mechanical circulatory support. (A) In adult patients (age 18 or greater) divided by primary diagnosis; (B) in paediatric patients (age less than 18) divided by primary diagnosis; (C) in adult patients (age 18 or greater) comparing congenital heart disease patients with all other diagnoses together; and (D) in paediatric patients (age less than 18) comparing congenital heart disease patients with all other diagnoses together. The log-rank P-value represents the result of an unadjusted comparison of the survival trend during follow-up across the groups.
Results of survival analyses comparing mortality for each other primary diagnosis to congenital heart disease in (A) the adult and (B) the paediatric cohorts.
Similarly, when comparing CHD patients against all non-CHD patients as a single group, higher hazards of mortality were observed among the CHD cohort (unadjusted log-rank P = 0.007), Fig. 2C. After stratification of Cox models at the 1-year mark due to non-proportional hazards, patients with CHD had higher hazards of mortality within the first year (HR 1.98 95% CI 1.35–2.91, P < 0.001, Fig. 4A). After additional adjustment for mediators, the association remained significant (HR 1.67 95% CI 1.07–2.62, P = 0.025), indicating that despite some potential mediation the higher 1-year mortality is not only related to these factors. Similarly, as displayed in Figs 4A and 5, patients with CHD had higher hazards of bleeding (sdHR 2.10, 95% CI 1.40–3.16, P < 0.001). This also remained significant after adjustment for potential mediators (sdHR 2.11, 95% CI 1.33–3.34, P = 0.001). The results of sensitivity analyses were consistent and are summarized in Supplementary Material, Results, and Supplementary Material, Tables S6 and S7.
Hazards of primary and secondary end-points during follow-up in (A) adult and (B) paediatric patients, comparing congenital heart disease patients with all other patients as a single category. Analyses for the primary outcome of mortality were stratified by time at 1 year due to non-proportional hazards of the exposure variable. Insufficient events were present in the paediatric cohort to evaluate association with mortality: after 1 year. All results are derived from Cox proportional hazards models, where patients were censored at the time of outcome occurrence, mortality, transplant, or last recorded date of follow-up.
Cumulative incidence of secondary outcomes over the first year from the time of implantation of mechanical circulatory support in adult patients (age 18 or greater) comparing congenital heart disease patients with all other diagnoses together. The log-rank P-value represents an unadjusted comparison of the survival trend across the groups.
Cumulative incidence of secondary outcomes over the first year from the time of implantation of mechanical circulatory support in paediatric patients (age less than 18) comparing congenital heart disease patients with all other diagnoses together. The log-rank P-value represents the result of an unadjusted comparison of the survival trend across the groups.
Paediatric
For paediatric patients implanted with MCS, the 1-month and 1-year mortality rates were similar in CHD patients compared with non-CHD patients: 9.4% vs 9.5% (P = 0.944) and 22.1% vs 17.3% (P = 0.403) respectively. At a median follow-up of 0.41 years (interquartile range 0.14 to 1.16), the mortality rates in CHD and non-CHD patients were also similar at 23.4% and 19.4% respectively (P = 0.700), summarized in Supplementary Material, Table S5.
Unadjusted survival analysis by underlying diagnosis demonstrated significant heterogeneity in mortality across diagnosis groups (log-rank P = 0.027, Fig. 2B). However, after adjustment for confounders, none of the other diagnoses displayed differences in hazards of mortality when compared to CHD. This remained consistent after accounting for potential mediators, as displayed in Fig. 3B.
Similarly, when comparing CHD patients against all non-CHD patients as a single group, hazards of mortality were similar (unadjusted log-rank P = 0.230, Fig. 2D). After adjustment for confounders, there were no significant differences in the hazards of primary or secondary outcomes for CHD patients when compared to all others, and this remained consistent after adjustment for mediators. The odds of recovery was apparently higher (sdHR 1.80, 95% CI 0.97–3.37, P = 0.065) and that of heart transplant was lower (sdHR 0.72, 95% CI 0.51–1.02, P = 0.062) but these differences did not reach statistical significance and were completely lost on adjustment for mediators. The full results are shown in Figs 4B and 6. The results of sensitivity analyses were consistent and are summarized in Supplementary Material, Results, and Supplementary Material, Tables S6 and S7.
