Perioperative extracorporeal membrane oxygenation in neonates with transposition of the great arteries: 15 years of experience

Abstract

OBJECTIVES

Extracorporeal membrane oxygenation (ECMO) can act as a bridge to recovery in both pre- and postoperative patients with transposition of the great arteries (TGA). However, literature on its use in these patients is scarce.

METHODS

Retrospective single-centre cohort study encompassing all TGA patients who received ECMO between January 2009 and March 2024.

RESULTS

Twenty-two neonates received ECMO during the study period, with an overall median age and weight at time of ECMO cannulation of 6.5 (1.8–10) days and 3.7 (3.3–4.0) kg, respectively. Twelve neonates received ECMO prior to the arterial switch operation because of severe persistent pulmonary hypertension (83%), respiratory failure due to severe pulmonary atelectasis (8%) or hypoxia after pulmonary arterial banding procedure (8%). Postoperative ECMO was used in 11 patients; of these, 1 (9%) had also received ECMO preoperatively. Postoperative indications for the remaining patients included failure to wean from cardiopulmonary bypass (50%), low cardiac output in Intensive Care Unit (20%), or after cardiopulmonary arrest (30%). Overall, median ECMO duration for all TGA patients was 75 (45–171) h, with a survival rate of 59% to hospital discharge. Among the preoperative ECMO patients, 5 patients (42%) died (4 preoperatively, 1 postoperatively performed while on ECMO). In the postoperative ECMO group, survival rate was 60%.

CONCLUSIONS

In this single-centre retrospective study, TGA neonates received ECMO preoperatively primarily for severe pulmonary hypertension and postoperatively for failure to wean from cardiopulmonary bypass. This study showed a 58% and 60% survival to hospital discharge in ECMO patients supported preoperatively and those supported postoperatively, respectively.

Open in new tabDownload slide

INTRODUCTION

The use of extracorporeal membrane oxygenation (ECMO) can be an effective method of mechanical circulatory support for the management of severe pulmonary or cardiopulmonary failure [1]. Over the years, ECMO has evolved as an essential treatment available for children with congenital heart disease (CHD), such as transposition of the great arteries (TGA). TGA is characterized by concordant atrioventricular and discordant ventriculo-arterial connection, leading to a circulation in which the pulmonary and systemic circulation run in parallel rather than in series. The presence of an adequate communication between the 2 circulations [i.e. atrial septal defect or ventricular septal defect (VSD)] to allow mixing, along with a smooth foetal-to-neonatal transition involving a physiologic decrease of pulmonary vascular resistance, is crucial for adequate arterial saturation after birth [2]. However, neonates with TGA are at risk of developing pulmonary hypertension of the newborn (PPHN) after birth, especially those with an intact ventricular septum (TGA-IVS). In cases with severe PPHN or refractory cardiorespiratory failure, ECMO can offer a bridge to recovery while ensuring adequate systemic blood supply [3, 4].

In general, ECMO fulfils 3 primary purposes for children with cardiopulmonary failure: serving as a bridge to recovery, decision-making or therapeutic intervention [3, 4]. For TGA-patients, indications include preoperative stabilization (e.g. TGA with PPHN and subsequent hypoxia), failure to wean from cardiopulmonary bypass (CPB), low cardiac output in the Intensive Care Unit (ICU) postoperatively or after cardiopulmonary resuscitation (CPR) [5, 6]. As there are no randomized controlled trials in neonates with CHD and ECMO, there are no evidence-based guidelines for ECMO initiation, and our knowledge of ECMO support in these patients depends solely on retrospective single-centre and Extracorporeal Life Support Organization (ELSO) registry studies. Characterization, outcomes and complications of the use of ECMO in TGA patients have not been widely studied [7]. Therefore, the aim of this study is to describe our experience and outcomes in neonates with TGA supported perioperatively with ECMO over the past 15 years.

