Extracorporeal life support outcome for 128 pediatric patients with respiratory failure

Abstract

Purpose: The aim of this study was to describe a single-center experience with pediatric extracorporeal life support (ECLS) and to determine variables predictive of outcome in pediatric patients, both before the institution of ECLS and while on support.

Methods: From October 1985 to September 1998 the authors supported 128 children with severe acute hypoxemic respiratory failure(n = 121, Pao2/FIo2 ratio = 58+/-29) or acute hypercarbic respiratory failure (n = 7, Paco2 = 128+/-37), despite maximal conventional ventilation. Mode of access included venoarterial bypass (VA, n = 64), venovenous bypass (VV, n = 53), and VV to VA bypass (n = 11). The techniques used included lung rest, pulmonary physiotherapy, diuresis to dry weight using hemofiltration if needed, minimal anticoagulation, and optimal systemic oxygen delivery.

Results: The median age was 1.4 years (range, 2 weeks to 17 years). The mean duration of ECLS was 288+/-240 hours (range, 4 to 1148 hours or 0.2 to 47.8 days). Lung compliance increased from 0.32+/-0.02 mL/cm H2O/kg to 0.59+/-0.03 mL/cm H2O/kg in survivors, but only increased from 0.34+/-0.02 mL/cm H2O/kg to 0.35+/-0.02 mL/cm H2O/kg in nonsurvivors (P<.002 comparing change between survivors and nonsurvivors). Mean body weight decreased from 9%+/-2% over dry weight to 4%+/-2% in survivors, whereas in nonsurvivors the mean body weight increased from 25%+/-5% over dry weight to 35%+/-7% (P<.001). Outcome results by diagnosis were pneumonia, 73%; acute respiratory distress syndrome, 67%; and airway support, 60%, with overall lung recovery occurring in 77%, and hospital survival in 71%. Multivariate logistic regression modelling of patients with hypoxemic respiratory failure found the only pre-ECLS variable significantly associated with outcome to be pH (P<.05). Variables during the course of ECLS significantly associated with decreased survival were the presence of creatinine greater than 3.0 (P<.01), the need for inotropes (P<.04), failure to return the patient to dry weight (P<.04), and lung compliance that did not improve significantly. (P<.01).

Conclusions: ECLS provides life support in severe respiratory failure in children, allowing time for injured lungs to recover. Pre-ECLS predictors, such as pH and variables during ECLS, such as presence of renal failure, improvement in compliance, return to dry weight, and the need for inotropes on ECLS, may be useful for predicting outcome.