Predictors of mortality
In order to provide clinically relevant information for prognosis, predictors of mortality among CHD patients who require MCS were evaluated. Overall, mortality rates numerically varied across different CHD types, as summarized in Supplementary Material, Table S8. When assessing the whole cohort, 5 predictors of adverse outcome were identified. Higher hazards of mortality were seen with temporary devices, requirement for ECMO or intra-aortic balloon pump, history of cardiac arrest, greater age, single ventricle circulation, requirement for ultrafiltration and ventilation. Conversely, having a left ventricular assist device (LVAD) only, a greater INTERMACS profile, higher right atrial pressure, and male sex were associated with lower hazards of mortality. The full results are presented in Table 1, and the survival curves comparing patients with single ventricle circulation with all other CHD patients are shown in Supplementary Material, Fig. S2.
Clinical factors associated with mortality in congenital heart disease patients receiving mechanical circulatory support
| Variable | Hazards ratio | 95% confidence interval | P-value |
|---|---|---|---|
| Temporary device type | 6.11 | 2.3–16.21 | <0.001 |
| LVAD only | 0.29 | 0.14–0.61 | 0.001 |
| ECMO | 2.66 | 1.45–4.87 | 0.002 |
| IABP | 10.82 | 2.32–50.5 | 0.002 |
| Age | 1.03 | 1.01–1.06 | 0.003 |
| Cardiac arrest | 2.79 | 1.38–5.63 | 0.004 |
| INTERMACS patient profile | 0.67 | 0.50–0.90 | 0.009 |
| Single ventricle circulation | 3.51 | 1.38–8.95 | 0.009 |
| Male | 0.47 | 0.26–0.86 | 0.014 |
| RA pressure | 1.12 | 1.02–1.24 | 0.019 |
| Ultrafiltration | 2.47 | 1.10–5.53 | 0.029 |
| Ventilator | 1.93 | 1.04–3.57 | 0.037 |
| Variable | Hazards ratio | 95% confidence interval | P-value |
|---|---|---|---|
| Temporary device type | 6.11 | 2.3–16.21 | <0.001 |
| LVAD only | 0.29 | 0.14–0.61 | 0.001 |
| ECMO | 2.66 | 1.45–4.87 | 0.002 |
| IABP | 10.82 | 2.32–50.5 | 0.002 |
| Age | 1.03 | 1.01–1.06 | 0.003 |
| Cardiac arrest | 2.79 | 1.38–5.63 | 0.004 |
| INTERMACS patient profile | 0.67 | 0.50–0.90 | 0.009 |
| Single ventricle circulation | 3.51 | 1.38–8.95 | 0.009 |
| Male | 0.47 | 0.26–0.86 | 0.014 |
| RA pressure | 1.12 | 1.02–1.24 | 0.019 |
| Ultrafiltration | 2.47 | 1.10–5.53 | 0.029 |
| Ventilator | 1.93 | 1.04–3.57 | 0.037 |
All analyses used Cox regression adjusted for age and sex.
ECMO: extracorporeal membrane oxygenation; IABP: intra-aortic balloon pump; INTERMACS: Interagency Registry for Mechanically Assisted Circulatory Support; LVAD: left ventricular assist device; MCS: mechanical circulatory support.