MATERIALS AND METHODS

Study population

This retrospective cohort study conducted at Leiden University Medical Centre includes all live-born neonates diagnosed with TGA who received ECMO support either before or after an arterial switch operation (ASO) between January 2009 and March 2024. Patients were either TGA-IVS, TGA with VSD or double-outlet right ventricle with subpulmonary VSD (Taussig-Bing anomaly). Coronary anatomy was scored as a usual coronary artery pattern (1LCx-2R and 1 l-2CxR) or alternate anatomies according to the Leiden Convention coronary coding system [8]. ECMO patients were divided into 2 subgroups: pre-arterial switch operation (pre-ASO) and post-arterial switch operation (post-ASO). Variables such as demographic data, gender, gestational age, morphological diagnosis, birth weight, surgical- and ECMO characteristics (procedural time, ECMO type and vascular access), haemodynamic course, extracorporeal CPR, ECMO-related complications and survival were obtained from the hospital records. Complications related to ECMO were assessed using complication codes established by the ELSO registry [4, 9]. The 1 patient who received both pre- and post-ASO ECMO support was assessed exclusively within pre-ASO group regarding the outcome parameters.

Extracorporeal membrane oxygenation approach in transposition of the great arteries

The timing to initiate ECMO remains challenging, yet it should be initiated before signs of severe oxygen deficiency, organ damage or cardiac arrest in accordance with the guidelines outlined by the ELSO [10]. Veno-arterial ECMO (V-A ECMO) is the main support mode for ECMO in CHD and augments systemic cardiac output and respiratory gas exchange to facilitate adequate tissue oxygen delivery [3, 4]. V-A ECMO can be performed via peripheral cannulation (most commonly drainage via right internal jugular vein, return via right common carotid artery) or central cannulation (most commonly drainage via right atrium, return via ascending aorta) [10]. In our centre, peripheral V-A ECMO cannulation is typically performed in TGA-patients pre-ASO, whereas central cannulation is preferred for post-ASO patients. Desired ECMO flow is 100–150 ml/kg/min or 3 l/m²/min, titrated to achieve adequate oxygen delivery and lactate clearance. In case of significant cardiac failure with myocardial stunning and cardiac dilatation on ECMO, an extra cannula in the left atrium (i.e. left vent) or balloon atrial septostomy may be required to provide left ventricular decompression [11]. The approach in our institution is generally to wean infants from ECMO prior to the ASO. Weaning from ECMO is a complex process and requires careful assessment and timing, but can be initiated with signs of myocardial recovery and adequate resolution of systemic inflammatory response or pulmonary problems as per ELSO guidelines [12]. We generally start weaning ECMO-flow rates when lactate has normalized, pulse pressure is at least 15 mmHg, echocardiography suggests normalization of myocardial function and/or an RV-pressure is less than half systemic in children with PPHN, and pulmonary function seems adequate to allow for decannulation. If weaning is not successful despite several attempts, emergency ASO during ECMO support can be considered.

Data management and analysis

Descriptive statistical analysis was performed using IBM SPSS (version 29; IBM SPSS Inc, Chicago, USA). All analyses were conducted using complete case analysis approach. Demographic and clinical data regarding procedure and complications were presented as frequency with percentages for categorical variables and mean with standard deviation or median with interquartile range for continuous data.

Ethics

Approval was obtained from the appropriate ethics committee of the Leiden University Medical Centre (No. 2023-028). Committee waived the need for individual written informed consent.

RESULTS

Patient and extracorporeal membrane oxygenation characteristics

During the study period, 163 TGA patients were admitted to our centre, of whom 25 patients had PPHN (25/163; 15%) pre-ASO. All neonates with PPHN had TGA-IVS. Overall, 22 neonates received ECMO support (22/163; 13%), of whom (12/163; 7.4%) were assisted with ECMO before ASO (pre-ASO group). Four patients were deemed ineligible for ASO and died preoperatively without receiving ECMO (Supplementary Material, Table S1). Among them, a premature infant (gestational age 28 + 4 weeks, birth weight 1230 g) with a known prenatal TGA diagnosis died from multi-organ failure secondary to severe sepsis. The other 3 infants were diagnosed with TGA after birth and included the following: 3.5-month-old infant who died due to circulatory shock, compounded by progressive kidney injury and a middle cerebral artery infarction following delayed diagnosis; 2 infants who suffered severe neurological complications. One experienced a large subdural haematoma and extensive cerebral ischaemia, while the other had a severe genetic infantile motor neuron disease (AIFM1 mutation). One hundred fifty-five patients underwent an ASO, of whom 11 (7.1%) patients received ECMO support postoperatively (post-ASO group). There was 1 patient who received ECMO both before and after ASO and was included in the pre-ASO group. Baseline patient characteristics are outlined in Table 1. Overall, median age and weight at ECMO cannulation were 6.5 (interquartile range 1.8–10) days and 3.7 (interquartile range 3.2–4.0) kg, respectively. Morphological diagnosis consisted of TGA-IVS in 17 (77%), TGA with VSD in 1 (4.5%) and Taussig-Bing anomaly in 4 (18%) patients. All patients underwent V-A ECMO.

