Outcome after extracorporeal membrane oxygenation therapy in Norwood patients before the bidirectional Glenn operation

Abstract

OBJECTIVES

Patients after the Norwood procedure are prone to postoperative instability. Extracorporeal membrane oxygenation (ECMO) can help to overcome short-term organ failure. This retrospective single-centre study examines ECMO weaning, hospital discharge and long-term survival after ECMO therapy between Norwood and bidirectional Glenn palliation as well as risk factors for mortality.

METHODS

In our institution, over 450 Norwood procedures have been performed. Since the introduction of ECMO therapy, 306 Norwood operations took place between 2007 and 2022, involving ECMO in 59 cases before bidirectional Glenn. In 48.3% of cases, ECMO was initiated intraoperatively post-Norwood. Patient outcomes were tracked and mortality risk factors were analysed using uni- and multivariable testing.

RESULTS

ECMO therapy after Norwood (median duration: 5 days; range 0–17 days) saw 31.0% installed under CPR. Weaning was achieved in 46 children (78.0%), with 55.9% discharged home after a median of 45 (36–66) days. Late death occurred in 3 patients after 27, 234 and 1541 days. Currently, 30 children are in a median 4.8 year (3.4–7.7) follow-up. At the time of inquiry, 1 patient awaits bidirectional Glenn, 6 are at stage II palliation, Fontan was completed in 22 and 1 was lost to follow-up post-Norwood. Risk factor analysis revealed dialysis (P < 0.001), cerebral lesions (P = 0.026), longer ECMO duration (P = 0.002), cardiac indication and lower body weight (P = 0.038) as mortality-increasing factors. The 10-year mortality probability after ECMO therapy was 48.5% (95% CI 36.5–62.9%).

CONCLUSIONS

ECMO therapy in critically ill patients after the Norwood operation may significantly improve survival of a patient cohort otherwise forfeited and give the opportunity for successful future-stage operations.

Open in new tabDownload slide

INTRODUCTION

Since the invention of extracorporeal membrane oxygenation (ECMO) therapy by Dr Robert H. Bartlett, its use has steadily risen, becoming a well-established method for both paediatric respiratory and cardiac failure [1]. The latest Extracorporeal Life Support Organization (ELSO) registry dashboard reports a total survival to discharge rate of 52% for paediatric patients and 61% for neonates, regardless of ECMO indication (https://www.elso.org/registry/elsoliveregistrydashboard.aspx) [2]. Outcome of patients with single-ventricle physiology, especially after stage 1 palliation receiving ECMO therapy, varies in the literature between individual centres and tended to improve over the years with hospital survival probabilities of 31–61% [3–5]. In this study, we report our experience with ECMO therapy in single-ventricle patients after the Norwood procedure and before the bidirectional Glenn (BDG) operation. Our investigation aims to determine risk factors associated with in-hospital mortality in this particular patient cohort, pinpoint subgroups at risk and compare our results with previously published data.

MATERIALS AND METHODS

We collected retrospective data from the hospital, perfusion and operative reports, paediatric intensive care data and outpatient records to compile demographic, operative and postoperative information. Informed consent was waived, given the study’s retrospective nature, and our study received approval from the local ethics committee (EK Nr: 1083/2023).

In our institution, over 450 Norwood operations were performed since 1997 of which 306 patients were operated since 2007 when ECMO therapy became available in our hospital. In the timespan from 2007 to 2022, ECMO therapy was initiated after the Norwood and before the BDG operation in 59 patients, who present our study cohort.

Patients were grouped according to the following 4 main anatomic features: aortic atresia (26 patiens; 44.1%), aortic stenosis (15 patiens, 25.4%), unbalanced complete atrioventricular canal (9 patiens, 15.3%) or classic single ventricle (double inlet left ventricle) or systemic left ventricle with tricuspid atresia and transposition of the great arteries with an hypoplastic aorta arising from a subaortic chamber (9 patiens 15.3%).

