The major purpose for the use of autotransfusion is to prevent the transmission of blood borne infectious agents, such as human immunodeficiency virus and non-A & non-B hepatitis virus. To evaluate the efficacy and quality of autolgous shed blood for autotransfusion, eighty patients who had total knee arthroplasty from Dec. 1992 to Mar. 1994, were included in one of two groups: Group I, who received the autotransfusion, or Group II, who did not. Each group included 20 patients of unilateral TKR and 20 patients of bilateral TKR. The Orth-evac system(Deknatel, USA) was used to salvage drained blood in the first six hours after the operation. All of the patients were evaluated for the postoperative blood loss, transfusion requirements, hemoglobin, hematocrit, platelet count, blood pressure and body temperatures. l. In bilateral TKR, the reinfusion of shed blood reduced the requirements for homologous blood by 41.4%(1.2 pints in group I versus 2.9 pints in group II). In unilateral TKR, it was decreased to 36.4%(0.4 pint in group I vs 1.1 pint in group II). 2. In bilateral TKR, the requirements for homologous transfusion was decreased from 95% of patients in control group to 55% in group I .In unilateral TKR, it was decreased from 60% to 20%. 3. There were four patients who had high fever above 39℃ after autotransfusion. 4. At the immediate postoperative period there were two patients who had hypovolemic shock in group I patients who had bilateral TKR. 5. There was no clotting abnormality, no transfusion reaction and no thromboembolic disease in group I patients. In conclusion, the reinfusion of autologous shed blood after TKR is an acceptable alternative to the homologous transfusion without untoward effect.