The main indication for TGA-patients with ECMO pre-ASO was PPHN (n = 10), with respiratory insufficiency due to severe pulmonary atelectasis (n = 1) or hypoxemia after pulmonary arterial banding (n = 1) as other indications. The incidence of ECMO in children with PPHN was 40% (10/25) with a mortality rate of 30% (3/10). All patients in the pre-ASO group were cannulated peripherally. Median weight, age and ECMO duration pre-ASO were 3.7 (3.3–4.2) kg, 2 (1.0–2.8) days and 83 (51–171) h, respectively (Table 1). Indications for ECMO after ASO were failure to wean off CPB (n = 5), low cardiac output syndrome in the ICU (n = 2) or after CPR in the ICU (n = 3). All patients were cannulated centrally, including the 1 patient who was also supported peripherally pre-ASO. Median weight, age and ECMO duration post-ASO were 3.5 (3.1–4.1) kg, 10 (8.5–16.0) days and 58 (30–191) h, respectively.

Outcomes

Overall survival to hospital discharge in the ECMO group was 59% with a median length of ICU stay of 26 (17–48) days. Supplementary Material, Tables S1 and S2 provide demographic comparison between the entire TGA-cohort and the pre-and post-ASO ECMO groups. In the pre-ASO group, 8/12 (67%) were weaned of ECMO and 7/8 (88%) later underwent ASO (Fig. 1). Three patients died on ECMO and 1 patient could not be weaned off ECMO and underwent ASO on ECMO, continued ECMO support post-ASO and died thereafter (Fig. 1). In the pre-ASO ECMO group, survival rate was 7/12 (58%).

In the post-ASO ECMO group, survival rate was 6/10 (60%). In the post-ASO ECMO group, 3/10 (30%) patients underwent reoperation while on ECMO for various indications: coronary revision due to severely impaired left ventricular function in 1 patient, re-exploration due to bleeding of the neo-pulmonary artery wall with clot formation causing pressure on the right coronary artery in another patient and thrombectomy in the coronary anastomosis after (modified) Yacoub aortocoronary flap technique for single sinus coronary artery with intramural LAD (2 l*CxR), causing myocardial ischaemia with subsequent mitral valve repair in a 3rd patient. Only 1 survived to hospital discharge. No patients died post-ASO without having received ECMO (Supplementary Material, Table S2).

Comparison of demographics between survivors and non-survivors are demonstrated in Table 2. In the survivor group, 46% of patients had no complications while on ECMO (Fig. 2); in the non-survivor group, 56% had complications in 3 or more organ domains. Complications across organ domains are outlined in Supplementary Material, Table S3. Non-survivors had more renal complications, as all neonates subjected to peritoneal dialysis or continuous veno-venous haemofiltration (4/22, 18%) were confined to the non-survivor group. There was a positive linear association between the duration of ECMO support and the number of ECMO complications across various complication categories (Supplementary Material, Fig. S1).

DISCUSSION

Over the years, ECMO has evolved as an essential support modality for neonates with TGA experiencing severe cardiorespiratory failure. ECMO can act as a bridge to recovery in both pre- and post-ASO TGA-management. This single-centre retrospective study showed a 58–60% hospital survival in neonates supported with ECMO before and/or after ASO, respectively.

