BackgroundTotal knee arthroplasty (TKA) is associated with significant perioperative blood loss and postoperative allogenic blood transfusion. Tranexamic acid (TXA) reversibly blocks lysine binding sites on plasminogen molecules and inhibits plasmin formation. Comparisons of the efficacy and safety of intra-articular and intravenous TXA in primary TKA have not previously been reported.MethodsA prospective randomized trial was conducted in 150 patients who underwent TKA, and these patients were randomized into three groups. Patients in Group A were injected by intra-articular TXA according to body weight (20 mg/kg). Patients in Group B received a standard dose of intra-articular TXA (2000 mg), and those in Group C were infused with TXA according to body weight (20 mg/kg) before tourniquet deflation and again 3 h later. Baseline characteristics and data collected at blood transfusion were compared. Differences among four time points (baseline, day 0, day 2, and day 5) were carried out using ANOVA.ResultsThe hemoglobin levels at postoperative day 5 were 10.6 g/dL for Group A, 10.6 g/dL for Group B, and 10.7 g/dL for Group C. The drain output was 399 ml for Group A, 314 ml for Group B, and 305 ml for Group C (p = 0.03). Group C had significantly less drain output than Group A after post hoc comparisons (p = 0.05), whereas no significant difference was observed between Group A and B (p = 0.09) or between Group B and C.ConclusionThe weight-adjusted dose of TXA administered intravenously significantly reduced the drain output but not the total blood loss when compared with the weight-adjusted dose of TXA administered intra-articularly. No significant difference was observed in the other parameters among the three groups.Trial registrationThe Joint CUHK-NTEC CREC, CRE-2013.644-T. Registered 1 March 2014.