In a series of 61 infants who had congenital diaphragmatic hernia (CDH) treated at our center from 1978 through 1996, 37 of 59 (61%) survived the perioperative period with two infants lost to follow-up. Nine (47%) of 19 infants survived before the introduction of extracorporeal membrane oxygenation (ECMO) into our region in 1986. Since 1986, 28 (70%) of 40 infants survived. Eighteen infants required ECMO, and 12 (75%) survived. A chart review was performed to determine whether infants surviving CDH are suffering from delays in neurological development, and, if so, whether this is attributable to ECMO. Of 12 ECMO survivors, 8 (67%) exhibited functional or anatomic evidence for neurological delay. Of 21 non-ECMO survivors, where adequate follow-up was available to make an assessment of neurological development, five (24%) exhibited evidence for delay. This difference was significant (P < .05, Fisher's Exact test). Of these five infants, three were premature, and one had DiGeorge syndrome. More ECMO survivors required diaphragmatic (67%) and abdominal (67%) patches at the time of diaphragmatic repair than non-ECMO survivors (4% and 12%, respectively; P < .05, Fisher's Exact test). In addition, more ECMO survivors required gastrostomy tube placement for feeding (50%) than non-ECMO survivors (16%; P < .05, Fisher's Exact test). A greater need for Nissen fundoplication in ECMO survivors (42%) than in non-ECMO survivors (12%) approached significance (P = .05, Fisher's Exact test). There were trends toward higher 1 and 5 minute APGAR scores and initial and best preoperative P(O2) in the non-ECMO survivors. A comparison between ECMO survivors who exhibited evidence of neurological delay with those who did not showed no differences in duration of ECMO, incidence of intracranial complications during ECMO, need for gastrostomy tube feeding or Nissen fundoplication, or incidence of carotid artery repair between the two groups. Infants surviving CDH who require ECMO have a greater incidence of neurological delay than those who do not. This is likely because of severity of the presenting illness as reflected by a greater need for diaphragmatic and abdominal patches during diaphragmatic repair, the need for Nissen fundoplication and gastrostomy tube feeding, and a trend toward poor APGAR scores and best preoperative P(O2) levels in these patients. However, there may be characteristics of ECMO, as yet unidentified, that may contribute to this outcome.