Clinical data for this study were extracted from the database of a multi-institutional retrospective registry, which included consecutive patients waitlisted for first-time, single-organ, high-emergency HT in 16 Spanish institutions. These patients were treated with various types of temporary mechanical circulatory support devices between January 1, 2010, and December 31, 2020. The study protocol, previously described elsewhere,12 was approved by the Committee for Ethics in Clinical Research of the Autonomous Community of Galicia, Spain, and ratified by the institutional review boards of all participating centers.

Registry data were collected through a retrospective, case-by-case review of medical records for each participant, with local investigators responsible for this task. No external monitoring of the recorded data were performed.

In this manuscript, we analyzed the clinical outcomes of patients included in the registry who underwent HT while supported by VA-ECMO. We compared those who received an additional mechanical strategy for LV unloading with those who did not. Throughout the study period, patients waitlisted under VA-ECMO support were given the highest priority in the Spanish organ donor allocation protocol, referred to as status 0. This priority meant that these candidates had absolute precedence in receiving the first suitable donor organ available within Spain. Specific details regarding the successive modifications of the Spanish organ donor allocation protocol during the study period have been discussed previously.12

As this was an observational, retrospective study, the decision to add mechanical LV unloading to VA-ECMO, including the type and timing, was left to the discretion of the attending clinical team based on local protocols and clinical experience.

All patients were followed up from the date of HT up to 1 year after HT, with 1-year posttransplant survival serving as the primary endpoint of the study. We also assessed other relevant adverse clinical outcomes that occurred during the in-hospital postoperative period, including excessive surgical bleeding, cardiac reoperation, postoperative graft dysfunction, postoperative infection, the need for postoperative mechanical circulatory support, and postoperative renal failure requiring dialysis. Specific definitions of all study outcomes are provided in the supplementary data.

Qualitative variables are presented as the number of patients and percentages, while quantitative variables are expressed as means±standard deviation or medians [interquartile range], as appropriate. Statistical comparisons between groups were conducted using the chi-square test for qualitative variables and the Student t-test or Mann-Whitney test, as appropriate, for quantitative variables.

Kaplan-Meier curves were constructed to graphically represent the cumulative probability of survival during the first year after HT in patients bridged with VA-ECMO alone or with VA-ECMO and LV unloading strategies. The survival functions were compared using the log-rank test.

Multivariable Cox regression was used to adjust for potential confounders in the observed association between LV unloading and 1-year posttransplant survival. A backward stepwise analysis with a P-out criterion >.10 was conducted to identify clinical variables independently associated with the outcome of interest. All baseline clinical variables were considered unless they had more than 10% missing values (bilirubin, albumin, aspartate aminotransferase, pH, PaO2/FiO2 ratio, LV end-diastolic diameter, cardiac index, central venous pressure, capillary wedge pressure, systolic pulmonary artery pressure, diastolic pulmonary artery pressure, mean pulmonary artery pressure, mean transpulmonary gradient).

Variables selected for inclusion in the first step of the backward stepwise process were those that had a univariate association with 1-year posttransplant survival at a P value <.20. Candidate variables entering the multivariable model included recipient age, history of stroke, preoperative infection, preoperative invasive mechanical ventilation, preoperative renal replacement therapy, preoperative vasopressor use, cold ischemia time, preoperative hemoglobin, preoperative creatinine, and mechanical LV unloading.

The final multivariable model from the backward stepwise process was used to estimate the adjusted hazard ratio (HR) for 1-year posttransplant mortality in patients with mechanical LV unloading vs controls. The proportional hazards assumption was verified through graphical representation of Schoenfeld residuals against time (P value for the global test=.820).

Further adjustments were made to validate the results by adding other covariates related to recipients, donors, and devices that were asymmetrically distributed between the study groups to the basic model.

Given the increased use of mechanical LV unloading strategies as an adjunct to VA-ECMO over time and the overall improvement in emergency HT outcomes in recent years, we conducted an exploratory analysis of 1-year post-HT survival in both study groups, stratified by temporal eras. This was done to rule out significant confounding effects of temporal changes on the observed associations. Temporal eras were defined according to historical modifications of the Spanish organ donor allocation protocol, as previously described.12 Era 1 spanned January 2010 to May 2014, Era 2 from June 2014 to May 2017, and Era 3 from June 2017 to December 2020. All statistical analyses were performed using SPSS version 25.

