We tested the hypothesis that hypoxic newborn piglets can be successfully resuscitated with lower O2 concentrations than 21%. Severely hypoxic, 2-4-d-old, anesthetized piglets were randomly divided into five resuscitation groups: 21% O2 (n = 10), 18% O2 (n = 9), 15% O2 (n = 9), 12% O2 (n = 8), all normoventilated, and a hypoventilated 21% O2 group (PaCO2; 7.0-8.0 kPa, n = 9). Base excess (BE) reached -20 +/- 1 mmol/L at the end of hypoxia. After 3 h of resuscitation, BE had risen to -4 +/- 1 mmol/L in the 21% O2, 18% O2, and hypoventilated groups, but was -10 +/- 2 mmol/L in the 15% O2 group (p < 0.05 versus 21% O2 group) and -22 +/- 2 mmol/L in the 12% O2 group (p < 0.05 versus 21% O2 group). Four animals died during resuscitation, all allocated to the 12% O2 group (p < 0.05 versus 21% O2 group). Somatosensory evoked potentials (SEPs) recovered in 39 of 45 piglets, and remained present during resuscitation in all except the 12% O2 group. SEP recovered initially even in six of eight animals in the 12% O2 group, but disappeared again in all later during resuscitation. The SEP amplitude recovered to levels not significantly different from the 21% O2 group in all groups except the 12% O2 group. Plasma hypoxanthine concentrations and extracellular hypoxanthine concentrations in the striatum decreased during resuscitation to levels not significantly different from the 21% O2 group in all but the 12% O2 group (p < 0.05 versus 21% O2 group). In conclusion, severely hypoxic newborn piglets were resuscitated as efficiently with both hypoventilation and 18% O2 as with 21% O2.