What’s new in neonatal and pediatric extracorporeal membrane oxygenation in 2022–2023? Insights from the EuroELSO congress 2023

Extracorporeal membrane Oxygenation (ECMO) has been successfully delivered in neonates and children for more than 50 years and currently is a well-established form of advanced life support for meconium aspiration syndrome, congenital diaphragmatic hernia (CDH), pediatric acute respiratory distress syndrome (PARDS), postcardiotomy failure, acute myocarditis, cardiomyopathy and finally for extracorporeal cardiopulmonary resuscitation (ECPR).

The last 2 years (2022- and beginning of 2023) have been particularly fruitful for the neonatal and pediatric ECMO research and in this draft we will summarize the main achievements.

After the release of the Pediatric Acute Lung Injury Consensus Conference (PALICC) guidelines in 2015, ¹ there have been several advances in ECMO research and ECMO indications have expanded. ² Newer neonatal and pediatric ECMO pumps have been released on the market and anticoagulation approach, alternative to heparin, have been tested successfully. Thus, in 2023, the PALICC group published the updated guidelines (PALICC-2) on the use of ECMO support in patients with PARDS. ³

Unfortunately, since the lack of randomized controlled trials comparing ECMO versus invasive mechanical ventilation in children, only conditional recommendations with low certainty of evidence were developed.

In particular, since the scientific rationale behind ECMO is that it may reduce both the risk of ventilation-induced lung injury and of multiorgan failure due to the biotrauma, the PALICC-2 consensus suggested that patients with a potentially reversible cause of PARDS should be evaluated for ECMO when lung protective ventilation strategies result in inadequate gas exchange. ⁴

Notably, since the absence of gas exchange thresholds to start ECMO in children, the PALICC-2 consensus suggested to use the neonatal and adult thresholds for pediatric patients.^5,6 The PALICC-2 consensus also suggested that serial clinical evaluations, rather than a single one, should be performed when evaluating ECMO candidacy at bedside and that veno-venous (VV)-ECMO, rather than veno-arterial (VA)-ECMO, should be used in patients PARDS with adequate cardiac function.

Once on ECMO, the PALICC-2 consensus suggested to maintain a physiologic level of PaO₂ and PaCO₂, and to slowly reduce the PaCO₂ level after the ECMO start. The PALICC-2 consensus also suggested to reduce the ventilator pressure settings below the lung protective ventilation limits (rest settings) once on ECMO. ⁴

A recent analysis of Extracorporeal Life Support Organization (ELSO) Registry performed by Shah et al. ⁷ showed that the rapid decrease of PaCO₂ levels early after ECMO deployment was associated with higher odds of neurologic injury and this effect was further increased when associated to the concomitant increase of the mean arterial pressure. Similarly, but in neonates with respiratory failure, Joram N et al. ⁸ showed how a significant decrease (>50%) in PaCO₂ levels after the ECMO start was associated with higher odds of acute neurologic events.

Currently, the correct management of the native lung during ECMO is still a matter of debate both in neonates and children.^9–11 Both Friedman ML et al. ¹¹ and Blauvelt DG et al. ¹⁰ showed that the maintenance of high ventilator settings during ECMO (e.g. FiO₂>60% and a PEEP >10 cmH₂O) was associated with higher odds of mortality. A recent survey ⁹ performed in the United States (US) on the ventilation strategies used during neonatal ECMO showed that a peak pressure between 16 and 20 cmH₂O, a positive end expiratory pressure (PEEP) between 9 and 10 cmH₂O, a respiratory rate between 10 and 15 breath/min and a FiO₂ between 20 and 30% were the most common ventilator settings used in many ECMO centers. Notably, non-ventilation practices such as flexible bronchoscopy (FB), exogenous surfactant administration were also used to speed up lung recovery. Rosner EA et al. ¹² showed that FB was also feasible during pediatric ECMO and improved dynamic respiratory system compliance. FB reported low complication rates (3.1%) when performed by experienced operators and could be beneficial when performed early in the course of ECMO.

Tracheostomy is a practice rarely performed during pediatric ECMO because of the risk of bleeding, ¹³ however, it may provide several benefits such as the reduction of sedation and early mobilization. In general, tracheostomy is performed when PARDS does not recover in a reasonable amount of time (e.g. 2–3 weeks) or in patients with acute on chronic respiratory failure (e.g. cystic fibrosis, interstitial pulmonary fibrosis) bridged to lung transplantation with ECMO to improve airways clearance and avoid deconditioning. ¹⁴

An analysis of the ELSO registry (years 2015–2019) performed by Kohne JG et al. ¹³ showed that the use of tracheostomy during ECMO was quite rare in children (2.6%). In this cohort, tracheostomy was performed after more than 7 days (median 13.2 days) of ECMO support. Surgical site bleeding (12%) remained the main concern; however, this study was not able to attribute bleeding to the tracheostomy placement.