Pre-arterial switch operation extracorporeal membrane oxygenation support

PPHN was the most frequently observed indication for pre-ASO ECMO support. Previous reports have associated TGA with PPHN, with incidence rates up to 21.5% [13, 14]. Similarly, the incidence of PPHN in our entire cohort of TGA patients was 15% (25/163), of which 10 (40%) received ECMO. As in our cohort, PPHN was predominantly observed in the subtype of TGA patients with intact ventricular septum [13, 14]. Exact causes for severe PPHN in these patients have not been elucidated so far. In literature, suggestions have been made indicating a complex multifactorial aetiology. Contributing factors might include: (i) hypoxemia and acidosis during the intrauterine or the early postnatal period; (ii) restrictive oval fossa or foetal ductal constriction resulting from increased oxygen content in the pulmonary artery in foetuses with TGA; (iii) increased pulmonary artery wall thickness and muscularity with intimal proliferation; (iv) mediating pathways such as endothelin-1, prostacyclin-cGMP, nitric oxide-cAMP and vascular endothelial growth factor; and/or (v) clinically unrecognized pulmonary microthrombi [13, 15–17].

Conventional treatment of PPHN includes adequate sedation and analgesia, mechanical ventilation, inhaled nitric oxide and haemodynamic support with fluids and inotropes including milrinone and/or balloon atrial septostomy. Our approach has not changed significantly over the course of the study period. In TGA-patients with PPHN, resistant to conventional treatment, ECMO can be used as cardiopulmonary support. Nevertheless, PPHN in neonates with TGA has been associated with significant morbidity and preoperative mortality rates up to 29% in series with and without ECMO use [13]. In a previous report of PPHN in TGA-patients by Sallaam et al. [14], 45% underwent ECMO support pre-ASO, with a mortality rate of 44% within this ECMO subgroup. This is similar to our cohort in which 40% of neonates with PPHN received ECMO with a mortality rate of 40%.

Pre-ASO ECMO in infants with PPHN is most often used as a bridge to ASO, yet the optimal timing of ASO in relation to ECMO remains unclear.

During ASO, CPB and stress of surgery and anaesthesia may cause a further increase of pulmonary vascular resistance, as well as a decrease of myocardial function, potentially leading to insufficient cardiopulmonary function postoperatively [14]. Therefore, the approach in our centre is to wean patients of ECMO prior to ASO as also advocated by Jaillard et al. [5], allowing for gradual reconditioning of the myocardium to prevent postoperative deterioration related to recurrent PPHN. However, other studies suggest to perform ‘rescue ASO’ while on ECMO to maintain end-organ function [18]. In our cohort, all but 1 patient was successfully weaned off ECMO before surgery. All our pre-ASO ECMO patients were supported with V-A ECMO and cannulated peripherally. V-venous ECMO has also successfully been described [19].

Post-arterial switch operation extracorporeal membrane oxygenation support

The ELSO guidelines emphasize the importance of timely ECMO indication postoperatively [3]. According to data from the Congenital Heart Surgery Database of the Society of Thoracic Surgery, the usage of ECMO support in neonates following ASO varies between 3.7% and 7.5%, depending on whether VSD repair was conducted (7.5%) or not (3.7%) [20]. In our cohort, 7.1% of all ASO patients received post-ASO ECMO, most of them (50%) because of failure to wean from CPB. All were cannulated centrally. Failure to wean from CPB may be influenced by several factors such as residual PPHN, inadequate myocardial function, or technical issues associated to the surgical procedure [21]. Causes for low cardiac output post-CPB are myocardial ischaemia (typically caused by technical problems of coronary artery transfer) or the inability of the left ventricle to adapt to systemic pressures (often referred to as a detrained left ventricle) post-ASO. Determining the exact timing for ECMO initiation after surgical repair of CHD in general remains challenging, with conflicting findings on hospital survival [22]. While some studies report improved survival rates when ECMO is initiated in the operating room, others indicate no disparity in survival based on the timing of ECMO initiation postoperatively [21].

Timely diagnosis and intervention of unexpected residual lesions are associated with improved outcomes in post-cardiotomy ECMO patients in general [23, 24]. Yang et al. [7] identified coronary abnormalities, such as intramural, solitary coronary artery and mono-coronary ostium, as influential factors on the outcomes of TGA-patients undergoing ECMO in the intraoperative and postoperative subgroup. In our series, 60% of the patients receiving post-ASO ECMO support exhibited a variant in coronary artery anatomy. This finding potentially reinforces the hypothesis that non-typical coronary anatomy contributes to increased surgical complexity.

