A noninvasive, simple, and continuous method to assess cerebral perfusion during cardiopulmonary bypass (CPB) could help prevent cerebral ischemia. Transcranial Doppler sonography (TCD) allows a noninvasive, on-line measurement of blood flow velocity in cerebral arteries. The correlation of TCD-estimated and actual cerebral blood flow (CBF) has not been well studied during CPB. We determined the correlation of middle cerebral artery (MCA) mean velocity and CBF determined by the Kety-Schmidt method during nonbypass and two hypothermic bypass flow conditions. Sixteen patients undergoing hypothermic (27 degrees C) CPB for coronary artery bypass grafting and/or valve replacement surgery were enrolled in the study. We were able to determine MCA velocity in only 12 patients. We determined CBF and MCA velocity in each patient during four 15-min study periods: 1) prebypass after sternotomy before aortic cannulation; 2) hypothermic (27 degrees C) CPB with 1.2 L.min-1.m-2 pump flow; 3) hypothermic CPB with 2.4 L.min-1.m-2 pump flow, and 4) 30 min after weaning from CPB. There was no difference in the mean arterial pressure between the two CPB pump blood flows. The pooled change in MCA velocity and CBF as percentage of baseline (prebypass) for all patients and at all time points had a correlation of 0.33 (r). A decrease or increase in MCA velocity did not necessarily indicate a corresponding decrease or increase in CBF. This technology may be of limited usefulness during the circulatory condition of hypothermic, nonpulsatile CPB.