ECMO initiation indication was classified as either pulmonary or cardiac failure. Patients were indexed as pulmonary failure if oxygen saturations were inadequate, and or a low pO₂ was the reason for ECMO support despite of initially good haemodynamics. Cardiac failure referred to hypotension with clinically evident low cardiac output and/or progressive decline in cardiac function despite optimal inotropic support, without impaired oxygenation, and/or arrhythmias causing haemodynamic instability. ECMO therapy was initiated for cardiac failure if mean arterial blood pressure could only be maintained above 30 mmHg with perfusion pump rates of epinephrine >0.3 μg/kg/min, dobutamine >10 μg/kg/min and norepinephrine >0.15 μg/kg/min. Other indications included the need for recurrent bolus applications and increasing lactate levels despite euvolaemia and no trend improvement. Criteria for pulmonary failure were met if blood gas analysis showed a paO₂ of <35 mmHg or saturations <65%, despite an unobstructed airway, absence of atelectasis, ventilation with nitric oxide and concomitant rising lactate levels. The decision to initiate ECMO therapy was jointly made by a team comprising a paediatric cardiac surgeon and a paediatric intensivist or anaesthesiologist and the same criteria for ECMO initiation were applied in the operating theatre such as on the intensive care unit (ICU). The Norwood operation was performed under full-body perfusion. This was achieved by sewing a 3.5-mm polytetrafluoroethylene (PTFE) prosthesis to the brachiocephalic artery to ensure upper body perfusion during aortic arch reconstruction. Additionally, a 2nd arterial cannula was placed directly above the diaphragm into the descending aorta to ensure perfusion of abdominal organs and lower extremities. Aortic arch reconstruction involved an extended end-to-end anastomosis of the descending aorta and distal aortic arch, with a curved patch enlargement using a pulmonary bifurcation homograft or curved PTFE prosthesis. The methodic details of this reconstruction and the double-arterial cannulation technique have been previously published [6–8]. Pulmonary perfusion was established using a 5-mm PTFE right ventricle to pulmonary artery conduit (Sano conduit) in 92% of patients, and a 3.5-mm Blalock–Thomas–Taussig shunt (BTTS) was implanted in 5 patients due to anatomical reasons. Sano conduits were always left open in contrast to BTTS, which were always clipped during ECMO therapy.

Extracorporeal membrane oxygenation cannulation strategy

We used veno-arterial ECMO therapy in all cases. For neonates and children under 4 weeks post-chest closure, central cannulation was performed using either the existing PTFE prosthesis at the brachiocephalic artery or directly the neoaorta, with venous cannulation via the right atrial appendage. If ECMO indication arose later than 4 weeks post-chest closure, cannulation was done via a PTFE prosthesis anastomosed to the right carotid artery and a venous cannula directly into the right jugular vein. Peripheral cannulation was always undertaken via surgical open approach.

Our ECMO systems included the Xenios™ Console with a Delta pump 3™ pump head, utilizing 2 sets: Novalung-MiniLung™ petite kit with Oxy Medos Hilite™ 800LT for neonates (0–5 kg) and Novalung-MiniLung™ kit with Oxy Medos Hilite™ 2400 LT for children (5–15 kg).

Cannula selection comprised FreeLife™ FLKA (8F) or Medtronic Biomedicus Cortiva coated™ (8F) arterial cannulas for neonates, or Medtronic BioMedicus Life Support™ (9F) based on aorta size. Medtronic DLP Carmeda coated™ (16F, 18F) cannulas were used for venous central cannulation. The surgeon selected cannula size based on vessel size, required perfusion volume and patient weight.

Extracorporeal membrane oxygenation flow rates

For ECMO flow rates, we calculated the cardiac index using the formula cardiac index (l/min) = body surface area (m²) × 3. Anaesthesiologists adjusted therapeutic ECMO flow based on this calculated reference value.

Anticoagulation management

For anticoagulation management, we administered 25–35 I.U. unfractionated heparin/kg/h to achieve target values of 45–65 seconds of partial thromboplastin time (PTT) or unfractionated heparin (UFH) monitoring of 0.3–0.5 I.U./ml. We aimed for antithrombin 3 levels over 70%, fibrinogen over 100 mg/dl and platelet count over 100 000/μl.

Criteria for ECMO circuit and oxygenator exchange were met if signs of haemolysis (e.g. rise in free haemoglobin >20 mg/dl, lactate dehydrogenase (LDH) >700 U/l, total bilirubin >2 mg/dl) or thromboembolism (e.g. D-dimer >25mg/dl, platelet count <100 000/μl, fibrinogen <100 mg/dl, antithrombin 3 < 70%) were detected. Cannulas were changed if significant thrombus formation or plaque was observed on their inner surfaces. Immediate cannula replacement was performed if a sessile thrombus became mobile or floating parts were visible, posing an imminent risk of embolism.

