OBJECTIVES Neurologic impairment, at least partly ischemic in origin, has been reported in up to 25% of infants undergoing cardiopulmonary bypass, with or without circulatory arrest. Controversy continues about the effect of pump flow, pulsatile or nonpulsatile, on the brain and in particular on cerebral blood flow. This study examines the relationship between pump flow rate and cerebral hemodynamics during pulsatile and nonpulsatile cardiopulmonary bypass. METHOD Near-infrared spectroscopy was used to determine cerebral blood flow and cerebral blood volume (measured as concentration change) in a randomized crossover study. Pulsatile and nonpulsatile flow were used for six 5-minute intervals at each of three different pump flow rates (0.6, 1.2, and 2.4 L x m2 x min(-1)) in 40 patients, median age 2 months (range 2 weeks to 20 years 5 months). The relations between pulsatile flow, pump flow rate, cerebral blood flow, hemoglobin concentration change (cerebral blood volume), mean arterial pressure, arterial carbon dioxide tension, and hematocrit value were prospectively examined by means of multivariate analysis. RESULTS Cerebral blood flow decreased 36% per L x m(-2) x min(-1) decrease in pump flow rate and was associated with changes in mean arterial pressure but did not differ according to pulsatility. Change in hemoglobin concentration was unrelated to changes in pulsatility of pump flow. CONCLUSION Cerebral blood flow is related to pump flow rate. Pulsatile flow delivered with a Stöckert pump does not increase cerebral blood flow or alter hemoglobin concentration during cardiopulmonary bypass in children.