The risks associated with transfusion can be minimized with autologous blood. The efficiency of preoperative deposit, preoperative hemodilution and intra- and postoperative autotransfusion in reducing homologous transfusions has been demonstrated. There seem to be few studies, however, that compared the different methods of autologous transfusion. This study was designed to evaluate the comparative efficiency of these methods. PATIENTS AND METHODS. Sixty-four patients scheduled for total hip arthroplasty were randomly divided into four groups: group I--preoperative autologous deposit: group II--preoperative hemodilution; group III--intra- and postoperative autotransfusion; group IV--control. Preoperative autologous donations were stored in CPDA-1 buffer. Three units of 450 ml were requested. A predonation hemoglobin (Hb) concentration of 11 g dl was required. Surgery was carried out in the 5th week after the first donation. Preoperative hemodilution to Hb 9 g/dl was carried out after induction of anesthesia and initial circulatory stabilization. A cell separator was used for intra- and postoperative autotransfusion. Postoperative autotransfusion of drainage blood was continued until 6 h after the beginning of the operation. Polygeline was used for volume resuscitation. If the Hb concentration fell below 9 g/dl in the operating room and intensive care unit or below 10 g/dl in the general ward, autologous blood or homologous packed red cells were transfused. Autologous blood collected with the cell separator was retransfused at the end of the operation and after the autotransfusion period irrespective of the actual Hb concentration. RESULTS. The general data of the patients, blood loss, and Hb concentration at the beginning of the study and postoperatively were comparable in the four groups. Homologous transfusion requirements amounted to 0 (0-1250) ml (median, range) packed red cells in group I (preoperative deposit). 500 (0-2000) ml in group II (hemodilution), 125 (0-1000) ml in group III (autotransfusion) and to 500 (0-1500) ml in group IV (control). In group I 14 of 16 patients, in group II 1 of 16, in group III 8 of 16 patients, in group IV 5 of 15 patients did not require homologous transfusion. The difference between group I and IV was significant (p = 0.004 and p = 0.003). Global coagulation tests, antithrombin III, and total serum protein were comparable in the four groups. DISCUSSION. The efficiency of preoperative hemodilution to reduce homologous transfusion requirements is limited]. In the present study, as in two other recent studies, hemodilution did not reduce homologous transfusion requirements. Autotransfusion with a cell separator can save approximately 50% of the erythrocytes lost during hip arthroplasty and 70% of the drainage loss. The homologous transfusion requirements for the autotransfused group reported here were less than in the control group; the difference, however, was not statistically significant. Patients participating in preoperative autologous deposit did not require homologous blood for hip arthroplasty in 62%-70% of cases in other investigations; in the present study 88% of the patients did not require homologous blood. CONCLUSION. Under the conditions studied, preoperative autologous deposit was the most efficient method of autologous transfusion for hip arthroplasty. It should be employed primarily.