Extracorporeal membrane oxygenation weaning

ECMO weaning occurred under optimal conditions (optimized inotropic support, ventilation with nitric oxide if necessary) with gradual ECMO flow reduction, closely monitoring haemodynamic and respiratory parameters. For initially stable patients, the ECMO circuit was clamped, cannulas removed with purse-string sutures left in place and the circuit allowed to recirculate. If stability persisted, purse-string sutures were tied, and the chest was either directly closed or left open for secondary closure based on fluid status and patient stability. If a patient was successfully weaned but required ECMO recannulation within 24 h, it was counted as 1 ECMO run until final weaning or therapy cessation.

If an anatomic cause or unexplained need for ECMO support was suspected, the patient was transferred to the cath. lab (4 cases; 7%). In 2 cases, pulmonary artery stenosis was repaired via operation and in 1 case, an intervention was performed. Another patient with pulmonary artery and Sano conduit stenosis underwent conduit replacement and intraoperative stenting of the pulmonary arteries. In 3 additional cases in which weaning from ECMO was not possible, reoperations were performed without previous catheterization (2 AV valve repairs, 1 neo aorta replacement with a homograft). Among the 7 patients (12%) who underwent interventions/operations during ECMO therapy, 4 could be successfully weaned from ECMO post-operation/intervention.

Statistical analysis

Statistical evaluation was performed using IBM SPSS Statistics Version 29.0. Categorical variables are expressed as quantity (%) and numerical variables as median (interquartile range). The study investigates only factors influencing hospital survival in case of ECMO use after the Norwood procedure. Later influencing factors on survival during pursued staged palliation were not analysed as in the vast majority of cases, they might not be associated with the extracorporeal support. Identification of hospital mortality risk factors was made using univariable binary logistic regression testing, in all variables of interest except ‘need for dialysis’ where Fisher’s exact testing was used. This was because all patients who developed kidney injury with need of dialysis died within the ECMO hospital stay, which made binary logistic regression testing an unqualified testing method in this particular variable. Our multivariable binary logistic regression model was created via a classical approach, by inclusion of all covariates possibly associated with hospital mortality in the 1st step and elimination of all covariates, which caused model instabilities in a 2nd step. Sub-group correlation analysis of cardiac and pulmonary indication and intraoperative and postoperative ECMO cannulation to hospital mortality was undertaken using Fisher exact testing. In addition, Kaplan–Meier plotting was used to illustrate long-term survival of all patients with ECMO therapy after the Norwood and before the BDG operation and we compared the survival data through Log-rank testing to all patients that received a Norwood operation without ECMO therapy before the BDG operation in our hospital—a group that is not part of our original study cohort. The observation period used for our Kaplan–Meier plot ranged from the date of surgery to the last sign of life.

RESULTS

Our patient group comprised 59 individuals, with demographic details outlined in Table 1. Median age and weight at ECMO initiation was 9 days (5–28) and 3.3 kg (2.8–3.7), and 37 patients (63%) were male. Most patients were neonates (44 patients, 75%) at the time of ECMO cannulation, followed by infants (14 patients, 24%), with only 1 child over 1 year old (2%). Pulmonary failure prompted ECMO in 17 patients (29%), while cardiac failure led to ECMO in 42 patients (71%), as depicted in Fig. 1. Cardiac arrhythmias resulting in ECMO cannulation represented a subgroup within the cardiac indication, occurring in 4 patients (7%). The median duration of ECMO therapy was 5 (4–7) days, with a range of 0–17 days.

Operative and ECMO cannulation data are depicted in Table 2. Of the whole cohort, 58 patients (98%) received ECMO therapy in the same hospital stay in which the Norwood operation took place. One patient needed ECMO therapy after a catheter intervention and concurrent Rhino-/Enterovirus pneumonia. Most patients received ECMO therapy after isolated Norwood operation [49 (83%)], 4 patients (7%) after Norwood and concomitant total anomalous pulmonary venous connection (TAPVC) correction, 4 patients (7%) after right ventricular to pulmonary artery conduit change before the BDG and 1 patient (2%) after Norwood and aortic root replacement using an 11-mm aortic homograft.

Among the 58 patients receiving postoperative ECMO support, 48% (28 patients) required initiation in the operating room due to an inability to be weaned from cardiopulmonary bypass. In the remaining 52% of cases (30 patients), ECMO was installed in the ICU. The ICU cannulations were categorized as early (within the first 24 h postoperation) in 29% (17 patients) and late (later than 24 h postoperation) in 22% (13 patients). Preceding ECMO cannulation, the median bypass time was 233 min (197–278), cross-clamp time was 76 min (69–100), and the median deep hypothermic circulatory arrest time for atrial septum excision was 3 (2–4) min.

