Tranexamic Acid for Acute Hemorrhage: When Is Enough Evidence Enough?

David Faraoni, MD, PhD, FAHA; Jerrold H. Levy, MD, FAHA, FCCM


Anesth Analg. 2019;129(6):1459-1461. 

In this issue of Anesthesia & Analgesia, Lier et al[1] present a critical reappraisal of tranexamic acid (TXA) utilization over the past decade and summarize their opinion of the potential downsides of a liberal usage of TXA. The authors discuss the lack of pharmacokinetic and optimal dosing studies in the adult noncardiac surgical population, the risk of thromboembolic complications associated with administration of TXA, and review the large randomized clinical trials, including Clinical Randomisation of an Antifibrinolytic in Significant Haemorrhage 2 (CRASH2) and World Maternal Antifibrinolytic Trial (WOMAN), and challenge their methodology and statistical significance.[2,3] They also advocate for a more selective and individualized TXA use based on viscoelastic hemostatic assays for real-time assessment of fibrinolysis. Because many of their arguments are controversial, we would like to examine some of the authors' statements.

Evidence-based medicine (EBM) is "the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients and integrating individual clinical expertise with the best available external clinical evidence from systematic research."[4] Although clinical expertise and expert opinions play a role in EBM, those remain at the bottom of the EBM pyramid for apparent reasons. Most clinicians and researchers must accept that prospective randomized trials remain the most reliable evidence because of the strictly controlled methodology allowing for an optimal comparison between 2 groups. When evidence is summarized into a review, a systematic design is often accompanied by a meta-analysis to avoid subjective interpretation of the literature. The question regarding the efficacy and safety of prophylactic or therapeutic administration of TXA in that regard is easy to analyze considering several large prospective studies that include thousands of patients have been published over the past decade.[5] Most of the smaller studies support TXA's effectiveness in reducing bleeding and transfusion, while CRASH2 demonstrated reduction in mortality when administered within the first 3 hours following trauma (16% vs 14.5%).[6] The same studies also reported that the use of TXA is not associated with increases in arterial or venous thrombotic events.[2] So, where does the idea of TXA-induced thrombotic events come from?

Any patient surviving from severe injury and massive hemorrhage is at increased risk of thromboembolic complications. Bleeding patients frequently require multiple allogeneic blood products and/or coagulation factors, present with hemodynamic instability, develop systematic inflammatory responses, and have prolonged intensive care and hospital stays. While the risk of thrombotic complications is important, there are little data suggesting that TXA increases this risk. TXA is a clot stabilizer that inhibits plasminogen conversion to plasmin but is not a clot promoter. Of note is that none of the prospective randomized studies published to date reported an increased risk of thrombotic complications when TXA was administered prophylactically or therapeutically to patients following an acute hemorrhage.

In their opinion piece, Lier et al[1] discussed the method used in those prospective studies to diagnose clinically relevant thrombotic events. The authors should acknowledge that the strategy used was similar in the treatment and control arms, and one could assume that the estimated risk was identical in each group. The authors also refer to a few small retrospective studies reporting an increased risk of thrombotic complications when TXA was used. The use of retrospective design in such studies is a major bias, as the authors were not able to control for the severity of the injury and the bleeding, the number of blood products transfused, the administration of coagulation factors, or that sicker patients may have been more likely to receive TXA than patients presenting less severe injury.

Lier et al[1] also refer to the concept of "fibrinolytic shutdown" that was developed by viscoelastic test users to stratify trauma-induced fibrinolysis into "hyperfibrinolysis" and "hypofibrinolysis" or "fibrinolytic shutdown."[7] This concept should be interpreted with caution as it is based on the assumption that viscoelastic hemostasis assays (eg, thromboelastography [TEG] or thromboelastometry [ROTEM]) are sensitive enough to detect variability in fibrinolytic activation or inhibition. Although viscoelastic hemostatic assays performed on whole blood were thought to be a promising alternative to the time-consuming standard laboratory testing performed on plasma to monitor clot stability and fibrinolysis, those assays are not sufficiently sensitive to detect nonsevere changes in fibrinolytic activation. Raza et al[8] demonstrated that fibrinolytic activation, as measured by increased plasmin α2-antiplasmin complexes, tissue-type plasminogen activator (t-PA), and D-dimer biomarkers occur in almost two-thirds of trauma patients and that the degree of fibrinolytic activation is associated with increased injury and with significantly greater transfusion requirements, morbidity, and mortality. Moreover, >90% of fibrinolytic activation in trauma patients is not detected with the current definition of hyperfibrinolysis measured by thromboelastometry.[8] The insensitivity of viscoelastic hemostatic assays to detect nonsevere hyperfibrinolysis is explained by the imbalance between the concentration of activators required to initiate clot formation and the absence of fibrinolytic activator in the test cup. Although efforts to restore the balance by using t-PA have been studied in TEG or ROTEM assays,[9,10] those assays have not been validated against the standard laboratory assays. Methods to measure fibrinolytic activity in blood or plasma are limited by laboratory definition of a "gold standard" assay that reflects overall fibrinolysis,[11] beyond the current gold standard of the euglobulin lysis assay. Until such an assay becomes available, recommending a monitoring-based administration of TXA is problematic as patients presenting with submaximal fibrinolytic activation will not receive therapy. As a result, until "fibrinolytic shutdown" can be readily demonstrated with a point-of-care assay, the controversy regarding TXA in administration in traumatic coagulopathy will continue. Although the pharmacokinetics of TXA has been described in children and adults undergoing cardiac surgery with cardiopulmonary bypass,[12,13] little is known regarding the optimal dose offering maximal efficacy without increasing the risk of potential side effects. As discussed above, to determine the minimal effective plasma concentration required to completely inhibit fibrinolytic activation, additional sensitive assays would be needed. Therefore, the dose schemes are mainly based on the assumption that a concentration of 20–50 μg/mL would allow complete inhibition of fibrinolytic activation, but that remains to be proven.[14] The only estimation of the "optimal" dose scheme to be administered in trauma patients was recently published by Grassin-Delyle et al.[15] In this study, the authors suggested that if the objective is to maintain an effective concentration of 20 μg/mL over 8 hours, the total doses with the body weight-adjusted scheme would be between 27.9 and 33.1 mg/kg for patients weighing between 50 and 120 kg. This means that the total dose will be lower or equivalent to the doses actually used in clinical trials (TXA 1 g over 10 minutes then 1 g over 8 hours).

In summary, prophylactic and therapeutic administration of TXA should be part of management strategies for acute hemorrhage in all international guidelines until the potential risk of adverse events is demonstrated in well-designed randomized clinical trials. Although the concept of more targeted strategy is interesting, bedside monitoring of fibrinolytic activation is limited by the sensitivity of viscoelastic hemostatic assays. Considering the apparent safety of TXA, the relative cost, and the potential effectiveness as demonstrated in a large number of prospective randomized trials, TXA is an important part of bleeding prevention and management, and as part of a multimodal strategy.