We conducted a retrospective analysis of the Spanish Registry of Heart Transplantation (SRHT), a prospective registry that has enrolled all patients receiving a transplant in Spain since the initiation of this activity in May, 1984. Data on recipient demographics, baseline clinical characteristics, donor characteristics, transplant procedure and immunosuppression, as well as outcomes, were collected from the registry database. Vital status was ascertained through December 31, 2021.

The SRHT is a comprehensive registry that encompasses all the transplant centers across Spain and includes all the patients who have undergone HT since 1984. Centers are linked to a public, nonprofit national health system. These centers have established a stringent protocol for patient monitoring and follow-up, benefitting from patients’ notable commitment to the health system. Data are entered by each center using a predefined form hosted within a specific application accessible via the internet. This procedure is supervised by a contract research organization and is under the oversight of the medical direction of the registry. Details about donors are supplied by the National Transplant Organization. Annually, in January, each center logs the previous year's transplants, and clinical data are regularly updated. Vital status checks are mandatory for new and historical cases.

All patients granted written informed consent upon their inclusion on the transplant waiting list, allowing the use of their deidentified data for research purposes. Additionally, the SRHT operations have received approval from the ethic committee of Hospital Universitario La Fe de Valencia (Spain).

We included patients aged 18 and older who underwent their first isolated HT between May 1984 and December 2021. The study focused on patients diagnosed with HCM or RCM. In the RCM group, we included patients with amyloidosis, as well as those with idiopathic/undefined, radiation/chemotherapy, endocardial fibrosis, and sarcoidosis etiologies. The comparison group comprised patients with a pretransplant diagnosis of dilated cardiomyopathy (DCM), which encompassed individuals with idiopathic/undefined, chemotherapy-induced, postpartum, familial, and toxic etiologies.

The primary outcome was all-cause mortality at 1 year, with 5-year all-cause mortality as the secondary outcome. Additionally, we analyzed specific causes of death within the first year, the incidence of primary graft failure, duration of posttransplant mechanical ventilation, and intensive care unit length of stay.

The quantity and percentage of these missing values, categorized by the study variables, can be found in table 1 of the supplementary data. In aggregate, 4.3% of the complete dataset had missing information. Variables exceeding 20% of missing data were excluded. For the remaining variables, a multiple imputation approach was used utilizing chained equations through an iterative Markov Chain Monte Carlo algorithm. Imputation was conducted using linear regression for quantitative variables, logistic regression for dichotomous variables, and multinomial logistic regression for categorical variables. In this process, all study variables and the outcome variable were included as independent variables. The regression models were executed on each of the 15 imputed datasets, and the outcomes were averaged using Rubin's rules.23

Quantitative variables are summarized as median [interquartile range]. Categorical variables are summarized as frequency (percentage). Differences among groups were analyzed using the Kruskal Wallis rank test or the Wilcoxon rank-sum test for continuous variables, and with the chi-square test for categorical variables. Bonferroni's correction was used for post hoc pairwise comparisons.

Survival was analyzed using the Kaplan-Meier method, and survival curves were compared using the log-rank test. Hazard ratios (HR) with 95% confidence intervals (95%CI) were estimated using Cox regression. Proportional hazard assumption was assessed graphically using log-log plots. Multivariable analyses included, apart from primary etiology, all variables that reached a P-value<.1 in the univariable analysis.24 A sensitivity analysis was conducted wherein the variable of “high priority transplantation” was excluded from the multivariable regression models.

To evaluate temporal trends in mortality risk, the past 20 years (2002-2021) were categorized into 4-year intervals, with the 2002 to 2006 period serving as the reference group. A linear regression was conducted between the HR and the transplant era, incorporating a weighted approach based on sample size and including an intercept.25 The adjusted R-squared value was employed to quantify the strength of the regression. The significance level of 2-tailed tests was set at α <.05. The analysis was performed using Stata 16.0.

Between May 1984 and December 2021, 8351 patients met the inclusion criteria. After we excluded etiologies not relevant to the present analysis, the study population comprised 3703 patients (median age, 53 [17] years; 991 females [26.8%]). The study groups were composed of 331 recipients with HCM and 260 recipients with RCM (comprising 104 patients with idiopathic RCM, 90 with amyloidosis, 19 with endocardial fibrosis, 5 with sarcoidosis, 10 with radiation/chemotherapy-related cardiomyopathy, and 32 with other causes). The reference group consisted of 3112 recipients with DCM. The percentage changes in the study groups relative to the entire transplant population are illustrated in figure 1. Approximately 57.4% of HCM transplants and 43.8% of RCM transplants occurred between 2012 and 2021. Indeed, during the last 5 years, HCM and RCM have collectively accounted for nearly one-seventh of all cases of transplant.

