Hyperoxic cardiopulmonary resuscitation (CPR) is associated with an increase in neurologic dysfunction upon successful resuscitation with much of the damage attributable to an increase in reperfusion oxidant injury. We hypothesized that by contrast, hypoxic ventilation during resuscitation would improve neurologic outcome by reducing available substrate necessary for oxidant injury. Specifically, this study investigated the effects of 2 levels of hypoxic ventilation during resuscitation: F1O2 = 0.085, PaO2 = 26.6 +/- 3.4 mmHg, (HY8), and F1O2 = 0.12, PaO2 = 33.0 +/- 4.2 mmHg, (HY12), and normoxic resuscitation: F1O2 = 0.21, PaO2 = 60.6 +/- 17.0 mmHg, (N) on survival and neurological outcome following 9 min of normothermic cardiac arrest. Concentrations of malonaldehyde (MDA) and 4-hydroxynonenal (4-OH) in plasma and concentrations of glutathione (GSH) in erythrocyte lysates were measured to quantify possible radical damage. Physiological variables including arterial blood gases were followed for 24 h after resuscitation. Neurologic outcome was assessed using a standardized scoring system. Hypoxically (HY8) resuscitated dogs tended to have a greater neurologic deficit than normoxically resuscitated dogs and had reduced overall survival (16.9 +/- 8.9 h) compared to N dogs (24.0 +/- 0.0 h). Overall survival time correlated negatively (-0.693) and significantly (P = 0.0018) with plasma glucose concentration. Arterial plasma glucose concentrations were higher in the HY8 group compared to the N group immediately (HY8, 312 +/- 86 mg/dL; N, 196 +/- 82 mg/dL; P = 0.17) and 30 min (HY8, 331 +/- 109 mg/dL; N, 187 +/- 74 mg/dL; P = 0.077) following resuscitation. No statistically discernible differences in markers of oxidant injury were apparent among the 3 groups, but pooled data increased significantly with time for MDA and 4-OH. Pooled data for GSH showed a significant drop at 1 h following resuscitation and returned to normal by 6 h. Data from these markers suggested attendant oxidant injury in all groups. Thus, hypoxic ventilation at 2 depths of hypoxia during resuscitation failed to improve neurologic outcome beyond that achieved by ventilation with air, suggesting that normoxia rather than hyperoxia or hypoxia is the ideal target for arterial oxygenation during resuscitation.