Cannulation details varied: in 56% (33 patients), the ascending aorta and right atrium were used, while in 41% (24 patients), the brachiocephalic artery and right atrium were connected to ECMO using the pre-existing 3.5 mm PTFE prosthesis. Peripheral cannulation occurred in 1 patient through the common carotid artery and jugular vein. Another patient, initially cannulated in the femoral artery and vein during CPR, required switching the venous cannula to the right atrium due to insufficient perfusion. The use of a BTTS was limited to 5 patients. Three shunts were clipped and redone at the time of weaning (2 patients died), while 2 received ECMO therapy due to shunt thrombosis, with 1 being converted into a Sano conduit.

A total of 46 patients (78%) were able to be weaned from ECMO after median 4 days (range 0–13 days) and 33 patients (56%) could be discharged from our hospital after median 45 days (range 26–207 days). In 18 patients, ECMO was installed during CPR of whom 14 patients (78%) could be weaned and 9 patients (50%) discharged. Of all patients who received intraoperative ECMO support (n = 28), 22 patients (79%) were weaned and 18 patients (64%) could be discharged from our hospital. The children who received ECMO in the ICU (n = 30) could be weaned in 77% (23 patients) and discharged in 47% (14 patients). Patients with cardiac ECMO indication (n = 42) were weaned in 73% (31 patients) and discharged in 45% (19 patients), whereas 15 (88%) and 14 patients (82%) could be weaned and discharged, respectively, with pulmonary indication (n = 17). Among the 4 patients with arrhythmias leading to ECMO cannulation, all were successfully weaned, and 2 were discharged. Comparison between the intra- and postoperative and cardiac and pulmonary indication hospital survival probabilities using Fishers exact testing showed significant higher survival probabilities for pulmonary indication patients and only a trend favouring intraoperative ECMO initiation, when comparing ECMO initiation timing. These data are depicted in Table 3.

Severe ECMO complications in our group included sepsis in 19%, bleeding in 32%, seizures in 10%, cerebral lesions in 14% and kidney injury requiring dialysis in 14% of patients. Bleeding instances involved rethoracotomy (13 patients), intracerebral bleeding (3 patients), subdural haematoma (2 patients) and 1 patient with diffuse bowel bleeding. Cerebral lesions comprised 5 ischaemic and 3 haemorrhagic cases. Univariable risk factor analysis identified cardiac indication (OR 5.6; 95% CI 1.4–25; P = 0.014), lower weight (OR 2.38; 95% CI 1.1–5.0; P = 0.03), longer ECMO duration (OR 1.34; 95% CI 1.1–1.6; P = 0.002), cerebral lesions (OR 11.8; 95% CI 1.4–103.3; P = 0.026) and the need for dialysis (OR 30.8; 95% CI 1.7–564.1; P < 0.001) as independent risk factors for hospital mortality (Table 3).

Multivariable binary logistic regression model (Table 4) consisting of the significant factors mentioned above, but excluding the factor ‘need for dialysis’ because of model instability, demonstrated ECMO duration (OR 1.4; 95% CI 1.1–1.6; P = 0.003) and cerebral lesions (OR 10.7; 95% CI 1.0–111.5; P = 0.048) as risk factors for hospital mortality.

Our follow-up consists of 33 patients. Median follow-up duration after discharge was 4.8 years (3.6–7.8). One patient (3%) is awaiting BDG operation (status post BDG take-down), 6 patients (18%) underwent stage II palliation and 22 patients (67%) already completed Fontan circulation. Unfortunately, 3 patients (9%) died during follow-up (2 of failing Fontan and 1 due to sudden cardiac death) after 27, 234 and 1541 days post-discharge and 1 patient was lost to follow-up.

Long-term survival analysis showed an expected 10-year survival probability of 51.5% of the patients treated with ECMO therapy after Norwood and before the BDG operation. Comparing this result to all Norwood patients without ECMO therapy (n = 247), Log-rank testing showed a significant difference in 10-year survival (51.5% vs 80.7%) (P < 0.001) as shown in Fig. 2. In total, 34 of our 247 Norwood patients who did not receive ECMO therapy within the Norwood hospital stay died. Of those, only 13 patients died during the Norwood hospital stay and are therefore comparable to our study cohort, in which ECMO was used during the Norwood hospital stay. We did not initiate ECMO in those 13 patients because no improvement of outcomes could be expected from ECMO therapy or fatal neurologic outcome due to long lasting CPR was to be expected (5 patients with irreversible cardiac failure, 4 patients with sudden cardiac arrest, 2 patients with neurologic contraindications, 1 patient with uncontrollable pulmonary hypertensive crisis and another patient with irreparable hypoplastic pulmonary artery bed).

