Patients who undergo cardiopulmonary bypass frequently have neuropsychologic dysfunction. This study was undertaken to determine whether altered cerebral perfusion and vascular responses may in part lead to these neuropsychologic changes. Pigs were placed on normothermic cardiopulmonary bypass for 2 hours. Basal cerebral blood flow and in vivo responses to administration by internal carotid artery of neuronally released vasoactive substances were evaluated before and 5 to 15 minutes after termination of cardiopulmonary bypass. Another group of pigs were placed on cardiopulmonary bypass for 2 hours and then perfused off bypass for 1 additional hour. In vitro responses of cerebral arterial microvessels (100 to 175 microns) from both groups were examined in a pressurized (40 mm Hg) no-flow state with videomicroscopy. Vessels from uninstrumented pigs served as control preparations for in vitro studies. Cerebrovascular resistance and cerebral perfusion were maintained constant during cardiopulmonary bypass and after separation from bypass. The internal carotid artery infusion of acetylcholine (cholinergic agonist) caused increased internal carotid artery blood flow before cardiopulmonary bypass but decreased blood flow after cardiopulmonary bypass. After 2 hours of cardiopulmonary bypass, the increase in internal carotid artery blood flow induced by isoproterenol (a beta-adrenoceptor agonist) was reduced, whereas the response to sodium nitroprusside (a guanylate cyclase activator) was unchanged. In vitro acetylcholine-induced microvascular vasodilation was converted to a contractile response and isoproterenol elicited less relaxation after 2 hours of cardiopulmonary bypass. One hour of cerebral perfusion after cardiopulmonary bypass caused a further reduction in isoproterenol-induced relaxation but had no further effect on the cholinergically mediated response. In vitro relaxation responses to sodium nitroprusside and forskolin (an adenylate cyclase activator) were similar in all experimental groups, suggesting that second-messenger mechanisms remain intact after normothermic cardiopulmonary bypass. In conclusion, basal cerebrovascular resistance and internal carotid artery blood flow are maintained if the systemic circulation and pressure are supported with fluid administration after cardiopulmonary bypass. Agonist-induced vasodilation of cerebral microvessels to cholinergic and beta-adrenoceptor stimulation are selectively impaired after normothermic cardiopulmonary bypass, whereas second-messenger mechanisms remain intact.