Study protocol Neonates were enrolled after informed consent was obtained from parents. At enrollment, postductal arterial blood samples were drawn for determination of pH, blood gas tensions, and methemoglobin saturation (270 Cooxime, Ciba-Corning, Diagnostics, Medfield, Massachusetts, USA) 10 min prior to the treatment with INO and every 2–4 h thereafter. The mean oxygenation index of all neonates [OI = mean airway pressure × fractional inspired concentration of oxygen (FiO2)/post-ductal partial pressure of arterial oxygen (PaO2)] during high-frequency ventilation and before starting INO was 46.3 ± 5 (mean ± SEM). The NO gas (AHG, Jeddah, Saudi Arabia) used in this study was certified at a concentration of 800 ppm NO with < 1% contamination by other oxides of nitrogen. NO gas was introduced into the ventilator circuit via an adaptor positioned on the inspiratory port of the Fisher and Paykel humidification chamber. Thus, NO was mixed with the bias flow gas of the oscillator and subsequently delivered to the neonate via the inspiratory limb of the ventilator circuit. The resulting concentration of the inhaled NO and NO2 was verified in-line by using an electrochemical sensor (Pulmonox, Tofield, Alberta, Canada). Exhaled gas was scavenged; the oxygen concentration was analyzed continuously before it reached the neonate's endotracheal tube. Nitric oxide was initially administered at 20 ppm for at least 2 h. If there was no response while the neonate was on high ventilatory support and FIO2 of 1.0, INO was increased gradually by 2 ppm to a maximum of 80 ppm. If there was a response, INO was maintained at 20 ppm and FIO2 was gradually decreased to 0.6, provided the PaO2 was 80–120 mmHg. Nitric oxide then was weaned to discontinuation. Ventilatory parameters thereafter were weaned and HFOV was replaced by a conventional ventilator. An arterial/alveolar oxygen (a/A) ratio less than 0.22 was used to define failure of HFOV and INO therapy, if INO reached 80ppm on high ventilatory support.