Clinical factors associated with mortality in congenital heart disease patients receiving mechanical circulatory support
| Variable | Hazards ratio | 95% confidence interval | P-value |
|---|---|---|---|
| Temporary device type | 6.11 | 2.3–16.21 | <0.001 |
| LVAD only | 0.29 | 0.14–0.61 | 0.001 |
| ECMO | 2.66 | 1.45–4.87 | 0.002 |
| IABP | 10.82 | 2.32–50.5 | 0.002 |
| Age | 1.03 | 1.01–1.06 | 0.003 |
| Cardiac arrest | 2.79 | 1.38–5.63 | 0.004 |
| INTERMACS patient profile | 0.67 | 0.50–0.90 | 0.009 |
| Single ventricle circulation | 3.51 | 1.38–8.95 | 0.009 |
| Male | 0.47 | 0.26–0.86 | 0.014 |
| RA pressure | 1.12 | 1.02–1.24 | 0.019 |
| Ultrafiltration | 2.47 | 1.10–5.53 | 0.029 |
| Ventilator | 1.93 | 1.04–3.57 | 0.037 |
| Variable | Hazards ratio | 95% confidence interval | P-value |
|---|---|---|---|
| Temporary device type | 6.11 | 2.3–16.21 | <0.001 |
| LVAD only | 0.29 | 0.14–0.61 | 0.001 |
| ECMO | 2.66 | 1.45–4.87 | 0.002 |
| IABP | 10.82 | 2.32–50.5 | 0.002 |
| Age | 1.03 | 1.01–1.06 | 0.003 |
| Cardiac arrest | 2.79 | 1.38–5.63 | 0.004 |
| INTERMACS patient profile | 0.67 | 0.50–0.90 | 0.009 |
| Single ventricle circulation | 3.51 | 1.38–8.95 | 0.009 |
| Male | 0.47 | 0.26–0.86 | 0.014 |
| RA pressure | 1.12 | 1.02–1.24 | 0.019 |
| Ultrafiltration | 2.47 | 1.10–5.53 | 0.029 |
| Ventilator | 1.93 | 1.04–3.57 | 0.037 |
All analyses used Cox regression adjusted for age and sex.
ECMO: extracorporeal membrane oxygenation; IABP: intra-aortic balloon pump; INTERMACS: Interagency Registry for Mechanically Assisted Circulatory Support; LVAD: left ventricular assist device; MCS: mechanical circulatory support.
DISCUSSION
This study aimed to provide a comprehensive analysis of the clinical characteristics and outcomes of MCS among adult and paediatric patients with CHD in the EUROMACS registry. The findings reveal that adult CHD patients have higher mortality in the first year after MCS implantation compared with non-CHD patients, an association which was not explained by the greater medical complexity and more critical condition of these patients at the time of admission. On the other hand, in the paediatric population, there were no differences in mortality between the CHD and non-CHD patients, but by the end of the first year children with CHD were more likely to still be on MCS and were less likely to have received a heart transplant.
The higher mortality among adult CHD patients in the first year of this study is in line with the findings of the only previously published paper on this subject to date, from the INTERMACS registry [6]. In this study, limited to adult CHD patients followed up for 1 year, those with CHD had higher mortality and adverse events compared with non-CHD patients, accompanied by a greater risk of bleeding events. Nonetheless, a difference between the 2 studies exists. In INTERMACS, the authors found that the mortality association was exclusively restricted to those on BiVAD/TAH support, and no difference was present after restricting to only those on LVAD. Though we did not replicate this exact analysis due to the limited number of TAH/BiVAD supported patients in the EUROMACS cohort, a mediation analysis was performed including the type of device as a potential mediator. After accounting for this, the difference in mortality was certainly attenuated, but remained statistically significant. This means that in our study, unlike the INTERMACS study, the difference in mortality could not be fully explained by differences in the type of device.
Differing from the results in adults, children with CHD seemed to have a similar mortality to that of patients with other diagnoses in our study. This result is not concordant with the recently published study by Ashfaq et al., who reported the clinical outcomes of paediatric patients receiving MCS in Paediatric Interagency Registry for Mechanical Circulatory Support [7]. In this analysis, aimed at identifying predictors of adverse outcome, CHD is one of the factors associated most strongly with hazards of early mortality. The discrepancy in results might relate to a difference in the underlying characteristics of the cohorts, but might also be due to limited power in the paediatric analysis in our study.
