RESUMO
OBJECTIVE: To test the hypothesis that routine use of a centrifugal pump in the cardiopulmonary bypass circuit would result in a lower incidence of early neuropsychologic deficit when compared with conventional roller pumps. DESIGN: Prospective, randomized, double-blind. SETTING: University teaching hospital. PARTICIPANTS: Patients (n = 103) scheduled for elective coronary artery surgery. INTERVENTIONS: Patients were randomized into group C (centrifugal pump for cardiopulmonary bypass; n = 54) and group R (roller pump for cardiopulmonary bypass; n = 49). MEASUREMENTS AND MAIN RESULTS: A neuropsychologic test battery of 6 standard tests was administered before surgery and 5 days after surgery. An abnormal test result was defined as deterioration by >1 group SD from an individual's preoperative test performance. There were no significant differences between groups in preoperative or surgical parameters, intensive care unit stay, or hospital stay. There were no significant differences in the incidence of neuropsychologic deficit for patients with a deficit in at least 1 test (group C, 33%; group R, 51%; odds ratio, 0.48; 95% confidence interval, 0.22 to 1.06) or patients with a deficit in >/=2 tests (group C, 6%; group R, 18%; odds ratio, 0.26; 95% confidence interval, 0.07 to 1.03). In group R, there were more individual test deficits per patient than in group C (p = 0.04). CONCLUSION: There was no significant difference in the incidence of neuropsychologic deficit postoperatively with routine use of centrifugal pumps. The larger number of individual test deficits in the roller pump group suggest that further studies to assess the potential neuropsychologic benefits of the use of centrifugal pumps are warranted.