CO2 kinetics during CPR was investigated in 15 anesthetized piglets. BP, blood gases, and acid-base balance were monitored through catheters in the carotid artery and a central vein, as well as in cerebrospinal fluid. Cardiac arrest was induced by a transthoracic direct current shock. CPR was begun immediately by artificial ventilation and simultaneous external chest compressions. Epinephrine was administered after 8 min of CPR. One group (n = 5) of animals received no buffer treatment while another (n = 5) received an infusion of 75 mmol sodium bicarbonate and a third group (n = 5) received an equivalent amount of tris-buffer mixture. The results of these experiments, as well as previously described circulatory variables during CPR, were analyzed using a computer model describing the CO2 kinetics of the pig. Our main finding was that PaCO2 was positively correlated to cardiac output during CPR; improved cardiac output during CPR resulted in more efficient tissue CO2 elimination and was associated with increased survival rates. PaCO2 was also somewhat reduced by efficient alveolar hyperventilation. The arterial PCO2 and pH did not reflect the acid-base balance in peripheral tissues. During CPR, bicarbonate and tris-buffer mixture both quickly passed through the blood-brain barrier. When buffer treatment is indicated during CPR, a buffer which does not increase tissue PCO2 may be the drug of choice.