We hypothesized that the central retinal artery color Doppler (CRAD) technique can be used as an alternative monitoring method to the conventional transcranial Doppler (TCD) technique which does not consistently provide good signals during cardiopulmonary bypass. This is the first study to compare CRAD with TCD during cardiac surgery. Twenty patients without stroke were examined at the seven preset stages of the two procedures. The maximum flow velocity (Vmax) of the right central retinal artery was measured with a 7.5-MHz Doppler system. The Vmax of the right middle cerebral artery was measured with a TCD system. To estimate the relationship between Vmax and maximum blood pressure (BPmax), CRAD-Vmax (6.6 +/- 2.2 cm/s), TCD-Vmax (44.0 +/-14.8 cm/s), and BPmax (114 +/- 28.5 mmHg) obtained at the initial stage were normalized. The Vmax values of CRAD were successfully obtained, at each of the seven study stages, in all 20 patients. The Vmax values of TCD were obtained at 125 points, and 15 missing values were seen during the stages of cardiopulmonary bypass. CRAD-Vmax was correlated to TCD-Vmax (r = 0.775, P < 0.0001). The normalized values of both TCD- and CRAD-Vmax were correlated to the normalized BPmax (r = 0.766, P < 0.0001) (r = 0.787, P < 0.0001). The accuracy of CRAD was confirmed by the significant correlation with TCD.