Outcomes

In our cohort, survival to hospital discharge was 59% and consistent in both pre-ASO (58%) and post-ASO (60%) ECMO patients. The average survival for neonates with all types of CHD supported with ECMO is 40–51% as reported by the ELSO registry database [3].

Various predictors for mortality on ECMO after surgery for CHD have previously been identified in literature and include young age, low weight, high inotrope score, high lactate, acidosis, duration of ventilation, presence of fluid overload, renal failure, dialysis, cardiac catheterization, CPR requirement and ECMO duration. In this cohort, we observed that non-survivors had more renal complications, compared to survivors. The association between ECMO, renal complications and mortality has been previously recognized; therefore, renal failure could serve as an indicator of poor prognosis while on ECMO support [25].

The timing and duration of ECMO play an important role, as prolonged ECMO duration has been associated with increased complications, such as renal failure, bleeding, thrombosis and infection, adversely affecting survival [26]. Based on the ELSO registry study, the overall survival rate of paediatric cardiac ECMO runs between 2000 and 2011 was 45%, but survival decreased to 23–25% for ECMO durations between 14 and 28 days, and further dropped to 13% for ECMO runs lasting longer than 28 days [27]. Consistent with these findings, we also observed a higher complication rate and involvement of multiple organs systems with prolonged ECMO duration. However, due to the cohort’s limited sample size and statistical power, we were unable to identify a definitive relationship between ECMO duration and outcomes.

Limitations

This study is subject to the limitations inherent to a retrospective single centre design, despite having no missing data. The small sample size limits statistical power leading us to provide only descriptive statistics. Additionally, the study period of more than a decade, along with technological advances in ECMO equipment and our increased experience with postcardiotomy ECMO, could have influenced ECMO-initiation thresholds and outcomes. In the first 5 years of the study period, we only supported 3 TGA patients with ECMO possibly indicative of our increased threshold at the start of our ECMO program, but in the last 10 years, we experienced a relatively steady number of annual ECMO runs for patients with TGA, suggesting a well-balanced approach to these specific patients. Unfortunately, the cohort is too small to analyse this effect of time on outcomes. However, for postcardiotomy, ECMO in general outcomes have also been relatively stable over the last 15 years according to the ELSO registry [4].

CONCLUSION

This single-centre retrospective study over a 15-year period of ECMO in neonates with TGA, showed a 58 and 60% survival to hospital discharge in patients supported pre-ASO and those supported post-ASO, respectively. Most pre-ASO patients received ECMO due to PPHN and post-ASO patients due to failure to wean off CPB.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

FUNDING

This work was funded by the Hartekind Foundation.

Conflict of interest: none declared.

DATA AVAILABILITY

Databases will be made available from the corresponding author upon reasonable request.

Author contributions

Jesse A. Weeda: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Resources; Software; Supervision; Validation; Writing—original draft; Writing—review & editing. Roel L.F. van der Palen: Conceptualization; Funding acquisition; Investigation; Methodology; Resources; Supervision; Validation; Writing—review & editing. Heleen E. Bunker-Wiersma: Investigation; Supervision; Validation; Writing—review & editing. Lena Koers: Investigation; Supervision; Validation; Writing—review & editing. Eelco van Es: Investigation; Supervision; Validation; Writing—review & editing. Mark G. Hazekamp: Investigation; Supervision; Validation; Writing—review & editing. Arjan B. te Pas: Investigation; Supervision; Validation; Writing—review & editing. Peter Paul Roeleveld: Conceptualization; Data curation; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Massimiliano Meineri, Aditya Patukale, Alister Joseph Thomas and the other anonymous reviewers for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• ASO

  Arterial switch operation

 
• CHD

  Congenital heart disease

 
• CPB

  Cardiopulmonary bypass

 
• CPR

  Cardiopulmonary resuscitation

 
• ECMO

  Extracorporeal membrane oxygenation

 
• ELSO

  Extracorporeal Life Support Organization

 
• ICU

  Intensive care unit

 
• PPHN

  Persistent pulmonary hypertension of the newborn

 
• pre-ASO

  Pre-arterial switch operation

 
• TGA

  Transposition of the great arteries

 
• TGA-IVS

  TGA with intact ventricular septum

 
• V-A ECMO

  Veno-arterial ECMO