Hospital policy of tranexamic acid to reduce transfusion in major noncardiac surgery – TRACTION trial 

24 July 2026

The TRACTION trial was a pragmatic, multicenter, double-blind, cluster-randomized, placebo-controlled trial. It evaluated whether a hospital policy of routine tranexamic acid (TXA) use in patients undergoing major noncardiac surgeries would safely reduce the incidence of red cell transfusion.  

Eligible adult patients were recruited across 10 hospitals (clusters) in Canada. Participating sites with patients undergoing surgery who were at high risk for red cell transfusion (5% or more) were randomised in a 1:1 ratio at 4-week intervals to provide either TXA (IV bolus at the start of surgery and an additional dose before skin closure with timing at the discretion of the anaesthetist) or matching placebo (0.9% sodium chloride).  

Exclusion criteria were active thromboembolic disease, cardiac and hip or knee total arthroplasty (in which TXA administration is standard practice), surgeries involving a free flap, trauma surgery with TXA in the previous 3 hours and pregnancy.  

The coprimary outcomes were transfusion of red cells during the index hospitalisation [effectiveness] and diagnosis of venous thromboembolism (VTE), within 90 days after surgery [safety].  

During the study, data linkages were successfully established for 8,273 (98.2%) patients. VTE outcomes were obtained from administrative sources (validated combination of physician billing and imaging codes or reports). The trial included a relatively diverse surgical population with a high proportion (around 60%) of oncology surgeries.  

Fewer patients in the TXA group received a red cell transfusion during hospitalisation compared to those in the placebo group (7.4% versus 9.8%; relative risk, 0.73; 95% C.I., 0.61 to 0.86, adjusted difference, −2.7 percentage points; 95% CI, −4.2 to −1.4).  

The percentage of patients who received a diagnosis of VTE within 90 days occurred in 2.1% of patients in both groups and met the prespecified criterion for noninferiority.  

Limitations of the study included analyses reliant on multisource data linkage and ascertainment of VTE missing sub-clinical events.  

The study’s findings are consistent with and add to the existing evidence supporting the use of antifibrinolytic therapy in surgery.  

 Reference:  

Health professionals news