BACKGOUND: Arterial blood gas analysis is essential on diagnosis and treatment of hypoxia and acid-base imbalance. It is important to decide the timing of arterial blood sampling as well as sampling method, sample transport, and analysis of the results. So we investigated to the adequate timing of sampling when inspired oxygen fraction is changed from 0.5 to 1.0. METHODS: 20 patients were anesthetized with enflurane-N2O-O2 (FiO2=0.5), and paralyzed with pancuronium 0.07~0.08 mg/kg. Ventilation was controlled with Ohmeda 7000 ventilator (BOC Health Care Inc., Madison, USA), using a constant tidal volume of 10 ml/kg and respiration rate of 12/min. After 1 hour of anesthesia, the nitrous oxide inhalation was stopped and 100% oxygen was inhaled, and then arterial blood gas values were measured at 2 min intervals for 20 min, 5 min intervals for next 30 min, and 10 min later. Blood samlpes were drawn from the radial artery and measured immediately on a blood gas analyzer (Civa-Corning 288 Blood Gas System, Civa-Corning Diagnostic Corp., Medifield, USA). Determining the optimal time of sampling was performed with the rate of variation of PaO2 according to time progression, then the point at which the slope decreased abruptly was regarded as statistically significant timing. RESULTS: After 12 minute, arterial oxygen partial pressure was not any more changed significantly. There were no change of pH, arterial carbon dioxide partial pressure, oxygen saturation, base excess, and bicarbonate. CONCLUSION: The timing of arterial blood gas sampling in change with inspired oxygen fraction from 0.5 to 1.0 is about 12 minute later.