The effects of pulsatile and nonpulsatile flow during cardiopulmonary bypass (CPB) with of centrifugal pump (Sarns) and membrane oxygenator, on blood cells, hemodynamics, and hormonal response were studied. In the pulsatile group (group P) in which pulsatile flow was generated by centrifugal pump and a 20 Fr arterial cannula was used, hemolysis and reduction of platelet count during CPB were more marked than in the nonpulsatile group (group NP), in which the same type of circuit was used. When the 20 Fr arterial cannula was replaced with a 24 Fr cannula (group Pc), the rate of hemolysis during CPB was significantly reduced compared with that in group P (p less than 0.05). The rate of rise in plasma free hemoglobin from 10 to 70 minutes CPB in group Pc was 15.0 mg/dl/hr, this value did not exceed that in either group NP or in group Pr, in which a roller pump rather than centrifugal pump was used to generate pulsatile flow. These findings show that pulsatile CPB with a centrifugal pump produces no deleterious hematologic effect in clinical use. The rise in the level of angiotensin II in group P was significantly smaller than that in group NP (p less than 0.05), and the rise in plasma renin activity and levels of angiotensin I, adrenalin and noradrenaline were smaller than those in group NP, although these differences were no significance. These findings indicate that the centrifugal pump generates pulsatile flow effectively, although not so effectively as to prevent the rise in peripheral vascular resistance. During CPB, there was no change in levels of thyroid hormones, including free T3, free T4 and reverse T3, in either pulsatile groups P and Pc or nonpulsatile group. TSH level in group Pc was significantly elevated in contrast with that in the nonpulsatile group (p less than 0.05), in which no change in TSH level was seen. It is suggested that pulsatile perfusion using a centrifugal pump might maintain sufficient hypothalamic-pituitary function to permit secretion of TSH in response to various stimuli.