Predictors of outcome for children requiring respiratory extra-corporeal life support: implications for inclusion and exclusion criteria

Intensive Care Med. 2008 Dec;34(12):2256-63. doi: 10.1007/s00134-008-1232-3. Epub 2008 Aug 1.


Objectives: A range of children receive extra-corporeal life support (ECLS) for respiratory failure, but there is little published data on this group. Our aims were: (1) to analyse predictors of outcome and (2) comment on inclusion and exclusion criteria.

Design: Retrospective review.

Setting: Tertiary ECLS centre.

Patients: A total of 124 children categorised as 'paediatric respiratory ECLS' from July 1992 to December 2005.

Results: Fifty-three percent of children had one or more co-morbid conditions; the median age was 10.1 (IQR 3-34) months; the median ECLS duration was 9 (IQR 5-17) days; survival to discharge was 62% and at 1 year was 59%. Although survival varied according to primary reason for ECLS (range 36-100%), after adjustment for this, the presence of a co-morbid condition was unrelated to mortality (OR = 1.49, 95% CI 0.65, 3.42, P = 0.34) Predictors of mortality were increased pre-ECLS oxygenation index (OR = 1.09, 95% CI 1.00, 1.18, P = 0.05) and shock (OR 2.53, 95% CI 1.21, 5.28, P = 0.01). The relationship between mortality and end organ dysfunction (OR 2.12, 95% CI 0.89, 5.02, P = 0.09) and greater number of pre-ECLS ventilator days (OR 1.10, 95% CI 0.99, 1.22, P = 0.08) was less conclusive.

Conclusions: Pre-existing co-morbid conditions may predispose children to develop severe respiratory failure but with careful case selection, do not appear to reduce the chance of survival. Severity of pulmonary dysfunction determined by OI and shock were key predictors of outcome and should remain important determinants of referral for ECLS.

MeSH terms

  • Adolescent
  • Child
  • Child, Preschool
  • Extracorporeal Membrane Oxygenation*
  • Female
  • Humans
  • Infant
  • Male
  • Pneumonia / therapy*
  • Respiratory Distress Syndrome / therapy*
  • Retrospective Studies
  • Survival Analysis