There has been increasing use of venoarterial (VA) extracorporeal membrane oxygenation (ECMO) for infants with respiratory failure, up to 92% of neonatal respiratory support in 2021. This study seeks to leverage the increased use of VA ECMO in this cohort to enrich an evaluation of the differences in rate of intracranial hemorrhage and ischemic stroke between venovenous (VV) and VA ECMO among infants with respiratory failure where clinicians may choose either strategy.

This project is a retrospective review of data in the ELSO registry.

From 2019-2021, there was increased use of venoarterial (VA) extracorporeal membrane oxygenation (ECMO) for infants with respiratory failure, up to 92% of neonatal respiratory support in 2021. The primary aim is to estimate the average effect on the rate of neurologic injury of VA ECMO versus venovenous (VV) among infants with respiratory failure over the period 2013-2018, during which clinicians could choose either cannulation strategy. This causal effect will be estimated using an inverse propensity weighted (IPW) approach. Secondarily, the investigators will project this estimated treatment effect forward into the period 2019-2021. The beginning of this period roughly corresponds to start of increased use of VA ECMO. Under the assumption of a homogenous treatment effect across both study periods, the rate of neurologic injury that would have occurred in 2019-2021 will be estimated, had the rate of VA ECMO not increased relative to pre-2019 levels. The hypothesis is that the results will point to an increased rate of neurologic injury starting in 2019 due to the increased use of VA ECMO.

Inclusion Criteria:

• Patient weighed less than or equal to 10kg at start of ECMO
• Pulmonary support was the indication for ECLS
• Initial cannulation strategy was VV or VA
• The run occurred during the period 2013-2023

Exclusion Criteria:

• Patient had CDH

• Patient was post-cardiotomy

• Non-conventional initial cannulation strategies were employed, such as

  • Central Cannulation (surrogate for inability to achieve peripheral cannulation)
  • Veno-veno-arterial ECMO
  • Initial cannulation approach reported as "other"

• Patient was transported into or out of ELSO center on ECMO support

• Patient had pre-ECLS Cardiac Arrest

• Patient did not have subsequent ECMO runs in the ELSO registry

All patients reported to the ELSO registry weighing less than or equal to 10 kilograms at the start of ECMO receiving ECMO for pulmonary support

Data received from the ELSO Registry is available to qualified researchers through an ELSO data request