Objective: To determine whether inhaled nitric oxide (NO) administered during conventional mechanical ventilation could produce improvements in oxygenation and reduce the incidence of meeting extracorporeal membrane oxygenation (ECMO) criteria in infants with hypoxemia. Design: Prospective, randomized, controlled trial: Enrolled infants were assigned to conventional treatment with or without adjunctive inhaled NO. Control infants meeting failure criteria (partial pressure of arterial oxygen (PaO2)<80 mm Hg (10.7 kPa)) were allowed to cross over. Caregivers were not masked to group assignment. Setting: Neonatal intensive care units at the University of Alabama Hospital and the Children's Hospital of Alabama, October 1993 to May 1994. Patients: Newborn infants, both term and near-term, with PaO2 less than 100 mm Hg (13.3 kPa) who were receiving mechanical ventilation with 100% oxygen. Exclusion criteria included major congenital anomalies, diaphragmatic hernia, profound asphyxia and significant bleeding. Interventions: Inhaled NO was initiated in the NO group at a dose of 20 to 40 ppm and advanced stepwise to 80 ppm if Pao2 remained less than 100 mm Hg (13.3 kPa). Outcome measures: Primary outcome variables were treatment failure and meeting of ECMO criteria before crossover. Improvement in oxygenation and ultimate use of ECMO or high-frequency oscillatory ventilation were secondary outcome variables. Results: Seventeen neonates with hypoxemia were enrolled; 16 had echocardiographic evidence of pulmonary hypertension, and eight had extrapulmonary shunting. At 1 hour of treatment, two infants in the NO group responded with increases in PaO2 of more than 100 mm Hg (13.3 kPa); after crossover, two had increases in PaO2 of more than 10 mm Hg (1.3 kPa) and one control infant had an increase in PaO2 of more than 10 mm Hg (1.3 kPa). All control infants met failure criteria and crossed over to receive NO; two had increases n PaO2 of more than 10 mm Hg (1.3 kPa) with NO treatment. Despite initial responses, all subjects in both groups eventually met failure criteria. There were no differences between groups in primary outcome variables Conclusions: Although inhaled NO produced a transient improvement in oxygenation in some infants, it did not reduce the incidence of meeting ECMO criteria in this population.