The aim of this study was to determine predictors of response whilst using high frequency jet ventilation (HFJV) for infants in intractable respiratory failure, in order to avoid the utilization of extracorporeal membrane oxygenation (ECMO). We reviewed patient demographics, ventilator parameters, blood gas values, length of oxygen therapy and use of surfactant and outcome, in infants given a 4 h trial of HFJV as the minimum to eliminate those infants where HFJV is used as a bridge to ECMO. The study was carried out in the neonatal intensive care nursery at Kosair Children's Hospital in Louisville, Kentucky, which provides high frequency ventilation and ECMO. Thirty infants who were eligible for ECMO and who met the study criteria were divided into two groups based upon response to HFJV. Twenty two infants responded to HFJV and eight nonresponders required ECMO after a trial of HFJV. Infants responding during HFJV demonstrated a significant decrease in oxygenation index without an escalation of mean airway pressure within 4 h after the initiation of HFJV. These infants had lower birth weights and an increased incidence of respiratory distress syndrome. No statistical differences were found in length of ventilation, days of oxygen therapy or duration of HFJV between the groups. Infants in intractable respiratory failure, who are eligible for extracorporeal membrane oxygenation, should receive a trial of high frequency jet ventilation, especially if the cause is respiratory distress syndrome unresponsive to surfactant therapy. During high frequency jet ventilation, the oxygenation index and mean airway pressure should be monitored serially, since they may predict the need for extracorporeal membrane oxygenation.