The 2022 and the begin of 2023 was also brilliant for the research on cardiac ECMO and ECPR.

Sperotto F. et al. ¹⁵ performed an analysis of the ELSO registry from 2000 to 2016 including children with biventricular physiology supported with VA-ECMO for failure to wean from cardiopulmonary bypass. Using a propensity score-weighted analysis, adjusting for baseline differences between patients who did or did not undergo left atrial (LA) decompression, the authors showed that LA decompression was independently associated with survival and decreased in-hospital adverse events.

Gutierrez ME et al. ¹⁶ using the data from ELSO Registry (years 2007–2018), aimed to describe the best practices and timing of VA-ECMO in children with myocarditis supported with ECMO. The authors confirmed that the use of VA-ECMO in this selected population was associated with a very high survival (72%) compared with other pediatric populations receiving VA-ECMO. Notably, in children with myocarditis who were cannulated, but did not suffer of cardiac arrest or undergo ECPR, a short time between intubation and ECMO cannulation was significantly associated with survival.

Sanford EL et al. ¹⁷ performed an analysis of the ELSO Registry including patients from 2011 to 2019 exposed to hypothermia (>24 h) after ECPR to evaluate the association between hypothermia and neurologic complications including brain death, seizure, brain hemorrhage and brain ischemia. This analysis demonstrated that hypothermia after ECPR was not associated with decreased odds of neurologic complications or mortality benefit at the time of hospital discharge.

The year 2022 and the begin of the 2023 have been important also for the research on neonatal ECMO. Notably, the majority of the scientific activity was on the use of ECMO in patients with CDH.

Rafat N et al. ¹⁸ performed a single center retrospective analysis of 243 CDH patients receiving ECMO between 2010 and 2020 to investigate if the early use of ECMO was associated with increased survival. The authors divided their cohort in 4 groups: (a) ECMO initiation < 12 h after birth, (b) ECMO initiation between 12 and 24 h after birth, (c) ECMO initiation between 24 and 120 h after birth and, (d) ECMO initiation > 120 h after birth. The authors showed that CDH patients supported with ECMO between 12 and 120 h after birth had a better survival compared with patients supported with ECMO immediately after birth or beyond 120 h. These results could suggest that the absence of the honeymoon period after birth or the prolonged use of injurious ventilation settings are associated with poor outcomes. Notably, the CDH Interest Group ¹⁹ analyzing the data of the ELSO registry in the years between 2000 and 2019 observed that the mortality rates of CDH patients supported with ECMO significantly varied across centers, with centers performing better (12%) and centers performing worse (6%) than the median standard. In centers performing better, the rate of complications was lower than in the centers performing worse. The same group also showed that high-volume centers were associated with lower odds of mortality. ²⁰ Thus, centralization in experienced centers should be recommended for these patients.

Over the past 20 years there has been an important improvement of the ECMO technology with a shift from the roller pumps to the centrifugal pumps and from the silicone and polypropylene hollow fiber membrane oxygenators to the polymethylpentene hollow fiber membrane oxygenators. ²¹ Contemporarily, attempts to change circuit membrane characteristics to dampen or eliminate the blood-membrane inflammatory response fostered the use of biomimetic (e.g., heparin bond, nitric oxide bond circuits) and biopassive ECMO circuits (e.g., phosphorylcholine, fluid repellant, etc.) in clinical practice. Despite these important achievements, Cai T. et al. ²¹ reported in his systematic review evaluating the impact of these new technologies on clinical outcomes, a considerable variability in mortality across studies. Unfortunately, many important ECMO outcomes (e.g. ECMO duration, the definitions of the complications were poorly reported, absence of reporting standards, etc.) were inconsistently reported in these studies and this made challenging to draw any robust conclusions.