During the study period, 245 patients (184 men and 61 women) underwent high-emergency HT while being supported by VA-ECMO at the participating institutions. A mechanical strategy for LV unloading was used in 133 patients (54.3%), who formed the intervention group. The remaining 112 patients (45.7%) were supported with VA-ECMO alone, constituting the control group (figure 1).

Figure 1.

Central illustration. Summary of the main results of the study. 95%CI, 95% confidence interval; ECMO, extracorporeal membrane oxygenation; HR, hazard ratio; HT, heart transplantation. LV, left ventricular.

(0.41MB).

Mechanical methods used for LV unloading included IABP support (n=112), left ventricular apical venting (n=12), left atrial venting (n=9), atrial septostomy (n=4), pulmonary artery venting (n=2), temporary left ventricular assist device implantation (n=2), and other or unspecified methods (n=1).

Table 1 shows a comparison of preoperative clinical characteristics of patients bridged to HT on VA-ECMO with or without a mechanical strategy for LV unloading.

Ischemic heart disease and cardiogenic shock related to acute myocardial infarction were more common among patients with LV unloading while atrial fibrillation and previous defibrillator implantation were more common in patients without LV unloading.

Femoral cannulation was the predominant access for arterial cannulation in both study groups (n=199; 81.2%). Among patients in whom alternative arterial cannulation sites were used, central cannulation predominated in the LV unloading group, while subclavian/axillary cannulation predominated in the control group.

The LV unloading group had a higher percentage of patients treated in the most recent era, and a lower proportion of patients with a history of cardiac arrest or cardiac resynchronization therapy, although none of these differences reached statistical significance.

Table 2 compares the clinical status, laboratory tests, and hemodynamics of recipients before transplantation. Patients bridged with ECMO and LV unloading were more frequently intubated at the time of HT than those in the control group. No other significant differences were observed between the groups regarding vital supportive therapies, the rate of pretransplant infection, or the duration of ECMO support. Patients with LV unloading also had lower mean serum albumin and international normalized ratio values than those in the control group.

Preoperative hemodynamic parameters were measured in a small proportion of patients. Mean values of LV ejection fraction, LV end-diastolic diameter, pulmonary artery pressure and pulmonary capillary wedge pressure were significantly lower in the LV unloading group than in the control group.

The most relevant clinical characteristics of implanted heart donors are detailed in table 2. Mean cold ischemia times were significantly longer in the LV unloading group than in the control group (217.9±69.4 vs 197.0±73.6minutes; P=.032). The sex distribution and mean age of donors were comparable between the 2 groups..

Figure 1 and figure 2 compare major adverse clinical outcomes during the in-hospital postoperative period after HT in both study groups. The cumulative rates of postoperative graft failure and the need for mechanical circulatory support were 17.3% and 15.8% in patients bridged with LV unloading, compared with 23.2% and 24.1% in the control group, respectively. However, these differences did not reach statistical significance (P for postoperative graft failure=.240; P for postoperative need for mechanical circulatory support=.102). The cumulative rates of isolated postoperative right ventricular failure, excessive surgical bleeding, cardiac reoperation, postoperative infection, and postoperative renal failure were comparable between the 2 groups.

Figure 2.

Cumulative incidence of in-hospital postoperative adverse clinical outcomes after heart transplantation in the 2 study groups. LV, left ventricular.

(0.28MB).

Overall, 79 patients (32.2%) died during the first year after HT. Causes of death included infection (n=28), primary graft dysfunction (n=27), unspecified multiorgan failure (n=9), bleeding (n=8), acute rejection (n=2), stroke (n=2), renal failure (n=1), and respiratory failure (n=1). Kaplan-Meier survival curves are shown in figure 1 and figure 3. The cumulative 1-year posttransplant survival was 74.4% in the LV unloading group (34 deaths) and 59.8% in the control group (45 deaths). According to the univariate log-rank test, patients bridged to HT with LV unloading had significantly higher 1-year posttransplant survival than those in the control group (P=.025).