Within 1 year of implantation, children with CHD had greater likelihood of still being on MCS and a lower likelihood of having received a heart transplant. Unfortunately, this is in line with well-recognized longer times on transplant waiting lists that CHD patients face [13], and the accompanying higher risk of wait list mortality compared with non-CHD counterparts [14, 15]. Though the exact reasons for this are not fully understood, it has been speculated that this may be related to the inherently greater risk of sensitization due to previous transfusions, potential differences in body size due to the tendency for CHD patients to be younger, limiting the donor pool, as well as the greater anatomical and surgical complexity [16].
In this paper, we identified multiple predictors of adverse outcome on MCS specifically applying to CHD patients. These included demographic characteristics such as age and sex, as well as markers of illness severity at admission (e.g. INTERMACS profile, the requirement for intra-aortic balloon pump or ECMO prior to MCS), which have previously been associated with adverse outcomes in patients on MCS[7]. Of note, we also identified that having a single ventricle circulation is a predictor of adverse outcomes. Though this is intuitive from a biological perspective, given the significant surgical and medical complexity involved in the management of these patients, it is a particularly interesting finding because previous studies have not demonstrated its association with clinical outcomes [6, 7]. However, it is important to note that among these studies, one had few single ventricle patients (n = 17) and might therefore have been underpowered [4], and the other evaluated the association in a non-select patient cohort (i.e. not CHD only) and in a single multivariable model including variables strongly correlated with single ventricle circulation (e.g. CHD and device type) [5]. This might have limited the likelihood of identifying the association due to collinearity, and given the different population subset investigated it cannot be considered truly comparable to our analysis.
From a clinical perspective, this study contributes to the expanding body of evidence concerning outcomes of advanced heart failure therapies in patients with CHD, offering insights that can aid shared decision-making and prognostication in clinical practice. This holds significant clinical relevance, given the increasing prevalence of CHD among those necessitating advanced heart failure therapy. Specifically, the study fills a void by presenting data from a European cohort, supplementing the previously available data solely from INTERMACS. Moreover, it initiates the identification of several factors that may contribute to prognosis, thereby facilitating the identification of individuals more or less likely to benefit from MCS. While yielding intriguing findings, this analysis is constrained by its current limited statistical power. Once the available cohort is sufficiently large, it will be important to perform a similar analysis and potentially expand this to the creation of a risk score to help clinical risk stratification and decision-making.
Limitations
This study has a number of limitations. Due to its registry-based nature, there was insufficient granularity of data to perform a full assessment of the classification of the CHD types based on international criteria. It was only possible to ascertain the underlying diagnosis of a patient (e.g. transposition of the great arteries) but in a significant proportion of cases, there was insufficient information to identify the type and/or previous surgical repair (e.g. arterial switch versus atrial switch versus congenitally corrected). Additional clinical data relating to the patients’ admission haemodynamics (e.g. pulmonary artery catheter findings) are not available in the EUROMACS Registry and this precluded the inclusion of these potentially important variables from the main analysis. In some cases, patients had records of device explanation without a clear record of recovery, transplant or mortality at the time of this event. Where possible, we examined the anonymized records in detail to seek alternative or closely coinciding mentions of mortality, recovery or transplant in other columns and labelled these events accordingly. Unfortunately, no additional information was available in some cases and therefore the reason underlying device explant could not be ascertained. In addition to this, despite being among the largest published studies of MCS in CHD, the participant numbers are still limited, and therefore repeated analysis at later date would be warranted as progressively larger numbers of patients are recruited to the EUROMACS cohort. This is especially true for the paediatric analysis, where the lack of association with early mortality might relate to limited power. Finally, in terms of generalizability, although the EUROMACS registry captures many of the MCS cases in Europe it does not have full coverage of the continent, and we therefore cannot infer whether characteristics of patients not captured in the cohort might differ from the ones that are included.
