Cases of heart transplantation (HT) bridged with mechanical circulatory support (MCS) have increased globally.1) Most MCS strategies as a bridge to HT involve the use of durable left ventricular assist devices (LVADs). However, durable LVAD implantation is not generally performed for patients with significant cardiogenic shock, who are typically first treated with temporary MCS.2)3) Veno-arterial (VA) extracorporeal membrane oxygenation (ECMO) is now one of the most popular temporary MCS approaches and is also used as a bridge to HT.

The use of VA-ECMO as a bridge to HT varies in different countries according to their donor heart utilization and allocation systems. It is known that the incidence of VA-ECMO-bridged HT is 1–5% in the United States,4) 5–10% in several European countries,5)6) and around 20% in Italy.7) In particular, the rate of VA-ECMO bridging is high in Korea, accounting for 25% of all transplanted patients,8) which may be explained by the highest priority given to ECMO in the allocation system, the strict governmental regulation of access to LVADs, and the relatively short waiting period for transplants.9)10) However, despite the widespread practice of VA-ECMO-bridged HT, the success of HT and post-transplant survival outcomes among patients bridged with ECMO have been reported to be low.5)11)12)13)14)

Prolonged VA-ECMO support may jeopardize post-transplant outcomes because of certain intrinsic complications. Peripheral-type VA-ECMO is relatively easy to apply to hemodynamically unstable patients; however, it has disadvantages such as an increased left ventricular (LV) afterload, limb ischemia, bleeding or thrombosis associated with the use of cannulas, and poor mobilization. Because of these non-physiologic effects, peripheral VA-ECMO is regarded as a short-term support option only. Central-type VA-ECMO, in which outflow from the pump directly returns to the ascending aorta, allows antegrade perfusion to end-organs with reduced LV loading, long-standing support with large bore cannulas, and sufficient patient mobilization.15) However, central-type VA-ECMO also has some disadvantages, such as a high bleeding risk and requirement for sternotomy to insert or remove the cannulas.16)

Thus far, the data regarding the optimal cannulation strategy for VA-ECMO are limited, especially for patients listed for HT. Most studies comparing the 2 types of VA-ECMO have been confined to cases of post-cardiotomy shock.15)17) In clinical settings, the VA-ECMO configuration type is determined at the discretion of the clinician with consideration of the type of ECMO that can keep the patient safe until HT. However, the effect of the VA-ECMO configuration on subsequent HT is still largely unknown. Therefore, in this study, a comparison was performed between central and peripheral VA-ECMO to assess their clinical implications in patients when used as a bridge to transplantation.

We identified 149 consecutive patients with a status 0 medical urgency who were on the KONOS HT waiting list between January 2015 and March 2020. Among these patients, we excluded 24 pediatric patients under the age of 18, 8 patients who had been weaned off ECMO and then re-evaluated as having a low urgency status due to an improvement in their cardiac function, and 6 patients who had been registered as status 0 for reasons other than VA-ECMO support, such as requiring continuous mechanical ventilation due to heart failure or serious complications of durable VAD requiring ICU admission. Finally, we analyzed a study cohort of 111 adult patients (22.5% female; mean age, 51.3±13.6 years) with status 0 due to a requirement for VA-ECMO support.

Peripheral VA-ECMO was used in 76 (68.5%) of our study patients, central VA-ECMO was used in 12 (10.8%) cases, and the conversion group contained 23 subjects (20.7%). The baseline characteristics among the 3 groups, other than the underlying etiology who received ECMO support (Table 1), showed no statistically significant differences. Extracorporeal Cardio-Pulmonary resuscitation was not required among the central group cases, and was not significantly different between the other 2 groups (peripheral 38.2%, central 0.0%, conversion 34.8%, p=0.005). The rate of ECMO support after cardiac surgery was higher in the central group (peripheral 11.8%, central 50.5%, conversion 8.7%, p=0.006).

Our analysis suggests that the likelihood of undergoing HT is higher when patients receive either peripheral or central VA-ECMO support without subsequent changes to this configuration. Notably however, the survival outcomes were found to be comparable between these 2 groups and the cases in our study that had undergone an ECMO conversion. Complication rates were also similar between the 2 ECMO types except for a higher incidence of surgical site bleeding in the central type and a significant occurrence of limb ischemia in the peripheral type cases. To the best of our knowledge, this is the first study to evaluate the prognostic differences between the central and peripheral VA-ECMO configurations of patients waiting for HT.