Notably, the improvement in ECMO technology was less important for neonates and children. This population represents the minority of the ECMO cases in the world and, the investments of the medical industries were poor in the last two decades. Currently, many pediatric centers are still using adult centrifugal pumps (3/8 inch) adapted to neonatal circuits (1/4 inch). ²² This circuit adaptation in some centers was associated to several mechanical complications (e.g. increase of oxygenator resistance due to clotting). ²³ However, the shift to the newer neonatal and pediatric magnetically suspended centrifugal pumps reduced this risk. In a single center retrospective study, Vermeer H. et al. ²³ described that when adult centrifugal pumps were used in children, recirculation and blood stagnation within the centrifugal pump increased the risk of clotting inside the pump and into the oxygenator.

In 2022 and 2023 we also got the first pediatric data from United States on the use of ECMO in patients with multisystem inflammatory syndrome associated to COVID-19 (MIS-C) and with severe acute respiratory syndrome coronavirus 2 (COVID-19). Using a hospital-based surveillance registry Bembea et al. ²⁴ described the clinical characteristics and outcomes of patients with MIS-C and with severe COVID-19 PARDS supported with ECMO between March 2020 till December 2021. In this public registry, the use of ECMO to manage these patients was uncommon (108 patients from 62 hospitals). The type, timing and ECMO duration was markedly different between patients with MIS-C and patients with COVID-19. Patients with MIS-C initiated ECMO earlier and had shorted ECMO runs compared with patients COVID-19 PARDS. VA-ECMO was more commonly used to manage patients with MIS-C. Overall survival was high in both groups and was similar to the one reported in Europe^25,26 previously.

In the last 2 years, we also got important data on anticoagulation management. Van Den Helm S. et al. ²⁷ performed a prospective observational study to investigate whether the routine monitoring of von Willebrand factor multimers could be useful in predicting the risk of bleeding. The authors showed that during the ECMO run children had a loss of high molecular weight multimers, an increase in intermediate molecular weight multimers and a reduced von Willebrand function compared to healthy children. Despite these characteristics, the occurrence of bleeding was not increased in these patients. Thus, the authors concluded that routine monitoring of von Willebrand multimers and function was not useful at bedside. Procaccini D. et al. ²⁸ conducted a single center retrospective study (years 2011 and 2020) to evaluate the clinical effects of two different practice of antithrombin (AT) supplementation (group A: supplementation when AT is <50% for neonates or <80% for infants and children vs group B: supplementation in case of heparin resistance classified as antifactor Xa <0.3 IU/mL or unfractionated heparin >40 IU/Kg/h or unfractionated heparin up-titration ≥ twice in 24 h). The authors showed that the prevalence of AT deficiency was high (43% of the measurements) in their cohort, nevertheless, AT replacement based on the evaluation of heparin resistance was associated with less AT administration, similar circuit changes and patient outcomes.

The use of anticoagulation strategies alternative to heparin (e.g. bivalirudin) has become more common in neonates and children, due to the limits of heparin (e.g. heparin resistance, heparin-induced thrombocytopenia, etc.). Nevertheless, even though bivalirudin can be considered a good alternative to heparin, its superiority has not been showed so far. Based on these considerations, Valdes CA et al., ²⁹ performed a systematic review to evaluate the outcomes of heparin-based versus bivalirudin-based anticoagulation during neonatal an pediatric ECMO. The authors showed that patients anticoagulated with bivalirudin had a lower rate of bleeding, transfusions and thrombotic complications. Mortality was not different between the two types of anticoagulation, however, overall costs associated with bivalirudin anticoagulation were lower.

In conclusion, these last 2 years were also fruitful to discuss on the end of life (EOL) care of the neonatal and pediatric ECMO patients. ECMO is still associated with a high mortality especially in high risk pediatric patients (e.g. ECPR, oncologic patients, etc.), thus, a special consideration on the EOL care is mandatory and different approaches are needed when dealing with neonates and children. Machado DS et al. ³⁰ carefully described all the aspects to be considered to provide a compassionate ECMO discontinuation (CED) when ECMO is no longer medically appropriate. First, high-quality communication with the family explaining the dying process and its unpredictability is essential; second, all the clinical aspects (symptoms management, psychosocial and spiritual support, team preparation, role assignment, etc.) should be addressed before starting the CED; third, all the practical aspects of the CED procedure (e.g. discontinue alarm, discontinuation of other life sustaining therapies, etc.) should be ready and available at bedside; fourth, the type of bereavement support should be agreed in advance to provide the post-mortem care according to the family’s wishes. Least but not last, a briefing with the health care team should be also considered to reduce the risk of burn-out.

Based on these achievements, we are sure that the next future for the neonatal and pediatric ECMO research will be even brighter!

Ad maiora semper.