CONCLUSION
In conclusion, this study provides data evaluating the characteristics and outcomes of CHD patients requiring MCS and outline multiple predictors of adverse outcome of MCS specifically among patients with CHD. Adult patients with CHD had higher mortality in the first year following implantation, but this was not observed in paediatric patients, who conversely were more likely to remain on MCS at the end of the first year and were less likely to have received a heart transplant. Predictors of adverse prognosis in CHD patients included older age, female sex, having a device other than LVAD only, requiring ECMO prior to MCS, and single ventricle circulation. The data provided may contribute to clinical prognostication and help inform shared decision-making in the growing number of CHD patients with advanced heart failure.
SUPPLEMENTARY MATERIAL
Supplementary material is available at EJCTS online.
ACKNOWLEDGEMENTS
The authors acknowledge all patients participating in EUROMACS and clinical teams who have contributed data to the EUROMACS registry.
FUNDING
Maddalena Ardissino is funded by an National Institute for Health Research (NIHR) Academic Clinical Fellowship.
Conflict of interest: Alec P. Morley serves on the board of directors for ARDS Alliance, Inc. and receives no financial renumeration for this position. All other authors report no conflicts of interest.
DATA AVAILABILITY
Data used for this study is not publicly available and was accessed on application to the EUROMACS registry committee. The authors are not authorized to share this data. Licence to publish the image included in the graphical abstract was granted to Alec P. Morley by Biorender.com under agreement number ZK26V1O9MI.
Author contributions
Maddalena Ardissino: Conceptualization; Formal analysis; Investigation; Methodology; Project administration; Visualization; Writing—original draft. Alec P. Morley: Conceptualization; Investigation; Methodology; Writing—original draft. Clive Lewis: Conceptualization; Investigation; Methodology; Supervision; Writing—review & editing. Catriona Bhagra: Conceptualization; Investigation; Methodology; Project administration; Supervision; Writing—review & editing. Victoria Stoll: Conceptualization; Investigation; Methodology; Supervision; Writing—review & editing. Evegnij Popatov: Data curation; Investigation; Methodology; Project administration; Writing—review & editing. Felix Schoenrath: Data curation; Investigation; Methodology; Project administration; Resources; Writing—review & editing. Jan Gummert: Data curation; Methodology; Project administration; Resources; Writing—review & editing. Piotr Przybyłowski: Data curation; Investigation; Methodology; Project administration; Resources; Writing—review & editing. Joanna Śliwka: Data curation; Investigation; Methodology; Project administration; Resources; Writing—review & editing. Bart Meyns: Data curation; Investigation; Methodology; Project administration; Resources; Writing—review & editing. Theo M.M.H. de By: Data curation; Funding acquisition; Investigation; Methodology; Project administration; Resources; Writing—review & editing. Nicola Jones: Conceptualization; Investigation; Methodology; Project administration; Supervision; Writing—review & editing. Steven Tsui: Conceptualization; Investigation; Methodology; Project administration; Resources; Supervision; Writing—original draft.
Reviewer information
European Journal of Cardio-Thoracic Surgery thanks Saeid Hosseini and the other anonymous reviewers for their contribution to the peer review process of this article.
REFERENCES
ABBREVIATIONS
- BiVAD
Biventricular assist device
- CHD
Congenital heart disease
- CI
Confidence interval
- ECMO
Extracorporeal membrane oxygenation
- HR
Hazard ratios
- INTERMACS
Interagency Registry for Mechanically Assisted Circulatory Support
- LVAD
Left ventricular assist device
- MCS
Mechanical circulatory support
- sdHR
Subdistribution hazard ratios
