“…In a previous studies, we found that USB return did not modify either the acute phase response or the cellular immune response to TKR [15,16], but it decreased the risk of receiving an ABT (48% vs. 11%, P < 0·01) and reduced the ABT index by 1 unit [9]. However, a transfusion protocol was not established, probably leading to unnecessary transfusions [17], as ABT rate in the control group was higher than that reported in recent TKR series [1,8,18,19]. In this present study, the introduction of a transfusion protocol (transfusion trigger, Hb < 9 g/dl) resulted in a reduction of ABT rate in patients managed without USB return with respect to our previous series (48% vs. 30·6%; relative risk (RR): 0·60, P < 0·01).…”