Probability of survival after prolonged extracorporeal membrane oxygenation in pediatric patients with acute respiratory failure

Objective: Extracorporeal membrane oxygenation (ECMO) has been used with increasing frequency for respiratory failure that is unresponsive to conventional therapy. We examined the relationship between duration of ECMO and outcome to understand whether prolonged ECMO (duration of the procedure for >14 days) was more commonly associated with futile therapy or eventual recovery. Design: A cohort study of all patients reported to the Pediatric ECMO Registry for Acute Respiratory Failure of the Extracorporeal Life Support Organization. Setting: Tertiary hospitals (n = 83) capable of providing extracorporeal support for pediatric patients. Patients: Children (n = 382) between the ages of 1 wk and 18 yrs of age with severe respiratory failure. Interventions: Extracorporeal membrane oxygenation. Measurements and Main Results: The death or live hospital discharge of ECMO-treated patients, together with the post-ECMO mechanical ventilation course, were examined as a function of duration of ECMO and of pre-ECMO respiratory status. The occurrence of complications and the causes of death were also noted. The criteria used to initiate ECMO, as well as the determination of the futility of further ECMO, were determined by local practice at individual centers. There were 382 patients treated with ECMO, of whom 184 (48%) survived. The proportional survival in the patients treated for the longest duration was similar to the overall group. The cause of death was given for 168 patients: 32 neurologic deaths; nine deaths due to ECMO complications; and 30 deaths due to nonpulmonary organ failure. There were 97 deaths due to elective ECMO termination; 80 of these deaths occurred after the determination of the futility of anticipating pulmonary recovery. The latter deaths occurred at widely varying durations of ECMO, with a median of 282 hrs. However, at that same duration, 47 eventual survivors (26% of all survivors) continued to receive ECMO. By discriminant analysis, the survival rate was independently related (r² = .18;p < .0001) to peak ventilator inspiratory pressure before ECMO and duration of intubation before ECMO, patient age, and the occurrence of several complications. Conclusions: While the survival rate in pediatric patients receiving ECMO appears related to the severity of lung disease and to the occurrence of ECMO complications, the survival rate in patients treated with ECMO courses of >2 wks was similar to the survival rate of patients treated for shorter periods of time. ECMO was terminated in some patients for pulmonary futility at durations of ECMO associated with survival in substantial numbers of patients in whom ECMO was continued.