ECMO-bridged HT was not widely performed in the past and was only used for around 1% of patients awaiting surgery.1)18) Most patients that have undergone VA-ECMO were classified as INTERMACS 1 or 2, and the accumulated evidence suggests that outcomes are poorer in these cases than in those with a higher grade based on this scoring system.19) Nevertheless, the use of VA-ECMO prior to HT has been growing, and many countries now give a higher priority for HT to transplant candidates on this bridging treatment than to those on long-term MCS without complications. Since changes to the United Network for Organ Sharing donor allocation system in the United States were made in the fall of 2018, patients on VA-ECMO have had a higher priority for transplantation.20)21) Some countries, including Korea, have regulated the transition of ECMO to durable LVAD implantation or other temporary MCS interventions such as Impella or CentriMag due to economic burden on the national health insurance system. In addition, there are a considerable number of patients awaiting HT with biventricular failure, who cannot be fully supported by LVADs alone. Therefore, ECMO bridging to transplantation is becoming more popular worldwide.

Nevertheless, VA-ECMO is a bridge therapy; thus, this treatment may be administered in a limited timeframe. Recent registry data indicate that survival rates after VA-ECMO decrease over time, and most recipients have been weaned off or have died by 3 weeks after its introduction.22) This may be attributed to an increased risk of ECMO complications with cumulative exposure. In our present study, the incidence of complications according to the ECMO type was largely similar, except for limb ischemia and surgical site bleeding. These 2 adverse events were found to arise in the early phase of VA-ECMO, whereas other complications manifested at different times. Hence, greater attention should be paid in the future to the duration of VA-ECMO based on its type.

In comparison with patients receiving peripheral ECMO support, patients receiving central ECMO support required significantly more surgeries to control bleeding complications. They also required more red blood cell and platelet concentrate transfusions; however, these differences were not statistically significant. The results are consistent with the findings of a previous meta-analysis on central ECMO. The reoperation rates of the studies included in the review ranged from 11% to 64%.15) Among the central ECMO patients in our study, the reoperation rate was 28.6%, indicating a relatively lower risk of bleeding. However, even though the risk of reoperation or transfusion was higher in cases of central ECMO than in cases of peripheral ECMO, there was no significant difference in the in-hospital mortality rate. Therefore, although careful monitoring for reoperation and meticulous hemostasis may be important, the decision to perform central ECMO surgery, if needed, should not be delayed due to fears about bleeding complications.

A total of 23 patients underwent a change in their ECMO configuration. The characteristics of these patients should be reviewed considering the lack of relevant data. By changing from a peripheral to a central ECMO configuration, it has been shown previously that the incidence of pulmonary edema due to elevated LV end-diastolic pressure is reduced,23) and the extremities become slightly more flexible. These effects would more readily facilitate awake ECMO and bedside rehabilitation for patients awaiting HT. Awake ECMO and the early physical rehabilitation of patients in the ICU are well-known to reduce the frequency of adverse clinical events such as ventilator-associated pneumonia, long-term use of antibiotics, and renal replacement therapy,21)24) which may improve the safety and longevity of VA-ECMO support. Despite these established benefits of central conversion, we found that the cumulative incidence of HT was significantly lower among patients with a change in the ECMO configuration (conversion group) than among peripheral or central patients without any configuration change (peripheral group, central group) (Figure 1). Even though we speculate that the poor outcome may be attributed to the urgent conversion of the ECMO type from peripheral to central, there is a limitation in identifying the exact cause of prolonged waiting time. Additionally, it is difficult to determine whether the extended waiting time was solely caused by the deterioration of the patient’s clinical condition or a shortage of suitable donors. Further research and analysis are needed to address this issue.

This study had several limitations. First, this was a retrospective observational study, and the number of study subjects was small. It is possible that there may be some biases in the conversion group, such as protopathic bias, where central conversion attempts to overcome a patient's deteriorating condition may actually increase waiting times for transplantation and lead to unfavorable outcomes, selection bias, where patients with pre-existing poor health may be more likely to undergo conversion due to rapid deterioration, or immortal bias, where patients who survive to conversion may be selectively included. As this is not a randomized trial, there may be limitations in minimizing these biases. However, considering that the conversion group did not show a significant difference in baseline characteristics and that we performed a landmark analysis to adjust for immortal bias, we can cautiously interpret that the result of this study may provide some clinical guidance in determining ECMO configurations for HT-waiting patients. All of factors may have influenced the observed clinical outcomes, and further confirmation of our results in a larger population by randomization is needed. Second, the criteria used to decide on central VA-ECMO were not predetermined, and the decision was made at the discretion of the attending physicians. Third, the lack of information on donor availability during the waiting period and the decision to decline transplantation due to the deterioration of the recipient’s clinical status are limiting factors in fully capturing the multifactorial nature of prolonged waiting times. Fourth, the central VA-ECMO configurations were not analyzed in detail. In most of the central VA-ECMO cases in our current study, a drainage cannula was placed in the left atrial loop or LV apex, and a return cannula was inserted into the ascending aorta. However, we did not perform any comparison according to the cannula position.

In conclusions, we observed no significant difference in the incidence of HT between the peripheral and central groups, although we did note a trend toward lower HT in the conversion group. Therefore, we suggest that if patients are being stably supported with their initial ECMO configuration, whether it is central or peripheral, it should be maintained, and ECMO conversion should only be cautiously performed when necessary.