Follow-up echocardiographic data were available in 29 patients, with signs of good systemic ventricular function in 21 patients (72%) and either grade 0 or grade I insufficiency of the atrioventricular valve in 25 patients (86%) as depicted in Table 5.

Of the 33 patients, with available follow-up data, 25 patients did not show any neurologic conspicuities at their last control. Two patients suffered from hypoxic encephalopathy, 1 after a truncal ganglion haemorrhage and 1 patient after already peripartal preoperative asphyxia. Another patient is diagnosed with genetically confirmed Kabuki syndrome. A developmental lag compared to children of the same age was seen in another 3 patients. Two more patients presented without clinically relevant neurologic conspicuities at last follow-up but 1 had history of radiologic mild brain atrophy and the other of an ischaemic cerebral lesion.

DISCUSSION

Single-ventricle physiology poses a known risk for mortality, during ECMO therapy, with reported rates varying from 39% to 69% [3–5, 9–11]. In our study, hospital mortality after ECMO therapy and Norwood operation is relatively low at 44%, with an estimated 10-year survival probability of 51.5%. Previously identified risk factors such as cerebral lesions, sepsis, acute kidney injury, cardiac failure, lower weight, and longer ECMO duration were all statistically significant in our risk analysis [9, 12]. Acute kidney injury requiring dialysis emerged as the most impactful factor on mortality, consistent with other studies [9, 13]. However, our data could not clarify whether haemodynamic insufficiency or other factors contribute to renal failure. Wolf et al. suggest continuous venovenous hemofiltration as a risk factor for mortality due to rapid intravascular depletion despite reversing fluid overload postoperation. To mitigate bias, early continuous venovenous hemofiltration (within 48 h of ECMO initiation) was performed [14]. These findings offer an alternative explanation for the 100% mortality rate observed in this subgroup of patients in our study.

In our cohort, ECMO therapy due to pulmonary failure had significant better outcome compared to low cardiac output. This might be explained by temporary impairment of pulmonary resistance, which can be triggered by older age, longer bypass time, preoperative restrictive foramen ovale or TAPVC. Optimized ventilation after a period of recovery on ECMO with reduced pulmonary perfusion might make a successful weaning more likely than in the operating theatre.

TAPVC alone carries a mortality rate of 10–30% according to literature [15, 16]. In single-ventricle patients, TAPVC is a significant risk factor, increasing 30-day mortality after stage 1 palliative surgery compared to those without TAPVC (44% vs 16%) [17, 18]. In our study, 5 patients with TAPVC, Norwood and ECMO therapy before the BDG were included. Of these, 4 patients (80%) were successfully weaned from ECMO, but 4 patients died during hospitalization (80%) and 1 patient died during follow-up 13 months after ECMO due to Fontan failure. All patients with TAPVC and concomitant Norwood procedure who required ECMO died during the study period, reinforcing previous findings that this anatomy poses a high mortality risk despite ECMO use.

Previous studies comparing the BTTS to the Sano conduit favour the Sano conduit regarding mortality, particularly in the 1st year postoperation [19, 20]. After this period, mortality rates become similar regardless of shunt type. However, between stages 1 and 2 palliation, patients with Sano conduit may require more interventions such as balloon dilation or stent placement in pulmonary arteries [20]. In our cohort, due to our previous findings of higher diastolic pressures, lower Qp/Qs, and higher dp/dt ratio in children with Sano conduit, leading to better survival, BTTS patients are underrepresented [21]. BTTS was reserved for cases where anatomical factors did not favour a Sano conduit. During arterio-venous ECMO, systemic perfusion remains unaffected by a right ventricle to pulmonary artery conduit, while a BTTS causes significant volume loss towards the lungs, leading to systemic hypoperfusion. Thus, BTTS patients required clipping during ECMO, necessitating recannulation during weaning, complicating the process and exposing them to additional bleeding and thrombosis risks compared to Sano shunt patients.

