How Low Should We Go? Outcomes from ECMO in Neonates with Lower Gestational Age or Birth Weight

Background: ECMO remains a critical life-saving support option in neonates despite the declining numbers in the past decade due to improvements in intensive care management. Initial recommendations for ECMO had relative contraindications in babies with a gestational age (GA) less than full term or birth weight (BWt) below 3 kg due to concerns for high incidence of intraventricular hemorrhage or technical issues with cannulation - and poor outcome. However, these relative concerns have changed over time and the envelope has been pushed further with smaller babies being supported. Previous reports have shown the application of ECMO in smaller neonates with better results. The purpose of this study was to leverage a large database to define the outcomes in the most recent decade, and understand both utilization and survival in patients with lower GA and BWt.

Methods: The Extracorporeal Life Support Organization (ELSO) registry is a voluntary international compilation of patients who underwent ECMO support. Data on all neonates from 2009-2019 were included, and those who had single and concluded runs of ECMO, were selected for final analysis. Demographic information, ECMO details, outcomes, complications, and diagnoses variables were collected where available. A composite morbidity outcome variable was created with the most severe complications, and a combined mortality / severe complication outcome. Univariate and Multivariable statistical tests were performed after multiple imputations.

Results: 15,304 neonatal ECMO runs in 14,773 patients were in the data set, and 14,167 met inclusion criteria. Overall, mean BWt was 3.2 kg, 42.4 were female, mean age at ECMO was 5.4 days, with an average GA of 38.4 weeks, and an overall survival of 61.3%. Univariate analysis noted that birth weight, gestational age, ECMO mode, pulmonary support type, pH and ventilator type / settings were highly significant predictors of survival. GA and BWt were found to be moderately correlated (spearman rho=0.46). The age range for the lowest quintile of GA was from 29-35 weeks. Multivariable assessment noted a significant linear relationship of both mortality and severe complications to GA and Bwt (p < 0.001, OR= 0.83 GA, 0.54 BWt). Analyses of the lowest quintile of GA revealed mortality and complications ranging from 65-86%.

Conclusion: ECMO support to smaller babies remained consistent during the study period, with significantly lower survival compared to older / larger neonates. Both GA and BWt were correlated, and had a strong linear relationship with mortality. There were very high rates of severe complications and death in the smallest group – with Cardiac support, VA ECMO and CDH being other predictors of mortality. These data suggest that ECMO support for neonates less than 34 weeks GA and / or 2.4 kg BWt needs further refining and should be considered in cases with higher chance of survival.