Figure 1.

Temporal trends in heart transplant in recipients with hypertrophic and restrictive cardiomyopathy (percentage of total procedures, 1984-2021).

(0.1MB).

The characteristics of the study groups are summarized in table 1 and table 2. The HCM and RCM groups displayed significant differences from the DCM group. These differences encompassed being younger, having a higher percentage of females, and showing a higher LVEF. For the remaining recipient characteristics, the differences compared with the DCM group varied depending on the specific study group. The HCM group had lower rates of cytomegalovirus (CMV) seropositivity, respiratory disease, pretransplant mechanical ventilation, circulatory support with extracorporeal membrane oxygenation, and pretransplant malignancies than the DCM group. In contrast, the RCM group exhibited lower pulmonary vascular resistance and body weight, lower percentages of diabetes, respiratory disease, and infection, and a higher pretransplant malignancy rate. The HCM and RCM groups differed only in terms of body weight (lower for RCM) and the rate of pretransplant malignancy (higher for RCM). Overall, the use of ventricular assist devices before HT was comparable across all groups. Nevertheless, this use showed an upward trend in the last decade, being more frequent in the DCM (21.9%) than in the HCM (13.4%) and RCM (9.7%) groups (P=.001).

Furthermore, the HCM and RCM groups received transplants from donors who were relatively older, with a higher proportion of female donors. Even so, the HCM group exhibited a significant trend toward a lower proportion of transplants performed with an unfavorable recipient/donor sex match. The cause of donor death was less frequently traumatic in the study groups, with statistical significance observed only in the RCM group compared with the DCM group. The HCM group less frequently received an urgent HT, and there was a nonsignificant trend toward longer ischemia times and a greater use of the bicaval technique in the study groups compared with the DCM group. The HCM and RCM groups showed a higher incidence of induction using anti-CD25 antibodies, whereas induction with OKT3 was less common in comparison with the DCM group.

The absolute and relative frequencies of all-cause death are summarized in table 3. In total, 746 deaths were observed in the first year, and 1078 deaths occurred by the fifth year of follow-up. Overall, the cumulative survival rate was 79.5%, 69.4%, 58.3% at 1, 5, and 10 years, respectively, with a median survival time of 12.8 years [2.8-22.1].

In the unadjusted analysis, 1-year survival was significantly poorer only for RCM compared with DCM (figure 2A). At 5 years, survival was significantly lower in RCM than in either DCM or HCM (figure 2B).

Figure 2.

Survival curves for 1-year (A) and 5-year (B) mortality.

(0.22MB).

Univariable and multivariable Cox regression analyses for 1-year and 5-year mortality are summarized in tables 2 and 3 of the supplementary data and multivariable adjusted comparisons between groups in table 4 of the supplementary data. The RCM group exhibited a significant 48% higher risk of 1-year mortality than the DCM group (figure 3A). Similarly, the HCM group also demonstrated a significant 38% higher risk of 1-year mortality compared with the DCM group (figure 3B). There were no differences in 1-year mortality between the RCM and HCM groups.

Figure 3.

Comparison of adjusted 1-year and 5-year mortality risks between restrictive (A) and hypertrophic (B) cardiomyopathy and dilated cardiomyopathy. One-year mortality risk adjusted for recipient age, renal dysfunction, pulmonary vascular resistance, bilirubin >2mg/dL, infection, mechanical ventilation, pretransplant circulatory support, malignancy, donor sex, urgency, cold ischemia time, transplant era and immunosuppressive induction. Five-year mortality risk adjusted for age, renal dysfunction, pulmonary vascular resistance, bilirubin >2mg/dL, infection, mechanical ventilation, pretransplant circulatory support, prior sternotomy, malignancy, urgency, surgical technique, transplant era and immunosuppressive induction.

(0.17MB).

Regarding 5-year mortality, the RCM group still showed a significant 52% excess mortality risk compared with the DCM group (figure 3A), whereas the 17% excess mortality risk observed for the HCM group compared with the DCM group was not statistically significant (figure 3B). A nonsignificant trend to higher 5-year mortality risk was observed for the RCM group compared with the HCM group (figure 3A).

