One hundred and fifty-two 5 year-old ECMO survivors and 53 normal controls were given a comprehensive neuropsychological test battery and kindergarten screening. Mean Full Scale IQ (FSIQ) for the ECMO group was significantly lower than that of the Control group (94 ± 21 vs 108± 19, p<.001). Eighteen ECMO (12%) and 0 Controls had FSIQ < 70 (Mentally Retarded range). Academic Risk (as defined by FSIQ <80 or presence of neuropsychological deficits or failed kindergarten screen) was significantly higher among the ECMO children compared to Controls (50% vs 34%, p<.05). We obtained reports of school performance from parents of 101 of the 115 (88%) ECMO survivors when they should have completed first grade. Thirty-seven of the 101 (37%) children had experienced Academic Problems, as defined by either grade retention or having received special education resources. Seventy percent of the children with Academic Problems had been classified as Academic Risk on neuropsychological testing at age five. Eighty percent of the children classified at age five as No-Academic Risk were promoted to second grade without special education services. Stepwise regression analyses were computed with Academic Problems as the dependent variable. FSIQ,Academic Risk, and Neonatal Neuroimaging Score were entered as independent variables. Within the ECMO group (N=101), only FSIQ accounted for a significant proportion of the variance (43%). When only ECMO subjects with normal IQ were included in the analyses (N=91), only Academic Risk status accounted for a significant portion of the variance (29%). We conclude that neuropsychological testing for five year-old ECMO survivors can significantly enhance our ability to predict school difficulties. This study was funded by NIH/NINDS.