No abstract

Neonatal and pediatric extracorporeal membrane oxygenation (ECMO) is carried out commonly using occlusive blood pumps. Centrifugal pumps provide simple and safe technology for transportation on ECMO. The assistence respiratoire extra corporelle (AREC) system enables single needle venovenous ECMO for infants. We report on our experience with neonatal and pediatric ECMO treatments using nonocclusive blood pumps. One-hundred forty-six ECMO treatments were performed for cardiac, neonatal, and pediatric indications in 54, 19, and 27% of cases. Centrifugal pumps were used in 99, and the AREC system in 42 cases. Hospital mortality was estimated retrospectively and influence of type of pump, type of ECMO belonging to indication group, and lactate at ECMO installation were estimated. Irreversible organ failure leading to ECMO termination was investigated within groups of indications. Survival (recent 50 ECMO treatments) was 80, 70, 43, and 30% after meconium aspiration syndrome, acute respiratory distress syndrome, cardiac surgery, and prolonged resuscitation. Lactate exceeding 100 mg/dl at ECMO installation predicted significantly worse outcome. Cerebral damage was the main reason for ECMO termination in all but persistent circulatory failure in the cardiac group. Myocardial recovery resulted in all except 2 cardiac cases. Nonocclusive blood pumps can be used safely in neonatal and pediatric ECMO. Early installation may improve outcome markedly. In cardiac cases results of surgery should be thoroughly investigated on the table before ECMO installation to prevent hopeless ECMO treatments.

While extracorporeal membrane oxygenation (ECMO) is being used increasingly after pediatric cardiac surgery, criteria are lacking for initiating ECMO after bypass weaning. To develop clinically useful ECMO entry criteria based on parameters readily available, children were examined at postoperative pediatric intensive care unit (PICU) admission. Using hospital mortality as the primary outcome, univariate and multiple logistic regressions were performed to estimate the predictive value of clinical (age, weight, and diagnosis) and laboratory (arterial blood pressure, pH, lactate, creatine kinase, and arterial and central venous oxygen saturation [ScvO2]) variables. Data from 218 children over a 2 year period were analyzed retrospectively. Univariate regression demonstrated that age, weight, diagnosis, blood pressure, venous and arterial saturation, and lactate were significantly associated with postoperative mortality (p < 0.05). In multiple regression, ScvO2 and lactate level were found to be independent predictors and were used in a predictive model (ScvO2 odds ratio: 2.03-828.6, p = 0.016) (lactate odds ratio: 1.58 -4.20, p = 0.0002) (R2 = 0.70). Applying an 80% risk of mortality to establish entry criteria as in neonatal ECMO, PICU admission values of lactate > 70 mg/dl if ScvO2 < 60% or lactate >163 mg/dl if ScvO2 > 60% are proposed to serve as postoperative ECMO entry criteria if bypass weaning has been possible but is followed by low cardiac output.