Aims The aim of this study was to examine whether tourniquet use can improve perioperative blood loss, early function recovery, and pain after primary total knee arthroplasty (TKA) in the setting of multiple-dose intravenous tranexamic acid. Methods This was a prospective, randomized clinical trial including 180 patients undergoing TKA with multiple doses of intravenous tranexamic acid. One group was treated with a tourniquet during the entire procedure, the second group received a tourniquet during cementing, and the third group did not receive a tourniquet. All patients received the same protocol of intravenous tranexamic acid (20 mg/kg) before skin incision, and three and six hours later (10 mg/kg). The primary outcome measure was perioperative blood loss. Secondary outcome measures were creatine kinase (CK), CRP, interleukin-6 (IL-6), visual analogue scale (VAS) pain score, limb swelling ratio, quadriceps strength, straight leg raising, range of motion (ROM), American Knee Society Score (KSS), and adverse events. Results The mean total blood loss was lowest in the no-tourniquet group at 867.32 ml (SD 201.11), increased in the limited-tourniquet group at 1024.35 ml (SD 176.35), and was highest in the tourniquet group at 1,213.00 ml (SD 211.48). The hidden blood loss was lowest in the no-tourniquet group (both p < 0.001). There was less mean intraoperative blood loss in the tourniquet group (77.48 ml (SD 24.82)) than in the limited-tourniquet group (137.04 ml (SD 26.96)) and the no-tourniquet group (212.99 ml (SD 56.35); both p < 0.001). Patients in the tourniquet group showed significantly higher levels of muscle damage and inflammation biomarkers such as CK, CRP, and IL-6 than the other two groups (p < 0.05). Outcomes for VAS pain scores, limb swelling ratio, quadriceps strength, straight leg raising, ROM, and KSS were significantly better in the no-tourniquet group at three weeks postoperatively (p < 0.05), but there were no significant differences at three months. No significant differences were observed among the three groups with respect to transfusion rate, thrombotic events, or the length of hospital stay. Conclusion Patients who underwent TKA with multiple doses of intravenous tranexamic acid but without a tourniquet presented lower total blood loss and hidden blood loss, and they showed less postoperative inflammation reaction, less muscle damage, lower VAS pain score, and better early knee function. Our results argue for not using a tourniquet during TKA. Cite this article: Bone Joint Res 2020;9(6):322–332.