Objective: To explore the influences on peri-operative blood loss and postoperative blood transfusion in primary total knee arthroplasty when using different methods to release the pneumatic tourniquet. Methods: From January 2015 to December 2016, 484 patients with osteoarthritis were enrolled, undergoing primary unilateral total knee arthroplasty in the Orthopedics Department of Peking University Third Hospital. Retrospective data was used, and all the patients had been operated on by three independent working groups. Group A was the strategic tourniquet release group. After the completion of all osteotomy and soft tissue balance, the tourniquet was released for five minutes for coagulation of bleeding tissue and was then inflated again before component cementation untill the end of the operation. In Group B, full-time tourniquet usage was applied to patients, and was just released post wound closure and pressure dressing. In Group C, the tourniquet was inflated from the beginning of incision to the end of cementation. It was then released after the prosthesis was cemented. A series of data in these three groups were observed and recorded, including output of drain, decrease of hemoglobin level in the first and third day after operation, the total measured blood loss, the hidden blood loss, the number of transfused patients, and the rates of allogenic blood transfusion. Results: The decrease of hemoglobin level in the first day after operation, the output of drain and the total level of measured blood loss were slightly higher in Group A. The decrease of hemoglobin level in the third day after operation and the hidden blood loss was slightly higher in Group B. Group C had the lowest total measured blood loss and hidden blood loss among the three groups. However, no significant differences of outcomes were observed among these groups. A total of 29 patients received blood transfusion(Group A 10/165, 6.1%; Group B 11/177, 6.2%; Group C 8/142, 5.6%). There was no significant differences among the three groups in output of drain(P=0.496), decrease of hemoglobin level(P=0.242), total measured blood loss(P=0.390), hidden blood loss(P=0.204), or postoperative blood transfusion rate(P=0.976). Conclusions: Under the routine application of the multimodal blood-loss prevention approach, including tranexamic acid, drain under atmospheric pressure, compression and cryotherapy after operation, the different choices of tourniquet releasing strategies have no significant effect on peri-operative blood loss or postoperative blood transfusion rate in total knee arthroplasty.