Objective: To examine the predictors of neurodevelopmental outcome (NDO) of ECMO and near-ECMO survivors at 2.5 to 6 years.

Design: Prospective longitudinal follow-up of ECMO and near-ECMO infants born between 1988 and 1992 who were evaluated at 9 to 12 months and at 2.5 to 6 years of age.

Patients: Thirty three neonatal ECMO survivors and 16 near-ECMO infants matched for birthweight, gestational age, birth head circumference, severity of cardiorespiratory compromise, duration of Ventilator and Oxygen therapy, and demographics were assessed at 9 to 12 months and at 2.5 to 6 years utilizing age appropriate standardized tests (CAT/CLAMS, Bayley Scales of Infant Development, Stanford-Binet-4th edition) and physical/neurological examination. Head CT scans were obtained in 94% of ECMO and 57% of near-ECMO patients.

Results: Neonatal risk variables (duration of metabolic alkalosis i.e. pH >7.45, hypocarbia i.e. PaCO2<30, mean highest pH, lowest pCO2, age at ECMO and duration) were determined. CT scans were score blinded to patient status utilizing a modification of Taylor's system, obtaining individual scores (ischemia, bleeding, and ventricular size) and total score. NDO status was determined by grading A to D based on cognitive developmental Quotient (CDQ) and number of domains affected, including motor, cognitive, language and behavior: A) CDQ>85, no domains affected; B) CDQ>85 and/or one domain affected; C) CDQ 75-85 and/or two domains affected; D) CDQ<75 and /or more than two domains affected. The significant neonatal risk variables for near-ECMO were duration of hypocarbia and metabolic alkalosis. The percentage of infants with NDO grade C+D at 9 to 12 months and 2.5 to 6 years in the ECMO group was 34% and 36%, whereas for near-ECMO group it was 25% to 19% respectively. At 9 to 12 months and 2.5 to 6 years examination, ECMO and near-ECMO had similar NDO status in the following areas: motor (45% vs. 38%), cognitive (18% vs. 13%), language and speech (39% vs. 38%), behavior (24% vs. 19%), vision (6% in both), and hearing (9% vs. 6%). C+D and A+B group significantly differed in age at initiation of ECMO (95 vs. 45 hours) and head CT scans (abnormal 90% vs. 24%, severe ischemia 67% vs. 20%) and had prolonged hypocarbia (23 hrs vs. 13) and metabolic alkalosis (51 hrs vs. 18). Of the CT scans obtained of the near-ECMO infants, 50% were abnormal.

Conclusion: NDO suggests that delaying initiation of ECMO may increase the risk of adverse neurodevelopmental outcome.