PURPOSE Tourniquets are still widely used in total knee arthroplasty (TKA), although they may be associated with several adverse effects. An observer-blinded, randomized, controlled trial was performed to evaluate the effects of tourniquet use in TKA. METHODS Fifty participants who underwent staged bilateral TKA were recruited for this study. The first-side TKA was randomly allocated to either long-duration tourniquet use
or short-duration tourniquet use followed by a 3-month washout period and crossover to the other tourniquet strategy for the opposite-side TKA. Blood loss was monitored perioperatively. The operating time, allogeneic blood transfusion rate, thigh pain, knee pain, limb swelling, clinical outcome as measured by the Likert-type Western Ontario and McMaster University (WOMAC) score, straight leg raising and knee active range of motion (ROM) were also recorded. RESULTS The long-duration tourniquet group exhibited reduced total blood loss [-99.1 ml, 95 % confidence interval (CI) -168.1 to -30.1, P = 0.0411] and intraoperative blood loss (-225.2 ml, 95 % CI -369.5 to -80.9, P = 0.0071) compared with the short-duration tourniquet group. However, there were greater postoperative blood loss (69.6 ml, 95 % CI 21.1 to 118.2, P = 0.0282) and hidden blood loss (52.8 ml, 95 % CI 10.5 to 95.1, P = 0.0332) in the long-duration tourniquet group. The short-duration tourniquet group showed better outcomes for thigh and knee pain, limb swelling, WOMAC score at 6-week follow-up, straight leg raising and knee ROM. Similar allogeneic blood transfusion rates were observed for both groups. CONCLUSION Total and intraoperative blood losses were reduced with the long-duration tourniquet use, whereas the short-duration tourniquet use would reduce postoperative and hidden blood losses without increasing the allogeneic blood transfusion rate. In addition, short-duration tourniquet use would result in faster recovery and less pain during the early rehabilitation period following TKA. LEVEL OF EVIDENCE I.