Five patients undergoing cardiopulmonary bypass (CPB) procedures were extensively monitored because of anticipated high risk for neurological complications. Arterial blood pressure (BP), central venous pressure, and epidural intracranial pressure (EDP) were continuously recorded throughout CPB; thus, information on the cerebral perfusion pressure (CPP) was also continuously available (CPP = BP - EDP). Cerebral electrical activity was recorded by a cerebral function monitor. The flow velocity in the middle cerebral artery (MCA) was recorded using a transcranial Doppler technique. During steady-state CPB (constant hematocrit, constant temperature, and constant flow from the heart-lung machine) partial pressure of arterial carbon dioxide (PaCO2) was repeatedly changed to study the effect of changes in this variable on MCA flow velocity during nonpulsatile bypass. During CPB with constant temperature, hematocrit, and PaCO2, the effect of changes in CPP on MCA flow velocity was recorded and analyzed. During nonpulsatile, moderately hypothermic (28 degrees to 32 degrees C), low-flow (1.5 L/min/m2) CPB, there was no evidence of cerebral autoregulation, with CPP levels ranging from 20 to 60 mm Hg. The CO2 reactivity, however, was clearly present and in the range of 1.9 to 4.1%/mm Hg, indicating that there was a dissociation between cerebral autoregulation and CO2 reactivity under these circumstances.