Shunt thrombosis as a reason for ECMO therapy was identified in 2 of 5 BTTS patients (40%) but in only 1 of 54 (2%) Sano conduit patients and was counted as pulmonary failure ECMO indication. Because BTTS is underrepresented in our group, no distinct conclusions can be drawn to which shunt type is superior regarding ECMO results. However, we assume that the fact that the Sano shunt can be left open and perfused on ECMO lowers the risk for shunt associated complications on ECMO and enables better pulmonary recovery.

In our study, patients with ECMO cannulated in the operating room showed a trend towards higher in-hospital survival compared to later cannulation (64% vs 47%), though this trend did not reach statistical significance (P = 0.197). A large multicentre analysis with over 2000 patients similarly found no correlation between timing of ECMO cannulation and mortality after paediatric cardiac surgery [22]. Gupta et al. found that with increasing time between operation and ECMO initiation, ECMO duration, ventilation times, ICU stay length and hospital stay length increased by 1–3% for every 1 day. However, they considered these results clinically insignificant and deemed their timing of ECMO initiation appropriate. Since our study also found no statistical significance regarding outcomes, we cannot advocate for either early or late ECMO cannulation. Nevertheless, we observed a trend suggesting that early ECMO cannulation is associated with higher survival probabilities, possibly due to avoiding prolonged periods of haemodynamic instability or extensive medical treatment before ECMO therapy is employed. In cases of doubt, early ECMO initiation should be considered, though our small sample size may contribute to the lack of significant correlations. Additionally, risks such as local thrombosis at the cannulation site, possible embolism or bleeding complications due to anticoagulation should be weighed when ECMO indication is uncertain.

Arrhythmia was an indication for ECMO therapy in 4 patients (7%), and ECMO was initiated under CPR in 18 patients (31%). Despite findings in other studies, we did not find a statistically significant correlation between these factors and mortality, likely due to our small study size [12, 13]. In our cohort, post-ECPR patients experienced a stroke rate of 11% and a hospital survival probability of 50%, which is comparatively high [13, 23, 24]. However, we lack comprehensive follow-up data for long-term neurological outcomes, particularly in the context of ECPR. A comparison of neurological outcomes between ECMO and non-ECMO Norwood patients would be valuable. Nonetheless, the discharge of half of the resuscitated patients encourages us to pursue all life-saving measures for these vulnerable children.

CONCLUSION

We believe ECMO therapy is essential in paediatric cardiac centres treating single-ventricle patients, particularly post-Norwood operation, due to potential haemodynamic or respiratory complications. Despite the inherent risks, achieving over 50% survival and successful Fontan completion in most survivors is promising. Pulmonary ECMO indication correlates with better survival compared to low cardiac output cases. However, while we observed high weaning rates, a notable proportion of cardiac indication patients succumbed during ICU stay. The small subgroup with concomitant TAPVC did not benefit from ECMO in our series.

Further studies on quality of life and neurologic outcomes comparing Norwood patients with and without ECMO are needed to better contextualize these results.

Limitations

Our study was limited by its retrospective single-centre design and the small number of patients. Even though some risk factors could be associated with statistical significant higher mortality, a larger cohort in future studies is required to enhance statistical power to further illustrate relevant factors regarding this topic. An advantage of pulmonary shunt type cannot be investigated as the cohort encompasses nearly only patients with Sano shunts.

FUNDING

None declared.

Conflict of interest: None declared.

DATA AVAILABILITY

The data underlying this article will be shared on reasonable request to the corresponding author.

Author contributions

Fabian Seeber: Conceptualization; Data curation; Methodology; Writing—original draft. Niklas Krenner: Data curation. Eva Sames-Dolzer: Conceptualization; Methodology; Supervision; Writing—review and editing. Andreas Tulzer: Formal analysis; Visualization. Ishita Srivastava: Data curation. Michaela Kreuzer: Data curation; Formal analysis. Roland Mair: Data curation. Gregor Gierlinger: Data curation. Mohammad-Paimann Nawrozi: Data curation. Rudolf Mair: Supervision; Writing—review and editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Jürgen Hörer, Amir-Reza Hosseinpour and the other anonymous reviewers for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• BDG

  Bidirectional Glenn

 
• BTTS

  Blalock–Thomas–Taussig shunt

 
• ECMO

  Extracorporeal membrane oxygenation

 
• ELSO

  Extracorporeal Life Support Organization

 
• ICU

  Intensive care unit

 
• PTFE

  Polytetrafluoroethylene

 
• TAPVC

  Total anomalous pulmonary venous connection