Why do children not survive extracorporeal membrane oxygenation?

Background: Extracorporeal membrane oxygenation (ECMO) is used in critically ill children with cardiac and/or respiratory failure. Use is increasing in children with high-risk comorbidities. Reasons children do not survive ECMO are poorly described.

Aims: Describe characteristics and cause of death, compare mortality in children with high-risk comorbidities, evaluate mortality trends over a decade.

Method: All children <18 years old who received ECMO at this institution from 1 January 2011 to 31 December 2020 were described and categorised by outcome: died on or <48 h post-ECMO, died ≥48 h post-ECMO, survived to hospital discharge. Children who did not survive ECMO (DNSE) were categorised to: ECMO withdrawal for irrecoverable original condition, withdrawal for poor prognosis neurological condition, brain death, withdrawal for poor prognosis with multiple complex conditions, and unsupportable. Poison regression was used to analyse survival trends.

Results: Four hundred twenty-eight children received ECMO, 19% DNSE, 14% died ≥48 h post-ECMO and 67% survived. ECMO was electively withdrawn for irrecoverable original condition (39%), poor prognosis for neurological condition (32%) or multiple complex conditions (18%). One hundred twenty-two children had ≥1 high-risk comorbidity. Children with genetic syndromes (58%), risk-adjusted congenital heart surgery score-1 ≥4 (53%), primary immunodeficiency (50%) had lower hospital survival. No children with malignancy/bone marrow transplant survived to hospital discharge. Overall hospital survival was 67%, with no significant change during the study period (P-trend = 0.99).

Conclusion: Children who DNSE have therapy electively withdrawn for irrecoverable disease or poor prognosis. Children with high-risk comorbidities have a reasonable chance of survival. This study informs clinicians ECMO may be a therapeutic option.