The efficiency of post-operative cell saving after major joint arthroplasty has been demonstrated in prospective studies focusing on blood management. In everyday practice, however, it is likely that transfusion policy is followed less rigorously because of a slackening in attention to blood management, with a reduced efficiency of post-operative cell saving. The primary research question of this retrospective study was whether the number of allogeneic blood transfusions administered to patients treated with a retransfusion system was similar to the results found in a preceding prospective study. A total of 438 patients treated with the Bellovac ABT retransfusion system were analysed in which the majority was operated on a total hip arthroplasty (THA) and total knee arthroplasty (TKA). The amount of retransfused shed blood, the perioperative haemoglobin levels and the number of allogeneic blood transfusions were registered. The average amount of retransfusion was 152 mL in THA and 410 mL in TKA, whereas the allogeneic blood transfusion rate was 8·4 and 5·1% in both groups, respectively. The average percentage of allogeneic blood transfusions administered in this study (i.e. 7%) proved to be marginally higher than the percentage found in a preceding prospective study (i.e. 6%) because of slackening of attention for transfusion policy in everyday practice. Limited bone resection procedures such as resurfacing THA or unicompartmental knee arthroplasty were associated with very limited shed blood and low risk of allogeneic blood transfusion, indicating the doubtful cost efficiency of using a retransfusion system in these patients. It can be concluded that the efficiency of the retransfusion system in everyday practice was similar to the efficiency shown in a preceding prospective study focusing on blood management. However, continual training of the clinical team is crucial.