Clinical techniques for artificial perfusion have not previously been examined directly for their effects on brain high-energy metabolism. Our study involved 24 large mongrel dogs that were anesthetized, instrumented for central venous intravenous access, and subjected to craniotomy to expose the dura and underlying parietal cortex. The animals were divided into the following six experimental groups of four animals each: 1) nonischemic controls; 2) 15-minute cardiac arrest without resuscitation; 3) 45-minute cardiac arrest without resuscitation; 4) 15-minute cardiac arrest plus 30 minutes resuscitation with conventional cardiopulmonary resuscitation (CPR); 5) 15-minute cardiac arrest plus 30 minutes resuscitation with interposed abdominal compression (IAC) CPR; and 6) 15-minute cardiac arrest plus 30 minutes resuscitation with internal cardiac massage. Cardiac arrest was induced by central venous injection of KC1 0.6 mEq/kg, and it was confirmed by continuous ECG monitoring. The three active resuscitation models included administration of NaHCO 3 and epinephrine, but no iattempt was made to restart the heart by defibrillation during resuscitation, At the indicated time in each group, a 4-to 5-g sample of brain was removed through the craniotomy, immediately cooled to 0 C and processed for isolation of mitochondria. The mitochondria were studied for their content of superoxide dismutase and for quantitative oxygen consumption with glutamate/ malate substrate during resting and ADP-stimulated respiration. Our results show a significant drop in brain mitochondrial superoxide dismutase activity during the first 15 minutes of cardiac arrest. There is minimal injury to brain mitochondrial oxygen consumption during both 15 and 45 minutes of complete ischemia. IAC-CPR, h~wever, produces severe inhibition of rnitochondrial oxygen utilization. In coP_trast, internal cardiac massage after the 15-minute cardiac arrest is not associated with significant additional mitochondrial injury. These data show that selection of an artificial perfusion technique associated with optimum brain blood flows may be critical to the status of the high-energy machinery in the postischemic brain.