Background: In spite of evidence suggesting that resuscitation with 100% O2 is detrimental, international guidelines still recommend its use. Clinical studies comparing 21% and 100% O2 included many infants with only mild and moderate asphyxia. Objectives: We aimed to investigate the effect of these oxygen fractions on haemodynamic parameters, arterial blood gases, oxygen saturation indices and markers of inflammation and hypoxic damage when resuscitating asystolic newborn pigs following asphyxia. Methods: Newborn swine (n = 32, age 12–36 h, weight 2.0–2.7 kg) were progressively asphyxiated until asystole occurred. Cardiopulmonary resuscitation was initiated with ventilation with either 21% (n = 16) or 100% O2 (n = 16). Return of spontaneous circulation (ROSC) was defined as a heart rate ≧100 min–1. Results: Mean time of hypoxia, pH, base excess and pCO2 at asystole were comparable between the groups. All animals except 2 in the 100% group achieved ROSC. One animal in the 21% group suffered bradycardia at baseline and was excluded. For the remaining 15 animals resuscitated with 21% O2, median time to ROSC (interquartile range) was 150 s (115–180), whereas animals in the 100% group achieved ROSC after 135 s (113–168); p = 0.80. There were no differences in the temporal changes in mean arterial blood pressure, heart rate, pH, pCO2, interleukin-1β or lactate/pyruvate ratios. However, systemic and regional cerebral oxygen saturations were higher in the animals resuscitated with 100% oxygen. Conclusion: In this animal model of severe perinatal asphyxia, resuscitation with room air seemed to be as safe and effective as the use of 100% oxygen.