Figure 3.

Cumulative probability of 1-year posttransplant survival represented by Kaplan-Meier curves in the 2 study groups. LV, left ventricular.

(0.11MB).

Among male candidates, the 1-year posttransplant survival was 76.5% in those treated with LV unloading vs 62.8% in those without unloading (P=.042). Among female candidates, the 1-year posttransplant survival was 68.6% in those treated with LV unloading and 50% in those without unloading (P=0.142)

Multivariate backward stepwise Cox regression analyses identified 5 clinical factors associated with the risk of 1-year mortality following HT in the study cohort (table 3): history of previous stroke (adjusted HR, 2.47; 95% confidence interval [95%CI], 1.12-5.44; P=.025), cold ischemia time (adjusted HR, 1.30; 95%CI, 1.07-1.59; P=.010), pretransplant renal replacement therapy (adjusted HR, 2.14; 95%CI, 1.05–4.38; P=.037), pretransplant invasive mechanical ventilation (adjusted HR, 2.58; 95%CI, 1.38–4.83; P=.003) and pretransplant LV unloading (adjusted HR, 0.50; 95%CI, 0.32-0.78; P=.003). The strength of the association between pretransplant LV unloading and 1-year mortality was similar in men (adjusted HR 0.48, 95%CI 0.28–0.83) and women (adjusted HR 0.49; 95%CI 0.21-1.12); however, in the latter group, the association did not reach statistical significance, likely due to the limited sample size.

The inverse association between mechanical LV unloading and 1-year mortality was observed in both patients managed with an IABP alone (adjusted HR, 0.52; 95%CI, 0.32-0.84; P=.007) and those managed with other strategies for mechanical LV unloading (adjusted HR, 0.43; 95%CI, 0.19–0.97; P=.042).

Further multivariable adjustments were performed to assess potential confounding bias by adding other covariables related to temporal eras and characteristics of devices, recipients, and donors that were considered potential confounders. However, no relevant changes in the observed statistical associations were noted (table 4).

Figure 4 shows a stratified analysis of 1-year posttransplant survival rates in the study population according to temporal eras. In the whole cohort, 1-year posttransplant survival increased from 63.9% in Era 1 to 66.3% in Era 2 and 73.9% in Era 3; however, this trend did not reach statistical significance (P=.240).

Figure 4.

Stratified analysis of 1-year posttransplant survival rates in the study population according to temporal eras. LV, left ventricular.

(0.12MB).

The 1-year posttransplant survival rates of patients bridged with LV unloading increased significantly over time (Era 1=63.4%; Era 2=75%; Era 3=82.7%; P=.036) and were higher than those of the control group, which remained unchanged (Era 1=64.3%; Era 2=57.5%; Era 3=56.7%; P=.496).

This study has several limitations. As a retrospective investigation, it may be subject to various sources of bias, including selection, information, and confounding biases. Therefore, its results should be considered hypothesis-generating only. A strength of this investigation is its inclusion of all activity from Spanish centers that maintained an active adult HT program during the evaluated period. However, this also means there may be potential differences in selection protocols and therapeutic management among centers that were not controlled for in the analysis, and clinical events were adjudicated by local investigators rather than by an independent committee. While our results can be directly applied to Spain, caution is warranted when applying them to other countries with organ-sharing donor systems that may differ from the Spanish system.

We acknowledge that the characterization of the study population could have been improved if additional important clinical variables, such as the vasoactive-inotropic score or the Society of Cardiovascular Angiography and Intervention stages, had been collected; however, this information was not available for the study. Moreover, the number of missing values in hemodynamic variables was too high to allow us to draw reliable conclusions. Additionally, we cannot provide specific information regarding the type and duration of posttransplant mechanical circulatory support for those patients who required it, so no conclusions can be drawn about its potential impact on study outcomes.

Finally, while significant effort has been made to adjust for the most relevant potential confounders, some variables that could affect these differences may be missing. Despite this, we have collected a large number of variables to minimize the influence of possible confounding factors.