Aprotinin effects on postoperative bleeding, redo operations due to excessive bleeding and postoperative blood products use have been retrospectively evaluated on a population of 894 consecutive patients operated upon between 1987 and 1991 for valvular surgery. In this period, aprotinin has been routinely used following the "High Dose" protocol starting from January 1990. To analyse results, all the patients considered (Group I) were divided into subgroups, following the heart valve operated upon, in "mitral" patients (391 patients, 43.7%, Group II), "aortic" patients (375 patients, 41.9%, Group III) and "mitro-aortic" patients (128 patients, 14.3%, Group IV). Each of these Groups has been subsequently split depending on whether or not they received aprotinin (Subgroups IA, IIA, IIIA, IVA with aprotinin, subgroups IB, IIB, IIIB, IVB without aprotinin). Considering the whole population (Group I), aprotinin determined a significant reduction of post-operative bleeding (499 +/- 634 ml Group IA versus 713 +/- 572 Group IB, p = 0.000), redo operations for bleeding (15/410, 3.7% Group IA versus 45/484, 9.3% Group IB, p = 0.0000) and consequently the percentage of patients exposed to blood products transfusion (37/410, 9% Group IA versus 163/484, 54.3% Group IB p = 0.0000). When the other Groups (II, III and IV) were considered, aprotinin determined a significant reduction of postoperative bleeding and of donor blood transfusions, while redo operations for bleeding, although less in total number, were not significantly reduced. Moreover, aprotinin has been effective in reducing postoperative blood losses, redo-operation for bleeding and blood use independently of the kind of oxygenator used: bubble vs hollow fibers.(ABSTRACT TRUNCATED AT 250 WORDS)