This was a single-center, prospective and randomized controlled study. Patients who underwent arthroscopic ACL reconstruction due to ACL rupture from December 2016 to December 2019 at our study location were enrolled consecutively. The exclusion criteria were as follows: previous knee procedures on the same side, renal disorder or insufficiency, abnormal coagulation profile, treatment with drugs interfering with coagulation or TXA clearance, multiple ligament injury, patients affected by blood, hemorrhagic or liver disease, and patients with nonsteroidal anti-inflammatory or antiplatelet therapy. This study was approved by the medical ethics committee of the Second Affiliated Hospital of Xi’an Jiaotong University. All participants gave written informed consent. Finally, 120 patients (74 males and 46 females) were included, with an average age of 31.0 ± 8.1 years (Fig. 1). The age, sex, body mass index (BMI) and time from injury to surgery of all patients were recorded.
Transtibial single-bundle reconstructions using autologous hamstring tendon grafts were performed in all patients; the graft was fixed using the RIGIDLOOP Adjustable Cortical System (DePuy Mitek, USA) for femoral fixation and MILAGRO Advance Interference Screws (DePuy Mitek) for tibial fixation. The intraoperative data registered were operative time, partial meniscectomy, meniscus repair, and tunnel diameter. General anesthesia or spinal/epidural block was selected according to the individual condition of each patient. The drainage tubes were clamped for 2 hours before release. A pneumatic tourniquet was used in all patients and was released after skin closure. All operations were performed by the same team of surgeons. Cefazolin was administered 30 minutes before surgery for short-term prophylaxis. The analgesia regimen was intravenous administration of ketorolac (30 mg every 8 hours) for the first day and oral administration of celecoxib (200 mg twice daily) for the next 5 days.
All patients were randomly divided into three groups (the intravenous group, intra-articular group and placebo group) by computer-generated randomization, with 40 patients in each group. Patients in the intravenous group received intravenously administered TXA (15 mg/kg in 100 mL of saline solution) 10 minutes before tourniquet release; patients in the intra-articular group received intra-articular TXA (15 mg/kg in 100 mL of saline solution) injected via the drainage tube; and patients in the placebo group received an equivalent volume of normal saline administered into the knee joint cavity and intravenously. The surgeon, anesthetist, patients and observer were blinded in regard of the use of tranexamic acid.
A standard protocol was used for the rehabilitation program. Muscular training included exercises for the quadriceps, hamstrings and core stability at each stage. Patients were discharged with a long brace that was locked in the fully extended position at rest for 6 weeks postoperatively, and they were instructed to partially to fully bear weight with crutches under the protection of the fully extended brace according to their tolerance. After 4 weeks postoperatively, a full range of passive activity was allowed, with an increase in the flexion ROM of 15° per day up to ≥ 120°.
The drainage tubes were routinely removed, and the volume of drainage was recorded 24 hours after surgery. All enrolled patients experienced a 4-week follow-up period. After 1, 2 and 4 weeks postoperatively, the visual analogue scale (VAS) score , patellar circumference, hemarthrosis grade  and Lysholm score  were recorded. The patellar circumference was measured 1 cm proximally to the superior border of the patella of both knees to evaluate joint swelling. To minimize interindividual variability in the patellar circumference, we considered the difference between the affected side and the contralateral side . The clinical grading of hemarthrosis was scored according to Coupens and Yates . Patients were accurately examined for the presence of leg swelling or calf pain, and those with suspected deep venous thrombosis (DVT) were recommended to undergo Doppler sonography. The postoperative evaluation was performed by a trained orthopaedic surgeon (R.M.) who was blinded to the treatment groups.
The sample size was determined at a website (url: http://powerandsamplesize.com/Calculators/). In the sample size determination, multiple comparisons were made using the drain output volumes. The effect size, standard deviation, study power, and significance level were 75, 40, 0.80, and 0.05, respectively, and 39 patients were required in each group. All continuous variables are expressed as the mean ± SD. All data were managed through SPSS (IBM SPSS Statistics 19, USA). All continuous variables were analyzed by a Shapiro-Wilk test to evaluate the normal distribution of quantitative data. Indicators conforming to normal distribution were expressed in as mean ± standard deviation; a one-way analysis of variance (ANOVA) test was used for comparisons within groups, and least significant difference (LSD) test was used for comparison between each pair of groups. Indicators that did not conform to normal distribution were described as median and quartile; a Kruskal-Wallis H test was used for comparison within groups, and a Mann-Whitney U test was used for comparison between each pair of groups. The χ2 test or Fisher exact test was used to compare the categorical variables. Differences were considered significant at p < 0.05.