From March 1990 to January 1993, a randomized prospective study was performed to determine the optimal timing of surgery for infants with high-risk congenital diaphragmatic hernia (CDH). Thirty-two CDH patients who presented with respiratory distress within 12 hours after birth were randomly divided into two groups: Group A had early repair (within 6 hours), and group B had delayed repair (at 96+ hours). Extracorporeal membrane oxygenation (ECMO) was initiated in both groups as necessary. Fourteen patients were assigned to group A, and 18 were assigned to group B. Two patients initially assigned to group A had acute deterioration, and their operations had to be postponed. Data were collected, but these patients were eliminated from the study. The two groups were comparable based on gestational age, birth weight, Bohn's criteria, and oxygenation and ventilatory index. Nine of 12 group A patients (75%) survived, and 13 of 18 group B patients (72%) survived (P > .05, not significant). The ECMO requirements for the two groups were not significantly different (8 of 12 (67%) v 16 of 18 (89%); P > .05). Surgical intervention for bleeding complications related to ECMO was required in three of eight (38%) with immediate repair and seven of 16 (44%) with delayed repair (P > .05). There was no difference in survival nor incidence of ECMO between the two groups. This is the first prospective study of timing of hernia repair that supports the conclusions of earlier reports of retrospective studies.