There is no reliable clinical indicator showing how long extracorporeal membrane oxygenation (ECMO) implementation can be delayed before the risk of death becomes unacceptably high in neonatal and pediatric respiratory patients. However, the late use of ECMO may be defined by the elevation of specific physiologic markers separate from pulmonary function and hemodynamic assessments that indicate when the optimal time for implementation of ECMO has past, resulting in a higher than normal mortality, possibly due to reperfusion injury. Neonatal patients were reviewed retrospectively to determine if later implementation of ECMO correlated to increased mortality. Neonatal and pediatric respiratory patients placed on ECMO were reviewed retrospectively to determine if the first adjusted anion gap (AGc), the first venoarterial CO2 gradient (p[v-a] CO2), or the first Viability Index (AGc + p[v-a]CO2 = INDEX) on ECMO could be used to identify a cut-point for increased mortality. Expired neonates (n = 31) were placed on ECMO an average of 2 days later than neonatal survivors (n = 163). The review of 210 respiratory neonatal and pediatric ECMO patients with an overall survival of 82% showed that all three markers were elevated in the expired patients (n = 38, p < .05). Cut-points were an AGc > or = 23 mEq/L, the p[v-a]CO2 > or = 16 mmHg, and the INDEX > or = 28. These values correlated with a significantly higher risk of mortality (p < .05); survival to discharge being 43% or less. Patients under the cut-points had survival rates of 84% or higher. Starting ECMO too late may cause reperfusion injury that reduces survival. This study describes specific physiologic markers taken soon after ECMO initiation that correlate with mortality. These markers, if assessed earlier, may allow for a more timely ECMO implementation and higher survival.