The purpose of our study was to determine the effectiveness of a postoperative autologous blood reinfusion system as an alternative to homologous, banked blood transfusions in total knee arthroplasty. We carried out a prospective randomised controlled study on 60 patients having unilateral total knee replacements. In all these patients, the same surgical team applied the same surgical technique, and all patients followed the same rehabilitation program. In 30 of these patients (group A), a reinfusion system of unwashed salvaged blood was applied, and they were supplemented postoperatively with banked blood transfusions when required. A control group of 30 patients (group B), in whom standard suction drains were used, received one unit of homologous banked blood transfusion intraoperatively and additional blood transfusions postoperatively when required. The administration of banked blood transfusion was determined by the haemoglobin value (<9 mg/dl) and/or clinical signs (blood pressure, pulse, etc.). The values of haemoglobin, haematocrit and platelets were recorded preoperatively and the first, fifth and 15th postoperative days, respectively. Five patients of group A required nine units of homologous blood (0.3 units/patient) postoperatively. Ten patients of group B required an additional 15 banked blood units postoperatively (in total 45 banked blood units for group B; 1.5 units/ patient). In the study group, the total homologous blood requirements were reduced by 80%, while the postoperative blood requirements were reduced by 50%. There was no significant difference in the postoperative haematocrit and haemoglobin values between the two groups. The cost of the blood management in the study group was reduced by 36%. The use of an autologous blood reinfusion system reduces highly effectively the demands of homologous banked blood transfusion in total knee arthroplasty.