Definitions of adverse events associated with extracorporeal membrane oxygenation in children: results of an international Delphi process from the ECMO-CENTRAL ARC

Extracorporeal membrane oxygenation (ECMO) is a high-risk and low-volume life support with increasing clinical study. However, heterogenous outcome definitions impede data assimilation into evidence to guide practice. The Academic Research Consortium (ARC), an international collaborative forum committed to facilitating the creation of stakeholder-driven consensus nomenclature and outcomes for clinical trials of medical devices, supported the ECMO Core Elements Needed for Trials Regulation And quality of Life (ECMO-CENTRAL) ARC. The ECMO-CENTRAL ARC was assembled to develop definitions of paediatric ECMO adverse events for use in clinical trials and regulatory device evaluation. An initial candidate list of ECMO adverse events derived from the mechanical circulatory support ARC was supplemented with a review of ECMO-relevant adverse event definitions collated from literature published between Jan 1, 1988, and Feb 20, 2023. Distinct teams of international topic experts drafted separate adverse event definitions that were harmonised to existing literature when appropriate. Draft definitions were revised for paediatric ECMO relevance with input from patients, families, and an international expert panel of trialists, clinicians, statisticians, biomedical engineers, device developers, and regulatory agencies. ECMO-CENTRAL ARC was revised and disseminated across research societies and professional organisations. Up to three rounds of internet-based anonymous surveys were planned as a modified Delphi process. The expert panel defined 13 adverse event definitions: neurological, bleeding, device malfunction, acute kidney injury, haemolysis, infection, vascular access-associated injury, non-CNS thrombosis, hepatic dysfunction, right heart failure, left ventricular overload, lactic acidaemia, and hypoxaemia. Definitional structure varied. Among 165 expert panel members, 114 were eligible to vote and 111 voted. Consensus was achieved for all proposed definitions. Agreement ranged from 82% to 95%. ECMO-CENTRAL ARC paired rigorous development with methodical stakeholder involvement and dissemination to define paediatric ECMO adverse events. These definitions will facilitate new research and the assimilation of data across clinical trials and ECMO device evaluation in children.