These results were not significantly impacted by the level of urgency with which the HT was performed. Thus, excluding the variable “high priority transplantation” yielded unaltered results consistent with those obtained from the original models (table 5 of the supplementary data). In contrast, the use of ventricular assist devices before HT could have influenced the early prognosis of HCM, as in this small subgroup of patients, the HR increased to 1.61 (95%CI, 0.88-3.33) compared with DCM, without statistical significance due to the limited sample size. In the case of RCM, the increase in the HR to 1.61 (95%CI, 0.61-4.19) appeared to be less pronounced due to the already poor overall prognosis in this group.

Figure 4 illustrates the temporal changes in mortality risk in our study groups. There was a significant decrease in 1-year mortality for DCM and RCM, while the HCM group showed little change, maintaining a mortality risk similar to that of the 2002 to 2006 period in the 2017 to 2021 period (figure 4A). Regarding 5-year mortality, all groups showed consistent improvements over time, with noteworthy progress for the RCM and HCM groups (figure 4B).

Figure 4.

Temporal trends for 1-year (A) and 5-year (B) adjusted mortality risk from 2002 to 2021. R2=adjusted squared coefficient of correlation. In each group, the reference group is the transplant conducted between 2002 and 2006.

(0.23MB).

Primary graft failure occurred in around 23% of the study groups, which was notably higher than in DCM cases. Prolonged mechanical ventilation and an extended intensive care unit stay indicated a more intricate immediate postoperative recovery in HCM and RCM.

While the overall comparison of causes of death was statistically significant, no specific pairwise comparisons were significant (table 3). Notably, the HCM group had more deaths from surgical complications, multiorgan failure, and strokes, but fewer deaths from acute rejection compared with the DCM group. In contrast, the RCM group had more deaths from primary graft failure and multiorgan failure, but fewer deaths from infection and strokes than the DCM group.

Table 6 of the supplementary data summarizes the baseline characteristics of the amyloidotic and nonamyloidotic etiologies of RCM. Patients with amyloidosis were significantly older, had a higher body mass index, and less frequently underwent the bicaval technique than patients with other etiologies. In addition, patients with amyloidosis showed a nonsignificant tendency to be more frequently male, to have a lower percentage of peripheral vascular disease and prior sternotomy, and a higher percentage of anti-CD25. The survival rates for individuals with amyloidosis were lower than those with RCM due to other etiologies (figure 5). This significant association persisted even after multivariable adjustment (figure 5 and table 7 of the supplementary data).

Figure 5.

Survival curves (Kaplan-Meier method) for 5-year mortality for amyloidosis and nonamyloidosis etiology of restrictive cardiomyopathy. Hazard ratio was adjusted for pulmonary vascular resistance, respiratory disease, mechanical ventilation, malignancy, cold ischemia time, surgical technique, transplant era, immunosuppressive induction.

(0.06MB).

A high percentage of recipients with underlying HCM had an LVEF <50% at the time of HT (70%). These were more frequently male and received chronic amiodarone before HT compared with those without reduced ejection fraction at baseline. There were no between-group differences in either the univariable or adjusted survival analysis (figure 6, table 8 of the supplementary data).

Figure 6.

Survival curves (Kaplan-Meier method) for 5-year mortality according to LVEF prior to heart transplant in recipients with hypertrophic cardiomyopathy. Hazard ratio was adjusted for age, Cytomegalovirus seropositivity, pulmonary vascular resistance, respiratory disease, mechanical ventilation, pretransplant sternotomy, urgency, transplant era and immunosuppressive induction. HR, hazard ratio; LVEF, left ventricular ejection fraction.

(0.07MB).

Our study presents an overview of a long-term nationwide registry in which we cannot exclude the potential presence of unaccounted-for confounding factors. A thorough characterization of pretransplant clinical status (time on waiting list, frailty markers, hemodynamic status), could be very useful in interpreting our findings. Although we had a modest amount of missing data, which were treated using standard techniques, the potential for data entry error persists. The relatively small sample size of some groups could have diminished the statistical power.

Inaccuracies in local adjudication of the primary etiology could have affected some cases. This could possibly explain the high prevalence of “idiopathic” or “other” causes of the restrictive phenotype. In addition, increased awareness of the diagnosis of certain cardiomyopathies, in particular amyloidosis, could explain its increasing prevalence as a cause of HT in recent years. Moreover, the database did not allow us to distinguish between different types of amyloidosis, and it does not consider the possibility of phenocopies in cases of hypertrophic/restrictive cardiomyopathies. Likewise, information on the presence and severity of intraventricular gradients in HMC is lacking. Our results should be interpreted in the context of the prolonged observation period, which entailed significant changes in the criteria for inclusion and care of patients on the waiting list as well as in posttransplant management.