ABSTRACT
AIMS: Tranexamic acid (TXA) has been shown to significantly reduce transfusion rates in primary total hip arthroplasties (THAs), but high-quality evidence is limited in the revision setting. The purpose of the current study was to compare the rate of blood transfusions and symptomatic venous thromboembolic events (VTEs) in a large cohort of revision THAs treated with or without intravenous (IV) TXA. PATIENTS AND METHODS: We performed a retrospective review of 3264 revision THAs (2645 patients) between 2005 and 2014, of which 1142 procedures received IV TXA (1 g at incision and 1 g at closure). The mean age in the revision group with TXA was 65 years (28 to 95), with 579 female patients (51%). The mean age in the revision group treated without TXA was 67 years (21 to 98), with 1160 female patients (55%). Outcomes analyzed included rates of transfusion and symptomatic VTEs between procedures undertaken with and without TXA. These comparisons were performed for the overall cohort, as well as within cases subcategorized for aseptic or septic aetiologies. A propensity score was developed to minimize bias between groups and utilized age at revision THA, sex, body mass index, American Society of Anesthesiologists (ASA) score, preoperative anticoagulation, and year of surgery. RESULTS: Tranexamic acid significantly and substantially reduced the rate of blood transfusions after revision THA overall from 54% to 26% (p < 0.001; adjusted relative risk (RR) 1.6; 95% confidence interval (CI) 1.3 to 1.9), with a significant reduction in both aseptic (49% to 18%; p < 0.001) and septic (73% to 53%; p = 0.04) revisions. The rate of VTE was minimal overall, with three events (0.3%) in the TXA group and four events (0.2%) in the non-TXA group. There were no significant differences in VTE rates based on TXA use or aetiology of revision. CONCLUSION: Intravenous TXA significantly reduced transfusion rates during all-cause revision THAs, including a subgroup analysis of both aseptic and septic cohorts. Adjusted risk using propensity modelling showed no statistical difference in rates of VTEs between either group. Cite this article: Bone Joint J 2019;100-B(6 Supple B):104-109.
